Language and Space: Theories and Methods: An International Handbook of Linguistic Variation, Volume 1

Language and Space: Theories and Methods HSK 30.1

Handbücher zur Sprach- und Kommunikationswissenschat Handbooks o Linguistics and Communication Science Manuels de linguistique et des sciences de communication Mitbegründet von Gerold Ungeheuer () Mitherausgegeben 19852001 von Hugo Steger

Herausgegeben von / Edited by / Edite´s par Herbert Ernst Wiegand

Subseries: Language and Space An International Handbook o Linguistic Variation Edited by Jürgen Erich Schmidt Band 30.1

De Gruyter Mouton

Language and Space An International Handbook o Linguistic Variation Volume 1: Theories and Methods Edited by Peter Auer and Jürgen Erich Schmidt

De Gruyter Mouton

ISBN 978-3-11-018002-2 e-ISBN 978-3-11-022027-8 ISSN 1861-5090 Library of Congress Cataloging-in-Publication Data Language and space: theories and methods : an international handbook of linguistic variation / edited by Peter Auer, Jürgen Erich Schmidt. p. cm. ⫺ (Handbooks of linguistics and communication science ; 30.1) Includes index. ISBN 978-3-11-018002-2 (hardcover : alk. paper) 1. Language and languages ⫺ Variation. 2. Linguistic geography. 3. Dialectology. I. Auer, Peter, 1954⫺ II. Schmidt, Jürgen Erich, 1954⫺ P120.V37L33 2010 417⫺dc22 2009048180

Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de. 쑔 2010 Walter de Gruyter GmbH & Co. KG, Berlin/New York Typesetting: META Systems GmbH, Wustermark Printing: Hubert & Co. GmbH & Co. KG, Göttingen Cover design: Martin Zech, Bremen ⬁ Printed on acid-free paper Printed in Germany www.degruyter.com

Introduction to the Language and Space series In 1982 and 1983 the renowned HSK series was launched with a two-volume handbook, Dialektologie: Ein Handbuch zur deutschen und allgemeinen Dialektforschung. Even though this first handbook played a significant role in the subsequent success of the series and was out of print by the mid 1990s, early plans to publish an updated edition (as was done with other successful titles) were soon shelved. Consensus instead settled around the view that advances in the understanding of the object of dialectological study, along with fast-paced developments in the relevant disciplines and the need to adequately represent the rich findings of this global research effort, made a completely fresh start necessary. Even though the variability of human language is in essential ways caused and constrained by the dimensions of time and space, and although most people today still speak with some form of distinct regional coloring, dialects isolated from supranational and standard varieties are increasingly becoming marginal phenomena, right across the world. Accordingly, a reorientation of research into language and space has begun ⫺ shifting from a discipline focused on the reconstruction of premodern language states (traditional dialectology) to approaches dedicated to a precise analysis of the dynamic processes at work within complex language systems and their explanation in terms of cognitive and interactive-cum-communicative factors. This shift in emphasis towards embodied and evolving language has led to a blurring of the established boundaries between dialectology, sociolinguistics and language contact studies and to the adoption of impulses from geography, sociology and anthropology as part of a wider reappraisal of the relationship between geographical place and cultural space. Additionally, a way has needed to be found to take account of significant differences in how language is “territorialized”. These range from traditional, sedentary settlement patterns to personally mobile and electronically delocalized postindustrial lifestyles, and from semiliterate, largely oral cultural traditions through, say, the formation and maintenance of immigrant communities and enclaves within multicultural and urbanized landscapes, to the inhabiting of pre-eminently social spaces in the increasingly fragmented and ad hoc milieus of contemporary society. Against the background of this reorientation, the idea of a subseries within the HSK range entitled Language and Space: An International Handbook of Linguistic Variation developed out of intensive discussions between representatives from various research fields, the editors and publishers. Inaugurating this subseries are two “foundation” handbooks, canvassing international developments in theory and research methods and, for the first time, interrogating the theoretical and practical foundations of linguistic cartography. These cross-linguistic foundational volumes are to be complemented by a loose sequence of volumes that each analyze the full dimensions of spatial variation within an individual language or language group whilst remaining guided by a uniform structure. This first introductory volume, Theories and Methods, directly addresses both the changes in the object of study (linguistic variation across “space”) and the attempts within the relevant disciplines to adjust to the concomitant reconceptualization of its nature. As intimated by its subtitle, the volume is divided into two halves. The first of

vi

Introduction to the Language and Space series these, the theoretical wing, encompasses a transdisciplinary discussion of the notion of space together with critical evaluations of linguistic approaches to it plus several articles on the structure and dynamics of (and between) language spaces. The second, methodological wing details and showcases traditional and contemporary methods of data collection, analysis and presentation in linguistic geography and language variation studies, with special emphasis on the methodological problems within the individual structural domains (phonology, prosody, morphology, lexis, syntax and discourse) and a series of illustrative and multifaceted case studies. The second volume, Language Mapping, addresses a striking deficit in the field of studies into language and space. To date there has never been a collected consideration of the many issues impinging upon the creation and use of maps in the investigation of language, its distribution and variation. Within various major languages, schools and traditions have emerged within which problems have been addressed and approaches have been refined, but there has been a dearth of exchange between these traditions. Starting from a thoroughgoing consideration of the conceptual, cognitive and cartographic fundamentals of committing languages to maps, the second foundational volume also explores the individual traditions, their origins, peculiarities and strengths, before considering numerous aspects of the revolutionary enabling impact of computing on language mapping and some of the intersections between the cartography of language and other fields of human endeavor. Naturally, given the topic, the volume will be accompanied by an extensive, separately bound collection of maps. These two foundation stones are to be followed by a series of works that will, while oriented to a uniform structure, thoroughly explore the current state of research into the spatial dimensions of particular languages or language areas. Given linguists’ increased awareness of the complexity of the relationship between language and physical space, and of the fiction of a single “authentic” variety per speaker, the volumes are focused on (groups of) languages rather than regions and attempt to chart their internal variational structure and dynamics, their interface with other languages and their distribution across physical, social and cultural space. Each volume will open with a section examining the history of investigation into the language(s) in question, the foci of current research and perceived deficits. Then the genesis of the (areal) linguistic constellations and variety spectra will be treated along with a complete anatomy of the language space. But the bulk of each volume will be devoted to a detailed description of linguistic subregions and domains including, obviously, those which transcend traditional bounded spaces as well as attitudes and social configurations, and to an exploration of aspects specific to the language (group), including its use in a range of locations as a postcolonial or an immigrant language, the roles of various media and the techniques and technology used to present results. For the noble HSK handbook tradition, Language and Space thus represents the revisiting, after more than a quarter of a century, of one of the fundamental dimensions of human language in all its variety and flux. The new series attempts to draw together and take account of the advances in our understanding of this dimension, broaching the boundaries between disciplines, questioning but not abandoning established traditions, drilling down into the concept of space itself, in order to bring to its readers some of the excitement of the scientific hunt for that most immanent quarry ⫺ language itself. Jürgen Erich Schmidt, Marburg (Germany) 19 June 2009

Introduction to this volume Theory and Methods, the introductory volume of the Language and Space series, is appearing at a point in time when the theoretical and methodological reorientation of research into the interplay of language and space is in full swing. It is illuminating to see this current reorientation against the background of the major research tradition and developments within the disciplines involved, since this also reveals how the present handbook positions itself. At the time when Georg Wenker and Jules Gillie´ron were establishing linguistic geography (the late nineteenth century), other linguists were already struggling to gain a theoretically and methodologically adequate understanding of the variability of language and expressing surprisingly modern theoretical and methodological concerns. As early as 1880, for instance, the neogrammarian Philipp Wegener set out a program with which the “unendliche menge” of “sprachformen” (‘infinite number of linguistic forms’; 1880: 465) of areally, socially and contextually determined linguistic variation should be approached, starting out from individual village dialects. In 1905 and based on detailed observations of numerous, carefully socially classified informants, Louis Gauchat undertook an analysis of linguistic variation within the village of Charmey (in Switzerland) before reaching the conclusion that “l’unite´ du patois … est nulle” (‘the unity of the dialect is zero’; 1905: 222). The critical methodological insight to emerge from his work was that the alleged homogeneity of a village dialect was in fact an artifact (1905: 179, 222; vs. Zimmerli 1899). However, methodical and practical constraints long prevented these early insights into the complexity of linguistic variation from playing a key role in the shaping of research undertakings. Wegener failed to develop a methodology that was adequate to the task of dealing with the complex research goals, and while Gauchat’s laborious methods were of use in providing a meticulous analysis of the linguistic dynamics in a single village (also cf. Enderlin 1910, who even employed covert data collection techniques), they were not suited to gaining an overview of the areal dimension of linguistic variation. In hindsight, it therefore appears almost inevitable that, despite these efforts, robust, relatively simple to execute, yet also reductionist methods have dominated research for so long: introspection (in the neogrammarian paradigm of Ortsgrammatiken) and, even more importantly, surveys featuring translation and naming tasks (traditional dialect geography). With the help of these methods, an enormous effort over several generations of European researchers produced an abundance of relatively precise descriptions of the phonology, phonetics, lexicon and morphology of the dialects spoken in specific locations, alongside monumental dialect atlases that provide an accurate but partial account of variation in spoken language across space. The methodological standards of this epoch of language and space research increasingly aimed at establishing strict comparability across space with the highest possible density of survey locations. As a consequence, the heterogeneous groups of informants originally selected for the national dialect atlases (of Wenker and Gillie´ron) were replaced in the later regional atlases by informants with identical social characteristics (elderly, sedentary men and women with little education and manual occupations), indirect survey methods based on the distribution of written

viii

Introduction to this volume questionnaires (Wenker) were replaced by direct (face-to-face) surveys conducted by phonetically trained fieldworkers, and so on. The enormous wealth of data on spatial variation in spoken language accumulated in this way opened up fascinating possibilities for the analysis of language change; it allowed the study of the effects of both internal and external factors on language and reconstructions of the processes that may have produced the areal distributions discovered. For structuralists, it allowed the defining of systems of contrasts between broad dialectal areas. All this obscured the fact that the mainstream of dialectology had taken a reductionist turn, ignoring the fundamental problems Gauchat had astutely recognized so early on. The intrinsic heterogeneity of language and its embedding in a range of social and other factors was excluded, so that in the end only the spatial dimension remained. While Wegener and Gauchat were confronted with the problem of how to develop adequate methods of data collection, later attempts to break away from the monodimensional and homogenizing approach of dialect geography suffered from the lack of appropriate analytic techniques; in the end, analysis was often replaced by pure documentation. The sound archives which began to be set up relatively soon after the emergence of viable recording technologies and which recorded “the individual language, at the same location, of persons of differing age, gender and social standing” (Wagner 1924⫺ 1925: 230; our translation) are one example. Another example is Hans Kurath’s Linguistic Atlas of New England, for which, between 1931 and 1933, direct survey techniques were used to collect variants from six “types” of informants who differed in level of education, degree of social contact, age and general attitude (“old fashioned” vs. “modern”). Significantly, the decisive methodological breakthrough finally came from outside, namely from the sociolinguistics of urban life. The complex sociolinguistic reality of contemporary, urbanized societies could not be adequately captured using traditional dialectological methods (cf. Chambers and Trudgill 1980: 55⫺56). Borrowing quantitative methods from the social sciences, William Labov (1966) grounded a new discipline that was able to observe and precisely analyze the facets of urban language use of representative social groups. Aside from Labov, it was the pioneering work of linguists such as John Gumperz, Joshua Fishman, Leslie Milroy, Peter Trudgill and many others which led to the rapid development of an independent and extremely successful new discipline: sociolinguistics. With increasingly refined data collection and analysis techniques, it was finally possible to isolate the social, interactional and attitudinal factors that steer the complex language use of disparate social groups, to uncover the regularities behind the variational registers and styles of groups of speakers and finally, through the combination of apparent and real-time studies, to capture the dynamics of language change. But, mirroring the origins of language and space research, it was the roaring success of sociolinguistics which ⫺ for methodological and practical reasons ⫺ at the same time led to a renewed narrowing of focus. Whereas traditional dialectology suppressed the complexity of linguistic variation beyond the areal dimension and favored the survey and analysis of rural informants’ dialect knowledge to the exclusion of language usage data, sociolinguistics tended to ignore both variation across space and competence data. The amount of effort required to gain valid spontaneous speech data in comparable contexts made the systematic investigation of the full spread of linguistic variation across space appear an impossible task. It was also not clear how a valid analytic connection

Introduction to this volume could be established between the new corpora of spoken everyday language and the data on dialect knowledge elicited by traditional dialectologists. The separate developments in traditional dialectology and modern quantitative sociolinguistics (with its equally rigid methodology) thus led to the isolation of dimensions of language variation that were in fact intimately interconnected and to the creation of apparently incompatible data classes. In comparison to the situation described above and documented in the earlier handbooks in this series on Dialektologie (1982⫺1983) and Sociolinguistics/Soziolinguistik (first edition 1987⫺1988), i. e., 30 years ago, the current landscape has shifted fundamentally. The strict boundaries between sociolinguistics and dialectology have fallen, and approaches have been developed in which theoretical and empirical linguistics can be interrelated in promising ways. It should be immediately obvious why research on language and space has a significant role to play here. No other dimension of variation so fundamentally shapes the diversity of human language as does space, both across and within languages. The spatial, social and contextual dimensions are inextricably linked to each other, and language diversity and variability are related in complex ways to interactional and attitudinal factors. The problem that shaped the very beginnings of language and space research, namely how to obtain data on language use and language competence that are both reliable and comparable (across space), still awaits a workable solution. But nowhere else in the field of linguistics are we (or have we ever been) offered such a great opportunity to analytically combine a wealth of data about well-chosen sectors of the linguistic knowledge of areally distributed groups of speakers at different times with a wealth of well-documented data about sectors of the variable speech behavior of speakers in such a way that empirical explanations for the fundamental questions posed by a theory of language (change) can be found. The Theory and Methods volume has set itself the goal of rendering the current methodological and theoretical reinvigoration of language and space research visible and, in so doing, of highlighting the innovative impulses this is bringing to the whole of linguistics. If we are right, it is the following lines of development which have characterized research into language and space over the last thirty years: ⫺ A breakdown of interdisciplinary barriers, as a consequence of which a field of study emerges which reaches far beyond classical dialectology and sociolinguistics to also encompass language contact studies, linguistic and areal typology, theoretical linguistics and cognitive sciences and which draws in and adapts impulses from geography and anthropology. ⫺ A turn to the exploration of the entire spectrum of language variation, in which the two fields of dialectology and sociolinguistics are being drawn closer together. On the one hand, traditional dialectology’s monodimensional surveys of linguistic competence have been expanded to include competence data from different social groups (starting with Fujiwara’s work in Japan, 1974⫺1976) and finally combined with the systematic collection of speech data across space (pluridimensional dialectology, cf. Thun in this volume). On the other hand, community studies centered upon specific locations have increasingly begun to take the spatial dimension of linguistic variation into account (thereby putting an end to the “sidelining of the spatial in early variationism”, cf. Britain in this volume), so that ⫺ at least in Europe ⫺ the broad development of regional varieties can be sketched (cf. Auer 2005).

ix

x

Introduction to this volume ⫺ The development of web-based resources which cross-connect data on the dialect knowledge of groups of speakers collected at different points in time with data from historical sound archives and more recent surveys of regional speech. This enables real-time analyses across space in which stability and change in regional varieties can be tracked and the effects of interactional and cognitive/linguistic factors accurately determined (Schmidt in this volume). ⫺ The end of the traditional focus on monolingual, immobile speakers from small regions or specific locations. More recent language and space research asks about the linguistic foundation of spaces (and places) of all sizes: from those which have traditionally formed the focus of dialectological research (villages or regions) to politically defined territories (such as nation-states) which assume a (standard) language as their correlate and ideological justification, from global spaces called into existence by European colonial expansion (overseas varieties of European languages, pidgins and creoles) via supranational regions in which languages have converged (sprachbund) to transnational spaces emerging with the support of electronic media in the age of globalization. ⫺ The emergence and development of folk linguistics, in which the subjective spatial structurings (including evaluations) that speakers develop are systematically investigated. It takes as its object those perceived differences between the ways in which people speak that enable them to locate conversation partners within larger frames of reference ⫺ frames which, however, are still dependent upon the “placing” participant’s perspective, i. e., on his or her life-world (cf. Niedzielski and Preston 2000). ⫺ The rapprochement between theoretical linguistics and language and space research. In the last decades, language and space research has begun to systematically reclaim the long-neglected research fields of areal syntax (e. g., SAND and SADS) and prosody. Given the importance of syntactical and phonological studies (OT, autosegmental phonology) for theoretical linguistics, this creates the preconditions for a systematic consideration of the areal dimension of linguistic variation in the ongoing attempts to theoretically model the cognitive processes supporting language. ⫺ A focus on postmodern views of the language and space connection. In the wake of globalization, especially the increasing speed of communication and enlarged communicative reach, studies have set out to explore the consequent changes in the degree to which language is spatially bounded. At the forefront of these investigations is a focus on the dissolution of traditional ties to space on the one hand, and on new ways of symbolizing belonging in spatial terms (cf. place-making activities) on the other. We will discuss this highly productive development in more detail below. The powerful impulses currently re-orienting language and space research are not exhausted by the empirical investigation of the relation of language to space in its full complexity. At the same time, the notion of space itself is no longer taken for granted and has become the object of theoretical reflection. Potential docking points are earlier and ongoing discussions in geography (cf. Johnstone 2004 and in this volume or Cresswell 2004 for a geographer’s point of view) or Simmel’s (1903) sociological theory of space, not to mention the phenomenological tradition spearheaded by Edmund Husserl (cf. Günzel 2006). Any linguistic theory of space will need to acknowledge the central importance of the empirical fact of a multilayered relationship between language and space together with its historical development, the contours of which can be sketched as follows.

Introduction to this volume The primary form of the relationship of language to space is the product of millennia of exclusively face-to-face interactions leading to the development of commonalities and differences in linguistic systems in pre-modern times. Speakers’ perception and recognition of language differences occurred within a spatial framework, which led to an evaluation of language in terms of the basic categories of own vs. other. Spatially differentiated speech therefore did not just provide the base medium for the interactive constitution of social and cultural systems, themselves perceived in relation to space. Rather, linguistic categorizations and evaluations were an integral part of these systems, and language differences an indexical (socially symbolic) expression of them. In two important transformations, the nature of this pre-modern “language⫺body⫺ place connection” (as Quist, this volume, puts it) has altered and become more complex. In a first, modern transformation, this connection was dissolved by the uniformitarian language ideology that the modern nation-state imposes upon its citizens: the individual no longer just belongs to the local Gemeinschaft, co-extensive with Schütz’s “world within … actual and potential reach” (Schütz and Luckmann [1983] 1989: 166) and characterized by strong network ties based on face-to-face communication; beyond this local Gemeinschaft, there is the imagined community of the nation-state, which is beyond the reach of the individual subject. This community is symbolically present through its national standard language, a language variety which is by definition distributed evenly over the territory of the nation-state, although it is not evenly distributed across the social layers of the population. The invention of printing plays a central role here. It made linguistic interactions in the absence of direct personal contact possible, which in turn enabled the emergence of written norms across larger areas and laid the foundation for wide-scale (national) pronunciation norms. The tension between the local vernacular (dialect) and a uniform state language established a new dimension of sociolinguistic variability; it became a motor and symbol of social differentiation and thereby defined a social space (Mæhlum in this volume), i. e., a vertical structure on top of the existing horizontal one. What used to be nothing more than the “natural” way of speaking in a given location (“first order indexicality” in the sense of Johnstone, Andrus and Danielson 2006), now became, in the worst case, the language of the underprivileged classes who had no access to education, a variety that needed to be avoided in out-group situations, or, in the best case, a symbol of regional or local belonging (“second order indexicality”; Johnstone, Andrus and Danielson 2006). The second social and medial transformation which has untied the body⫺language⫺ place connection is the post-modern one. Its wider context is the process of globalization, which accelerated during the last quarter of the last century, a period which Bauman (1998: 8) calls the “Great War of Independence from space”. There has been not just an enormous increase in the speed at which capital and information flow freely around the world (beyond the control of the nation-state), but also, for expanding groups of speakers (especially migrants and elites), a fundamental shift in the spatial boundedness of life and language. As a consequence of the effective overcoming of distance (for both faceto-face and mediated communication), the individual’s communicative reach is enlarged. Where people reside and where they are socialized can have less influence on their communicative practices than forms of communication that transcend spatial separation. Alongside the typical, historically anchored, complex, yet monolingual registers of groups of speakers, clusters of speakers can increasingly be observed whose linguistic repertoires are composed of variants typical of a region, urban speech forms that have

xi

xii

Introduction to this volume arisen among linguistically heterogeneous peer groups and pan-ethnolectal forms originating from different contact languages, etc. But not only has this untying of the language-body⫺space connection led to a complex, multilayered situation in which pre-modern and modern groups of speakers coexist with post-modern ones; it has also evoked counter-tendencies. The age of globalization has given rise to a new interest in symbolizing belonging in spatial terms, in turning abstract space into places, which are impregnated with meaning and which symbolize belonging. People living in a location ⫺ whether born there or (more often) not ⫺ may choose to construe a local identity for themselves. These place-making activities use the symbols of (local) language(s): ⫺ multilingual street signs and graffiti (Auer 2009) colonize public spaces and symbolize their producers’ claims to them; ⫺ dialects and autochthonous minority languages are revitalized in order to mark local belonging; ⫺ dialects and minority languages may also become folklorized and commodified; they then become part of the way in which a location presents itself to its inhabitants and to outsiders as “special”, “genuine” or “authentic”, in order to attract tourists, etc.; ⫺ dialect stereotypes help to create identity-rich places (Johnstone, Andrus and Danielson 2006 call this “third order indexicality”), usually reinforced by the media (cf. Androutsopoulos in this volume); ⫺ fragments of both the international lingua franca (English) and some immigrant minority languages become available as resources for creating new regional (“glocalized”) ways of speaking, new (supra)regional styles and lects. Space still matters. The theoretical challenge for the future will be to analyze how, to what degree, and why it matters for language and how language matters for space. In order to do so, we need to model (a) the interactional and social bases of the spatial categorization of linguistic variation under pre-modern, modern and post-modern conditions, (b) the conversion of heterogeneous linguistic practices into consolidated language change in intergenerational transmission, (c) the relevance of ethnodialectological representations of language spaces and of place-making language-related activities for language change and (d) the relationships between large-scale (global, international), medium-scale (national) and small-scale (regional) spatial frames and language contact on all these levels. We offer up this volume in the hope that it will provide a rich foundation for the necessary theoretical advance. Let us now turn to an overview of the structure of the handbook. Vis-a`-vis the relation of language to space, the first half of the handbook serves the dual goal of (comprehensively) documenting the theoretical achievements to date and offering impulses for necessary future development. The introductory part is dedicated to a transdisciplinary discussion of current conceptualizations of space. The concepts of “geographical space”, “social spaces”, “political spaces” and “transnational spaces” are examined in historical and theoretical terms with reference to the neighboring academic disciplines, whereby their constitutive and interactively mediated relevance to language is made clear. Insights from the “linguistic turn” in the social sciences are productively reflected back upon their source to reveal implicit economic and ideological dimensions, particularly in relation to territoriality, identity and standard languages. The extent to which place is a linguistic and cultural construct becomes manifest.

Introduction to this volume As a counterfoil, Part II introduces the impressive abundance of genuinely linguistic approaches developed over the long history of the study of the spatial dimension of language. These range from the foundation of an exact, empirically based investigation of linguistics (by the neogrammarians and in early linguistic geography), through the development of structuralism (including its generative variants) and variation linguistics, to those approaches which are currently shaping the ongoing theoretical development and refinement: social anthropology and interactional sociolinguistics, perceptual dialectology (folk linguistics) and the new linguistic dynamics approach, which attempts to integrate linguistic-cognitive and interactional explanatory factors. The discussion of these approaches is decidedly critical, i. e., the specific deficits of an approach are also made visible (e.g., the tendency towards an artificial isolation of an Ortssprache ‘village variety’ in the neogrammarian period or the virtually unexamined relation to space in generativist theories) and the question of potential ideological exploitation is also discussed (as with the kulturmorphologische approach). Part III addresses the consequences that emerge from the critique of spatial monodimensionality in the classification of varieties and spaces. It attempts to provide insight into the anatomy and dynamics of variety formations. The concept of a language space here is construed broadly, i. e., as a complex experiential space constituted out of interactions in a particular language and its varieties. A catalogue of the basic dimensions of variation is followed by studies exploring the unfolding of the key recurrent dynamic patterns of development (horizontal and vertical convergence vs. divergence and stasis of existing varieties and the emergence of new varieties). Language spaces are generally perceived as contiguous but, due to past or more recent migration, they can also be discontinuous, a situation that is discussed in separate articles, as are languages which have developed into minority languages as a result of the horizontal diffusion of standard varieties. The perspective shifts in Part IV. While the focus on confluent and divergent developments across time and space is maintained, in contrast to the concentration on the complex “internal” structure of a single dynamic language space of Part III, Part IV turns the attention outward, to the interplay between language spaces. In addition to foundation articles from the perspectives of language contact studies and areal language typology, there are contributions that explore specific consequences of migration and colonialism (pidgins and creoles, overseas varieties, new minorities) and the special case of nonconvergence in the face of continuing language contact. The second major division of the handbook is devoted to the methods of language and space research. Since excellent discussions of the general methods of empirical research into language variation are already available in the HSK handbooks Sociolinguistics / Soziolinguistik, Quantitative Linguistik / Quantitative Linguistics and Corpus Linguistics, it has been possible to refrain from revisiting such topics here. Instead, given that the effectiveness of methods and the limitations of methodological approaches can only be properly assessed on the basis of concrete research findings, we have included a number of illustrative case studies. These represent the various schools (and traditions) that characterize the current research climate; they reveal the interplay between the differing methods and the concrete object of study and, taken together, offer a good overview of the entire span of contemporary research into language and space. Part V takes as its topic the basic problems of data collection and corpus-building in areal linguistic research. The articles first address the problem of how to maintain empir-

xiii

xiv

Introduction to this volume ical standards in traditional survey methods. Part V is then rounded off with an overview of contemporary methods for collecting linguistic and attitudinal data with varying degrees of (un)obtrusiveness, each assessed with respect to the linguistic observer paradox. Part VI, “Data analysis and the presentation of results”, can also be compact, since map-based data analysis will be the topic of a separate volume (Language Mapping). The focus is confined to developments with specific relevance to the investigation of language and space. For instance, there have been have marked improvements in the various methods for measuring dialectality in recent years, and these are now becoming something of a standard analytical tool. Advances have also been made in the modes of data presentation through linguistic atlases and dictionaries. The printed versions have been joined by internet-based counterparts that combine and integrate numerous sources and, by making it possible to directly compare different classes of data across time (dynamic atlases) and space (digital networks of dictionaries), open up new and more exact analytical possibilities. This part is rounded out by an article which attempts to draw together the entire sweep of analytical approaches from classical dialectology via traditional sociolinguistics to speaker-oriented interpretative social dialectology. Part VII demonstrates the full spectrum of research topics and methodological approaches by means of exemplary studies. Three articles are devoted to different types of regional atlases. The Swiss German dialects are taken as an example with which to illustrate the potentially rich findings a theoretically informed analysis of the “static” maps of a classical monodimensional linguistic atlas can offer: the diffusion of innovations emerges as neither a random process nor one that is in any simple way reducible to the prestige enjoyed by groups of speakers. Far more decisive for the development and maintenance of regional types are language-internal structural constraints. The principles steering the transformation of the old European base dialects into modern regional dialects can be demonstrated using the example of the bidimensional atlas of the German dialects of the Middle Rhine, which systematically took account of social and areal dimensions. The effectiveness of pluridimensional atlases is demonstrated using the example of Portuguese varieties (Fronterizo) in the border regions between Uruguay and Brazil, where clear zones characterized by differing innovation rates and orientations can be seen to have emerged under the influence of the contact languages (Spanish and Brazilian Portuguese). In contrast to these studies, all of which illustrate the ongoing development of genuinely linguistic geographic methods of data collection and analysis, three further contributions present exemplary studies of urban and transnational spaces that make use of methods developed in and adapted from sociolinguistics, social dialectology and anthropology. Taking small multiethnic groups of school pupils in Copenhagen as an example, the establishment of “communities of practice” and the social styles that shape them are illustrated. In contrast, a methodologically diverse Swedish project studying multiethnic groups of school pupils in different major cities reveals the problems raised in identifying higher order urban varieties, styles or practices on the basis of a wealth of differences at different linguistic levels. It also shows how the perception of linguistic differences is independent of observable language use. How and why ethnographic methods can be applied to the investigation of transnational language spaces is demonstrated using an example from the francophone world, in which the circulation of language, identities and resources on a global market are elucidated. This part of the handbook is rounded off by a re´sume´ of various studies on the role of both the mass media and new media in the construction and perception of “linguistic locality”.

Introduction to this volume While the investigation of phonetic/phonological, lexical and, to a more limited extent, morphological variation has been of central relevance from the very beginnings of dialectology, other linguistic levels were long neglected. There are various reasons for this: for instance, although the significance of prosody was recognized from the outset, there was a lack of widely accepted and manageable survey and analysis procedures. Research into areal syntax is yet another story. Here, traditional linguistic geography underestimated the degree to which syntactic structures played a role in the formation of language spaces and theoretical linguistics long overlooked the analytic potential which non-written areal varieties offered for syntax. The handbook thus concludes, in Part VIII, with a systematic consideration of the methodological problems specific to the various structural domains. On the one hand, this makes apparent the advances that have been made in the field since the 1982 Dialektologie handbook was published. In phonology and morphology new analytic procedures have been developed that are capable of being applied alongside established methods; research into areal syntax has evolved into an international hot topic from which landmark publications continue to emerge. On the other hand, it has also become apparent where deficits and unfulfilled wishes remain; for instance, although areal prosody studies have profited from the by now manageable tools of instrumental analysis and the outlines of a consensual descriptive system have been developed, we are still far from an even sketchily complete description of prosodic spatial structures (but cf. Gilles 2005 and Peters 2006). More markedly, the investigation of areal variation in discourse structures has yet to progress beyond the stage of initial excursions into this field of study. It remains to be hoped that, precisely through the explication of such specific research problems and the detailing of deficits and desiderata, the necessary impulses for future research efforts will become clear.

Reerences Auer, Peter 2005 Europe’s sociolinguistic unity, or: A typology of European dialect/standard constellations. In: Nicole Delbecque, Johan van der Auwera and Dirk Geeraerts (eds.), Perspectives on Variation, 7⫺42. (Trends in Linguistics 163.) Berlin/New York: Mouton de Gruyter. Auer, Peter 2009 Visible dialect. In: H. Hovmark, I. Stampe Sletten, A. Gudiksen (eds.), I mund og bog. 25 artikler om sprog tilegnet Inge Lise Pedersen pa˚ 70-a˚rsdagen d. 5. juni 2009 [In Mouth and Book. 25 Articles on Language Dedicated to Inge Lise Pedersen on her 70th Birthday, 5 June 2009], 31⫺46. Copenhagen: Nordisk Forskningsinstitut. Bauman, Zygmunt 1998 Globalization: The Human Consequences. New York: Columbia University Press. Chambers, J. K. and Peter Trudgill 1980 Dialectology. Cambridge University Press: Cambridge. Cresswell, Tim 2004 Place: A Short Introduction. Malden, MA: Blackwell. Enderlin, Fritz 1910 Die Mundart von Kesswil im Oberthurgau. Mit einem Beitrage zur Frage des Sprachlebens. (Beiträge zur Schweizerdeutschen Grammatik 5.) Frauenfeld: Huber & Co.

xv

xvi

Introduction to this volume Fujiwara, Yoichi 1974 A Linguistic Atlas of the Seto Inland Sea. 3 vols. Tokyo: University of Tokyo Press. Gauchat, Louis 1905 L’unite´ phone´tique dans le patois d’une commune. In: Aus romanischen Sprachen und Literaturen. Festschrift Heinrich Morf zur Feier seiner fünfundzwanzigjährigen Lehrtätigkeit von seinen Schülern dargebracht, 175⫺232. Halle a.d.S.: M. Niemeyer. Gilles, Peter 2005 Regionale Prosodie im Deutschen: Variabilität in der Intonation von Abschluss und Weiterweisung. Berlin/New York: de Gruyter. Günzel, Stefan 2006 Einleitung zu Teil II/Phänomenologie der Räumlichkeit. In: Jörg Dunne and Stephan Günzel (eds.), Raumtheorie: Grundlagen aus Philosophie und Kulturwissenschaften, 105⫺ 128. Frankfurt: Suhrkamp. Johnstone, Barbara 2004 Place, globalization, and linguistic variation. In: Carmen Fought (ed.), Sociolinguistic Variation ⫺ Critical Reflections, 65⫺83. Oxford: Oxford University Press. Johnstone, Barbara, Jennifer Andrus and Andrew E. Danielson 2006 Mobility, indexicality, and the enregisterment of “Pittsburghese”. Journal of English Linguistics 34(2): 77⫺104. Labov, William 1966 The Social Stratification of English in New York City. Washington, DC: Center for Applied Linguistics Peters, Jörg 2006 Intonation deutscher Regionalsprachen. Berlin/New York: de Gruyter. Niedzielski, Nancy A. and Dennis R. Preston 2000 Folk Linguistics. Berlin/New York: Mouton de Gruyter. Schütz, Alfred and Thomas Luckmann [1983] 1989 The Structures of the Life-World. Translated by Richard M. Zaner and H. Tristram Engelhardt, Jr. Evanston, IL: Northwestern University Press. [Die Strukturen der Lebenswelt, vol. 2. Frankfurt: Suhrkamp.] Simmel, Georg 1903 [1995] Soziologie des Raums. [In: Otthein Rammstedt (ed.), Gesamtausgabe, vol. 7: Aufsätze und Abhandlungen 1901⫺1908, 132⫺184. Frankfurt am Main: Suhrkamp.] Wagner, Kurt 1924⫺1925 Grammophonische Aufnahmen deutscher Mundarten. Theutonista 1: 229⫺231. Wegener, Philipp 1880 Über die deutsche Dialectforschung. Zeitschrift für deutsche Philologie 11: 450⫺480. Zimmerli, Jacob 1899 Die deutsch-französische Sprachgrenze in der Schweiz. Part 3: Die Sprachgrenze im Wallis. Basel/Geneva: H. Georg.

Peter Auer and Jürgen Erich Schmidt 1

1

Editorial responsibilities for the present volume were divided among the two editors. Parts I⫺IV were edited in Freiburg, Parts V⫺VIII in Marburg. Peter Auer wishes to thank Elin Arbin and Carolyn Mackenzie for their help; Jürgen Erich Schmidt thanks Mark Pennay for his editorial assistance with the project.

Contents Introduction to the Language and Space series . . . . . . . . . . . . . . . . . . . Introduction to this volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I. 1. 2. 3. 4.

II. 5. 6. 7.

8. 9. 10. 11. 12.

Introduction: Language and space Barbara Johnstone, Language and geographical space Brit Mæhlum, Language and social spaces . . . . . . . Susan Gal, Language and political spaces . . . . . . . . Marco Jacquemet, Language and transnational spaces

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

1 18 33 50

Robert W. Murray, Language and space: The neogrammarian tradition Renate Schrambke, Language and space: Traditional dialect geography Clemens Knobloch, Language and space: The kulturmorphologische Ansatz in dialectology and the German language and space ideology, 1920⫺1960 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sjef Barbiers, Language and space: Structuralist and generative approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . David Britain, Language and space: The variationist approach . . . . . Penelope Eckert, Who’s there? Language and space in social anthropology and interactional sociolinguistics . . . . . . . . . . . . . . . . . . . . Dennis R. Preston, Language, space and the folk . . . . . . . . . . . . . Jürgen Erich Schmidt, Language and space: The linguistic dynamics approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

70 87

Linguistic approaches to space

III.

Structure and dynamics o a language space

13.

Gaetano Berruto, Identifying dimensions of linguistic variation in a language space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Beat Siebenhaar, Horizontal convergence of linguistic varieties in a language space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Unn Røyneland, Vertical convergence of linguistic varieties in a language space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rüdiger Harnisch, Divergence of linguistic varieties in a language space Alexandra N. Lenz, Emergence of varieties through restructuring and reevaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reinhild Vandekerckhove, Urban and rural language . . . . . . . . . . . Claudia Maria Riehl, Discontinous language spaces (Sprachinseln) . . Johan Taeldeman, Linguistic stability in a language space . . . . . . . . Claus D. Pusch, Old minorities within a language space . . . . . . . . .

14. 15. 16. 17. 18. 19. 20. 21.

v vii

107 125 142 163 179 201

226 241 259 275 295 315 332 355 375

xviii

Contents

IV.

Structure and dynamics across language spaces

22.

27.

Shana Poplack and Stephen Levey, Contact-induced grammatical change: A cautionary tale . . . . . . . . . . . . . . . . . . . . . . . . . . Walter Bisang, Areal language typology . . . . . . . . . . . . . . . . . Christian Mair, The consequences of migration and colonialism I: Pidgins and creoles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Daniel Schreier, The consequences of migration and colonialism II: Overseas varieties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thomas Krefeld, The consequences of migration and colonialism III: New minorities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Göz Kaufmann, Non-convergence despite language contact . . . . .

V.

Data collection and corpus-building

28.

Werner König, Investigating language in space: Methods and empirical standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Guido Seiler, Investigating language in space: Questionnaire and interview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tore Kristiansen, Investigating language in space: Experimental techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

23. 24. 25. 26.

29. 30.

VI.

Data analysis and the presentation o results

31. 32. 33. 34.

John Nerbonne and Wilbert Heeringa, Measuring dialect differences . Alfred Lameli, Linguistic atlases ⫺ traditional and modern . . . . . . Claudine Moulin, Dialect dictionaries ⫺ traditional and modern . . . Juan Andre´s Villena-Ponsoda, Community-based investigations: From traditional dialect grammar to sociolinguistic studies . . . . . .

. .

391 419

.

440

.

451

. .

468 478

494 512 528

. . .

550 567 592

.

613

VII. Exemplary studies 35.

36. 37. 38.

39. 40. 41.

Pia Quist, Untying the language-body-place connection: A study on linguistic variation and social style in a Copenhagen community of practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Walter Haas, A study on areal diffusion . . . . . . . . . . . . . . . . . . . Joachim Herrgen, The Linguistic Atlas of the Middle Rhine (MRhSA): A study on the emergence and spread of regional dialects . . . . . . . . Sally Boyd and Kari Fraurud, Challenging the homogeneity assumption in language variation analysis: Findings from a study of multilingual urban spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Harald Thun, Variety complexes in contact: A study on Uruguayan and Brazilian Fronterizo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Monica Heller, Language as a process: A study on transnational spaces Jannis Androutsopoulos, The study of language and space in media discourse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

632 649 668

686 706 724 740

Contents

xix

VIII. Structural domains: Methodological problems 42. 43. 44. 45. 46. 47.

Index

Peter Gilles and Beat Siebenhaar, Areal variation in segmental phonetics and phonology . . . . . . . . . . . . . . . . . . . . . . . Peter Gilles and Beat Siebenhaar, Areal variation in prosody . Stefan Rabanus, Areal variation in morphology . . . . . . . . . Dirk Geeraerts, Lexical variation in space . . . . . . . . . . . . Bernd Kortmann, Areal variation in syntax . . . . . . . . . . . . Norbert Dittmar, Areal variation and discourse . . . . . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

760 786 804 821 837 865

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

879

I. Introduction: Language and space 1. Language and geographical space 1. 2. 3. 4. 5. 6. 7. 8.

Introduction “The science of empire”: Geography to 1918 Region and culture Superorganicism and social-scientific geography The 1960s onwards: Towards new paradigms New concerns Conclusion References

1. Introduction My goal in this chapter is twofold: to encourage readers to think critically about terms such as “geography” and “space”, and to provide an overview of the ways in which geographers have used these terms and others like them. Linguistics and geography have ridden the same political and intellectual currents over the past two centuries. During that time, linguists concerned with variation and change have imagined our object of inquiry ⫺ language ⫺ in various ways, as structure or process; as relatively orderly and predictable or fundamentally stochastic and unpredictable; as autonomous or distributed; as a cognitive or a social phenomenon. We have paid less and more belated attention to defining the variables we use to account for the distribution of linguistic forms across time and space, variables such as “society”, “gender”, “age” or “region”. Investigating how geographers have imagined the object of their inquiry enriches our sense of the conceptual possibilities suggested by the phrase “geographical space”. Thus this chapter traces the history of geographical inquiry with a particular eye to the parallel history of dialectology and sociolinguistics. Where there are explicit, cited links between sociolinguistics and geography, I point to some of them, although I do not claim to be tracing all of the ways in which geographers and sociolinguists have drawn on each other’s work.

2. The science o empire: Geography to 1918 In British and American geography departments, the principal division is between “physical” and “human” geography. Physical geographers overlap with geologists, ecologists, hydrologists, biologists, chemists and physicists in their interest in landforms and the flora and fauna they support. Human geographers overlap with archaeologists, anthropologists, sociologists and economists (and occasionally with linguists) in their interest in how humans interact with the environment, shaping “the world” as the world shapes us. I concentrate in this chapter on human geography. But the two branches have common origins.

2

I. Introduction: Language and space The sixteenth-century European transition from feudalism to capitalism both enabled and was enabled by the development of a trading system linking Europe with resources in Africa and the “New World” (Heffernan 2003; on the history of geography in general, see also Livingstone 1992; Dunbar 2002). This became possible in the context of advances in navigation technology. Early Modern geographers helped develop such technology as well as techniques for map-making and systems for describing the flora, fauna, geology and peoples of the rest of the world, some of which would become economic resources. In Britain, geography began to acquire credibility as an academic discipline during the seventeenth and early eighteenth centuries. Geographers began to present their work in more scientific ways, and they played a role in the intellectual conversation of the day about the relative merits of societies and civilizations. Geography was no longer seen as just a by-product of navigation. The origin of this shift in focus is sometimes dated to 1769, the year of James Cook’s first Pacific exploration (Heffernan 2003). But academic geography was bound, often tightly, to the imperial project (Godlewska and Smith 1994; Bell, Butlin and Heffernan 1994). The link between academic geography and the aristocracy was forged in part through the rise of the aristocratic “grand tour” of the world, which often included observing, describing and sometimes collecting artifacts and people. The grand tour helped spark the development of geographical societies which institutionalized and underwrote journeys of exploration. These societies were founded by wealthy businessmen with economic interests in the findings of such journeys, in collaboration with aristocratic amateur scientists and academic geographers. The first such society was the Association for Promoting the Discovery of the Interior Parts of Africa, founded in London in 1788. This was followed by the Socie´te´ de Ge´ographie de Paris in 1821, the Gesellschaft für Erdkunde zu Berlin in 1828, and, in 1830, the British Royal Geographic Society (RGS), which became especially influential. The Royal Geographic Society’s “fellows” were mostly amateurs, although some professional scientists were included. The society provided funding and equipment and set goals for the journeys they sponsored. The Royal Geographic Society was involved in all of the major nineteenth-century journeys of exploration in Africa. The maps and other results of these journeys fuelled the later nineteenth century European rush to colonize Africa, both as a way of getting access to new raw materials at the lowest possible cost and as a way to gain political influence in Europe. During the nineteenth century, as part of the shift from often amateur Wissenschaft or knowledge-building to expert “science” (Woolgar 1988), geography was institutionalized in academia, particularly in Germany and France. One argument for this was that training in geography would help inculcate patriotism. Nineteenth-century geography was exemplified by the German Alexander von Humboldt, who “sought to establish a systematic science of geography that could analyse the natural and the human worlds together” (Heffernan 2003: 7). He was the brother of Wilhelm von Humboldt, whose linguistic theory was linked with nineteenth-century Romanticism and nationalism in similar ways. Like his brother, Wilhelm von Humboldt sought to describe the essential link between peoples, landscapes and language by virtue of which shared language was seen as one of the “pillars” of nationalist theory. According to Wilhelm, each language has its own elements of innere Sprachform, “a peculiar property of the nation or the group that speaks it” (Robins 1979: 174⫺175). By the end of the century, British geographer Hugh Robert Mill could claim that geography was “absolutely essential for our well-being, and even for the continuance of

1. Language and geographical space the [British] nation as a Power among the states of the world” (Mill 1901, quoted in Mitchell 2000: 17). Such a claim was based on the environmental determinism suggested by the von Humboldts and further developed in neo-Lamarckian thought. Neo-Lamarckism was based on geneticist Jean-Baptiste Lamarck’s theory that acquired characteristics could be inherited by an organism’s offspring. In geography, this idea was used to argue that environmental conditions create habits which are biologically transmitted to subsequent generations. Neo-Lamarckian environmental determinism holds, in other words, that there is an organic relationship between people and place, such that the natural, inherited ability of a people to better themselves, shaped by their physical environment, determines the ability of a state to prosper. Since European and EuropeanAmerican peoples came from climates and continents that were particularly likely to produce positive behavioral traits, their intervention in parts of the world with less advantageous climates and physical environments (hotter, more tropical places, for example) was thought to be justified, if not morally imperative. Environmental determinism (or “scientific racism”, as its detractors label it) continued to shape geography in the early twentieth century. In the United States, geographer Ellen Churchill Semple (1911) articulated the theory most influentially. But by the 1920s the theory collapsed, in part because it failed to provide an accurate account of variation across space, and partly because it was no longer needed to justify imperialism and colonization (Mitchell 2000: 19).

3. Region and culture At the end of the nineteenth century, with the “blank spots” on the European and EuroAmerican maps of the Earth mainly filled, intellectuals feared that the twentieth century would bring a new era of conflict. World War I fulfilled some of those fears. The organic relationship between peoples, places and languages was called into question by attempts by nation-states to conquer others, and the brutality of the war made many feel that human life was essentially disorderly. Geographers played important roles during the war, particularly in developing maps for intelligence work, and geography played a larger role in British and American school curricula after the war than it had before. Post-war disillusionment with the idea that the nation-state was the natural condition for global peace led to a new concern with region. One approach, called the “science of regions”, is associated particularly with Richard Hartshorne (1939). This descriptive approach to physical geography continued the nineteenth-century and earlier tradition of cataloguing the characteristics of “undiscovered” places, and it both developed from and created links between geographers and their colleagues in geology and other physical sciences. In response to the failure of environmental determinism, another group of geographers turned to culture to explain regional variation in human behavior. This approach was pioneered by American Carl Sauer in The Morphology of Landscape (Sauer 1925). For Sauer, “culture, working with and on nature, created the contexts of life” (Mitchell 2000: 21). Human life ways are not shaped directly by nature, as environmental determinists claim. Rather, by means of agriculture, architecture and other activities, humans shape places into the landscapes that form the environments for their lives and the object of study for geographers. The activities that

3

4

I. Introduction: Language and space shape landscape are cultural activities, determined by human societies’ ideas about how to use and live in nature. Landscape is thus a manifestation of the culture that produced it; to find out about local culture, geographers read the landscape. So while it is important to describe the processes that create the natural (i. e., pre-human) environment, this is preliminary to the job of exploring cultural landscapes, which should be the primary task of geography. Mitchell (2000: 22⫺23) links Sauer’s work with that of nineteenth-century German linguist Johann von Herder. Herder attributed the unique characteristics of languages not directly to nature but to the expression of the “spirit of a culture” (Mitchell 2000: 23); cultural variation was due not to the physical environment but to the relationships people created with places as they interacted with them. Thus the authentic roots of the nation-state can best be seen in vernacular practices such as folklore, folk music, folk poetry and fairy-tales. The influence of these ideas can be seen clearly in nineteenthcentury European dialectology, particularly in the form of people like the Grimm brothers, who were both folklorists and students of language change. But Sauer and his students were motivated less by nineteenth-century Romantic nostalgia for the authentic, unitary roots of the natural nation than by the sense that Western civilization was destroying what was particular about places and peoples, erasing the differences that would allow geographers and others to trace human history though the distribution of human artifacts and traditions. Influenced by anthropologists Boas and Kroeber, Sauer claimed that the job of geography was to “[chart] the distribution over the earth of the arts and artifacts of man, to learn whence they came and how they spread” (Sauer 1952: 1, quoted in Mitchell 2000: 25). Sauer’s rejection of “causal geography” and his focus on the landscape, an object of study that no other discipline claimed, helped make geography more respectable as a science (Mitchell 2000: 26⫺29). Like his colleagues in anthropology, Sauer went beyond the description of landscape. He attempted to explain the natural and cultural processes that create the meanings of human environments. For Sauer, cultural landscapes are constantly created and recreated, and geographers seek to trace their histories by describing such things as human artifacts, population patterns, building types, systems of production ⫺ and systems of communication. Geography textbooks thus often mention the work of the dialect atlas surveys about regional differences in lexicon as they reflect differences in culture and its interaction with nature. Just as do regional patterns of barn-building (Ensminger 1992), regional patterns of word usage track differences in how people interact with the physical environment. Other regional geographers were less concerned with the processes of divergence and change that underlay regional differences, and their work sometimes consisted of descriptive compendia of observations about regions and their landscapes. This gave rise to one critique of Sauerian regional geography: it was too concerned with material artifacts and too mistrustful of modernity, hence overly focused on the rural, the primitive and the archaic. Similar anti-modernist sentiment can be said to underpin (explicitly or not) the American dialect atlas projects of the twentieth century, which are vulnerable to the same critique. For example, both the Linguistic Atlas of the Middle and South Atlantic States in the 1930s and 1940s and Ensminger’s exploration of Pennsylvania barn types were published as or well after the kinds of regionally based cultural differentiation they describe were beginning to break down due to such influences as radio, motorized vehicles, urbanization and the standardization of agricultural practices and vocabularies.

1. Language and geographical space

4. Superorganicism and social-scientiic geography Another critique of Sauer’s approach had to do with its use of the concept “culture”. Because he was in the end more interested in the effects of culture than with the processes that led to these effects, Sauer relied on a common-sense, uncritical understanding of this idea (Mitchell 2000: 29). This critique was addressed in Wilbur Zelinsky’s (1973) work, first published in the early 1970s. Zelinsky, a student of Sauer’s, attempted to define “culture” more systematically. For Zelinsky, culture could be seen as (1) “an assemblage of learned behavior”, or, more abstractly, (2) “a structured, traditional set of patterns for behavior, a code or template, for ideas and acts”, or, more abstractly yet, (3) a totality which “appears to be a superorganic entity living and changing according to a […] set of internal laws” (Zelinsky 1973: 71). The idea of culture as superorganic, existing “both of and beyond its participating members” (Zelinsky 1973: 40), means that individual behavior does not predict culture or vice versa. To describe culture in this sense is to describe what unifies it, the “internal laws” which operate behind the diversity in behavior that can actually be observed. Zelinsky acknowledged that modernization tends to make individuals freer to decide how and where to live and that “spurious” or “synthetic” regions created by sports fans or public relations professionals, like “the Steeler Nation”, “Chicagoland”, or “Pennsylvania Dutch Country”, can sometimes seem more real than more authentic regions do. However, Zelinsky’s model of culture as superorganic fits best with the “traditional region”, the region into which people are born and in which they live their lives and in which “an intimate symbiotic relationship between man and land develops over many centuries, one that creates indigenous modes of thought and action, a distinctive visible landscape and a form of human ecology specific to the locality” (Zelinsky 1973: 110). Critical geographer Don Mitchell (2002: 30⫺33) suggests that Zelinsky’s view of culture as superorganic was a liberal reaction to the conflict and fragmentation of the 1960s, an attempt to reassert a fundamental unity underlying differences in behavior. Critiques of superorganicism point to how, by reifying culture and making it agentive, the theory positions the individual as an automaton. Critics like James Duncan (1980) ask how culture is constructed through human activity. Echoing many sociolinguists, they wonder how culture actually works in human society and life. With increased state regulation of national economies and the increasing professionalization of business in the post-World War II years, many academic disciplines also “professionalized”, making new connections with government, industry and the military in return for new sources of research funding (Woolgar 1988; Johnston 2003; see Johnston 1997 on geography after 1945). The time was right for the emergence of new ways of working that could produce “scientific” information about human behavior, and disciplines like economics, political science and sociology became more influential as they turned increasingly to quantitative “social-scientific” methods, and to the research questions such methods could best answer. Beginning in the 1950s and 1960s, some geographers began to re-orient their work towards social science. They liked the rigor of social scientific methods and thought that scientific rigor required quantitative methods (Gregory 1962). They also realized that quantitative research would raise geography’s status. The quantitative approaches to geographical inquiry that became dominant during this period were known as “locational analysis” or “spatial science”. They drew on models from economics and other social sciences, including assumptions like the principle of

5

6

I. Introduction: Language and space least effort and that of rational choice ⫺ the assumptions, that is, that people will do the things that require the least effort and best serve their interests. Widely used textbooks that brought such ideas into geography from economics, sociology and the like included Haggett (1965) and Morrill (1970). One influential theory that drew on socialscientific models was “central place theory” (Christaller [1933] 1966; see also Berry 1967). Central place theory assumes that potential consumers moving through space want to minimize their time and effort and business owners want to maximize consumers’ spending and the rate at which the consumer population turns over. The resulting model describes the ideal arrangement in space of smaller and larger central places that function as business and service centers. Sociolinguists have drawn explicitly on models of spatial flow and diffusion from quantitative human geography. (For a useful overview of this work, see Bailey et al. 1993.) The most influential model from the point of view of sociolinguistics has been that of Torsten Hägerstrand ([1953] 1967). Location theory is concerned with the simulation and modeling of processes of change across space, and it has been applied to various patterns of change, including the spread of disease. Hägerstrand’s “gravity model” of hierarchical change was adopted by Trudgill (1974) in his work in Norway and by Callary (1975) in a study of the spread of urban speech forms in Illinois, USA. Bailey et al. (1993) and Horvath and Horvath (2001) have also used elements of location theory to account for “contra-hierarchical” patterns of change and for “space effects” and “place effects”. In Hägerstrand’s view, “innovation spreads in a community through a network of face-to-face interpersonal communication such that the likelihood of adoption at a given site is higher when it is close to a site of previous adoption” (Yapa 1996: 238). The assumption underlying this model is that interaction becomes less frequent as a function of distance. Diffusion can also be blocked by such things as economic and class differences or geographical factors that make face-to-face communication less likely. There are also amplifiers of diffusion, such as tightly knit social networks or population density. In general, according to Rogers (1983), the factors that influence diffusion across space include the phenomenon itself (for a phonological change, this might include whether it is a merger or a split), communicative networks, distance, time and social structure. One effect of the need for face-to-face interaction is that innovations can either move from cities to suburbs to rural areas, or bypass the rural areas near cities, “cascading” to further-away urban centers where city dwellers are more likely to have contacts. Both are types of “hierarchical” diffusion, models of the spread of change that begins in “central places”. The social-scientific, quantitative “revolution” in geography was codified by David Harvey in a 1969 textbook (Harvey 1969). But by the 1970s, Harvey, along with other geographers, had begun to look to Marxist theory for a better understanding of the social inequality that so often serves as a barrier to spatial diffusion (Harvey 1973: 1982). Location theory paid too little attention, it was thought, to social structures and social systems. As Johnston explains it (2003: 62), “By reducing all decision-making to economic criteria, subject to immutable “laws” regarding least-costs, profit-maximization and distance-minimizing, geographers, it was claimed, were ignoring (even denigrating) the role of culture and individuality in human conditioning and behavior”. One of the most influential critics of social-scientific, quantitative geography was Derek Gregory (1985), who pointed out that location theory presupposes that each member of a society

1. Language and geographical space has the same likelihood of adopting change, assuming contact with it. In particular, quantitative geographers’ “laws” were thought to describe, and thus help perpetuate, the capitalist status quo. One important reaction to the critique was L. A. Brown’s (1981) “market and infrastructure perspective”, a method for paying systematic attention to differences in access to resources. Studies taking this more socially conscious perspective came together with studies of the diffusion of modernization from anthropology, sociology and political science. One such model is that of Lakshman Yapa (1996), who proposes a way of talking about innovation based on the idea that what spreads from place to place is not simply information but a “nexus of production relations and biased innovations” (Yapa 1996: 232).

5. The 1960s onwards: Towards new paradigms Just as in many other social-scientific and humanistic disciplines, quantitative approaches began to be questioned during the later 1960s and 1970s. New or newly discovered strands of philosophy and social theory were brought to the table, many with origins in nineteenth and earlier twentieth-century European philosophy and political theory. Geographers skeptical of the static, consensual quality of social-scientific models began to explore Marxist and neo-Marxist social theory, with its focus on power, class struggle and change. Feminist geographers including Gillian Rose (1993) and Suzanne Mackenzie (1989) moved beyond describing how capitalism pins individuals into places along an economic axis, describing “positionality” along multiple social dimensions including that of gender and theorizing identity in terms of performance rather than social determination. Feminists’ attention to the hybridity, mutability and particularity of identity led to a more generally postmodern geography (Gregory 1989; Soja 1989; N. Duncan 1996). Others turned to the discourse-centered philosophy of Michel Foucault, exploring how place and space are socially constructed via a variety of knowledge-building practices. Methodologically, all these approaches suggested that qualitative, particularistic research aimed at describing social process and practices needed to supplement or even replace quantitative work aimed at making generalizations about the outcomes of such processes. In the two sections that follow, I focus on two of the post-1960s approaches to geography that have found direct application in work about language and place: Marxist and neo-Marxist geography and humanistic geography. I then touch on their common concern with the ways in which places are created and their meanings shared through discourse, sketching some current approaches that draw on Foucauldian thought.

5.1. Marxist-inluenced geography In the UK, Marxist and then feminist “cultural studies” (first institutionalized at the University of Birmingham’s Centre for Contemporary Cultural Studies) began to influence geographers in the late 1970s and early 1980s (Jackson 1980; Cosgrove 1983). What Denis Cosgrove called “radical cultural geography” required increased attention to is-

7

8

I. Introduction: Language and space sues such as race, gender, power, dominance (for geographers, often enacted in the control of space), and the production of “social space”. It also encouraged renewed interest in landscape in the Sauerian sense, but with a more direct focus on diversity and conflict in the creation and regimentation of landscape. Edward L. Soja (1989) traces the beginning of “critical human geography” to the work of the French Marxist scholar Henri Lefebvre ([1974] 1991). According to Lefebvre, the production of space is a consequence of modern capitalism. Space is produced, in Lefebvre’s sense, when it is the result of labor; the production of space is contrasted with its creation by nature. When a space is a product, it can be reproduced and it takes on the characteristics of an object. For example, it can be bought and sold. Spaces that are produced may bear traces of how they were produced ⫺ how the original, natural raw material was modified. “One might say”, according to Lefebvre, “that practical activity writes upon nature, albeit in a scrawling hand” (Lefebvre 1991: 117). But such traces tend to fade, and produced spaces eventually tend to become detached from the conditions under which they were produced, including the labor involved. The processes that “naturalize” produced space Lefebvre calls “forgetfulness” or “mystification”. People forget, for example, what residential neighborhoods looked like before they were developed; various practices at home and at school encourage children to think of the neighborhood as the natural environment for life and discourage them from wondering about its history. Eventually, physical traces of the neighborhood’s production out of farmland or forest ⫺ flattened hills or artificial contours, non-native trees and flowers, new roads and driveways ⫺ become invisible, part of a new “natural” background. The study of the history of space, then, looks at how the “spatio-temporal rhythms of nature” are transformed by “social practice” (Lefebvre 1991: 117). Space can be “appropriated” or “dominated” by human activity. The appropriation of space occurs when space is minimally modified to serve the needs of humans. This may result in a structure ⫺ a hut or an igloo, for example ⫺ whose form stays close to what the raw materials dictate. Space is dominated, on the other hand, when it is transformed by modern technology which introduces new forms, which are often rectilinear, closed, emptied and sterilized. A fort or a highway would exemplify the domination of space via technological modes of production. Lefebvre claims that most appropriation of space in the modern era also involves its domination (Lefebvre 1991: 164⫺168). The terms space and place and their derivatives have had many uses in geography. In general, space is seen as the raw material for the construction, by human societies, of landscape, or, more abstractly, place. But geographical theories differ on whether it is possible to speak of space independent of human activity. For those who carry social constructivism to its extreme, everything we observe is seen through one cultural lens or another, so it makes no sense to try to describe pre-cultural spaces. Geographers like Soja (1989) and J. Nicholas Entrikin (1991) come down in the middle, claiming that humans’ space (or, for Soja, “spatiality”) is always interconnected with the “physical space of material nature” and the “mental space of cognition and representation” (Soja 1989: 120). Entrikin similarly speaks of the “betweenness” of place, partly physical, but always also partly the result of social activity. The focus of Marxist geography is, however, on the social process through which human spaces are created, maintained, struggled over, and for Marxists, these processes are massively shaped by economic relations and their historical residue. Sociolinguist David Britain (1997, 2005) draws on the work of Marxist-influenced geographers to add explanatory detail to the more abstract model of social networks

1. Language and geographical space that Lesley Milroy (1987) and others have used in describing processes of language change. Research by geographer John Urry (1985) for example, shows how social relations and spatial structures are connected in different ways depending on the local economic and political system. Nineteenth-century capitalism, in England, led to the foundation of towns near natural resources like coal and iron. These towns attracted laborers, who needed inexpensive housing near where they worked; the resulting row houses led to new social relations among members of the working class, including, perhaps, denser, more multiplex social networks than they experienced in the countryside where they came from. Beginning in the 1980s, critics of “radical” Marxist geography began to claim that Marxist approaches were too deterministic and too dependent on a static model of social structure as shaped by an unchanging set of economic relations between labor and capital. These geographers “reemphasized the need for more sophisticated conceptions of the forces shaping people’s subjective experiences, of the role of these experiences in shaping people’s practices and struggles, and of the significance of subjective knowledge in scientific accounts of social change” (Chouinard 1996: 389⫺390). Doreen Massey’s (1984) “critical realism” is considered a classic formulation of this position. According to Massey, there are general tendencies, in a capitalist economic system, for socio-spatial relations to take certain forms, but the particular processes that lead to these relations are implemented by individuals. As a result, both the situation and the individual actor are changed. In other words, there is a continuous interplay between “structure” and “agency”, terms drawn from sociologist Anthony Giddens (1984), who calls this dialectic process “structuration”. Methodologically, for Massey, this means that a geographer can explain particular locational events but not make general laws. Geographers Allan Pred (1984, 1990) and Nigel Thrift (1991) were particularly influential in bringing structuration theory into geography. Drawing on social theorists like Zigmunt Bauman (1992), who suggest that the mass media are now among the main technologies for the production and distribution of culture, some neo-Marxist geographers have explored how the media function in contests over the meaning of space. Robert D. Sack (1988, cited in Entrikin 1991) shows how advertising can make strategic use of (and help perpetuate) nostalgia for neighborhood, community, or region. Sociolinguists have begun to notice that linguistic forms people think of as local can be used in such advertising, and in other planned attempts to capture the “heritage” or “authentic” aspects of places (Beal 1999), and that these uses may have ramifications for the trajectory of linguistic change. Critics of Marxist-influenced geography included feminists such as Mackenzie (1989) and Rose (1993), who pointed out that humans are positioned not only in an economic system but along many other axes as well, one being gender. They encourage attention to the hybridity and particularity of identity and of people’s relations with space and place. They advocate new methods of interpretation, sometimes less theory-driven ones such as text analysis and the analysis of narrative. For Courtice Rose (1980), doing geography is more like reading than like traditional scientific work. Sociolinguist Greg Myers (2006) draws on Massey’s (1994) feminist work in his discourse analysis of how members of focus groups “formulate place” (Schegloff 1972) as they introduce themselves and support their positions. The turn away from the formulation of general laws about how economic systems shape space and how it is experienced is encapsulated by Margaret Rodman (2003: 208), “It is time to recognize that places, like voices, are local

9

10

I. Introduction: Language and space and multiple. For each inhabitant, a place has a unique reality, one in which meaning is shared with other people and places. The links in these chains of experienced places are forged of culture and history”.

5.2. Humanistic geography Another new approach with origins in the 1960s and 1970s draws on the phenomenological strand of German philosophy. Like Marxist, feminist and other critical approaches, it originated as a reaction to the “old” cultural geography and the abstract law-finding of quantitative social-scientific geography. Humanistic geographers, drawing on phenomenologists such as Heidegger, are concerned with how individuals experience and “inhabit” the world, describing human interactions with the environment from the humans’ perspective, in other words our “senses of place”. Key humanistic geographers include Yi-Fu Tuan (1977), Entrikin (1976), Edward Relph (1981) and Anne Buttimer (1993). According to humanistic geographers, the physical aspects of place are always mediated by subjective experience. We experience a house not as a set of geographical coordinates or a particular arrangement of building materials or furniture, but as a set of smells, sounds, scenes and emotions that are shaped by repeated ways of interacting with houses. We “inhabit” a house, making it our own, by experiencing and/or manipulating it in a variety of ways ⫺ walking through it, touching the walls, looking out the window, turning the water on and off, rearranging the furniture, maybe even writing a poem about it. We experience a region, sometimes, partly through the sounds of a dialect (Mugerauer 1985). Physical places are sometimes designed with particular human experiences in mind, as when people build structures or post signs that make places more distinctive and memorable or make sure a house is well-lit and fresh-smelling before showing it to a potential buyer. Such manipulations have become more obvious in the past few decades as attention to “heritage” and “authenticity” in the marketing of place has led to the increasing commodification of senses of place: shopping malls that incorporate elements of the industries that once occupied their sites, towns designed to feel “like home”, with nostalgic street names like “Cherry Lane”. Technological developments like satellite navigation and online mapping systems have also changed how we experience space. A sense of place comes both from immediate experience and from more abstract modes of knowledge. Spatiality is constructed directly through bodily experience that results from motion and sensation: vision, smell, touch and hearing (Thrift 2000; Urry 2000). Among the indirect sources of knowledge about places are art (for example, landscape painting, travel writing, television documentaries), education (school lessons, guide books, maps and brochures and so on) and politics (debates over urban redevelopment or public transit, for example) (Tuan 1977: 161⫺164). These make places visible and encourage people to see them in common ways, or make them aware that they see them differently.

5.3. Discourse and place Neo-Marxist and humanistic approaches to geography, as well as other strands of postmodern geographical thought, have in common the idea that spaces become human

1. Language and geographical space places partly through discourse. “Discourse” in this sense (Foucault 1972: 1980) refers to talk, writing and other practices involving language, as well as to the ideology that is produced and reinforced through talk. In other words, it is through ways of talking that arise from and evoke particular linked sets of ideas that people come to share or attempt to impose ideas about what places mean and how to behave in them. The “linguistic turn” to which Marxist, Foucauldian and feminist theory gave rise has shifted many scholars’ focus throughout the humanities and social sciences, and geography is no exception. As Bridge and Watson (2000) put it in connection with urban studies, “Cities are not simply material or lived spaces ⫺ they are also spaces of the imagination and spaces of representation” (Bridge and Watson 2000: 7). Among the “technologies” (Benjamin [1936] 1968) or “practices” (de Certeau 1984) that shape how places are offered to the imagination are the mass media, with their particular economic goals and technical constraints, as well as linguistic practices like narrative, which arise in particular social and situational contexts. Students of language and space have begun to draw on humanistic and discursive geography (Johnstone 2004). For example, they explore how storytelling and other genres of discourse can evoke and shape the meanings of places and ways of speaking, encouraging people to experience them the same way and learn the same lessons from them (Johnstone 1990; Modan 2007). Recent work on the “enregistration” of dialects, drawing on the semiotic theory of Michael Silverstein (1993, 2003) and Asif Agha (2003), explores how sets of linguistic forms that are hearable or visible in an area can coalesce, in people’s minds, into “dialects”, and how dialects get linked with cities and regions through practices like newspaper feature-writing (Johnstone, Bhasin and Wittkofski 2002; Johnstone, Andrus and Danielson 2006), the telling of travel stories (Johnstone 2007), and nostalgic online chat (Johnstone and Baumgardt 2004). Interrogating how linguists ourselves construct dialects and places as we talk about them may add a muchneeded layer of reflexivity to our work.

6. New concerns It has been argued that economic and cultural developments have diminished the relevance of place in human lives. The threat to meaningful places in the modern world is often said to be the result of rapid change and mobility. Edward Said (1978: 18), for example, speaks of the “generalized sense of homelessness” experienced by the globally mobile. According to Giddens (1991), the dynamism of modern life has the effect of separating place from space, removing social relations from local contexts via “abstract systems” such as currency, therapy and technology (Giddens 1991: 14⫺21). Once social life becomes “disembedded” in this way, “place becomes phantasmagoric” (Giddens 1991: 146); an individual’s experiential world is no longer the physical world in which he or she moves. The electronic media are often associated with a sort of liberation from place, and new attention to what happens on the borders and at the boundaries and to heterogeneity and adaptiveness calls into question the idea that “cultures” in the traditional sense ever existed (Bhabha 1994; Urciuoli 1995). But local, space-based community may still have a role to play. People sometimes attempt to “re-embed the lifespan within a local milieu” (Giddens 1991: 147), via such

11

12

I. Introduction: Language and space activities as attempts to cultivate community pride. Cultural geographers who focus on traditional cultures and traditional aspects of culture recognize the continued persistence and importance of traditional sources of meaning such as localness (Entrikin 1991: 41). Local contexts of life may be tied to human identity in more immediate ways, too. As Stuart Hall points out (1991: 33⫺36), “the return to the local is often a response to globalization”. Face-to-face community is knowable in a way more abstract communities are not. Anthropologist Ulf Hannerz (1996: 26⫺27) proposes that the local may still be an important source of continuity, because “everyday life” is local. People’s earliest experiences usually take place in a local context, and local encounters tend to be faceto-face and long-term. Furthermore, the local is sensually immediate and immersive in a way that more distanced forms of experience are not. Thus the principal vehicles for the production and transmission of culture may still be local ones. Geographers come down on both sides of this debate about whether place still matters. For Castells (1966), economic, political and cultural globalization mean that “flows” are replacing places and geography is in danger of losing its traditional object of inquiry. Castree (2003) argues, to the contrary, that the role of place in human life has changed, but not vanished. To be sure, geographers can no longer think of the world as a mosaic of bounded places, each of which can be studied on its own. For contemporary human geographers, places are both unique ⫺ the result of particular, small-scale interactions and experiences ⫺ and the same ⫺ shaped by the same large-scale global forces. The identities of people and places are thus “glocal” ⫺ both global and local. Global forces play out differently in different sets of local circumstances, because particular local circumstances constrain how such forces are experienced and dealt with. Sociolinguists are likewise noticing how the larger-scale leveling effects of language and dialect contact can be counteracted by particular regional loyalties and patterns of interaction, media consumption and such, which can lead to the preservation of variant forms and the development of new differences. Like sociolinguists, geographers are now being reminded of the importance of scale, of thinking about change not as a singular process but as a series of incremental shifts in patterns that emerge at different grain sizes (Herod 2003). Like sociolinguists, geographers are thinking about identity, power and resistance (Katz 2003) and about the technology that standardizes space as well as the local ecological systems that such technology can disrupt (Simmons 2003).

7. Conclusion In choosing to organize this chapter chronologically, I mean to highlight the variety of ways in which “geography” and “space” have been imagined over time. I do not mean to suggest that subsequent approaches have always displaced earlier ones or that chronology necessarily represents progress. Geographers today do many different kinds of work; different research questions call for different conceptions of space and its role in human life and different methods for answering these questions. The same is true of linguists exploring language and space. We can learn from several strands of geography, and I hope this chapter has provided a framework for doing so.

1. Language and geographical space

8. Reerences Agha, Asif 2003 The social life of a cultural value. Language & Communication 23: 231⫺273. Bailey, Guy, Tom Wikle, Jan Tillery and Lori Sand 1993 Some patterns of linguistic diffusion. Language Variation and Change 3: 359⫺390. Bauman, Zygmunt 1992 Intimations of Postmodernity. London/New York: Routledge. Beal, Joan C. 1999 “Geordie Nation”: Language and regional identity in the north-east of England. Lore and Language 17, 33⫺48. Bell, Morag, Robin Butlin and Michael Heffernan (eds.) 1994 Geography and Imperialism, 1820⫺1940. Manchester, UK: Manchester University Press. Benjamin, Walter 1968 [1936] The work of art in the age of mechanical reproduction. In: Hannah Arendt (ed.), Illuminations: Essays and Reflections, 217⫺251. New York: Schocken Books. [First published in Zeitschrift für Sozialforschung 5: 40⫺63] Berry, Brian Joe Lobley 1967 The Geography of Market Centers and Retail Distribution. Englewood Cliffs, NJ: Prentice-Hall. Bhabha, Homi K. 1994 The Location of Culture. London: Routledge. Bridge, Gary and Sophie Watson 2000 City imaginaries. In: Gary Bridge and Sophie Watson (eds.), A Companion to the City, 7⫺17. Oxford: Blackwell. Britain, David 1997 Dialect contact and phonological reallocation: “Canadian raising” in the English Fens. Language in Society 26, 15⫺46. Britain, David 2005 Innovation diffusion, “Estuary English” and local dialect differentiation: The survival of Fenland Englishes. Linguistics 43: 995⫺1022. Brown, Lawrence A. 1981 Innovation Diffusion: A New Perspective. New York: Methuen. Buttimer, Anne 1993 Geography and the Human Spirit. Baltimore: Johns Hopkins University Press. Callary, Robert 1975 Phonological change and the development of an urban dialect in Illinois. Language in Society 49: 155⫺70. Castells, Manuel 1966 The Rise of the Network Society. Oxford: Blackwell. Castree, Noel 2003 Place: Connections and boundaries in an interdependent world. In: Sarah L. Holloway, Stephen P. Price and Gill Valentine (eds.), Key Concepts in Geography, 165⫺185. London: Sage. Chouinard, Vera 1996 Structure and agency: Contested concepts in human geography. In: Carville Earle, Kent Mathewson and Martin S. Kenzer (eds.), Concepts in Human Geography, 383⫺410. Lanham, MD: Rowman & Littlefield. Christaller, Walter 1966 [1933] Central Places in Southern Germany. Translated by Carlisle W. Baskin. Englewood Cliffs, NJ: Prentice-Hall. [First published: Jena, Fischer].

13

14

I. Introduction: Language and space Cosgrove, Denis 1983 Towards a radical cultural geography. Antipode 15(4): 1⫺11. de Certeau, Michel 1984 The Practice of Everyday Life. Berkeley: University of California Press. Dunbar, Gary. S. (ed.) 2002 Geography: Discipline, Profession, and Subject Since 1870. Amsterdam: Kluwer. Duncan, James 1980 The superorganic in American cultural geography. Annals of the Association of American Geographers 70: 181⫺198. Duncan, Nancy 1996 Postmodernism in human geography. In: Carville Earle, Kent Mathewson and Martin S. Kenzer (eds.), Concepts in Human Geography, 429⫺458. Lanham, MD: Rowman & Littlefield. Ensminger, Robert F. 1992 The Pennsylvania Barn, Its Origin, Evolution, and Distribution in North America. Baltimore: Johns Hopkins University Press. Entrikin, J. Nicholas 1976 Contemporary humanism in geography. Annals of the Association of American Geographers 66: 615⫺632. Entrikin, J. Nicholas 1991 The Betweenness of Place: Towards a Geography of Modernity. Baltimore: Johns Hopkins University Press. Foucault, Michel 1972 The Archaeology of Knowledge and The Discourse on Language. Edited by Alan M. Sheridan Smith. New York: Harper. Foucault, Michel 1980 Power/Knowledge: Selected Interviews and Other Writings, 1972⫺1977. Edited by Colin Gordon. New York: Pantheon. Giddens, Anthony 1984 The Constitution of Society. Cambridge, UK: Polity. Giddens, Anthony 1991 Modernity and Self-identity. Cambridge, UK: Polity. Godlewska, Anne and Neil Smith (eds.) 1994 Geography and Empire. Oxford: Blackwell. Gregory, Derek 1985 Suspended animation: The stasis of diffusion theory. In: Derek Gregory and John Urry (eds.), Social Relations and Spatial Structures, 269⫺336. London: Macmillan. Gregory, Derek 1989 The crisis of modernity? Human geography and critical social theory. In: Richard Peet and Nigel Thrift (eds.), New Models of Geography, Volume 2, 348⫺385. London: Unwin Hyman. Gregory, Stanley 1962 Statistical Methods and the Geographer. London: Longman. Hägerstrand, Torsten 1967 [1953] Innovation Diffusion as a Spatial Process. Translated by Allan Pred. Chicago: University of Chicago Press. [First published: Lund: Gleerup] Haggett, Peter 1965 Locational Analysis in Human Geography. London: Edward Arnold. Hall, Stuart 1991 The local and the global: Globalization and ethnicity. In: Anthony D. King (ed.), Culture, Globalization, and the World-system, 19⫺39. Basingstoke, UK: MacMillan.

1. Language and geographical space Hannerz, Ulf 1996 Transnational Connections: Culture, People, Places. London: Routledge. Hartshorne, Richard 1939 The Nature of Geography: A Critical Survey of Current Thought in Light of the Past. Lancaster, PA: Association of American Geographers. Harvey, David 1969 Explanation in Geography. London: Edward Arnold. Harvey, David 1973 Social Justice and the City. Baltimore: Johns Hopkins University Press. Harvey, David 1982 The Limits to Capital. Chicago: The University of Chicago Press. Heffernan, Mike 2003 Histories of geography. In: Sarah L. Holloway, Stephen P. Price and Gill Valentine (eds.), Key Concepts in Geography, 3⫺22. London: Sage. Herod, Andrew 2003 Scale: The local and the global. In: Sarah L. Holloway, Stephen P. Price and Gill Valentine (eds.), Key Concepts in Geography, 229⫺247. London: Sage. Horvath, Barbara M. and Ronald J. Horvath 2001 A multilocality study of a sound change in progress: The case of /l/ vocalization in New Zealand and Australian English. Language Variation and Change 13: 37⫺57. Jackson, Peter 1980 A plea for cultural geography. Area 12: 110⫺113. Johnston, Ron 1997 Geography and Geographers: Anglo-American Human Geography since 1945. London: Arnold. Johnston, Ron 2003 Geography and the social science tradition. In: Sarah L. Holloway, Stephen P. Price and Gill Valentine (eds.), Key Concepts in Geography, 51⫺71. London: Sage Publications. Johnstone, Barbara 1990 Stories, Community, and Place: Narratives from Middle America. Bloomington: Indiana University Press. Johnstone, Barbara 2004 Place, globalization, and linguistic variation. In: Carmen Fought (ed.), Sociolinguistic Variation: Critical Reflections, 65⫺83. New York: Oxford University Press. Johnstone, Barbara 2007 A new role for narrative in variationist sociolinguistics. In: Michael Bamberg (ed.), Narrative: State of the Art, 57⫺67. Amsterdam/Philadelphia: John Benjamins. Johnstone, Barbara, Jennifer Andrus and Andrew Danielson 2006 Mobility, indexicality, and the enregisterment of “Pittsburghese”. Journal of English Linguistics 34: 77⫺104. Johnstone, Barbara and Dan Baumgardt 2004 “Pittsburghese” online: Vernacular norming in conversation. American Speech 79: 115⫺ 145. Johnstone, Barbara, Neeta Bhasin and Denise Wittkofski 2002 “Dahntahn Pittsburgh”: Monophthongal /aw/ and representations of localness in southwestern Pennsylvania. American Speech 77: 148⫺166. Katz, Cindi 2003 Social formations: Thinking about society, identity, power and resistance. In: Sarah L. Holloway, Stephen P. Price and Gill Valentine (eds.), Key Concepts in Geography, 249⫺ 265. London: Sage. Lefebvre, Henri 1991 [1974] The Production of Space. Translated by Donald Nicholson-Smith. Malden, MA: Blackwell. [First published: Paris: Editions Anthropos]

15

16

I. Introduction: Language and space Livingstone, David N. 1992 The Geographical Tradition: Episodes in the History of a Contested Enterprise. Oxford: Blackwell. Mackenzie, Suzanne 1989 Restructuring the relations of work and life, Women as environmental actors, feminism as geographical analysis. In: Audrey Kobayashi and Suzanne Mackenzie (eds.), Remaking Human Geography, 40⫺61. London: Unwin Hyman. Massey, Doreen 1984 Spatial Divisions of Labor: Social Structures and the Geography of Production. London: Methuen. Massey, Doreen 1994 Place, Space, and Gender. Cambridge, UK: Polity Press. McDavid, Raven Ioor (ed.) 1980 Linguistic Atlas of the Middle and South Atlantic states. Chicago: University of Chicago Press. Mill, Hugh Robert 1901 On research in geographical science. Geographical Journal 18: 407⫺424. Milroy, Lesley 1987 Observing and Analyzing Natural Language: A Critical Account of Sociolinguistic Method. Oxford: Basil Blackwell. Mitchell, Don 2000 Cultural Geography: A Critical Introduction. Malden, MA: Blackwell. Modan, Gabrielle 2007 Turf Wars, Discourse, Diversity, and the Politics of Place. Malden, MA: Blackwell. Morrill, R. L. 1970 The Spatial Organization of Society. Belmont, CA: Wadsworth. Mugerauer, Robert 1985 Language and the emergence of environment. In: David Seamon and Robert Mugerauer (eds.), Dwelling, Place and Environment: Towards a Phenomenology of Person and World, 51⫺70. Dordrecht: Martinus Nijhoff. Myers, Greg 2006 Identifying place: Where are you from? Journal of Sociolinguistics 10: 320⫺343. Pred, Allan 1984 Place as historically contingent process: Structuration and the time-geography of becoming places. Annals of the Association of American Geographers 74: 279⫺297. Pred, Allan 1990 Making Histories and Constructing Human Geographies: The Local Transformation of Practice, Power Relations, and Consciousness. Boulder, CO: Westview. Relph, Edward 1981 Rational Landscapes and Humanistic Geography. London: Croon Helm. Robins, Robert H. 1979 A Short History of Linguistics, 2nd ed. London: Longman. Rodman, Margaret C. 2003 Empowering place: Multilocality and multivocality. In: Setha M. Low and Denise Lawrence-Zu´n˜iga (eds.), The Anthropology of Space and Place: Locating Culture, 204⫺223. Malden, MA: Blackwell. Rogers, Everett M. 1983 Diffusion of Innovations. 3rd ed. New York: Free Press. Rose, Courtice 1980 Human geography as text interpretation. In: Anne Buttimer and David Seamons (eds.), The Human Experience of Space and Place, 123⫺134. New York: St. Martin’s Press.

1. Language and geographical space Rose, Gillian 1993 Feminism and Geography: The Limits of Geographical Knowledge. Minneapolis: University of Minnesota Press. Sack, Robert D. 1988 The consumer’s world: Place as context. Annals of the Association of American Geographers 78: 642⫺664. Said, Edward 1978 Orientalism. New York: Vintage Books. Sauer, Carl 1925 The morphology of landscape. University of California Publications in Geography 2: 19⫺54. Reprinted in, Leighly, John (ed.) 1963 Land and life, A selection of the writings of Carl Sauer. Berkeley: University of California Press. Sauer, Carl 1952 Agricultural Origins and Dispersals. New York: American Geographical Society. Schegloff, Emmanuel A. 1972 Notes on a conversational practice: Formulating place. In: David Sudnow (ed.), Studies in Social Interaction, 75⫺119. New York: The Free Press. Semple, Ellen Churchill 1911 Influences of Geographic Environment on the Basis of Ratzel’s System of Anthropo-Geography. New York: Russell and Russell. Silverstein, Michael 1993 Metapragmatic discourse and metapragmatic function. In: John A. Lucy (ed.), Reflexive Language, 33⫺58. Cambridge: Cambridge University Press. Silverstein, Michael 2003 Indexical order and the dialectics of sociolinguistic life. Language and Communication 23: 193⫺229. Simmons, Ian G. 2003 Landscape and environment: Natural resources and social development. In: Sarah L. Holloway, Stephen P. Price and Gill Valentine (eds.), Key Concepts in Geography, 305⫺ 317. London: Sage. Soja, Edward W. 1989 Postmodern Geographies: The Reassertion of Space in Critical Social Theory. New York: Verso. Thrift, Nigel J. 1991 For a new regional geography. Progress in Human Geography 15: 456⫺465. Thrift, Nigel J. 2000 With child to see any strange thing: Everyday life in the city. In: Gary Bridge and Sophie Watson (eds.), A Companion to the City, 398⫺409. Oxford: Blackwell. Trudgill, Peter 1974 Linguistic change and diffusion: Description and explanation in sociolinguistic dialect geography. Language in Society 3: 215⫺246. Tuan, Yi-Fu 1977 Space and Place: The Perspective of Experience. Minneapolis: University of Minnesota Press. Urciuoli, Bonnie 1995 Language and borders. Annual Review of Anthropology 24: 525⫺546. Urry, John 1985 Social relations, space and time. In: Derek Gregory and John Urry (eds.), Social Relations and Spatial Structures, 20⫺48. London: Macmillan.

17

18

I. Introduction: Language and space Urry, John 2000 City life and the senses. In: Gary Bridge and Sophie Watson (eds.), A Companion to the City, 388⫺397. Oxford: Blackwell. Woolgar, Steve 1988 Science: The Very Idea. London/New York: Tavistock Publications. Yapa, Lakshman 1996 Innovation diffusion and paradigms of development. In: Carville Earle, Kent Mathewson and Martin S. Kenzer (eds.), Concepts in Human Geography, 231⫺270. Lanham, MD: Rowman & Littlefield. Zelinsky, Wilbur 1973 The Cultural Geography of the United States. Englewood Cliffs, NJ: Prentice-Hall.

Barbara Johnstone, Pittsburgh (USA)

2. Language and social spaces 1. 2. 3. 4. 5. 6.

Introduction The conventionalization of linguistic practices: The linguistic market The dialect user versus the standard user ⫺ a matter of the polarization of values Tradition ⫺ modernity: A nexus between geographical and socio-cultural space Concluding remarks References

1. Introduction Language is not only an instrument of communication or even of knowledge, but also an instrument of power. A person speaks not only to be understood but also to be believed, obeyed, respected, distinguished. (Bourdieu 1977: 648) The main topic of this article is the indexicality of language in a social and cultural space. Language functions as a sign system with reference to significant sociocultural features and qualities. Hence, the use of a specific variety, or any given variety, in social interaction will serve as an indicator of the interlocutors’ social standing and identity. The relationship between dialect on the one hand and standard spoken language (henceforth called the standard for short) on the other, is one of the core contrasts in modern linguistics. These two opposites have constituted a core linguistic axis in sociolinguistic research from the 1960s up until today, and a substantial part of the present discussion in sociolinguistics revolves around this dichotomy. At the same time, the contrast between dialect and standard may serve to illustrate some of the complex ways in which we experience social space ⫺ in contrast to geographical space. As will be discussed in close detail below, the distinction between dialect and standard was originally

2. Language and social spaces a geographical one, but in many contexts it is interpreted primarily as an expression of social identity. The use of dialect, or of certain dialect markers, is often conceived of as an indicator of a specific social position rather than as an indication of geographical origin. In this respect, the sociocultural references of language cannot be seen as detached from or independent of the geographical dimension (cf. Johnstone in this volume). In other words, the territoriality of language is intimately connected and subtly interwoven with other aspects of language which are often regarded as expressions of certain social features and characteristics. This article covers three general topics, which successively support each other and which partly overlap. The first theme addressed is the conventionalization of linguistic practices, and this section introduces a key term and metaphor ⫺ the linguistic market. Next follows a description of some of the most prominent values associated with two of the “products” on this market ⫺ dialect and standard, respectively. A crucial point in this context is how these contrasting social values can ultimately be related to the dichotomy between tradition on the one hand and modernity on the other. The third and final section is a discussion of how these sociohistorical and sociocultural dimensions can be used to elucidate and explain the dialect leveling we have witnessed over the last decades ⫺ a development which primarily involves a shift from the traditional dialects towards more standardized spoken varieties.

2. The conventionalization o linguistic practices: The linguistic market Discourse is a symbolic asset which can receive different values depending on the market on which it is offered. Linguistic competence (like any other cultural competence) functions as linguistic capital in relationship with a certain market. (Bourdieu 1977: 651) The role of language as a sociohistorical phenomenon serves as the theoretical basis of the present discussion. A given language community, with specific, given linguistic practices, is a historical product of a complex set of social, cultural and political conditions. In this regard, the linguistic choices made by the individual language users can be treated on a par with a number of other social practices in any given society, and consequently language practices can be subjected to the same analysis as other social phenomena, such as lifestyles, leisure activities, general style preferences, etc. However, the various linguistic expressions cannot be regarded exclusively as a reflection of society’s social structures; language in itself is a constituting factor in the social stratification. This juxtaposition of language and other social forms of expression is among the core ideas in Pierre Bourdieu’s sociological theories of culture (see, e.g., Bourdieu 1977, 1994, 1995; for a more comprehensive account of Bourdieu’s ideas about language, see Bourdieu 1991). The present exposition, and also the terminology employed, is very much inspired by Pierre Bourdieu and his theoretical works. Bourdieu’s notion of conventionalized social practice constitutes a focal part of the present theoretical approach. We all take part in various communities of practice ⫺

19

20

I. Introduction: Language and space communities with unique discourse traditions. To a community of practice the repetition of certain patterns is a vital factor in the establishment and maintenance of structures, routines, and norms. Linell describes this dialogistic shaping of linguistic practice as follows (1998: 59⫺60): [L]inguistic structures, cultural routines, norms etc. do exist prior to interactions (but only in and through the interactants’ being acquainted with them). At the same time, however, these structures, routines and norms are interactionally generated, traded down and reconstructed. That is, they exist prior to individual interactions, yet would not exist without a living historical continuity of interactions. Social structures are (re)created, tried out, tested, negotiated and modified every time they are instantiated or drawn upon. Habit is modified by accommodation, while accommodation is enabled and constrained by habit.

As members of a community we share preferred social representations, stereotypes, and discursive practices; these, in turn, become the backdrop against which the individual experiences and perceives the world and which ultimately influence her actions in this world. Put differently, every person is an actor who interprets and acts ⫺ both at the same time. In this context Bourdieu’s concept of “habitus” (cf. Thompson 1991: 12) is of central heuristic value. It is difficult to pinpoint the exact meaning of the term as Bourdieu uses it, and it will suffice to say that it is related to our collectively acquired and habitual patterns of cognition, including evaluations, social preferences and concrete actions ⫺ as these are acquired in the process of socialization: Endowed with this habitus, which consists in universal schemas of perception and cognition, principles of interpretation and evaluations, we live our lives. We ascribe meaning to what we experience and we act “adequately”. Our habitus generates a form of structurally adapted “practical knowledge” which works perfectly in all aspects of everyday life. But we are not alone in this. Since we share the habitus with other people in the same class and class fraction, we also become class members and social agents. Our individual actions become parts of the social practices of classes and class fractions. This constitutes the basis for the social process of recreation. (Rosenlund 1991: 28; my translation)

Consequently, we may say that habitus consists of a set of dispositions that make us inclined both to “[…] act and react in certain ways” (Thompson 1991: 12). These dispositions steer our actions and preferences in a certain direction; however, they do not predetermine them, “It gives them [i. e. individuals] a ‘feel for the game’, a sense of what is appropriate in the circumstances and what is not, a ‘practical sense’ (le sens pratique)” (Thompson 1991: 13). The practical implication is that through socialization, the various dispositions become internalized in each individual, at the same time as the internalized dispositions constitute the premise for the individual’s externalized actions. Through these largely conventionalized practices, society is “constructed”, and can be perceived as an objective reality that exists more or less independently of the people who “construct” it. In other words, society is created by human beings, at the same time as human beings are created by society; cf. Berger and Luckmann’s (1967: 61) statement that “[…] society is a human product. Society is an objective reality. Man is a social product”. It is within these theoretical frames that all forms of linguistic interaction take place ⫺ and are conventionalized. Certain verbal conventions are often preferred in the context of particular social activities, or social fields, and as such these conventions are

2. Language and social spaces intimately connected with the social categories that are associated with or dominate the various fields. Through these processes the various codes are assigned a symbolic value on the linguistic market: Linguistic utterances or expressions are always produced in particular contexts or markets, and the properties of these markets endow linguistic products with a certain ‘value’. On a given linguistic market, some products are valued more highly than others; and part of the practical competence of speakers is to know how, and to be able, to produce expressions which are highly valued on the markets concerned. […] For different speakers possess different quantities of ‘linguistic capital’ ⫺ that is, the capacity to produce expressions a` propos, for a particular market. Moreover, the distribution of linguistic capital is related in specific ways to the distribution of other forms of capital (economic capital, cultural capital, etc.) which define the location of an individual within the social space. (Thompson 1991: 18)

In other words, a condition of hegemony develops in a particular market, regardless of whether it is a local or global market, whereby one variety attains a certain social authority, legitimacy and prestige, while other varieties are situated lower in the hierarchy. Naturally, there will be an intimate connection between this language hierarchy and the historical and sociopolitical structures of the society in question. Recognition of the existence of linguistic hierarchies, together with knowledge about the social values attached to the different codes, is usually part of the sociolinguistic competence of the actors that operate in a linguistic market. In other words, such knowledge is part of the actors’ habitus and contributes to sustaining a relative consensus within a given social community about whether the use of certain language codes is a symbolic asset, with the prospect of future social gain, or alternatively a social disadvantage. Over time, the values and connotations associated with different varieties often come to appear “natural”, i. e. obvious and incontestable. In other words, they become part of the doxa of a speech community, that is, part of the cultural ideas that are taken more or less for granted and thus normally not subject to reflection or critical discussion. Conceptions concerning the relationship between dialect and standard are examples of a doxic field. The concept “linguistic market” should primarily be understood as a metaphor for forces and mechanisms that apply within a socially defined community. The key issue is how the social interaction that takes place in the metaphorical market can be described by the use of economic terminology ⫺ a form of analogy typical in Bourdieu’s theories. Following this train of thought, all language use will be part of the “economy of everyday living”, in which language ⫺ on a par with the goods on offer in other markets ⫺ is continuously assessed in terms of values, investments, profits and capital. One particularly important dimension of the conditions of linguistic production is what Bourdieu refers to as “reception envisaged” and “self-censorship”: In reality, the conditions of reception envisaged are part of the conditions of production, and anticipation of the sanctions of the market helps to determine the production of the discourse. This anticipation, which bears no resemblance to a conscious calculation, is an aspect of the linguistic habitus which, being the product of a prolonged and primordial relation to the laws of a certain market, tends to function as a practical sense of the acceptability and the probable value of one’s own linguistic productions and those of others on different markets. […] In the case of symbolic production, the constraint exercised by the market via the anticipation of possible profit naturally takes the form of an anticipated

21

22

I. Introduction: Language and space censorship, of a self-censorship which determines not only the manner of saying, that is, the choice of language […] or the ‘level’ of language, but also what it will be possible or not possible to say. (Bourdieu 1991: 76⫺77)

In other words, the mechanisms operating on the linguistic market presuppose a certain consensus about the linguistic “economy” and the symbolic dominance that some varieties have over other varieties on the linguistic market, and these expectations about the sanctions that accompany certain language choices are a critical factor in the overall picture of the conditions of linguistic production. On a more or less conscious but often automatic level, the individual language user will make “calculations” based on those experiences that constitute the habitus and that have implications for the values and acceptability of the various codes in the market. The symbolic relations of power and dominance that are outlined here are intimately connected to the historical, political and sociocultural conditions of society. Therefore, it must be stressed that these structures and mechanisms are by no means static or even necessarily the same across different societies. On the contrary, what we are dealing with are values, attitudes and patterns of action that can be negotiated and renegotiated at any point in time. This implies that the linguistic market is not invariable and constant; rather, it is dynamic and changeable, depending on varying sociopolitical conditions (see, e.g., Kerswill 2007 for a discussion on language and social class, and Wolfram 2007 on ethnic varieties).

3. The dialect user versus the standard user  a matter o the polarization o values ‘The vulgar and the refined’, ‘the particular and the general’, ‘the corrupt and the pure’, ‘the barbaric and the civilized’, ‘the primitive and the arbitrary’ were socially pervasive terms that divided sensibility and culture according to linguistic categories. The baser forms of language were said to reveal the inability of the speaker to transcend the concerns of the present, an interest in material objects, and the dominance of the passions. Those who spoke the refined language were allegedly rational, moral, civilized, and capable of abstract thinking. (Smith 1984: 3) “Only before God and linguists are all languages equal”, states a sociolinguistic adage. This is particularly valid for the relationship between dialect and standard, a relation that is colored by asymmetry, incongruence, and hierarchical structures. That is to say, the contrast is typical for how the dialect user and the standard language users are perceived ⫺ in different linguistic market places. Before examining what these conceptions consist of, and exploring their basis more closely, we need a basic definition of the terms “dialect” and “standard variety”. It is important to note that we are dealing with these concepts as idealized concepts, as two extremes that primarily exist at the level of ideal norms. The actual development of

2. Language and social spaces spoken language over the last decades in fact demonstrates variable forms of interference between dialects and a standard variety. The relationship between the two is first and foremost a matter of degree and extent (see, e.g., Auer and Hinskens 1996 and Berruto in this volume). In this context, the term “dialect” will be used in the traditional meaning of “geolect” ⫺ that is, as a variety that has a specific geographical basis. A dialect is thus to be understood as a language system that is found within local, regional or otherwise defined territorial boundaries ⫺ a language system which is unique to a certain geographical area within a national state. A “standard variety”, on the other hand, is a variety that serves as a norm or ideal standard for a larger speech community, usually a national state, and which is often codified ⫺ that is, given a standard definition in grammar books and/or dictionaries. One crucial premise for assigning the role of supraregional ⫺ that is, supra-local ⫺ role to a speech variety is that the development of a spoken variety is accompanied by the development of a written standard for this variety. The difference referred to here in terms of the geographical affiliation of the two varieties is, however, often understood as a social and cultural difference. The very essence of the relation that conventionally exists between dialect and standard is that the latter ⫺ compared to other varieties ⫺ is the most prestigious variety at the national level, and it is the language that is most likely to be associated with social elites ⫺ groups or persons with higher levels of education, authority and status. Consequently, the standard language is often associated with domains of society that are connected to authority ⫺ for example within universities, the media or the public scene in general. Before we continue to investigate which values are usually ascribed to dialect and standard language in more detail, let us first look briefly at the historical conditions for the emergence of the social asymmetry between the two. Standard speech is based on a national written standard. In addition, it usually originates historically from a regional vernacular: Naturally the spoken national standard is not a mere copy of the written language. One of the main requirements is a geographical concentration of people from the higher classes of society in administrative and/or ideological centers. With the introduction of the monetary economy, these centers became our capitals. (Teleman 1979: 43; my translation)

The continuous dialectic between the written standard and its spoken variant in politically and culturally dominant regions, often in and surrounding the capital, has made it possible for varieties that were initially considered dialects to attain the status of spoken standard language. The key issue is the process by which certain regional language features are “decoded” and attain a geographically unmarked, but socially marked, value. In other words, the emergence and codification of a standard spoken language is historically a sociopolitical process: “Successful standardisation involves the creation (or acceptance) of a variety as the most prestigious one, on account of its use by those who have status and power in the society” (Mesthrie et al. 2000: 21). The fact that one particular variety is “chosen” also necessarily implies that this variety stands in a hegemonic relationship to all other varieties. This hierarchy of prestige is subsequently confirmed and enhanced through different social conventions; the hierarchy becomes part of the habitus of the language users and appears as incontestable and “natural”:

23

24

I. Introduction: Language and space If the dominant group expects to remain dominant, then its members will assume, without even thinking about it, that subordinate-group members will make any necessary linguistic adjustments. As a result, they will speak their own prestige language or dialect in the normal fashion without attempting any convergence. (Fasold 1984: 190)

Accordingly, the relationship between speakers of a standard variety and speakers of a regional variety will be regarded in several contexts as an asymmetric social relation. However, the exact nature of this contrast between the standard variety and a dialect will vary a great deal across different national communities, much as different dialects within one national community may signal highly deviating values. Still, there are good reasons to assume that several of the basic aspects of the dialect⫺standard opposition are to a large degree universal and shared by most language communities. Historically, all processes of standardization are motivated by certain selection mechanisms, which in turn are the outcome of certain sociopolitical conditions. And these basic structures undoubtedly give rise to some premises that give the dialect⫺standard relation a universal symbolic content, which is contingent to only a minor degree on the individual language society (see, e.g., Joseph 1987; Milroy and Milroy 1985; Milroy 2007; Mugglestone 1997). How can we describe the more concrete sociocultural values that dialects and standard varieties, respectively, represent? One model that attempts to illustrate this contrast has been developed by Maurand (1981, see below). He provides a schematic overview of the distribution of standard French on the one hand, and Occitan on the other. In this context it must be noted that the sociolinguistic stratification in France is known to be extremely hierarchical and rigid, and thus the contrasts that appear in this model are likely to be more prominent than in other language communities. Nevertheless, as a model, which gives a schematic, idealized image of reality, it reflects some fundamental connotations which can be connected more or less universally to the dichotomy dialect⫺ standard (see Figure 2.1, next page). Occitan is primarily associated with the rural ⫺ with primary industries and smallscale communities, and it is often seen as old-fashioned and associated with older people ⫺ primarily elderly men. Standard French, on the other hand, is associated with urban culture, higher education and public, modern society, and it also signals values connected to youth and femininity. One indication of the currency of these sets of values can be found in what was for a long time the dominant methodological approach in dialectology in the western world: Perhaps the most typical feature shared by all of the major projects in dialect geography is the type of informant selected. No matter how diverse the culture, how discrepant the socioeconomic climate, and how varied the topography, the majority of informants has in all cases consisted of nonmobile, older, rural males [… i. e.] NORMs […]. (Chambers and Trudgill 1980: 33)

A slightly modified version of the ideas presented in Maurand (1981) is presented as Table 2.1, next page (following Mæhlum 2007: 57). In this approach, the geographical origin of the two codes is the primary concern (cf. the first point), followed by a set of contrastive social and cultural values. Again, it is important to note that we are dealing with idealized and stereotypical values ⫺ for example, the model does not explicitly

2. Language and social spaces

25

Fig. 2.1: Maurand’s model (1981), English version adopted from Grillo (1989: 80)

Tab. 2.1. Dialect

Standard

specific geographic origin stigmatized; not prestigious informal, plain, everyday rural, remote traditional

supra-regional; geographically neutral prestigious formal, official urban, central modern

show that there are actually a number of different urban dialects (cf. especially the penultimate point). The table gives an overview of the “market value” which on fairly general and somewhat simplified grounds may be ascribed to dialect and standard, respectively. The strength of conventionalized representations is first and foremost the sociolinguistic patterns that emerge from them. These patterns seem to apply more or less independently of speech communities and are especially relevant in relation to questions like who do we traditionally assume to be dialect users, and who do we generally assume uses a standard variety? And in which contexts is it generally regarded as most appropriate to use dialect or standard, respectively? Indeed there is great variation across national speech communities (see below); nevertheless, the tendencies are so pervasive and compelling that they should be considered as general. The extent of dialect use tends to be

26

I. Introduction: Language and space inversely proportional to the social status and prestige of the language user. Consequently, the higher the position in the social hierarchy, the more powerful the social expectations, and hence also the probability, that one uses a standard variety. The second basic principle, which follows almost as a consequence of the first, relates to contextual variation: The more formal or official the situation, the stronger the expectation that one uses a standard variety. A selection of examples from the Norwegian speech community can give us an indication of how strong and above all how internalized such expectations are. Sociolinguistically speaking, dialects are known to have an exceptionally high status in Norway. The use of dialect is not confined to local or private spheres but extends to a number of different public arenas (see, e.g., Wiggen 1995; Jahr and Janicki 1995). However, a study of attitudes towards different Norwegian dialects in the capital Oslo (Strømsodd 1979) shows that the use of dialect was generally far more acceptable among typical low status workers than among people in more prestigious professions. One of the arguments used to support this line of reasoning was that, for example, it would be necessary for a lawyer “[…] to use the standard variety since it is not possible to accommodate dialects to judicial language and legal terminology […] and to theoretical subjects that demand a high level of abstraction” (Strømsodd 1979: 86; my translation). The notion that dialects are “best suited” for people with a lower social status is in fact justified by the argument that the extent of dialect use stands in an inverse relationship to the degree of theoretical abstraction (see also the quotation from Smith 1984 at the beginning of this section). The subjects of Strømsodd’s study were largely users of a variety which is close to what is considered the standard variety in Norway. Similar attitudes ⫺ or acknowledgement of such common attitudes ⫺ are found among people who are socialized into a different dialect community. The following quotation is from an autobiographical essay by the Norwegian author Karin Sveen (2001) in which she talks about moving to the capital to embark on university studies. Sveen’s linguistic background is a dialect which is very much associated with the rural and rustic and which is traditionally grouped among the less prestigious Norwegian dialects: I had left my home and I was now in academia, where a different language law applied, the law of the placeless language, the language with universal validity. I was on a new ‘linguistic market’, in Bourdieu’s terms, and on this market my language was not in demand. It was good currency at the laundrette and on the street corner, but not ‘here’. The solution at the time was not to integrate these language economies, but to separate them, and there is no doubt as to which one had to yield in order for me to win back the command of speech. (2001: 212; my translation)

This exposition is a striking example of how linguistic varieties that originally signal geographical identity on certain linguistic markets are indexed with values that are explicitly social in nature. In other words, the use of dialect can be perceived of as a class marker. A number of studies have been concerned with how dialect users or speakers with a regional vernacular on the one hand and speakers of standard varieties on the other tend to be associated with completely different personality traits (see, e.g., Edwards 1985; Garrett 2007, both with rich bibliographies). People who use a standard variety usually

2. Language and social spaces receive high scores in intelligence and competence, which are both classic high-status qualities. Dialect or regional vernacular speakers, on the other hand, signal personal integrity, friendliness and dependability ⫺ qualities associated with loyalty and solidarity.

4. Traditionmodernity: A nexus between geographical and sociocultural space Language shift occurs as the result of choices made by individuals, […] in accordance with their own motivations, expectations and goals which they may or may not share with other members of their community. Taken together, individual choices form collective choices that impact on the future of a speech community and its language. It lies in the nature of collective choice that the individuals involved in it do not overlook all consequences of their acts. (Coulmas 2005: 168) Up until now the main focus has been on social mechanisms that influence language use here and now ⫺ that is, at a synchronous level. Another crucial dimension throughout the argumentation has been how these mechanisms must be regarded as historical products, that is, as products of certain social, cultural and political conditions. Within such conventionalized social frames, some structures seem more or less natural, and partly because of this collective habitus, certain linguistic practices are used repeatedly, and in this way gain social currency. In the following we will explore this explicitly diachronic dimension. An important question in this context is how the values related to the dialect⫺standard dichotomy can contribute to explaining linguistic changes that appear to take place along the dialect⫺standard axis. Or to put it in another way, how do the changes that have been documented in dialect landscapes worldwide over the past decades connect to more fundamental changes in the corresponding social spaces? A characteristic and basic aspect of dialect development throughout the western world over the past decades is the phenomenon of dialect leveling (see, e.g., Hinskens 1996; Røyneland 2005, and in this volume). Dialect leveling refers to the fact that language features that have a limited geographical range, or that for some other reason are considered as marked, are replaced by variants that have a wider range of use. As a result, the leveling processes contribute to the emergence of varieties that either conform to the standard variety or that are at least closer to the standard than the traditional dialects used to be. What we witness here are different processes of linguistic homogenization. They lead to the abandonment of local differences and make the language users ⫺ partly within a specific region, partly at the national level ⫺ appear more similar than was the case one or two generations ago. At a more fundamental level, these processes imply that language features that indicate tradition are replaced by those that are regarded as modern. Which social changes can be held responsible for this development? The following quotation from the Norwegian dialectologist O. T. Beito (1973: 225) provides an indirect answer to this question:

27

28

I. Introduction: Language and space The development from dialect to standard must be seen against the background of and in connection with the more general cultural development. […]. Dialects are linguistic expressions for local cultural communities, and it seems a reasonable development that differences related to dialect level out after the boundaries for these tiny cultural units burst and are erased (my translation).

The focus here is on small-scale local communities and to the boundaries ⫺ linguistic and other ⫺ that contribute to defining and sustaining these sociogeographical units as frames for people’s lives. This brings us to the crux of the matter: Which sociogeographical units function as the primary basis of identification in today’s modern world? What is the significance today of what Beito refers to as small “cultural units”? Does the literally local environment still serve as a primary sociogeographical frame of reference for the modern language users, or is this function taken over by other unities ⫺ at a regional, national or at a global level? Needless to say, there are no simple or definite answers to these questions (see, e.g., Johnstone 2004 for a detailed discussion). However, it is still possible to identify some general tendencies in the development of societies, especially in the industrialized part of the world. One of the most prominent developments is that local communities have become more integrated into a supra-local society than was the case only one or two generations ago. Two terms that are often used more or less as analogous to the pair local society⫺greater society, going back to Ferdinand Tönnies, are Gemeinschaft⫺Gesellschaft. This idealized dichotomy refers to the small-scale community, or local society, as opposed to a bureaucratic, modern, urban and industrial society (for a detailed discussion, see, e.g., Pløger 1997). The opposition Gemeinschaft⫺Gesellschaft is often linked directly to the rurality versus urbanity distinction (see, e.g., Thuen 2003: 72). One of the anticipated consequences of this integration of the local communities will be, at one level, social and cultural homogenization. Traditional local differences and distinctive features ⫺ material and immaterial ⫺ tend to be eliminated through these processes, giving way to cultural elements from a shared national or global universe of consumption and meaning. However, at a different level, the same forces are socially and culturally differentiating. An important factor in this context are the processes that contribute to the merging of local elements into new, synthesizing forms of cultural expressions in which different supra-local dimensions constitute an important component. The term “glocalization” is often used to express this ambiguity, but then it refers to the tension between the local on the one hand, and the global on the other. The crucial aspect for us is the social and cultural process of diversification that has taken place in particular local communities, a process that can take place when the supralocal is made relevant within a local context. This development has caused an increased differentiation of the local communities into many different social segments ⫺ cultures that co-exist, with their respective lifestyles and value systems, within the frames of the same local society. The Norwegian sociologist A. Hompland (1991) has coined the term “rurbanisation” to describe a striking demographic and sociocultural outcome of these processes. “Rurbanisation” refers to a special synthesis of the rural and the urban; an urbanization of lifestyles, business and consumption patterns without the concomitant rise of the cities. The effects of these processes are so diverse, both in terms of their strength and nature, that they contribute to an increase of social and cultural diversification.

2. Language and social spaces Under such conditions, each individual is exposed to a number of competing norms and must constantly choose between alternative forms of practice. The individual becomes a participant in a number of different arenas ⫺ at a local and supra-local level. Accordingly, the social freedom of choice is far greater today than it was a few generations ago. We are not tied to the same extent by loyalty and obligations to the local communities we are born into. “We have fewer attachments, weaker supra-individual control and more freedom. The rural community does not inhibit us as it used to do”, to quote Hompland again (1991: 62; my translation). This brings us to the kernel of our argumentation: the relation between tradition and modernity. Because of their range and complexity, a general consideration of the meaning of these two terms is beyond the scope of our present discussion. In what follows the relation of tradition⫺modernity will be explored in relation to language and language change only, with the focus on those aspects that can widen our understanding of the connection between dialect and standard. This analytical approach will also help us to see language change in a general sociohistorical perspective. Tradition refers to the inherited, local practices that are passed down through the generations. A “local tradition” refers to any process in which customs, beliefs and values and social forms of practice are associated with a specific geographical location ⫺ a place. The term modernity, on the other hand, can almost be defined as the opposite of inherited and place-bound values. Modernity transgresses what is specifically local and refers to practices whose historical origin is either absent or irrelevant. What characterizes modernity is a breach with the past ⫺ a breach with the more or less unconscious commitments to tradition and traditional ways of living. Accordingly, one of the most important corollaries of modernity is what Giddens (e.g., 1990) refers to as “disembedding”, by which he means the “[…] ‘lifting out’ of social relations from local contexts of interaction and their restructuring across indefinite spans of time-space” (1990: 21). The modern ways of living have distanced us from the traditional “order”; the traditional ties of time and space loosen and social relations become increasingly disconnected from local codes of conduct and practices. In this way modernity may give rise to processes such as individualization and de-localization. If we compare this to the focal semiotic values that are usually associated with dialect and standard (cf. the figure and table above), some interesting patterns appear: Originally a dialect is associated with a geographically defined community, and the use of a dialect signals belonging to a specific local, traditional culture. A standard variety, on the other hand, is often associated with a supra-local and geographically unspecified culture, and hence it can be interpreted as a severance from local and traditional contexts. This de-localization is in itself an important precondition for the standard variety’s intrinsic potential to signal values that are perceived as modern. The connection between these two axes may thus contribute to explaining the changes that have taken place in a number of dialect landscapes throughout the world: The general shift from dialect toward a standard variety must be seen in close connection with those changes in society which have transformed most parts of the western world from a traditional society to an increasingly modern society. As an integrated ⫺ and at the same time constitutive ⫺ aspect of modernity, language users have altered their speech in a direction that is perceived as more modern; or, put differently, as less traditional. On a social macro level, the shift from dialect to standard can therefore be interpreted as part of a set of collective mechanisms that have influenced the “winding up” of tradi-

29

30

I. Introduction: Language and space tional society. However, on a social micro level we need to focus on other aspects in relation to the consequences of these social and cultural “disembedding” mechanisms. At the individual level, modernity firstly involves more freedom of choice. The individual language user may choose to distance herself from local norms and practices in a manner that was not previously possible to the same degree. This increased acceptance of a break with tradition multiplies the number of possible linguistic choices, as Coulmas has argued (2005: 31): Postmodern societies are more mobile, have fewer class markers and are more tolerant of heterogeneity. […] Variation is acceptable and identities are multidimensional. Dialects in postmodern society are not distributed in the same way as in modern industrial societies where patterns of covariation of class and non-standard speech forms are relatively clear-cut.

This also applies to the traditional forms of expressions, such as dialects. Within the frame of modernity the use of dialect does not exclusively express affiliation with traditional values and ways of living. In many local communities today the local dialect is one among several codes available from a wider selection of choices, much as traditional and modern cultural features coexist in other areas as well, with boundaries between them that aren’t always intuitive (see Røyneland 2005 for a discussion). Because of our increased freedom to choose and combine different forms of expression, there is no longer a contradiction between being “modern” and at the same time using a dialect.

5. Concluding remarks The values ascribed to the various language varieties on the linguistic market is, as has been emphasized in the preceding sections, the result of certain historical, political and sociocultural conditions. These values are important components in the mechanisms that regulate linguistic interaction in a given speech community. Because they are part of the doxa of the community, and as such are taken for granted, they have a conserving function ⫺ that is, they contribute to consolidating current conditions and practices. Nonetheless, these values are still subject to social negotiation. The limits of what is considered appropriate, or what will potentially yield social gain, are constantly challenged ⫺ and transgressed. Modernity has at least to some extent released us from traditional ideas and practices and given us room for individual choices. The freedom that modernity represents therefore functions as a counterbalance to the forces that still bind us to inherited, traditional value sets. This is the source of a continuous conflict for the individual language user which requires a significant amount of strategic consideration and “socioeconomic” calculations. As such, modernity’s liberating potential is in constant battle with the conserving forces of convention.

6. Reerences Auer, Peter and Frans Hinskens 1996 The convergence and divergence of dialects in Europe. New and not so new developments in an old area. Sociolinguistica 10: 1⫺30.

2. Language and social spaces Beito, Olav T. [1958] 1973 Drag av utviklinga i hallingma˚let dei siste mannsaldrane. In: O. T. Beito and I. Hoff (eds.), Fra˚ norsk ma˚lføregransking, 190⫺226. Oslo: Universitetsforlaget. Berger, Peter L. and Thomas Luckmann 1967 The Social Construction of Reality. A Treatise in the Sociology of Knowledge. New York: Doubleday. Bourdieu, Pierre 1977 The economics of linguistic exchanges. Social Science Information 16(6): 645⫺668. Bourdieu, Pierre 1991 Language and Symbolic Power. Oxford/Cambridge: Polity Press. Bourdieu, Pierre 1994 Kultursociologiska tekster. I urval av Donald Broady och Mikael Palme. Stockholm/ Stehag: Brutus Östlings Bokförlag Symposion. Bourdieu, Pierre 1995 Distinksjonen. En sosiologisk kritikk av dømmekraften. Oslo: Pax Forlag. Chambers, J. K. and Peter Trudgill 1980 Dialectology. Cambridge: Cambridge University Press. Coulmas, Florian 2005 Sociolinguistics. The Study of Speakers’ Choices. Cambridge: Cambridge University Press. Edwards, John R. 1985 Language, Society and Identity. Oxford: Basil Blackwell. Fasold, Ralph 1984 The Sociolinguistics of Society. Oxford: Basil Blackwell. Garrett, Peter 2007 Language attitudes. In: Llamas, Mullany and Stockwell (eds.), 116⫺121. Giddens, Anthony 1990 The Consequences of Modernity. Cambridge: Polity press. Grillo, Ralph D. 1989 Dominant Languages. Language and Hierarchy in Britain and France. Cambridge: Cambridge University Press. Hinskens, Frans 1996 Dialect Levelling in Limburg. Structural and Sociolinguistic Aspects. Tübingen: Max Niemeyer. Hompland, Andreas 1991 Det store kappeskiftet. In: R. Alma˚s (ed.) A˚tte perspektiver pa˚ bygdeutvikling, 55⫺74. (NLVF-publikasjon 3.) Oslo: Norsk landbruksvitenskapelig forskning. Jahr, Ernst Ha˚kon and Karol Janicki 1995 The function of the standard variety: A contrastive study of Norwegian and Polish. International Journal of the Sociology of Language 115: 25⫺45. Johnstone, Barbara 2004 Place, globalization, and linguistic variation. In: C. Fought (ed.) Sociolinguistic Variation. Critical Reflections, 65⫺83. Oxford: Oxford University Press. Joseph, John E. 1987 Eloquence and Power. The Rise of Language Standards and Standard Languages. London: Frances Pinter. Kerswill, Paul 2007 Social class. In: Llamas, Mullany and Stockwell (eds.), 51⫺61. Llamas, Carmen, Louise Mullany and Peter Stockwell (eds.) 2007 The Routledge Companion to Sociolinguistics. London/New York: Routledge. Linell, Per 1998 Approaching Dialogue: Talk, Interaction and Contexts in Dialogical Perspective. Amsterdam: John Benjamins.

31

32

I. Introduction: Language and space Mæhlum, Brit 2007 Konfrontasjoner. Na˚r spra˚k møtes. Oslo: Novus forlag. Maurand, Georges 1981 Situation linguistiqe d’une communaute´ rurale en domain occitan. International Journal of the Sociology of Language 29: 99⫺119. Mesthrie, Rajend, Joan Swann, Anna Deumert and William L. Leap 2000 Introducing Sociolinguistics. Edinburgh: Edinburgh University Press. Milroy, James 2007 The ideology of the standard language. In: Llamas, Mullany and Stockwell (eds.), 133⫺139. Milroy, James and Lesley Milroy 1985 Authority in Language. Investigating Language Prescription and Standardisation. London/ New York: Routledge. Mugglestone, Lynda 1997 “Talking Proper”. The Rise of Accent as Social Symbol. Oxford: Oxford University Press. Pløger, John 1997 Byliv og modernitet. Mellom nærmiljø og urbanitet. Oslo: NIBR. Rosenlund, Lennart 1991 Om smak og behag. En introduksjon av Pierre Bourdieus kultursosiologi (Tidvise Skrifter 4, Kultur og kommunikasjon). Stavanger: Avdeling for kultur- og samfunnsfag ved Høgskolesenteret i Rogaland. Røyneland, Unn 2005 Dialektnivellering, ungdom og identitet. Ein komparativ studie av spra˚kleg variasjon og endring i to tilgrensande dialektomra˚de, Røros og Tynset. Acta Humaniora, Det humanistiske fakultet, Universitetet i Oslo. Smith, Olivia 1984 The Politics of Language 1791⫺1819. Oxford: Clarendon Press. Strømsodd, Svein Arne 1979 Dialektholdninger blant folk i to bydeler i Oslo. Unpublished Masters thesis, University of Oslo. Sveen, Karin 2001 Klassereise. Et livshistorisk essay. Oslo: Oktober. Teleman, Ulf 1979 Spra˚krätt. Om skolans spra˚knormer och samhällets. Lund: Liber. Thuen, Trond 2003 Stedets identitet. In: T. Thuen (ed.) Sted og tilhørighet, 59⫺76. (Kulturstudier 27.) Kristiansand: Høyskoleforlaget. Thompson, John B. 1991 Editor’s Introduction. In: P. Bourdieu, Language and Symbolic Power, 1⫺31. Oxford/ Cambridge: Polity Press. Wiggen, Geirr 1995 Norway in the 1990s: A sociolinguistic profile. International Journal of the Sociology of Language 115: 47⫺83. Wolfram, Walt 2007 Ethnic varieties. In: Llamas, Mullany and Stockwell (eds.), 77⫺83.

Brit Mæhlum, Trondheim (Norway)

3. Language and political spaces

3. Language and political spaces 1. 2. 3. 4. 5. 6.

Introduction Language ideologies and spatial representations Standard languages in national space Other models, other maps Locality in places References

1. Introduction Speaking overcomes distance: We communicate across stretches of space, relying on sound to travel. Linguistic messages go considerably further when they are written or transmitted through ever more sophisticated media. Yet speaking also seems to be anchored in territory, that is, in geographical space as mediated by political practices. Many forms of cultural logic link political claims to language and to territory. The most widespread of these in the contemporary world is that of the nation-state, with its presumed ethnolinguistic unity. In this hegemonic political form, speakers are considered authentic members of nations by virtue of their linguistic competence, and nations are supposed to be distributed over territories that are organized politically into states. Whatever specific political system a state adopts, its linkage to a posited nation is most often established through the medium of a standardized language with a literary tradition, a norm of monolingualism and the assumption of linguistic homogeneity in the polity. This presumed cultural-linguistic connection is a fundamental feature of the state’s legitimacy. Nation-states of this type are the political ideal in the contemporary world. Rarely do actual states show the full form. There are also many other cultural configurations ⫺ either long-standing or newly emergent ⫺ that draw on different links between linguistic forms and political space. In describing these less familiar patterns, we find differences in the cultural definitions of ‘polity’ ‘language’ and even ‘space.’ Thus, I argue that the relation of language to political spaces is always mediated by cultural systems ⫺ language ideologies ⫺ that define the very terms we are discussing. To make this case, the second section examines the cultural assumptions about language and location embodied in the nation-state, ones that also underlie eighteenth and nineteenth-century linguistic science. The internationally conventional language map is built on the perspective of the contemporary nation-state as a political form. This becomes clear when we consider the linguistic phenomena omitted by such maps. Maps are constitutive of cultural notions of space. When taken to be objective representations, they help to construct the linguistic world they claim merely to reflect. Nation-state maps remain persuasive representations of the ethnolinguistic world even when their elisions are revealed because there are many other practices that buttress or produce the configurations such maps display. Section three discusses some of these practices ⫺ standardization, purification, authentication ⫺ as institutionally orchestrated semiotic processes. Even the exceptions to nation-state imagery provide evidence of the current international strength of these ideas: When state boundaries and linguistic territories do not match ⫺

33

34

I. Introduction: Language and space as is often the case around the world ⫺ the result is political tension, or the threat of political mobilization. Nevertheless, there are limits to this hegemony. Section four discusses alternative ways of thinking about space and language by presenting cartographies that are motivated by strikingly different political perspectives and cultural values. Finally, section five takes up cases where perspectives clash and definitions of what counts as language are explicitly contested among speakers, analysts and governments. At stake in these conflicts is the perception of “place” or locality. The study of locality highlights the sensuous, experiential aspects of places ⫺ rather than abstract spaces ⫺ and how their structures of feeling and belonging are constituted by linguistic form.

2. Language ideologies and spatial representations It is customary to trace the language ideology of ethnonationalism to German philosophy in the late eighteenth century. Johann Gottfried Herder was perhaps the most influential of the many who argued that each language is unique in its poetic expression; each has the capacity to express all the human capacities, including feeling, reason and will. For Herder, the most salient contrasts among linguistic forms are those based on cultural inheritance and that differentiate national languages. They are the means of giving expression to the distinctive spirit of a people, its Volksgeist. He assumed uniformity of speech among the Volk when he equated one people, one fatherland, one language. Herder feared that the verbal traditions of Germans were increasingly endangered by the incursions of French among the aristocracy and by the tendency of ordinary peasants and simple city people to forget their past. Traditions must therefore be safe-guarded by a special sort of intellectual, one who shuns foreign models and is specially attuned to the artistry of “his” Volk. Herder’s views invoke the past to authorize current nationalism in which language and folklore become key symbols of identity, unity and persistence (Jankowsky 1973: 20: 53). These ideas about language continue to be characteristic habits of thought in Europe, among elites and the general populace (Blommaert and Verschueren 1998). Recent historical scholarship has stressed a further point: When Herder’s philosophy is seen within its historical context as a language ideology ⫺ a metadiscourse about discourse on language ⫺ its emphasis on intertextual links with a known or imagined past stands in direct contrast to the language ideology of John Locke, over a century earlier. For Locke, political unity rested on reason used for public discussion, which was made possible by the inherent properties of language. These included the primacy of referential over emotional meaning, the arbitrariness of signs in the service of philosophical expression and the rigorous precision necessary for rational discourse in politics and science. These properties were corrupted by appeals to history. Thus for Locke, political legitimacy depended on severing intertextual links with the past. To be able to use language in this way required certain forms of character and education. It is these two opposed positions, one authorizing the nation through links to “tradition” in the linguistic past, the other locating political authority in the timeless-placeless systematicity of language itself, that constitute the tense ideological formation we now call “modernity” (Bauman and Briggs 2003: 190⫺196). Although Locke’s theory argues that all languages are the same in their semiotic potential, and Herder’s theory stresses

3. Language and political spaces the differences between languages in the spirit and vision they impart, the two theories agree that a viable polity must be based on uniformity of linguistic practices among speakers, and the guidance or mediation of properly inclined and trained intellectuals. In this, both contribute to legitimating the ethnolinguistic perspective characteristic of the modern nation-state. As we will see in section 3, the nation-state is a product of “modernity” in a further sense: it relies for authorization as much on Herderian authenticity as on Lockean universal reason.

2.1. Locational technologies Herder considered geography to be an influence on national language, but how space as an abstract notion came to seem significant is a more complex story. Though often credited to Herder, the connection between language and nation had been stressed earlier by French, Spanish and English as well as other German philosophers. Most scholars agree that it was well established in European intellectual circles by the early eighteenth century (Aarsleff 1982); others give it a much older, even biblical origin (Olender 1992; Trautmann 1997). Leibniz and then Condillac explicitly noted that language was the repository of a people’s history and character. The famous English contribution came later, during Herder’s lifetime. Speaking from his post in British colonial India, William Jones announced in 1788 the discovery of similarities between Sanskrit, Greek and Latin. From these correspondences he deduced “family” relations among these languages and ⫺ presuming the language/nation equation ⫺ among the peoples who spoke them. Since the parallels Jones announced had long been known, the intellectual excitement came rather from Jones’s explanation of these similarities. He constructed a single story of kinship and geographical movement for Indo-European peoples that would later be shaped into the Stammbaum theory of historical linguistics, providing a model and evidentiary basis for the study of language change more generally. To do this, Jones relied implicitly on three “technologies of location,” that is, three frames of cultural assumptions guiding a project of mapping, ones that define abstract representational spaces and then show how entities are situated within those spaces (Trautmann 2006: 2⫺21). First, Jones assumed the mapping of the surface of the earth according to latitude and longitude, a Ptolemaic practice from the second century AD that ⫺ though often challenged ⫺ has survived the vast changes in geographical knowledge since that time. Indeed, the power of the framework lies in its invisibility. Despite their small size, their flatness, their lack of relief and paucity of other information, we have come to take such maps to be faithful pictures of geographical space itself. Second, like most thinkers of the period, in presenting new ethnological material from the colonies, Jones had recourse to biblical categories. His terminology ⫺ Hamites, Jasephites, Semites ⫺ invoked the schema of kinship in the first book of the Hebrew Bible that charts the relations among the male descendants of Noah after the flood. The patrilineal logic of these relations predicts a continual segmentation of branches over time, so that new and related tribal units are formed. The dispersion of such patrilineal segments, when projected onto Ptolemaic geographical space, created a concrete image of migration (Trautmann 2000: 6⫺9). Finally, the early and simple word list, initiated by Leibniz, was a technology for the spatial redefinition of language itself. Drawing on Locke’s notion of complex words built on earlier and simpler ones, it imagined language as

35

36

I. Introduction: Language and space having a center and a periphery. It divided the lexical stock into what was taken to be core elements as opposed to others that, because they were supposedly later developments, could be ignored in comparative research (Aarsleff 1982). In sum, the scenario of a family of nations with a single source, their different branches increasingly dispersed geographically, each with its own core language, relies on the lamination of these three locational technologies, and ultimately allowed their projection onto contemporary linguistic evidence as well as ancient textual sources. Assumptions about language, kinship and space ⫺ each coming from different historicopolitical contexts ⫺ contributed to discussions in philosophy, comparative philology and language history. These intellectual endeavors in turn have been in constant engagement with political projects, and contributed to the idea of nation-states as ethnolinguistic units, as well as of supranational units based on linguistic “kinship” that have authorized political movements such as pan-Slavism and pan-Germanism.

2.2. Language maps and their limits The culturally apprehended relationship between language and political space was further transformed by another locational technology: the nineteenth-century census. Meanwhile, global maps of latitude and longitude had accommodated new ideas about sovereignty. In contrast to feudal kingship, which was organized around a center, the modern nation’s sovereignty is “fully flat, and evenly operative over each square centimeter of a legally demarcated territory” (Anderson 1991: 19). Maps represented this conception iconically with solid lines marking boundaries and spaces within these lines filled with contrasting colors to show the extent of states and their colonial holdings. The new census created a form of information that could “exist” within those colored spaces. But language has been an enduring problem for this technology. Early European census categories did not include language at all. Their purpose was tax assessment. They distinguished among categories such as merchant, serf, aristocrat, Jew, German, thereby putting into a single series a set of classifications we would now consider to be on contrasting dimensions (occupation, social status, religion, nationality). By the early eighteenth century there were attempts, such as the famous Völkertafel of Styria, to organize a consistent chart of European national types and their characteristics: clothing, favorite activity and mode of religiosity (Stanzel 1999). In contrast to these earlier forms, the great innovation of the nineteenth-century census was to focus not on types but on individuals and to include every individual regardless of tax-status. Each person was made “visible” as the intersection of a set of bounded, non-overlapping classifications representing whatever social dimensions the state recognized. As a result of such a grid, each person could be enumerated as an instance of each exclusive category. Operating as a paper-and-pencil panopticon, the census grid aspired to distinctively locate anything ⫺ regions, people, products ⫺ in terms defined by the administrative interests of states. When disseminated through print capitalism, this model enumerated populations that stretched from metropoles to their colonies, and were organized for bureaucratic surveillance and control (Anderson 1991: 163⫺185). The attempt to include language within such a totalizing grid ⫺ and to represent it on maps ⫺ produced technical difficulties and political battles for the many state-sponsored statistical bureaus created in nineteenth-century Europe (Hobsbawm 1990). In a world

3. Language and political spaces that assumed language ⫽ national spirit ⫽ political autonomy, multilingual empires were deeply threatened by evidence of linguistic diversity within their borders. Aspiring movements for independence were reliant on the same evidence, to be placed before sympathetic international audiences. Such information remains crucial today for regional and migrant populations soliciting support from the European Union (Urla 1993). Closely related, and no less problematic, were the technical difficulties: what would “count” as a language statistic. Dialectologists had amply documented that linguistic forms do not have neat spatial boundaries; even borders between nation-states are crossed by dialect chains, and social strata differ even though located in a single space. Census-takers argued bitterly over whether and how to formulate questions about language. Should speakers register their “mother tongue”, “community language”, “language of the home”, “Verkehrssprache”, “Umgangssprache”, or some other category? Each of these would produce results with different political implications. The concept of standard language got a boost from its compatibility with the census: ideally homogeneous, spoken by everyone in a polity, commensurable with all other standards, bounded, deliberately non-overlapping, one-to-a-person, easy to count. To see how language maps work, let us consider an excellent, indeed exemplary, minority-friendly and technologically advanced case published by the European Bureau of Lesser Used Languages, the major clearing house for information about minority, regional and immigrant languages in the European Union. The EBLUL map recognizes regional and even some immigrant languages, and marks areas of bilingualism by using stripes of two colors. It does not presume that language areas (“Sprachgebiet” “area linguistique”) are isomorphic with nation-state boundaries. Yet, much of linguistic reality is erased. For example: colors operate like those that represent “evenly operative” sovereignty in modern political maps, giving the impression that languages show no internal differentiation, and that speakers are evenly distributed across territory without variation in language(s) used. Yet we know the density of speakers varies importantly, as does the density and type of bilingualism: do the striped segments represent bilingual speakers or territories inhabited by monolingual speakers of two languages? Also invisible are degrees and modalities of linguistic competence (fluency, literacy, pronunciation), their relative prestige and the situational distribution of the languages. Knowledge of English and French by educated speakers of other languages is not represented; nor is knowledge of Latin, Arabic, Church Slavonic and Hebrew by clerics. The map provides little information about the place of a language in public life, nor speakers’ often varied stances towards the standard languages they encounter or are said to speak. Like any representation, this map is inevitably partial. Judging by what it omits, and despite efforts to the contrary, it presents more a vision of what ethnolinguistic diversity should look like from the eye-view of a nation-state or through a European supra-state’s standardizing gaze, than a picture of actually existing linguistic practices.

3. Standard languages in national space Language maps remain remarkably believable despite their elisions because many social institutions support the sociolinguistic picture displayed in such maps. As Bakhtin (1981) argued, heteroglossia ⫺ diversity of varieties, a centrifugal proliferation of styles, ac-

37

38

I. Introduction: Language and space cents, registers, languages ⫺ is ubiquitous, the ordinary condition of linguistic life. The centripetal, standardizing processes, however, require active construction by state-related institutions. To be sure, print capitalism has been crucial in spreading books and newspapers written in standardized orthography and creating a public to read them (Anderson 1991). But at least as significant has been the work of centralized school systems and curricula, unified labor markets, and legions of language planners, linguists, teachers and poets who have created dictionaries, academies, grammars and literatures. The resulting corpus of standardized texts is indispensable to the process. But even more consequential has been the creation of a set of naturalized beliefs about standard language. Cultural workers have inculcated in speakers the notion of respect for standard forms of “correctness,” and indeed an entire ideology of the standard that has had effects on how speakers use language, what they accept as ideal models of speech, and what the existing differences in pronunciation, grammar, lexicon and prosody come to signal for interactants in face-to-face interaction and in mass mediated talk. Standardization does not happen by itself, nor simply through the magic of a capitalist market. Nation-states have differed in the timing and rate of standardization. They vary in how strictly their pedagogical systems are centralized, how thoroughly they penetrate the whole country, and how stringently they inculcate and enforce the standard. But the semiotic processes of language ideology that authorize one linguistic variety as the “standard” are more uniform across cases and as important as the institutional histories. In general, semiotic processes consist of sign relations that link linguistic forms to social facts; the signs are interpretable within cultural presuppositions about language, i. e. language ideologies (Woolard and Schieffelin 1994). Such sign relations include indexicality (a pointing to or assumed contiguity between the sign and its object), iconicity (resemblance between sign and object), fractal recursion and erasure (Irvine and Gal 2000; Silverstein 2003). There are several ways in which sign relations create standardization by drawing on political geography and on abstract spatial metaphors. First, in standardized regimes, linguistic variation is visualized ⫺ by ordinary speakers and often by linguists too ⫺ as an abstract space in which the standard “covers” other varieties, is superimposed on them, and therefore imagined to be located “above” them. Other forms are not simply different, or typical of different geographical regions and social strata. They are seen to be “lower,” and therefore worse. These judgments are inscribed in the workings of institutions such as schools. Students may fail to learn the forms designated as standard by language planners. But they invariably acquire the disposition to accept that some forms are “better” than others. They are taught that the standard forms have greater intrinsic, linguistic value ⫺ e.g., that they are more rational, more precise, more beautiful. But, in a sociological view, these values are cultural endowments, not inherent in linguistic form (Bourdieu 1991). Because linguistic varieties are indexical signs of those who speak them and of the situations in which they are used, devaluing a form means devaluing its speakers. Therefore, those who accept the high value of a standard language that they do not themselves control usually find that they devalue themselves. This is a process Bourdieu has called “linguistic domination.” One aspect of standardization seen in semiotic terms is exactly this switch in perspective: seeing oneself from the viewpoint of judgmental others. A second kind of spatialization concerns the specific indexical significance of linguistic varieties. Meanings associated with linguistic form are far more elaborate than differential acceptance or prestige. Linguistic varieties also come to index what Bakhtin

3. Language and political spaces (1981) has called “chronotopes”: cultural categories identifying a nexus of space-time and person. “Modernity” is one such chronotope, usually associated with standards, which are among the many insignia of modernity that nation-states are required to display if they are to be acceptable members of the contemporary international community of states. Anthems, flags, national airlines and ⫺ in the nineteenth century ⫺ even statistical bureaus, were among the many such signs, forced on states as the price of recognition by national ideology and its “coercive isomorphism.” By contrast with the standard language’s modernity, regional and “non-standard” forms are assigned a chronotope of rural distance and historical past, sometimes inflected as authenticity and old-fashioned simplicity, at other times as country bumpkin ignorance. Hence the supposed discovery of “Elizabethan English” in the isolated mountains of Appalachia, or the nostalgic attractiveness for urban Germans of touristic visits to the villages of German-speakers in rural Hungary. Although chronotopes of tradition/modernity are common, there is a more complex world of meaning, a “culture of standardization,” particular to each standardized speech community (Silverstein 1996). In the United States, for instance, standard English is seen as neither too precise nor too lax, neither emotional nor rational; just right for communicating political truths. It is believed to be a product of skill, achievable through costly training, and ⫺ like American commodities in advertisements ⫺ sold as the means to improving one’s life chances. The regional and socially marked forms are seen, in contrast, as lacking this balance, and often stigmatized as emotional or lazy, old-fashioned commodities ready to be traded in. On another dimension, American and British English are thought to be welcoming of foreign words and phrases, just as the countries are “open” to migrants (Crowley 1989). The French standard is also seen by its speakers as clear, rational and pure, but to retain its special clarity ⫺ so the ideology goes ⫺ standard French should not accept foreign words, or only after considerable domestication. This is an attitude that contrasts with the American and, as it turns out, echoes long-standing French immigration policy. In each case, borrowing ⫺ that is, the defense and policing of linguistic boundaries ⫺ is regimented by ideology, and thereby constrains linguistic structure. In each case, cherished national self-images are iconically projected onto the standard language and its speakers; internal differentiation carries the stigma of negative self-images. The recursive nature of standardization provides a third way in which it is connected to spatialization. Because standards are tied to nation-states as linguistic emblems, standard ideology creates expectations that standard languages will be evenly distributed over the state’s territory, just as modern sovereignty is. Contrary to this commonsense view, however, standardization creates not linguistic uniformity within a state but even more, and hierarchical, heterogeneity in speech. The creation of a standard necessarily recasts the relation among existing varieties, bringing all forms into a unified field of judgment and surveillance. Every attempt to standardize a regional form ⫺ e.g., to create a standard Friulian, Basque, or Corsican ⫺ will require picking out and regularizing a newly standard register. That step will bring with it the stigmatization of other forms used among the very speakers whose regional linguistic practices the new standardization was supposed to valorize. To be sure, what is seen as stigma from the perspective of the standard and its institutions can be re-evaluated. Novel regional or class-based forms (anti-languages) are often created and celebrated as resistance to standards. Because standardization is fractally recursive in this way, it brings not homogeneity but more heterogeneity (Gal 2006).

39

40

I. Introduction: Language and space Dialectologists working in countries with standardized regimes (now virtually everywhere), can no longer simply track regional differences in speech, because “regionalism” itself is created as a contrast with standard. How people define and evaluate regional, non-standard, local or indeed any socially-marked linguistic form will depend, in part, on how the standard is viewed. The indexical significance of the standard ⫺ like that of any linguistic form ⫺ emerges as the product of a cultural project: intellectuals of all kinds, politicians and pedagogues actively label and analyze the significance of linguistic differentiation as supposed clues to the character, intelligence and other qualities of speakers. Given these omnipresent efforts to lend meaning to variation, it is not geographical or social distance between speakers that creates differentiation. On the contrary, differentiation arises from the juxtaposition of contrastively valued linguistic varieties within a single person’s repertoire, or the interaction of neighboring speakers whose different forms signal social position. Even single sounds gain contrastive significance as indexical and iconic signals of identity or social situation. Thus linguistic change is best understood as a consequence of intimate contact among speakers rather than distance (Irvine and Gal 2000). Standards display a fourth kind of spatialization. Although historical accounts often show that a nation-state’s major administrative center is the source of many linguistic features adopted later as standard, the locus of the “best” and most admired linguistic forms is not necessarily a major urban center. Often it is some particular region of the countryside that is designated as the repository of authentic and “historically national” speech. As with the cultural meaning of linguistic variants, so too in the location of the “best” forms, there are cultural brokers involved in creating and sometimes changing the ideals. The region seen as the source of the best American speech was deliberately re-located in the early years of the twentieth century to the rural Mid-west of the country when elites stigmatized the cities of the East Coast as corrupted by the presence of foreign immigrants (Bonfiglio 2000). Similarly, the recent debates about the polycentric nature of Hungarian are also arguments about what regions should be considered the “center” or norm of Hungarian-speaking, and for what historical-political reasons (Lanstyak 1995). Finally, the forms of authority granted to standards can be understood as the result of conflicting spatializations. As exemplary modern phenomena, standard languages are simultaneously credited as both “traditional, authentic” and “universal,” even though (or because) these values are both present, though usually opposites, in modernist ideology. The apparent contradiction is resolved by spatially distinct, comparative contexts. When compared to regional or class dialects or minority languages internal to the nation-state, the standard language is understood to be a universal solvent: it is the means of translation among the other forms; it is the variety in which everything can be said clearly and truthfully. By contrast, in the larger comparative context of other national standards, emblems of other nations, the standard language is heard as specially authentic to its nation, providing ways of expressing supposedly untranslatable, culturally specific concepts.

4. Other models, other maps Standard ideology is now a common phenomenon across the globe. Nevertheless, there are alternate visions of political geography and linguistic practice that rely on different cultural values. As with the other representations we have discussed that claim simply

3. Language and political spaces to describe the world ⫺ kinship charts, word lists, latitude/longitude maps, census grids ⫺ these too are performative: creative, prospective models for, as well as partial models of, the systems they describe. It is illuminating to consider some of those that are the product of long-standing cultural principles as well as others that are relatively novel schemas in active competition with nation-state maps. Given the acceleration in economic globalization, it should not be surprising if some of these mappings are visions born of migration and displacement.

4.1. Southeast Asian polities A striking contrast to contemporary European maps is the Southeast Asian “galactic polity” (Tambiah 1985), for which the mandala provided the basic schema. According to Indo-Tibetan cosmology, part of Hindu-Buddhist thought, a mandala is composed of two elements, the core (manda) and the container or enclosing element (la), with satellites of increasingly complexity around a center. Across Southeast Asia up to the twentieth century, this general model organized activity in virtually every cultural domain. Architectural projects and textile decoration were shaped as mandalas; the cosmos was understood in myth as constituted by Mount Meru in the center surrounded by oceans and mountain ranges. At a philosophical and doctrinal level, the relation between consciousness and perception was supposed to be organized like a mandala, as was the heavenly, spiritual order. Mandalas were geometrical, topographical, cosmological, societal blueprints for polities. In fact, the physical layout of political units, on the ground, often looked astonishingly like mandalas, with the most powerful ruler located in the geographical center, surrounded by his feuding sons or heirs who were themselves surrounded by lesser rulers. The constant conflict among rulers precluded firm boundaries, but produced an oscillation of disintegrating and reconstituting politics, each fractally replicating the central one in organization and activity, and arranged in mandala shapes. In each region, the most powerful court, the “exemplary center” (Geertz 1980), gained and legitimated its power over lesser courts and their hinterlands not by military or administrative means, but through the elaborate display of the key value of “refinement” in pageantry, festivals and theatrical self-presentations. This demonstrated the court’s ever more precise imitation of the delicacy, civilization and perfection of the celestial, supernatural order and thereby justified and demonstrated that court’s worthiness and ability to rule. Most important for our purposes, the linguistic aspect of exemplary political centers was crucial to legitimate governance. In the cultural system that created power of this kind, refinement, civilization and minute gradations of power were expressed and reproduced primarily through linguistic and other expressive distinctions that required the complex ritualization of everyday interaction. The higher the lord, the more refined, perfect and elaborate was his verbal etiquette, his comportment and sartorial display, all of which then proved his capacity to rule. Although there was usually only one named language involved in any one court, the rulers of Java, Bali and the lesser kingdoms of Southeast Asia created and elaborated linguistic distinctions now called “speech levels” that are characteristic across all the languages of this region. Errington (1988) reports as many as five-to-seven such speech registers in Javanese. Lexicon and morphology distinguished “levels” that precisely marked each utterance for very fine degree of polite-

41

42

I. Introduction: Language and space ness or refinement. These distinctions created ⫺ and did not merely reflect ⫺ the hierarchical social order. Linguistic delicacy and the complex etiquette of displaying it increased as one moved from the peasantry who spoke simply, to the elaborations of ever higher-placed aristocrats.

4.2. A global constellation The model of a galactic polity bears only superficial resemblance to a quite different “global language system” proposed some years ago as a theory of linguistic demography by a Dutch political scientist. De Swaan (2001) calls his model “a galaxy of languages” in which constellations or concentric circles represent the relationship between languages in an abstract, relational space. In contrast to the Southeast Asian model, the units charted here are (mostly) standard languages, not ritualized registers; the value at issue is profit, not refinement. If the galactic polity is a form with a long past, the global system is supposed to map and understand a global future. Though presented as a social scientific theory, de Swaan’s model is recognizable as an ideological perspective that views economic globalization and the distribution of languages as parallel phenomena, both growing out of the presumed universal tendency of humans to maximize benefits. The “language galaxy” model recognizes that there are thousands of rarely-written languages in the world, spoken by only hundreds of people each. These are termed peripheral languages and likened to the moons that travel around planets in a solar system. Examples would be Sulawesi, Sumba, Bayak and Javanese in Southeast Asia. More like planets are the central languages, those that have writing systems, literatures, many more speakers, official status in states and participate in the workings of international organizations (i. e. standards). In Southeast Asia, Indonesian and Chinese would be such languages. Finally the model proposes that there are supercentral languages, with even more speakers, that are like suns in a solar system, around which the planet languages revolve. In this view, Arabic, Russian, English and French (in some parts of the world), Spanish (in other parts), are among the 12 such languages of the world. And finally there is a single center ⫺ English ⫺ for the 12 solar language systems. English with its huge number of (mostly second language) speakers is the hub of the linguistic galaxy as described by de Swaan. According to this model, speakers try to learn languages that are further in towards the center orbits than their native tongues, but do not learn languages further out. Bilingual individuals connect the orbits of moons to planets, planets to suns, so there is a key but limited role for multilingualism. Those at the center (i. e. the sun) of each solar system speak to each other in English. Although the model/scheme is hierarchical, as well as demographically and functionally oriented, and “constellation” is largely a metaphor, the system is not entirely abstract. There is a geographical factor, in that different parts of the world can be mapped by specifying which languages fulfill each of the hierarchically organized functions in a region. This model erases much of what the jigsaw map and the galactic polity highlight. It ignores nation-states and their boundaries, elides the problems of defining language, linguistic difference and the significance of indexical signals. It is uninterested in the ritual modes, genres and values (such as refinement or correctness) by which communities of speakers, within cultural systems, organize and evaluate their practices. In this vision, all languages are merely denotational codes, all speakers are individual decision-

3. Language and political spaces makers, and languages have chiefly instrumental, economic value. Speakers make rational choices in language learning, within what their circumstances allow. This is a social scientific schema, to be sure, but one written from the lofty perspective of transnational corporations and their aspiring executives, whose presumed career goal is always to move closer to headquarters, to the center. The model projects ⫺ and celebrates ⫺ a neoliberal world in which businesses and their mobile employees make choices in idealtype free markets.

4.3. Diverse diasporas Migrant populations are likely to have different visions of space, politics and language than do nation-states, which can imagine mobile people only as problems, unless they assimilate. However, it is not deterritorialization that migrants envision, but rather novel, spatial connections. A case in point is the language and spatial ideology of Hadrami men, who are merchants and traders with origins in the region that is now Yemen. They have been part of a centuries-old migratory pattern around the Indian Ocean (Ho 2006). Since the 13th century and continuing today, Hadrami traders have made their way southwest, down the coast of Africa, spreading Islam and commerce, marrying the daughters of local leaders, establishing families and often becoming local political elites. Others in the same lineages moved to the east, through the southern coast of presentday India and to Malaysia and Indonesia. The merchants retained individual and familial memory of where they and their male kin came from, retaining as well a tradition of ideal return, and a set of practices ⫺ including Islam ⫺ that tied them to others from Hadramawt. Other practices that maintained ties among merchant families included circum-Indian Ocean visiting, business relations and the careful ritual tending, in Hadramawt, of the graves of returned migrants. Most informative for our purposes is the custom of keeping genealogies, which were always written in Arabic, the language that has continued to be the sacred if not the everyday usage of these multilingual traders. In the calligraphically elaborate genealogical charts, written in Arabic, each man’s name is followed by the name of his place of death (also in Arabic). One genealogical tree reproduced by Ho (2006: xvi) names male relatives over six or seven generations ⫺ fathers, sons, brothers, uncles and cousins of a single lineage ⫺ who died in places as diverse as India, Say’un, two different parts of Java, the Sawahil (East Africa) and Mocha (Yemen). The values inscribed here are tellingly selective: the vast and diverse territories around the Indian Ocean are represented simply as points of death, unified by a single written language (Arabic), a single lineage and the locational practice of genealogical representation. For other populations, different visions of language and space create connection in diaspora. The island of Mauritius is a predominantly Mauritian Creole-speaking society, but over two thirds of the population are of Indian origin, mostly Hindu. Although Hindi is officially recognized by the state as the “ancestral language” of Indo-Mauritians, their everyday usage consists mostly of Mauritian Creole. Some, however, speak Bojhpuri, which is a northeastern Indian variety that ⫺ although part of the long dialect chain of northern India ⫺ is quite different from Standard Hindi, which drew on western dialects and in any case was regularized well after the ancestors of Indo-Mauritians left their homeland. How then do Indo-Mauritians come to recognize Hindi as “their” ancestral language? Politically and religiously motivated efforts to establish diasporic

43

44

I. Introduction: Language and space Hindu identity in Mauritius depend on strengthening culturally imagined ties to mainland India. These rely on a number of spatio-linguistic practices quite different from those of the Hadrami merchants. Of central importance is the annual pilgrimage to a mountain lake in the southwest of Mauritius on the occasion of the Hindu festival of Shivratri. The event is heavily attended and state-supported. Through the shape and location of the temples around the lake, and the rituals enacted there, Indo-Mauritians spatially and performatively create a sacred geography closely resembling that of Hindu pilgrimage sites on the sacred river Ganges in North India. The continuity with a sacred Hindu landscape in India is emphasized in legends that claim the lake is a tributary of the Ganges itself, the water coming through subterranean sources across the ocean. In 1972 a vessel of water from the Ganges was deposited in the lake, thereby officially consecrating the lake as a part of the Ganges. Equally important is the use of Hindi in the religious events at the pilgrimage site, which are understood as re-enactments of what the ancestors did at the site. But since Hindi is not a part of the daily repertoires of Indo-Mauritians, the use of it during the pilgrimage relies on schooling in the “ancestral language” and on the use of Hindi in associations that organize temple activities at villages around the island (Eisenlohr 2006). Yet the usage in each of these contexts is not what would be recognized as Hindi by speakers from the subcontinent. Rather, what characterizes these contexts is the use of everyday Bhojpuri that is purified of Creole lexical elements, which are replaced by sanskritized Hindi items. Many grammatical and phonological changes also appear, so that the effect, for speakers of Bhojpuri, is a new and purified register of Bhojpuri, one that is used to signal the start of Hindu genres of moral discourse. This register has the effect of blurring the boundaries between Bhojpuri and Hindi, making persuasive through recurring linguistic usage the connection between Indo-Mauritians and Hindispeakers. It also makes the ancestral Hindi taught in schools and used in sacred performances more accessible for Indo-Mauritians (Eisenlohr 2006: 66⫺110). The Hadrami merchants created a diaspora by picturing their patrilineage as an Arabic-Islamic web connecting disparate spaces. On their genealogies, they reduced the conventional space of longitude-and-latitude to points made significant by family members’ lifespan. By contrast, Hindu Mauritians created a diaspora by replicating a revered model. Indo-Mauritians become diasporic Hindus through representing events in such a way that they bear a close, iconic similarity to events elsewhere. Through the pilgrimage and through a hindiized register of Bhojpuri, experiences of the “homeland” become available on Mauritius, further verifying the constructed fact that Indo-Mauritians are indeed Indians. Spatial and linguistic replications operate as parallel synecdoches: a piece of Mauritian landscape recreates India’s sacred geography, just as India’s linguistic materials are recreated in Mauritian linguistic practice.

5. Locality in places The discussion so far has foregrounded representational practices in order to highlight that a vision of space is always constructed from some imagined or actual physical point or social perspective. Most human schemas of orientation are ego- and event-centered,

3. Language and political spaces as is suggested by the universal presence of deictic systems in the languages of the world. In English, for instance, “here/there” and “now/then” refer not to any clock time or geographical space but rather to a point of time/space defined in relation to the speech event in which these terms are uttered. Tense is temporally indexical in this way. Like tense, deictics of space can be transposed from the speaker onto objects in the physical and narrative surround. So we can talk not only about our own “front” and “back” but also that of trees, toasters and trucks; “here/there” can be centered inside the story we tell, as well as in the event of telling the story. Some cultural groups, such as the Guugu Yimithirr of Northern Australia have ⫺ in addition to such event-centric forms ⫺ what are called ‘absolute systems of coordinates.’ Instead of instructing a child to go to the left of a tree, they will tell him to go south or east of it (Haviland 1993). The invention of absolute space in the European tradition was a disputatious affair, linked to but not identical with the cartography of latitude and longitude discussed earlier. It has been singularly effective. Relying on the scientific tradition since Newton, most educated westerners assume an absolute abstract space which is believed to be “out there” as a constant set of fixed points, and to pre-exist our interactions and our representations. Phenomenologists, sophisticated Whorfians and most social scientists, by contrast, think that absolute space is itself a projection from deictically anchored, “origo-centered”, relative space (Levinson 1996). Terminologically, “space” is most often used ⫺ as it has been in this essay ⫺ to talk about these relatively abstract notions and their instantiation in semiotically organized cultural objects such as pictures and charts. By contrast, “place” is the term employed for the more immediate apprehension or experience of locality as a structure of feeling, mediated by bodily, ritual and ⫺ most ubiquitously ⫺ by linguistic practices (de Certeau 1984). Two questions arise from these distinctions: First, how do people project a deictic “origo” (“here”) of face-to-face interaction onto a more fixed physical location that is then considered their center, the “place,” and thus social home from which they look out on the rest of the world? Although this emplacing is not necessarily political, in the guise of “making community” it has surely functioned as the first step in many political projects. Second, how is a sense of locality sometimes displaced, so that the speakers consider themselves to be not at the center but at the margin or periphery of some social world to which they orient (either positively or negatively)? This second is surely a political-economic effect, the result of domination. By starting with the second question, we can see how linguistic ideologies and linguistic practices are crucial to both.

5.1. Displaced centers Linguists have long designated as “local” those languages or dialects that are demographically small, seem geographically isolated, culturally remote or exotic. Yet, the fact of “locality” in the contemporary world is not a matter of scale, space or strangeness in itself, but is a relational and contextual issue. No matter how remote, a linguistic form is reproduced by its practitioners with a self-aware sense, perhaps even an explicit theory, of what it is produced “from, against, in spite of, and in relation to.” In short, the local ⫺ while fragile and always in need of careful tending to assure its survival ⫺ is never separate or alone. It is always created by becoming the object of comparison to some “elsewhere” or center which is itself culturally apprehended. “Locality” emerges as

45

46

I. Introduction: Language and space a social fact in a national or globalized world (Appadurai 1996: 178, 184). We have already seen one example of this process as part of standardization, in which speech varieties are made “regional” or “deviant” once they are subsumed as the chronotopically distant or bottom of a standardized (metaphorical) space of linguistic varieties. In this same process, geography is always involved, as is political economy, but with surprising results. For many centuries, Europeans have perceived the continent as a cultural territory on which prosperity, progress and civilization are distributed in a gradient: concentrated in the west and decreasing as one moves east. Arguably, the political economic relations between European regions over the last three hundred years created the inequalities reflected in this gradient. The Cold War division of the continent solidified the differences. Currently, the cultural image itself has political and economic consequences. In the realm of language, the imagery helps to explain exceptions to the general rule that minority languages are indexical of local solidarity but less prestigious than the national languages of the states in which the minority is located. Through their codeswitching practices in the 1970s, the German-speaking minority in Romania demonstrated that for them Romanian was not a prestige language, although it was the language of state. Viewing themselves and their languages not from Romania, where they lived, but from the perspective of the political-economic-cultural gradient to which German-Romanians were finely attuned, they were persuaded that German trumped Romanian in prestige (Gal 1987). A more complex case comes from Indonesia. The Weyewa-speaking villagers on the island of Sumba were settled until the 1970s in village clusters. Each such cluster had a geographical center at which the greatest of rituals were performed, the ones at which the voice of ancestors would be heard through men’s performance of a complex genre of ritual speech. Spreading out from these centers like spokes on a wheel were other settlements that were increasingly less efficacious as the location of rituals, and at which only minor genres of ritual speech were performed. In the late nineteenth century, the largest of the rituals fell into disuse when the arrival of Dutch colonialists led to an increase in population, and a consequent dispersal of people away from the large central villages. The people found themselves removed from their own ceremonial centers. The net result has been a decrease in the importance and sheer presence of ritual speech, especially the most important kinds. An orientation to Dutch administration easily followed. The language itself, which was in the past considered “perfect” and “a source” of skill and knowledge, has been devalued because it is now seen as incomplete, lacking in elaborate ritual genres which were understood to be its most powerful and essential characteristics (Kuipers 1998).

5.2. Place-making through linguistic practices Locality is not simply relationality, however. It is important to analyze as well how people make a physical location or social group into a “place” that ⫺ in comparison to an “elsewhere” ⫺ is phenomenologically dense with meaning, familiar and legible for its inhabitants. The creation of places in this sense implies the making of subjects with patterns of action, ritual performances, ways of speaking and constructing events that are recognizable for others who are thereby identified as local subjects in the same place. The practices that make places are always in part linguistic, indeed a chronotope, and

3. Language and political spaces often named or ceremonialized. We can see in retrospect that the ethnography of speaking was concerned to document the linguistic details of just those everyday routines that “emplace” their practitioners (Bauman and Sherzer 1974). It is noteworthy that varied definitions of language are crucial in constituting place. In the example of Weyewa-speakers in Indonesia, ritual speech was defined as the most perfect Weyewa, the form without which the language was incomplete. In Northern California, among Native American Tolowa-speakers, the structural patterns that linguists consider the building blocks of language are of little importance. Recognition of each other as Tolowa-speakers, and the making of a dense and meaningful Tolowa “place,” comes from speakers’ knowing the Tolowa names of geological and geographical features in the landscape. Native speakers, linguists and government agents charged with maintaining this endangered language are in continual conflict because of these diverse definitions. Despite an ostensibly shared aim of maintaining the language, they work at cross-purposes (Collins 1998). Similarly, for the Western Apache, the names of features in the landscape are important as evidence of knowing the language. But even more significant are the specific narratives linked to landmarks, which operate as a communal, unwritten memorial of past activities and personages. It is the knowledge of these stories and the ability to judge when and how to tell them that constitute the knowledge necessary to be a complete person. The notion that “wisdom sits in places” (Basso 1996) is an explicit ethno-theory of the relationship between language and landscape that lends significance to everyday life, makes the social group visible to itself, and stands implicitly against American nation-state imagery. It is an exemplary case of “making place.” Knowing what to call a place is as an aspect of belonging, one that makes people members of a social group, recognizable as local persons. In the same way, the choice of designation for all those cities and regions in the world that have names in many languages (e.g., Bratislava, Pozsony, Pressburg) immediately signals one way of telling history as opposed to others and thereby a political stance for the speaker in a world of contrasting positions. This occurs in less obvious ways too, through the indexicality of reference that is a fundamental feature of all languages. Two English-speakers exchanging stories of their winter vacations must select lexical items to identify where they have been. The words ⫺ which need not be place names or even nouns ⫺ must be chosen from among the huge number that can accurately achieve reference. Has the speaker been “in the mountains”, “skiing”, “in Obergurgl”? The selection creates solidarity for interactants, or it can exclude. The first choice presumes the hearer and speaker know together which mountains count; the second conjures a world divided into practitioners of contrasting sports. The final one creates a “place” of connoisseurship which can either invite the other to celebrate a shared expertise or humiliate the hearer as ignorant (Schegloff 1972). Although place names are not indispensable for place-making, the loss of words for places can index political changes and show how dense, ritualized and embodied senses of place ⫺ as distinct from geography ⫺ can be erased. After the Second World War, land was collectivized in Hungary as part of the communist era. There were rarely any occasions on which to use the centuries-old proper names ⫺ specific to particular villages ⫺ for sections of arable land, meadow and woods. The names were remembered in the communist era by the men and women who had participated as youth in rituals of inheritance to which these names were crucial for identifying parcels of land. The

47

48

I. Introduction: Language and space names were also significant as archaic sounds suggesting an ancient rural continuity. Knowledge of the names identified elder generations who were able to evoke, through the casual use of them, their own former selves, values, customs and sense of place. The villages have remained, agriculture has continued and private property in land was restored at the end of communism. Men and women younger than 80 may have heard the archaic terms but cannot use them to refer to the land that has now regained its value. Those names index a profound political shift in the countryside, one that separates the “place” of the elders from that of co-resident, younger generations. The present owners of the land (often the same families) are creating place in novel ways.

6. Reerences Aarsleff, Hans 1982 From Locke to Saussure: Essays on the Study of Language and Intellectual History. Minneapolis: University of Minnesota Press. Anderson, Benedict 1991 Imagined Communities. London: Verso. Appadurai, Arjun 1996 Modernity at Large: Cultural Dimensions of Globalization. Minneapolis: University of Minnesota Press. Bakhtin, Mikhail 1981 The Dialogic Imagination. Austin: University of Texas Press. Basso, Keith 1996 Wisdom Sits in Places: Landscape and Language among the Western Apache. Albuquerque: University of New Mexico Press. Bauman, Richard and Joel Sherzer (eds.) 1974 Explorations in the Ethnography of Speaking. New York: Cambridge University Press. Bauman, Richard and Charles Briggs 2003 Voices of Modernity: Language Ideologies and the Politics of Inequality. New York: Cambridge University Press. Blommaert, Jan and Jef Verschueren 1998 The role of language in European nationalist ideologies. In: Bambi Schieffelin, Kathryn Woolard and Paul Kroskrity (eds.), Language Ideologies, 189⫺210. London: Oxford University Press. Bonfiglio, Thomas 2002 Race and the Rise of Standard American. Berlin, New York: Mouton de Gruyter. Bourdieu, Pierre 1991 Language and Symbolic Power. Cambridge, Mass.: Harvard University Press. Collins, James 1998 Our ideologies and theirs. In: Bambi Schieffelin, Kathryn Woolard and Paul Kroskrity (eds.), Language Ideologies, 256⫺270. London: Oxford University Press. Crowley, Tony 1989 Standard English and the Politics of Language. Urbana: University of Illinois Press. De Certeau, Michel 1984 The Practice of Everyday Life. Berkeley: University of California Press. De Swaan, Abram 2001 Words of the World: The Global Language System. Cambridge UK: Polity Press. Eisenlohr, Patrick 2006 Little India: Diaspora, Time, and Ethnolinguistic Belonging in Hindu Mauritius. Berkeley: University of California Press.

3. Language and political spaces Errington, Joseph J. 1988 Structure and Style in Javanese. Philadelphia: University of Pennsylvania Press. Gal, Susan 1987 Codeswitching and consciousness in the European periphery. American Ethnologist 14: 637⫺53. Gal, Susan 2006 Contradictions of standard language in Europe: Implications for practices and publics. Social Anthropology 14(2): 163⫺181. Geertz, Clifford 1980 Negara: The Theater State in 19th Century Bali. Princeton: Princeton University Press. Haviland, John 1993 Anchoring, iconicity and orientation in Guugu Yimithirr pointing gestures. Journal of Linguistic Anthropology 3(1): 3⫺45. Irvine, Judith T. and Susan Gal 2000 Language ideology and linguistic differentiation. In: Paul Kroskrity (ed.), Regimes of Language, 35⫺84. Santa Fe: School of American Research. Ho, Engsen 2006 The Graves of Tarim: Genealogy and Mobility across the Indian Ocean. Berkeley: University of California Press. Hobsbawm, Eric 1990 Nations and Nationalism since 1780, Programs Myths, Realities. Cambridge: Cambridge University Press. Jankowsky, Kurt R. 1973 The Neogrammarians: A Re-Evaluation of their Place in the Development of Linguistic Science. Washington, D.C.: Georgetown University Press. Kuipers, Joel 1998 Language, Identity and Marginality in Indonesia: The Changing Nature of Ritual Speech on the Island of Sumba. New York: Cambridge University Press. Levinson, Stephen 1996 Relativity in spatial conception and description. In: John J. Gumperz and Stephen Levinson (eds.), Rethinking Linguistic Relativity, 177⫺202. Cambridge: Cambridge University Press. Lanstyak, Istvan 1995 A magyar nyelv ko˝zpontjai. Magyar Tudoma´ny 10: 1170⫺85. Olender, Maurice 1992 The Languages of Paradise: Race, Religion, and Philology in the 19th Century. Cambridge, Mass.: Harvard University Press. Schegloff, Emanuel 1972 Notes on a conversational practice: Formulating place. In: David Sudnow (ed.), Studies in Social Interaction, 75⫺119. New York: Free Press. Silverstein, Michael 1996 Monoglot “standard” in America: Standardization and metaphors of linguistic hegemony. In: Donald Brenneis and Ronald Macaulay (eds.), The Matrix of Language, 284⫺ 306. Boulder: Westview. Silverstein, Michael 2003 Indexical order and the dialectics of sociolinguistic life. Language and Communication 23: 193⫺229. Stanzel, Franz K. (ed.) 1999 Europäischer Völkerspiegel: Imagologisch-ethnographische Studien zu den Völkertafeln des frühen 18. Jahrhunderts. Heidelberg: C. Winter. Tambiah, Stanley 1985 The galactic polity in Southeast Asia. In: Stanley Tambiah, Culture, Thought and Social Action: An Anthropological Perspective, 252⫺286. Cambridge, Mass.: Harvard University Press.

49

50

I. Introduction: Language and space Trautmann, Thomas 1997 Aryans and British India. Berkeley: University of California Press. Trautmann, Thomas 2006 Languages and Nations: The Dravidian Proof in Colonial Madras. Berkeley: University of California Press. Urla, Jacqueline 1993 Cultural politics in an age of statistics: Numbers, nations and the making of Basque identity. American Ethnologist 20(4): 818⫺843. Woolard, Kathryn and Bambi Schieffelin 1994 Language ideologies. Annual Review of Anthropology 23: 55⫺82.

Susan Gal, Chicago (USA)

4. Language and transnational spaces 1. 2. 3. 4. 5. 6. 7.

Introduction: Media and migration Deterritorialized spaces: Indexicality Reterritorializated spaces: Preservation vs. recombination Digital spaces Transidiomatic practices Conclusions References

1. Introduction: Media and migration One of the most significant aspects characterizing late modernity has been the high degree of space-time compression caused by the increasing mobility of people, commodities, texts, and knowledge (Giddens 1990; Appadurai 1996). This compression has transformed the geography of social relations and communication, leading many scholars to focus their studies on the transnational nature of late-modern communicative environments. These studies have linked the emergence of transnationalism with the post-industrial wave of migration, a wave characterized by people able to forge and sustain multistranded social relations across geographic, cultural, and political borders. Transnational migrants sustain a multiplicity of involvements in both home and host countries, a multiplicity made possible by their networks of interpersonal relationships. In a transnational space, social and symbolic ties reach beyond face-to-face relationships, connecting people who are involved in distant interaction through the same faith, class, ideology, ethnicity, or nationality (Basch, Glick Schiller and Szanton Blanc 1994; Faist 2000; Fist and Özveren 2004). As Pries (1999) pointed out, transnational spaces are characterized by triadic relationships involving the host state, the sending state, and the migrants’ networks. Structured by social networks, transnational spaces are multidimensional and multiply inhabited

4. Language and transnational spaces (Jackson, Crang and Dwyer 2004). Their inhabitants develop a “triadic geography of belonging” (Vertotec 1999) composed of their relation to place of residence, their creation of myths of homeland, and their imagination of the diasporic community. Transmigrants thus develop their sense of a place not only by engaging in their own surroundings (developing cultural and communicative competencies in relation to these surroundings), but also by activating wider flows and circuitries (allowing them to stay in touch with distant social realities and alternative social imaginations). Moreover, transnational spaces are experienced differently by people in different social groups and classes. As Massey (1994) discussed, transnational spaces can be imagined as articulated, hierarchical moments in networks of power relations and social understandings. The globalization of social relations is yet another source of geographically uneven development, giving rise to what Massey called a “global sense of place,” that is, the sense that one’s spatial-temporal coordinates are constantly redefined by economic, political, and socio-cultural relations which are stretched out over the planet at every different scale, from the local to the global. But how are the “global sense of place” and the “triadic geography of belonging” enacted, produced, and reproduced? In order to answer this question, we need to explore the communicative regimes in charge of injecting sense into the transnational situation. New research is needed that focuses directly on the development across transnational territories of communicative practices produced (and reproduced) by two social actors: the transnational migrants themselves and the local and global media. Global communication, and in particular electronic media, is much more than an enabler of people’s interactivity and mobility: it alters the very nature of this interactivity, transforming people’s sense of place, belonging, and social relations. Furthermore, electronic media provide people with distinct ways of expressing themselves ⫺ “media idioms”, in other words, which have yet to be fully explored in transnational studies. But it is not only transnational studies that have neglected this phenomenon. Language studies have also been slow to respond to the challenge of examining transnational media idioms, although the literature on language and migration is well-developed (and growing). The study of electronic media has been left to the socio-cultural anthropology of media (the product of a serious rethinking of the field of visual anthropology; see Ginsburg, Abu-Lughod and Larkin 2002) and to the more “ethnographic” approaches to cultural studies (which emerged out of the encounter between sociological analysis of the media and textual analysis, see Baudrillard 1984; Gitlin 1983; Mankekar 1999). However, Spitulnik’s critique (1993) still stands: these studies suffer from inattention to linguistic details and an overdependence on textual criticism. These studies, as well as the study of the Internet in diaspora, which has attracted considerable interest in media and cultural studies, had paid little attention to sociolinguistic issues (Androutsopoulos 2006a, 2006b). Even linguistically oriented media analysis, which focuses principally on the lexical components of media messages, has neglected to analyze the indexical/pragmatic aspects of media communication and the global spread of media idioms. It is only in recent years that language studies ⫺ spurred by the growth of cultural studies (Bhabha 1994; Garcı´a Canclini 1995; Hannerz 1996; Jameson and Miyoshi 1998; Tomlinson 1999; Nederveen Pieterse 2003) ⫺ started to engage with some of the concepts developed by the analysis of globalization and produced some significant scholarship (Woolard 1999; multiple contributions in Coupland 2003; Aravamudan 2006; Pennycook 2007). Unfortunately, even in this more recent scholarship, the link between media and migration is still poorly understood.

51

52

I. Introduction: Language and space In this article, I seek to remedy this situation by proposing a research paradigm that can account for both migrants’ communicative practices and media idioms. After discussing the three basic characteristics of transnational spaces (deterritorialization, reterritorialization, and digitalization), this article introduces the concept of transidiomatic practices to describe the communicative practices of transnational groups that interact using different languages and communicative codes simultaneously in a range of communicative channels. This focus on transidiomatic practices leads to a discussion of the social indexicalities and semiotic codes structured by (and structuring) the social zones of late-modern communication.

2. Deterritorialized spaces: Indexicality In the 1990s, the anthropological notion of “culture” gradually moved away from identification with a naturally contiguous territory or society and towards an association with social environments freed from the constrictions of face-to-face interactions in the localities of pre-modern societies. This move allowed anthropologists and social scientists to study social relations stretched across time and space by the movement of people between places, institutions, and groups (Gupta and Ferguson 1992, 1997; Giddens 1990). This anthropological turn was heavily influenced by the philosophical concept of deterritorialization developed by Deleuze and Guattari (1983, 1987). The notion of deterritorialization provided an alternative vision of the subject that contrasted with the dominant understanding of subjectivity as contained within the territorial confines established by centralized powers. Deleuze and Guattari explicated the displacement and dispersion of a subjectivity unrestrained by territorial control. Deterritorialization served as the cornerstone of a “nomadic” theory, in which the “nomad”, “migrant” and “gypsy” became the figures for a generalized poetics of displacement. This concept has been criticized for its metaphorical, universalistic bias (Mignolo 2000) as well as for its inability to account for transnational power relations and its indulgence in nostalgic, colonial fantasies about nomadic movements (Spivak 1988; Kaplan 1996). Nevertheless, it remains a powerful metaphor for representing globalization’s central social fact, the dissolution of the supposedly “natural” link between a geographical territory and cultural practices, experiences, and identities. As such, the concept of deterritorialization has been adopted by social theory to account for the cultural dynamics of people and practices that either no longer inhabit one locale (finding themselves in borderlands, diasporic groups, or mixed cultural environments) or inhabit a locality radically transformed by global cultural phenomena (Basch, Glick Schiller and Szanton Blanc 1994; Hannerz 1996; Clifford 1997; Tomlinson 1999a and b). Through the synergy of mobile people and electronic media, the vast majority of the planet is confronting new rules and resources for the construction of social identity and cultural belonging. As Appadurai (1996) pointed out, when the rapid, mass-mediated flow of images, scenarios, and emotions merges with the flow of deterritorialized audiences, the result is a recombination in the production of modern subjectivity. When Moroccan families make videotapes of their weddings to send to relatives who migrated to Italy (Jacquemet 1996), when Hmong refugees in the United States produce documentaries about their “homelands” by staging them in China (Schein 2002), or when Paki-

4. Language and transnational spaces stani taxi-drivers in Chicago listen to sermons recorded in mosques in Kabul or Teheran (Appadurai 1996), we witness the encounter of mobile media practices and transnational people. In this encounter, a new, deterritorialized social identity takes shape, light-years away from the corporate logic of the nation-state. This new identity coagulates around a sentiment of belonging that can no longer be identified with a purely territorial dimension. The media, by operating in the interface between global forces and local processes, have produced an intense acceleration in the dynamics of social interaction ⫺ an acceleration that does not progress along straight lines, but follows the centrifugal, rhyzomatic, and chaotic interweaving of multiple channels, voices, and audiences. What is most impressive in the cultural/linguistic landscapes of late modernity are the communicative practices, such as those described above, of people engaged in simultaneous interactions across a multiplicity of technologically mediated sites. In the deterritorialized world of late-modern communication, interlocutors can no longer take the spatio-temporal contexts of an interaction for granted, as a result the indexical information of any message becomes problematic. By indexicality, I mean the existential relationship between a sign (called index) and its referent, based upon a physical link of “spatio-temporal contiguity” between the two. Indexes, in order to be interpreted, need a clear spatio-temporal context, and it is only thanks to context that we can process statements such as pass me that book, since the demonstrative pronoun “points” to a book nearby. This context-dependent property of an index is not only found in demonstratives, personal pronouns, and temporal and spatial expressions, but also in sociolinguistic markers such as phonological variations (pointing to a particular relationship between speaker and socio-economic position) or pronominal choices (the T/V, informal/formal system pointing to contextually relevant social coordinates of respect, politeness, and hierarchy). The central property of indexicality is thus this ability to “point” communication toward the spatio-temporal contexts and the socio-cultural environments, to direct an addressee to attend to speech settings (for a more detailed discussion, see Silverstein 1976; Hanks 1990). In this logic, long-distance phone calls require a realignment of the caller’s indexical ground in order to match the time zone of the called to avoid embarrassing, face-threatening acts (such as waking the called in the middle of the night or interrupting morning prayers). This adjustment of one’s phenomenological field carries consequences for one’s sense of place, in some instances producing radical reframings of what constitutes “being here.” These translocal requirements have been complexified by the widespread diffusion, starting in the 1990s, of mobile telecommunication technologies (especially cellular technology), which have the capacity to produce a simulacrum of interactional contiguity no longer depended upon physical proximity. Nowadays, a considerable amount of interactional work must be spent in aligning the indexical maps of interlocutors, i. e., their spatial references contextualized in the particular environments occupied by the speakers. Consider, for instance, the case of calling a cellular phone, in which the precise location of the call’s recipient is obviously problematic. The caller’s typical opening move (where are you?) in this case is not a simple request for territorial information, but the start of a sophisticated process of matching indexical maps. It also initiates a communicative negotiation in which the caller seeks to find out whether the environment of the called is conducive to talk, thus avoiding potential face damage if the call has intruded either upon a private, “back-stage” area not usually open to interaction, or upon public

53

54

I. Introduction: Language and space space where the recipient’s talk would be limited by social rules. It must be noted that the “public/private” divide is rapidly being transformed by mobile technology, which stretches the range of permissibility and possibility for phone calls, as anyone who has taken a train next to a babbling cellularized individual knows (see Cameron 2000 on the demise of Goffman’s “civil disattention” caused by cell phones). This in turn has created new codes of conduct and new rules for the use of cellular phones (including their banishment from concert halls, lecture halls, and some restaurants). The hermeneutical implications of the link between indexicality and deterritorialized communicative practices are even more dramatic if we consider the phenomenon of reading texts accessed through the Internet. Let me relate a recent experience that I had while reading the on-line version of the Italian newspaper Il Manifesto. An article by the paper’s US correspondent on a socialist scholars’ conference held in New York discussed similarities between American and Italian debates over the conference topics: La discussione sulla crisi della forma partito, sull’egemonia dell’ideologia del mercato, sulla latitanza dei movimenti, tutto risuona in un ritornello tante volte sentito al di qua dell’Atlantico. ‘The debate over the crisis of the party-structure, on the hegemony of market ideology, on the lack of social movements, all this sounds like a refrain heard many times on this side of the Atlantic.’ (my translation, M. J.)

As an Italian-speaking transmigrant working and living in the San Francisco Bay Area, I felt a deep sense of estrangement when I attempted to anchor the phrase on this side of the Atlantic within his spatial environment. The article assumed the location of writer and audience to be in Italy (even if the writer was reporting from New York). It also assumed the location of the medium: Italian-language media were produced in Italy to be consumed locally by an Italian public (very much in line with Anderson’s [1983] understanding of the role of media in the construction of the nation-state as an imagined community). A transmigrant’s reading of this article in a location outside of Italy (let’s say an office on a U.S. campus) problematizes this easily taken-for-granted connection between territory, language, and media ⫺ resulting in a serious case of indexical vertigo. This sense of vertigo is intrinsically related to the nature of reading on-line. The consumption of foreign newspapers is not a recent development of globalization, but the cultural practice of accessing them through “traditional” means (going to a specialty newsstand carrying foreign journals, locating the newspaper, checking its date ⫺ usually at least a day earlier ⫺ buying the newspaper, holding it at eye level, leafing through the pages) evokes a physicality that allows the reader to participate in the media event “as if” one were still in Italy ⫺ and thus able to process “on this side of the Atlantic” easily and without confusion. On the other hand, the consumption of online newspapers transforms the phenomenological condition of “being here.” The fact that the reader could be located almost anywhere in the world (at least, anywhere with reliable Internet connection) produces a different processing of spatial and temporal locators. The online versions of newspapers come with their own modalities of reception that clearly differentiate them from the printed versions (which are, in many cases, also available for downloading as pdf files from the papers’ websites, to make things even more confusing). By relying on specific techniques, such as the use of hyperlinks to other publications, websites, blogs, or image and video banks; by providing comment space for on-going discussion among readers;

4. Language and transnational spaces and by using specific markers, such as providing the time of the last update (usually given in GMT), online newspapers alter the taken-for-granted relationship between reader and text and project a globalized but language-specific audience, able to access the text from virtually anywhere in the world. Moreover, given the global dominance of American operating systems (such as Windows, which in 2006 had a 84.7 percent share of all computer operating systems, and Internet Explorer, which had a 78 percent share of Internet traffic), even non-English e-journals accessed from an English-based environment are entextualized within English frames. These frames shape the phenomenological appreciation of on-line texts. In sum, online media allow the experience of reading to be purely deterritorialized from any point of geographical reference, and thus problematize indexical understanding of electronically produced and disseminated textual materials. Practices of communicative deterritorialization, such as cellular phone calls or online reading, are usually found at the nexus of mobile people and mobile texts. The most important new phenomenon facing the study of late modernity is, as Appadurai (1996) discussed, the encounter between media and migration ⫺ an encounter that provides an ideal entry point for studying the social and cultural mutations of the contemporary world. Ethnographers of communication, in particular, have much to gain and to contribute by turning their attention to processes of deterritorialization, especially those produced by the interaction of mobile people and electronic media. A linguistic analysis of media communication would provide new evidence of the indexical dynamics of such phenomena as multi-channeled communication (such as a cellular call to a radio talk show about events which the caller is experiencing directly, Scollon 1998: 2001), asynchronous interactions (such as voice-mail, email, or cellular messages), the linguistic ideology of media idioms, metacommunicative techniques (as in the use of techno-political devices in broadcast interviews), intertextuality and the construction of audiences, and the creolization of media idioms. Finally, a focus on deterritorialization should prompt linguistic studies to examine electronic media in the context of the late-modern human condition, and confront the phenomenon of people whose allegiances and sense of belonging are divided between multiple places ⫺ and whose linguistic practices reflect their experience with multiple, transnational sites. At the same time, a focus on processes of deterritorialization should be combined with attention to parallel processes of reterritorialization.

3. Reterritorializated spaces: Preservation vs. recombination The most important social implication of deterritorialization is not the dissolution of identities, cultures, or nation-states in a global environment (as some critics of globalization theory would lead us to believe; see Held and McGrew 2000; Lechner and Boli 2000), but the interplay between global and local processes and the reconstitution of local social positionings within global cultural flows (a phenomenon labeled “glocalization,” see Robertson 1995). People are continuously involved in practices of reterritorialization: the anchoring and recontextualizing of global cultural processes into their everyday life (Tomlinson 1999). These practices have a broad range of social effects. At one extreme, they can produce an ideological hardening of the local, “indigenous” identity/

55

56

I. Introduction: Language and space code/language in opposition to translocal phenomena. At the other, they can just as well initiate a much more creative process for the production of recombinant identities. Social groups faced with transnational flows of people and cultural practices not easily accepted locally may react with an ideological hardening of the social boundaries of their “community.” Locally dominant ethnic groups strengthen in-group identities by raising the membership bar through practices of intolerance and exclusion. As noted by Garcı´a Canclini in his ethnography of Tijuanians, “the same people who praise the city for being open and cosmopolitan want to fix signs of identification and rituals that differentiate them from those who are just passing through.” (1995: 239). Among the most pernicious practices is the imposition by socially dominant groups of limits on the linguistic rights of transnational, lower class, or minority subjects. From the English Only movement to the worldwide phenomenon of eradicating minority languages in favor of national ones, we find a motley crew of different forces and groups animated by the desire to resist globalization. Their resistance takes various and at times antagonistic forms, such as preserving “a common language,” avoiding ethnic strife, imposing a sense of national unity and civic responsibility, or exploiting a national mood of isolationism and xenophobia. An interesting example of this process of linguistic chauvinism by nation-states is offered by Singapore’s initiatives for a perfect society: the “Speak Good English Movement” and the “Speak Mandarin” campaigns to eliminate Singlish, “a patois that has spread through our nation like a linguistic virus” according to Lee Kuan Yew, Singapore’s senior minister (quoted in the New York Times, July 1, 2001). In particular the “Speak Mandarin” campaign has become a serious class struggle pitching Singapore’s middle class and state bureaucrats against the working poor, and is causing the shift in most domains, in particular the home, from other Chinese languages to Mandarin, producing a generation of children not being able to understand their grandparents (Ansaldo and Lim 2006). Meanwhile, minority groups may respond to globalization fears with a strategic retreat to ideologically defensive positions, such as re-identification with cultures of origin, reliance on symbolic membership in strong counter-ethnicities, revival of cultural integralism and traditionalism, and defense of the “purity” and “integrity” of the local language (Hill and Hill 1986). In all these cases, we find people who, feeling threatened by the diversity (among other unsettling changes) brought about by deterritorialization, activate an exclusive linguistic ideology to raise the membership bar (Anderson 1983; Crowley 1989; Crawford 1992; Silverstein 1996; Errington 2000). On the other hand, there is some evidence that global/local interactions may also produce a new form of reterritorialization that gives rise to recombinant identities, usually produced through encounters between global and local codes of communication. Anyone present in transnational milieus, whose talk is mediated by deterritorialized technologies and who interacts with both present and distant people, will find herself immersed in practices of recombinant identities. Diasporic and local groups alike renegotiate their identities whenever they engage in maintaining a simultaneous presence in a multiplicity of sites or participate in elective networks spread over transnational territories. These recombinant identities are based on multi-presence, multilingualism, and decentered political/social engagements. Consider, for instance, the growing phenomenon of customer call centers located thousands of miles away, multiple times zones apart, and in different nations from the

4. Language and transnational spaces serviced area, such as the call centers in South Asia, in particular in India, that serve customers in the United States or United Kingdom. The key to the success of Indian calling centers is, for the most part, the phone operators’ ability to reterritorialize their cultural and communicative practices to match the callers’ expectations of being served in a nearby location by a peer. In a quest for seamless connection with clients, these operators ⫺ most of whom are young college graduates ⫺ are required by their managers to study American or British popular culture, including food, habits, and popular TV shows in their customers’ areas (one reported exercise consists of listening to the soundtrack of the American TV-show Friends and the like, and then reconstructing the dialogue). Most importantly, operators’ talk must be contextualized and experienced within the spatio-temporal environment of their customers. For instance, operators’ computer screens show not only a customer’s time zone but also the local weather and traffic report, so that each call can be answered with the appropriate temporal greeting (“good afternoon” when India is already in the dark) and with small talk about, for instance, the miserable snow storm and the resulting horrible commute (Landler 2001). At the same time, Indian operators interact with co-workers in the next cubicle, take breaks to eat local food, and may occasionally check local news and personal email. This mixing of multiple languages and simultaneous local (with co-workers) and distant (with clients) interactions binds these subjects to a world of recombinations of their communicative, cultural, and social practices. Many other countries have followed in India’s footsteps and developed such call centers. In spring 2001, the New York Times published an article about how Ireland was becoming the phone message center for many German firms. Germans (from East Germany, where there is still high unemployment) went to Ireland to work and answer phone calls from German customers. Since it was too expensive to pay Germans German wages in Germany, firms paid workers Irish wages in Ireland, and saved money. America Online, on the other hand, currently prefers an Asian country with an English-educated population: the Philippines. Its voice hotline (800 number) takes you to an office somewhere in the Philippines, where all questions are answered by local people. (On the other hand, a more recent development reported in the press in 2007 is India outsourcing its economic model to countries like the Czech Republic, Morocco and China). While individual creativity must be acknowledged in this process of recombination, I do not want to idealize phone operators’ agency (both social and communicative) in working tasks which are imposed upon them by managers, together with often substandard and cramped working conditions. The imperative to speak proper English still points to the linguistic dominance of this language over many local languages, as contemporary capitalism pushes further into all areas of production worldwide. Political, social and cultural supremacy is in the hands of transnational governmental bodies, multinational companies, international relief organizations and churches, and various multinational military forces ⫺ in other words, in the hands of a globalized governmentality characterized by “mobile sovereignty” (Sassen 2001; Cheah and Robbins 1998; Ong 1998; Hardt and Negri 2000). Mobile sovereignty is tied to the emergence of two new layers in the international division of intellectual labor: a class made up of cosmopolitan elites (such as multinational corporate executives, UN bureaucrats, the staff of international NGOs, and international media producers) and a class made up of the local workers (phone operators, secretaries, computer technicians, interpreters, local politicians, and so on) under the

57

58

I. Introduction: Language and space elites’ direct or indirect control. Phone operators still occupy the lowest echelon in the hierarchy of cognitive labor: they are the “chain workers” of the information system, servicing the needs of others under the direct supervision of “symbolic analysts,” the “brainworkers” in charge of the production of immaterial goods and services (Reich 1992; Gee, Hull and Lankshear 1996; Ritzer 1996; Escobar 2001). Among the skills most desired in these chain workers, knowledge of global languages takes center stage. Such linguistic knowledge constitutes the best ⫺ and sometimes the only ⫺ opportunity available to many bright people (especially the youth in the global south) for social and geographical mobility. Nevertheless, these phone operators represent a new, moving frontier in class, language, and power relations. As such, their communicative practices have a significant impact on their everyday life and in the lives of people with whom they interact, both near to them and in the deterritorialized environments of the virtual economy. They, like many other workers involved in cognitive labor, are at the frontline of contemporary capitalism’s advance upon local environments, and on the cutting edge of the creation of new ways of speaking and communicating. Their communicative practices are an example of how new discourses and modes of representation are reterritorialized within the local environment and situated amid the hierarchies that structure the global political economy. In this light, the concept of reterritorialization provides a valuable perspective on intercultural communication in the transnational spaces of late modernity.

4. Digital spaces The third, and most neglected, dimension of transnational spaces is situated not in specific territorial sites but in the virtual reality of mediascapes. The transnational sensibilities and multilingual communicative dynamics that we have explored so far in this paper are made possible by global media and digital technologies. This section addresses communicative implications of three types of digitally-mediated interactions: virtual meetings, digital navigation systems and online social networking. Although a rich literature on virtual teams and their role in transnational spaces has emerged in recent years from the field of organization studies (Hinds and Kiesler 2002; Marquardt and Horvath 2001), very few studies have focused on the communicative aspects of these virtual interactions, in particular on the communicative dynamics of telepresence in videoconferencing or video calls. In the late 1980s, Virilio introduced the concept of telepresence to account for the spacializing role images would have in longdistance video communication (1988, 1990). Interactive, live transmission of video images over great distances becomes in itself a new kind of place. The experience of being in this place is what Virilio called telepresence, which supersedes in real time the real space of objects and sites. Telepresence is the experience of the continuity of real time overcoming the contiguity of real space. The impact of fiber optics, monitors and video cameras on our vision and on our surroundings is far greater than that of electricity in the nineteenth century. “In order to see,” Virilio observes, “we will no longer be satisfied in dissipating the night, the darkness of the outside world. We will also dissipate time lapses and distances, the outside world itself” (1990: 72).

4. Language and transnational spaces In consonance with Baudrillard’s understanding of the new informational landscape, Virilio advanced the notion that we no longer inhabit or share a physical public space. Our domain of existence or socialization is now the public image, with its volatile, functional and spectacular ubiquity. For Virilio, one of the most important aspects of the technologies of digital imaging and of synthetic vision made possible by optoelectronics was the “fusion/confusion of the factual (or operational) and the virtual,” the predominance of the “effect of the real” (1988: 128) over the physical world. The shortest distance between two points is no longer a straight line, as it was in the age of the locomotive and the telegraph. Today, in the age of satellites and fiber optics, the shortest distance between two points lies in real time transmission of information between the two points. The concept of telepresence has been picked up by the computer industry to describe the latest innovation in videoconferencing. In this virtual environment, users scattered around the planet communicate live via integrated voice, video and data networks, at a speed and quality of transmission paralleling face-to-face interactions. High-definition, multi-panel video screens reproduce the actual size of conference participants. This technology creates the illusion that the two (or more) parties to a conversation are not continents apart but at opposite sides of the same table. Initial reports suggest that participants in such meetings very quickly forget, or at least stop caring, that they are not in the same room. In order to achieve this telepresence effect, rooms in remote locations are equipped with matching conference tables that face wide screens, as well as matching de´cor, so that the local table blends seamlessly with the remote table projected on the screen. The rooms are equipped with multiple cameras and speakers, supported by enormous computing power, so that meeting participants can make eye contact and overlap or interrupt each other (in order to do so, the delays in audio and video transmission must be negligible, i. e., below 250 milliseconds, the threshold at which the human brain starts to notice the delay, see Cisco 2007). While meetings that make use of telepresence video conference technologies may be the wave of the future, many virtual meetings that take place today combine use of teleconferencing with the Internet to facilitate collaboration on shared projects, but do not simulate face-to-face interactions. Participants are still present to each other, nonetheless, as voices on the phone ⫺ and thus experience a form of telepresence. The only ethnography of communication to date that has explored telepresence focused on this sort of teleconferencing. Wasson (2006), in her study of the meetings of virtual teams in a corporate workplace, analyzed how participants in such teleconferences multitasked in more complex ways than they would during face-to-face interactions. As in face-toface interactions, a central situational focus was maintained within the interactional space of the virtual meeting, but participants were also simultaneously located in their local spaces. In their local spaces, they were often simultaneously engaged in other activities unconnected to the meeting. These multitasking skills seem to be a common theme of all digitalized environments, pointing at the complex communicative practices occurring simultaneously in the transnational space created by local and virtual environments. A second type of digitally-mediated interaction is the use of digital communications to provide directions for spatial navigation. Digital navigation has given rise to a new way transnational agents can move in unknown territories and thus acquire the appropriate indexical pragmatics. Since 2000, in-dash navigation systems using Global Positioning System (GPS) technology have become either standard or optional features in many car and truck models, and third-party systems have become widely available for

59

60

I. Introduction: Language and space all vehicles. Portable units have been developed that could be moved from car to car, and navigation software can now be added to laptop computers. Navigation systems read digital maps on a CD-ROM, DVD or hard disk and display on a screen one’s current location on a street map in real-time. Using GPS devices mobile, mobile people, and in particular transnational ones, no longer need clear knowledge of the territory in which they move ⫺ knowledge that was, until recently, one of the signs of belonging to a “home” territory. Now cellular communication allows them to rely on remote, digital interlocutors to provide the pertinent information. This offers great advantages to deterritorialized subjects. Consulting maps or obtaining directions in advance is no longer necessary. Instead, drivers can count on real-time directions while driving. A digital navigator can match its knowledge of the territory with the driver’s progress through it ⫺ duly described to her by voice (currently available only in “global” languages) or via a graphic interface. A related technology, mobile positioning, allow services to pinpoint the location of a mobile caller or vehicle in transit. These location-based services (LBS) are used for emergency purposes as well as enhanced business applications such as location-sensitive billing, traffic updates, fleet management and goods and people tracking. This technology is also expected to become ubiquitous, in particular in conjunction with cellular telephone communication. For instance, a person could be notified by cell phone of a sale going on in a store he is walking by, find out whether any of his friends are in the area, or track the location of his children’s cell phones. In particular, the use of such technologies for social networking should become an area of concern for scholars interested in social interaction, and in particular in human-machine interactions. A third type of digital communication has an especially powerful impact on transnational spaces: the social interactions taking place in social networking sites (such as MySpace or Facebook) and virtual worlds (such as Second Life). Social networking sites are becoming one of the primary communication channels for many young adults, and the numbers are staggering: in summer 2007 MySpace counted 200 million users, and the rapidly expanding Facebook reached the 40 million mark, having added 36 million users in just over a year (one of the fastest rates of growth ever recorded by an online service). In these environments, people, mostly young, learn to expand their cultural and communicative capacities by constructing online subjectivities in an open-ended process of becoming. These virtual communities are sites for the expansion of the cultural, communicative, and subjective capacities of their users, who are engaged in an exponential expansion of discrete nodes of both affect and affinity. They do so by allowing users to communicate with people worldwide through the creation of their personal space, where users can personalize their profile pages as they wish with texts, images, music, varying layouts and various other items (such as links to other websites). According to Gamble (2007), the communicative world of the Internet and other digital domains is doing much more than just developing in a manner that replicated communities of practice based on speech patterns: the communicative practices played out in these virtual communities are advancing into the domain of speech in the physical world. It is now possible to determine which speaker has an active digital life merely by exploring his/her vocabulary and sentence construction. In other words, the digital domain’s contribution to the evolution of language does much more than just blending the virtual world with the physical one (as suggested by Dube´, Bourhis and Jacob 2006):

4. Language and transnational spaces digital communication is now a home base for the development of language for a broad section of the general population. This is clearly visible in the communicative practices emerging from Facebook’s communities. Facebook is proving to have a great impact on the offline world through its offer of communicative tools, such as the “wall” and the “poke”. The wall is a space on each user’s profile page that allows friends to post messages for the user to see. One user’s wall is visible to anyone with the ability to see their full profile, and different users’ wall posts show up in an individual’s News Feed. Many users use their friends’ walls for leaving short, quick notes, whereas more private communication is produced in a different environment, called “Messages”, where notes are visible only to the sender and recipient(s), much like email. The act of constantly browsing through friends’ pages, pictures and walls, but doing so without posting a comment is referred as “creeping”. In July 2007, Facebook allowed users to post attachments to the wall, whereas previously the wall was limited to textual content only, thus increasing the multimodality and digital complexity of the communication. The “poke” feature, on the other hand, allows a user to attract the attention of another one. Friends often engage in what is known as a “poke war”, where the poke is exchanged back and forth continuously between two users by using the “poke back” feature. The user who neglects to return the poke promptly while still remaining an active user of Facebook is said to be the loser (Wikipedia 2007). It is now frequent to encounter uses of “poke”, “wall posts”, or “creeping” in youth speech patterns which could be understood as markers of cultural belonging to this online community. Online gaming and its capability to produce real time virtual communities deserves a closer look. Every day millions of pc gamers log in to online worlds to play and interact within a virtual community. Massively multiplayer online role playing games (MMORPGs) ⫺ such as Second Life, Ultima Online, World of Warcraft, or Everquest ⫺ allow people to don online personalities, or avatars, and duke it out in make-believe environments. These environments typically appear similar to those in the real world, with similar rules for gravity, topography, locomotion, and communication. Communication has, until recently, been in the form of text, but recently real-time voice communication via the internet has become available. For instance, in the simulation game Second Life, not only can entrepreneurs buy and sell digital real estate, create their own lines of avatar clothing and accessories, and hold virtual concerts, lectures, and sporting events; but in addition many users have come to see these games as an enhanced communications medium for staying in touch with friends and to make new connections. The creation of elective communities spread across continents and able to interact in real time in the virtual confines of cyberspace is pushing the limits of what constitutes social interaction and group communication. While by now the study of computer-mediated communication (CMC) is a well established field (Herring 1996, 2003; Crystal 2006), the indexical-pragmatic functions of CMC in transnational spaces have been largely overlooked. Moreover, existing research on CMC has focused almost exclusively on emergent practices in English whereas, as Danet and Herring (2007) pointed out, roughly two-thirds of Internet users are nonEnglish speakers (one notable exception is the study of code-switching among expatriate South Asians on Usenet in Paolillo 1996, see also the edited volume by Danet and Herring 2007). What are the communicative practices taking place in transnational, digitally-mediated environments, from virtual meetings to social networking sites? Ethnographically-

61

62

I. Introduction: Language and space based, linguistically-oriented fieldwork of digital transnational spaces is still quite sketchy and lacks in-depth analysis. As a result we are left to speculate on the complex communicative dynamics people activate when engaging in digital interaction. As Jones (2004) pointed out in his study of the shift from face-to-face to virtual interaction “traditional sociolinguistic conceptualizations of the terms of interaction and the contexts in which it takes place may need to be radically rethought in light of new communication technologies” (Jones 2004: 21). This awareness of a transformed context must inform our investigation of these spaces. The lenses we usually adopt in looking at language must be significantly altered to accommodate for communicative phenomena produced by people present in transnational contexts, whose talk is mediated by deterritorialized technologies, and who interact with both present and distant people. We need to study the communicative practices of these subjects, even if these communicative practices cannot be recognized as part of a single linguistic standard. To elucidate a different approach to the study of language in transnational space, the next section will introduce the concept of transidiomatic practice to study precisely this intersection between mobile people and digital communication.

5. Transidiomatic practices One of the most significant breakthroughs in language studies in the late twentieth century was the introduction of the notion of communicative practice. Under the influence of European political philosophers such as Foucault and Bourdieu, language and communication scholars adopted the notion of practice to deal not only with communicative codes and ways of speaking (some of the rallying concepts of the first wave of the ethnography of communication), but also with semiotic understanding, power asymmetry and linguistic ideology. By focusing on the “socially defined relation between agents and the field that ‘produces’ speech forms” (Hanks 1996: 230), a practice-oriented approach can then explore speakers’ orientations, their habitual patterns and schematic understandings and their indexical strategies. Hanks defined communicative practice as constituted by the triangulation of linguistic activity, the related semiotic code or linguistic forms and the ideology of social indexicality. He invoked a poetic image of practice as “the point of conversion of the quick of activity, the reflexive gaze of value, and the law of the system” (1996: 11). This triangulation of linguistic activity, semiotic codes and indexicality needs to be complexified to account for how groups of people that are no longer territorially defined, think about themselves, communicate using an array of both face-to-face and long-distance media, and in so doing produce and reproduce social hierarchies and power asymmetries. I propose to use the term transidiomatic practice to describe the communicative practices of transnational groups that interact using different languages and communicative codes simultaneously present in a range of communicative channels, both local and distant. Transidiomatic practices are the result of the co-presence of digital media and multilingual talk exercised by deterritorialized/reterritorialized speakers. They operate in contexts heavily structured by social indexicalities and semiotic codes that produced

4. Language and transnational spaces relatively stable power asymmetries and cultural hegemonies. Anyone present in transnational environments, whose talk is produced by both biological and digital means, and who interacts with both present and distant people is engaged in transidiomatic practices. Given the nature of economic globalization, many contemporary work environments ⫺ from the offices of international organizations to airport lounges, from international call centers to the board meetings of multinational companies ⫺ can be classified as transidiomatic. In addition, a great number of social settings ⫺ from living rooms to hospital operating rooms to political meetings ⫺ experience a translocal multilingualism interacting with the electronic technologies of contemporary communication. The world is now full of settings where speakers use a mixture of languages in interacting with friends and co-workers; read English and other “global” languages on their computer screens; watch local, regional or global broadcasts; and listen to pop music in various languages. Much of the time, they do so simultaneously. Moreover, transidiomatic practices are no longer solely restricted to areas of colonial and post-colonial contact, but flow through the multiple channels of electronic communication and global transportation, over the entire world, from contact zones, borderlands, and diasporic nets of relationships to the most remote and self-contained areas of the globe. These communicative resources are activated by people needing to operate in multiple, co-present and overlapping communicative frames. The language they use to communicate depends on the contextual nature of their multi-site interactions, but is necessarily mixed, translated, “creolized” (Hannerz 1992: 1996). Transidiomatic practices usually produce linguistic innovations with heavy borrowing from English (a reminder of the global impact of contemporary English ⫺ see Pennycook 2003 and 2007; Crystal 1997; De Swaan 2001), but any number of other languages could be involved in these communicative recombinations, depending on specific processes of reterritorialization in which the speakers are engaged. Through transidiomatic practices, diasporic and local groups alike recombine their identities by maintaining simultaneous presence in a multiplicity of sites and by participating in elective networks spread over transnational territories. These recombinant identities are based on multi-presence, multilingualism, and decentered political/social engagements. While individual creativity must be acknowledged in this process of recombination, I do not want to idealize the agency (both social and communicative) of most people involved in transidiomatic practices. These practices are still inserted into a global indexical order which assign superior values to certain systems of communication at the expenses of others. The fact that most aspiring musicians feel now compelled to have a presence on MySpace, that managers will soon have to endure interminable videoteleconferencing with long-distant bosses, or that the lack of knowledge on how to compose a proper e-mail message could restrict people’s work opportunities (Shipley and Schwalbe 2007) are all indications that communicative inequalities will continue to shape power relations in the age of globalization. On the linguistic level, we will still encounter a stratified, layered ideology of what to consider a legitimate communicative code; and this legitimacy will be in the hands of the mobile sovereignty discussed above, which is already increasingly preoccupied in assessing what kind of transidiomatic practices are welcomed (such as those of Indian phone operators forced to speak the local English of the area they serve) and what are considered “broken English” or gibberish.

63

64

I. Introduction: Language and space An area of particular sociolinguistic interest for the study of transidiomatic practices resides in tracing the transnational flow of particular discourse markers from one communicative environment to another. Elsewhere (Jacquemet 2005), I called such discourse markers transidiomatic floaters. In that article, I traced the drift of the transidiomatic floater don’t worry over multiple nation-states: from its English context to Italian advertising language (where it was manipulated into representing a testimonial for a website called “Don Uorri,” who had the typical features of a Mafia Don), then to an Italian telenovela (where Don Uorri was in charge of the romantic troubles of Albania immigrants in Italy). Later, in partly thanks to the broadcast of this Italian show in Albania, it jumped the Adriatic sea to be used in interactions between Albanians and foreigners (drivers who lost the way, restaurateurs dealing with hungry pleas during a black out, or local staff of international NGOs responding to the distressed calls of their supervisors). In little more than five years, don’t worry/don uorri came to play a major role in the inter-cultural repertoire of Albanian stranger-handlers. More recently, this floater seems to have reached St. Petersburg, where it is colloquially used in the expression “dont vori bi v kepi” (“don’t worry, wear a kepi” ⫺ a word play with the refrain don’t worry, be happy, pointing to the penetrating power of American pop music and the pervasive influence of Bobby McFerrin’s rather annoying vocalizations into Eastern Europe) (Anna Yatsenko 2007, personal communication at Reed College). The world is traversed by such floaters: Pennycook (2003) analysis of the expression by the way as used in English-Japanese rap points to the symbolic efficacy of such terms for Japanese rappers (and it would be interesting to compare it to the use of the online expression btw, which has also drifted into a multiplicity of discourse contexts). Similarly, Aravamudan’s (2006) discussion of the uptake of the term karma by Westerners (especially by practitioners of Eastern philosophy, such as the Beatles studying transcendental mediation) could be considered another transidiomatic floater now enjoying world wide recognition. The transidiomatic practices and floaters of both global elites and local/transnational semiotic workers represent a new, moving frontier in class, language, and power relations, and as such, they have a significant impact not only on the everyday life of these mobile agents but also in the lives of people with whom these mobile people interact, both near to them and in the deterritorialized environments of late-modern economy. These practices, especially the floaters, found themselves at the frontline of contemporary capitalism’s deterritorialized advance, on the cutting edge of the creation of new ways of speaking and communicating. They are instances of how new discourses and modes of representation are reterritorialized within the local environment, and as such must be taken into account in any assessment of the impact of transnational spaces on languages.

6. Conclusions In this article I have argued that our notion of transnational space must be expanded to take into account the emerging communicative geographies of the twenty-first century: international call centers, online social networking, multilingual chat rooms, videoconferencing, or Web 2.0 cellular telephones are all communicative environments which are poorly understood. As such, they deserve our full attention.

4. Language and transnational spaces A serious investigation of these communicative environments will necessarily problematize our taken-for-granted, common-sense knowledge of what is a “language” (Jacquemet 2005). It is now time to examine communicative practices based on disorderly recombinations and language mixings until now ordinarily overlooked. We should rethink the concept of communication itself, no longer embedded in national languages and international codes, but in the multiple transidiomatic practices of transnational flows. This will allow our imagination of linguistic exchanges to take shape within the discourse of local cultural becoming, social mutations, and global identities. This linguistic shift in turn entails a reconceptualization of what constitutes a “place.” People’s relationships to places are primarily interactional: the sense of being in a place is created through the interaction between the physical setting and the person experiencing it. We know that spatial perception is intrinsically tied to language, the primary tool shaping our understanding of space (Hanks 1990; Basso 1996; Gumperz and Levinson 1996). Transidiomatic practices, as much as any single territorialized language, alter perception of such constructs as “home”, “local community” or “public place”. Places are no longer defined in terms of scale and size, but instead in terms of interconnections and networks. Places become interactional hubs, charged with the linguistic organization of social and cultural flows in the context of global capitalism (Castells 2007). The synchronic, granular and multimodal characters of transidiomatic practices allow people to experience places no longer confined in local territories but rather constructed from the wider flows and circuits of contemporary transnational mediascapes. While transidiomatic practices will have necessarily lost some of the basic communicative characteristics we associate with locality and rootedness, they nevertheless force us to focus our attention on the communicative role of transnational networks ⫺ and in particular on the role of these practices in connecting multiple transnational nodes in the discontinuous space of late modernity.

7. Reerences Anderson, Benedict 1983 Imagined Communities. London: Verso. Androutsopoulos, Jannis 2006a Introduction: Sociolinguistics and computer-mediated communication. Journal of Sociolinguistics 10(4): 419⫺438. Androutsopoulos, Jannis 2006b Multilingualism, diaspora, and the Internet: Codes and identities on German-based diaspora websites. Journal of Sociolinguistics 10(4): 520⫺547. Ansaldo, Umberto and Lisa Lim 2006 Globalisation, empowerment, and the periphery. In: R. Elangaiyan, R. McKenna Brown, Nicholas D. M. Ostler and Mahendra K. Verma (eds.), Vital Voices: Endangered Languages and Multilingualism. Proceedings of the FEL X Conference, 39⫺46. Bath: Foundation for Endangered Languages. Appadurai, Arjun 1996 Modernity at Large. Minneapolis: University of Minnesota Press. Aravamudan, Srinivas 2006 Guru English. Princeton, NJ: Princeton University Press.

65

66

I. Introduction: Language and space Basch, Linda, Nina Glick Schiller, and Cristina Szanton Blanc 1994 Nations Unbound: Transnational Projects, Postcolonial Predicaments, and Deterritorialized Nation-States. London: Routledge. Basso, Keith H. 1996 Wisdom Sits in Places. Albuquerque: University of New Mexico Press. Baudrillard, Jean 1984 Simulacra and Simulation. Ann Arbor: University of Michigan Press. Bhabha, Homi K. 1994 The Location of Culture. London: Routledge. Cameron, Deborah 2000 Language: Small incivilities. Critical Quarterly 42(2): 127⫺131. Castells, Manuel 2007 Communication, power and counter-power in the network society. International Journal of Communication 1: 238⫺266. Cheah, Pheng and Bruce Robbins (eds.) 1998 Cosmopolitics: Thinking and Feeling beyond the Nation. Minneapolis: Minnesota University Press. Cisco Corporation 2007 Cisco TelePresence Trends. Retrieved 20 July 2007 from . Clifford, James 1997 Routes. Cambridge, MA: Harvard University Press. Coupland, Nikolas (ed.) 2003 Sociolinguistics and globalization. Journal of Sociolinguistics 7(4): 465⫺472. Crawford, James 1992 Hold Your Tongue: Bilingualism and the Politics of English Only. New York: AddisonWesley. Crowley, Tony 1989 Standard English and the Politics of Language. Urbana: University of Illinois Press. Crystal, David 1997 English as a Global Language. Cambridge: Cambridge University Press. Crystal, David 2006 Language and the Internet. Cambridge: Cambridge University Press. Danet, Brenda and Susan C. Herring 2007 The Multilingual Internet: Language, Culture, and Communication Online. Oxford: Oxford University Press. De Swaan, Abram 2001 Words of the World. London: Polity Press. Deleuze, Gilles and Felix Guattari 1983 Anti-Oedipus: Capitalism and Schizophrenia. Minneapolis: University of Minnesota Press. Deleuze, Gilles and Felix Guattari 1987 A Thousand Plateaus: Capitalism and Schizophrenia 2. Minneapolis: University of Minnesota Press. Dube´, Line, Anne Bourhis, and Re´al Jacob 2006 Towards a typology of virtual communities of practice. Interdisciplinary Journal of Information, Knowledge, and Management 1: 69⫺93. Errington, Joseph 2000 Indonesian(’s) authority. In: Paul Kroskrity (ed.), Regimes of Languages, 205⫺228. Santa Fe: School of America Research Press. Escobar, Arturo 2001 Culture sits in places: Globalism and subaltern strategies of localization. Political Geography 20: 139⫺174.

4. Language and transnational spaces Faist, Thomas 2000 The Volume and Dynamics of International Migration and Transnational Social Spaces. Oxford, UK: Clarendon/Oxford University Press. Faist, Thomas and Eyüp Özveren (eds.) 2004 Transnational Social Spaces: Agents, Networks, and Institutions. Burlington: Ashgate. Gamble, Jennifer M. 2007 Holding environment as home, Maintaining a seamless blend across the virtual/physical divide. M/C Journal, 10(4). Retrieved 21 July 2007 from . Garcı´a Canclini, Nestor 1995 Hybrid Cultures: Strategies for Entering and Leaving Modernity. Transl. Christopher L. Chiappari and Silvia L. Lo´pez. Minneapolis: University of Minnesota Press. Gee, James Paul, Glynda Hull, and Colin Lankshear 1996 The New Work Order. London: Allen & Unwin. Giddens, Anthony 1990 The Consequences of Modernity. Stanford, California: Stanford University Press. Ginsburg, Faye D., Lila Abu-Lughod and Brian Larkin (eds.) 2002 Media Worlds: Anthropology on New Terrains. Berkeley: University of California Press. Gitlin, Todd 1983 Inside Prime Time. New York: Pantheon Books. Gumperz, John J. and Stephen C. Levinson (eds.) 1996 Rethinking Linguistic Relativity. Cambridge: Cambridge University Press. Gupta, Akhil and James Ferguson 1992 Beyond “Culture”: Space, Identity, and the Politics of Difference. Cultural Anthropology 7(1): 6⫺23. Gupta, Akhil and James Ferguson (eds.) 1997 Anthropological Locations. Berkeley: University of California Press. Hanks, William F. 1990 Referential Practices: Language and Lived Space among the Maya. Chicago: University of Chicago Press. Hanks, William F. 1996 Language and Communicative Practice. Boulder, CO: Westview. Hannerz, Ulf 1992 Cultural Complexity: Studies in the Social Organization of Meaning. New York: Columbia University Press. Hannerz, Ulf 1996 Transnational Connections. London: Routledge. Hardt, Michael and Antonio Negri 2000 Empire. Cambridge, MA: Harvard University Press. Held, David and Anthony McGrew (eds.) 2000 The Global Transformations Reader. Malden, MA: Polity Press. Herring, Susan C. (ed.) 1996 Computer-Mediated Communication: Linguistic, Social and Cross-Cultural Perspectives, 81⫺106. Amsterdam: John Benjamins. Herring, Susan C. 2003 Computer-mediated discourse. In: Deborah Schiffrin, Deborah Tannen and Heidi E. Hamilton (eds.), The Handbook of Discourse Analysis, 612⫺635. London: Blackwell. Hill, Jane H. and Kenneth C. Hill 1986 Speaking Mexicano. Tucson: University of Arizona Press. Hinds, Pamela J. and Sara Kiesler (eds.) 2002 Distributed Work. Cambridge: MIT Press.

67

68

I. Introduction: Language and space Jackson, Peter, Philip Crang and Claire Dwyer (eds.) 2004 Transnational Spaces. London: Routledge. Jacquemet, Marco 1996 From the Atlas to the Alps. Public Culture 8(2): 377⫺388. Jacquemet, Marco 2005 Transidiomatic practices: Language and power in the age of globalization. Language & Communication, 25: 257⫺277. Jameson, Fredric and Masao Miyoshi (eds.) 1998 The Cultures of Globalization. Durham: Duke University Press. Jensen, Jürgen 2004 Plural societies and transnational social spaces. In: Jürgen Oßenbrügge and Mechthild Reh (eds.), Social Spaces of African Societies: Applications and Critique Concepts about “Transnational Social Spaces”, 35⫺76. Münster: Lit Verlag. Jones, Rodney H. 2004 The problem of context in computer-mediated communication. In: Philip Levine and Ron Scollon (eds.), Discourse and Technology, 20⫺33. Washington, DC: Georgetown University Press. Kaplan, Caren 1996 Questions of Travel: Postmodern Discourses of Displacement. Durham: Duke University Press. Landler, Mark 2001 Hi, I’m in Bangalore (but I Dare not Tell). New York Times. 3(21). Lechner, Frank J. and John Boli (eds.) 2000 The Globalization Reader. London: Blackwell. Mankekar, Purnima 1999 Screening Culture, Viewing Politics: An Ethnography of Television, Womanhood, and Nation in Postcolonial India. Durham: Duke University Press. Marquardt, Michael J. and Lisa Horvarth 2001 Global Teams. Palo Alto: Davies-Black Pub. Massey, Doreen 1994 Space, Place, and Gender. Minneapolis: University of Minnesota Press. Mignolo, Walter D. 2000 Local Histories/Global Designs: Coloniality, Subaltern Knowledges, and Border Thinking. Princeton, NJ: Princeton University Press. Nederveen Pieterse, Jan 2003 Globalization and Culture. Lanham, MD: Rowman and Littlefield. Ong, Aihwa 1998 Flexible Citizenship. Durham: Duke University Press. Paolillo, John C. 1996 Language choice on soc.culture.punjab. Electronic Journal of Communication 6(4). . Pennycook, Alastair 2003 Global Englishes, Rip Slyme, and performativity. Journal of Sociolinguistics 7(4): 513⫺ 533. Pennycook, Alastair 2007 Global Englishes and Transcultural Flows. London: Routledge. Pries, Ludger (ed.) 1999 Migration and Transnational Social Spaces. Aldershot, UK: Ashgate. Pries, Ludger (ed.) 2001 New Transnational Social Spaces: International Migration and Transnational Companies in the Early Twenty-First Century. London and New York: Routledge.

4. Language and transnational spaces

69

Reich, Robert B. 1991 The Work of Nations: Preparing Ourselves for the 21st Century. New York: Knopf. Ritzer, George 1996 The McDonaldization of Society. Los Angeles: Pine Forge Press. Robertson, Roland 1995 Glocalization: Time-space and homogeneity-heterogeneity. In: Mike Featherstone, Scott M. Lash and Roland Robertson (eds.), Global Modernities, 25⫺44. London: Sage. Sassen, Saskia 2001 The Global City. Princeton, NJ: Princeton University Press. Schein, Louisa 2002 Mapping Hmong media in diasporic space. In: Ginsburg, Abu-Lughod and Larkin (eds.), 229⫺245. Scollon, Ron 1998 Mediated Discourse as Social Interaction: A Study of News Discourse. London: Longman. Scollon, Ron 2001 Mediated Discourse: The Nexus of Practice. London: Routledge. Shipley, David and Will Schwalbe 2007 Send: The Essential Guide to Email for Office and Home. New York: Knopf. Silverstein, Michael 1976 Shifters, linguistic categories, and cultural descriptions. In: Keith Basso and Henry Selby (eds.), Meaning in Anthropology, 11⫺55. Albuquerque: University of New Mexico Press. Silverstein, Michael 1996 Monoglot “standard” in America. In: Donald Brenneis and Ronald K. S. Macaulay (eds.), The Matrix of Language, 284⫺306. Boulder, CO: Westview. Spitulnik, Debra 1993 Anthropology and mass media. Annual Review of Anthropology 22: 293⫺315. Spivak, Gayatri Chakravorty 1988 “Can the Subaltern speak?” In: Cary Nelson and Lawrence Grossberg (eds.), Marxism and the Interpretation of Culture. Urbana: University of Illinois Press. Tomlinson, John 1999a Globalization and Culture. Chicago: University of Chicago PressVertotec, Steven 1999b Conceiving and researching transnationalism. Ethnic and Racial Studies 22: 447⫺462. Vertotec, Steven 1999 Conceiving and Research Transnationalism. Ethnic and Racial Studies 22: 447⫺462. Virilio, Paul 1988 [1994] La Machine de Vision. Paris: Galile´e. [Published in English as The Vision Machine. Bloomington, IN: Indiana University Press]. Virilio, Paul 1990 L’Inertie Polaire. Paris: Christian Bourgois. Virilio, Paul 1997 Open Sky. London: Verso. Wasson, Christina 2006 Being in Two Spaces at once: Virtual Meetings and their Representation. Journal of Linguistic Anthropology 16(1): 103⫺130. Woolard, Kathryn A. 1999 Simultaneity and bivalency as strategies in bilingualism. Journal of Linguistic Anthropology 8(1): 3⫺29.

Marco Jacquemet, San Francisco (USA)

II. Linguistic approaches to space 5. Language and space: The neogrammarian tradition 1. 2. 3. 4. 5. 6. 7.

The neogrammarians The neogrammarians and their approach to contemporary dialects Idiolect, dialects and languages A neogrammarian program for dialect description The Ortsgrammatik tradition Sound change and messy dichotomies References

1. The neogrammarians Neogrammarian (a misleading translation of the German term, Junggrammatiker ‘young grammarian’) was the name adopted by a dynamic group of brilliant philologists and linguists originally based in Leipzig in the 1870s. The original group included Karl Brugmann, 1849⫺1919; Berthold Delbrück, 1842⫺1922; August Leskien, 1840⫺1916; Hermann Osthoff, 1847⫺1909; and later, Wilhelm Braune, 1850⫺1926; Hermann Paul, 1846⫺1921; and Eduard Sievers, 1850⫺1926, among others (see Jankowsky 1972 for an overview). The neogrammarians were the dominant force in linguistic science for over four decades until the ascendancy of various forms of linguistic structuralism, particularly following the publication of Saussure (1916). Although the neogrammarians are probably now best known for their claim that sound change is governed by laws that admit no exception (Osthoff and Brugmann 1878: xiii) ⫺ a heuristic that still plays an important role in the historical comparative reconstruction of languages and language families ⫺ their enduring contributions are also found in the many historical grammars of the earlier stages of the Indo-European languages that they produced, many of which are still in use in revised editions (for example, Braune’s [1880] Gotische Grammatik, twentieth edition published in 2004), and in the numerous detailed dialect grammars that have been written under a neogrammarian rubric. Anyone writing about the neogrammarians must pause and reflect on Lehmann’s (1967: 257) warning that “students will derive greater profit by reading the outstanding works of the neogrammarians and their predecessors than by reading about them.” There are many factors leading to Lehmann’s comment, and the issues are complex. We can only note here that the neogrammarians did not form some rigid scholarly monolith, but rather formed a relatively loose collection of scholars who did not necessarily agree on all points, and whose ideas and perspectives did not remain stagnant but rather evolved over decades of scholarship (see Putschke 1969: 20). We are also faced with the legacy of a polemical exchange ⫺ sometimes devoid of any intellectual rigor ⫺ in which rival scholars could be criticized not only for things they had never claimed, but even for things they opposed. It is only in such an atmosphere that a linguist publishing an attack on the neogrammarians in the journal Language could state: “It is true that sev-

5. The neogrammarian tradition eral [most? ⫺ see Jankowsky 1972: 234, RWM] of the opinions here attributed to the neogrammarians have never been openly asserted by them, and have even been sometimes strongly denied” (Bonfante 1947: 377, note 23). In an attempt to provide a well-grounded coherent overview, the present Handbook entry draws directly on the work of a few neogrammarian scholars in discussing the theoretical backdrop for the neogrammarian approach to the study of dialects in time and space: particularly Paul ([1886] 1920), a codification of the neogrammarian approach that includes extensive discussion of language change and dialects, and Sievers ([1876] 1901), which lays the foundations for the science of phonetics and provides the basis for the detailed dialect studies of the period.

2. The neogrammarians and their approach to contemporary dialects In spite of earlier claims to the contrary, evolution, not revolution seems to be the appropriate characterization of both the rise and fall of the neogrammarians (see Koerner 1981). The early period is best viewed as an attempt to formalize and give methodological rigor to the best practices of the period (Putschke 1969: 21). Indeed, even one of the neogrammarians’ own contemporaries and one of their most outspoken critics, Hugo Schuchardt, argued that there was nothing new in the neogrammarian approach. In his view, “the exceptionless operation of the sound laws is the only proposition the so-called neogrammarian school can consider to be its very own” (1885 [1972]: 41), but, according to Schuchardt, the proposition is false. With waning influence at the end of their period, it is often stated that structuralism ⫺ beginning with Saussure and developed in various streams such as the Prague School, American Structuralism, etc. ⫺ represented a dramatic shift away from neogrammarian theory. In fact, however, the foundations for structuralism are explicitly evident in works such as Paul (1920) and, especially, Sievers 1901 (see Putschke 1969: 20), as well as in the dialect descriptions, particularly in seminal works such as Winteler (1876) and Heusler (1888). In addition, of course, many neogrammarian positions were also accepted by the structuralists, for example, the insistence on the study of spoken language. In one respect, however, it is fair to say that the neogrammarian approach did represent a dramatic departure from most past practice; namely, in the importance that was placed on the study of contemporary spoken dialects. In the German-speaking world from the seventeenth to the early nineteenth century, linguistic discussion tended to focus on the development of a standard language, a Schriftsprache, and in this discussion dialects played, at best, a marginal role (Knoop 1982). Although the production of dialect dictionaries was a common activity during this period ⫺ and even as early as 1680 Leibniz argued that a developing standard language could be enriched through the incorporation of vocabulary from the dialects (see Schirmunski 1962: 57) ⫺ dialects generally came to be denigrated and even conceived of as degenerate forms of an idealized standard language. With the publication of the first edition of Jacob Grimm’s Deutsche Grammatik (1819) came the foundations of historical linguistics and the basis for a historical understanding of dialects, but Grimm’s grammar was primarily a work of philology that did not draw

71

72

II. Linguistic approaches to space on an investigation of the contemporary spoken dialects. At first independently of Grimm, and then strongly influenced by him, Schmeller (1821) produced the first dialect study on a historical linguistic foundation, Die Mundarten Bayerns grammatisch dargestellt. However, in general, work during this period ignored the contemporary dialects and remained primarily philological and historical (with some notable exceptions, such as Weinhold 1853; Regel 1868; see Schirmunski 1962: 56⫺62 for an overview). Old ideas die hard, and the conceptualization of language evolution that underlay much of the early nineteenth century research carried out by pre-neogrammarian philologists and Indo-Europeanists did nothing to favor the study of contemporary dialects. Schleicher (1871), for example, assumed two stages in the “life of a language:” in the pre-historic period there is language “development” and “growth,” whereas in the historic period there are only “declines both in sound and in form” (⫽ Lehmann 1967: 91) Such a view could only contribute to the scholarly neglect of the contemporary spoken dialects, especially since one of the primary goals of the period was the reconstruction of Proto-Indo-European. Given that the contemporary dialects were considered the product of a very long period of decline, it was assumed that their study could shed little light on Proto-Indo-European. The neogrammarians reacted very strongly against this aspect of earlier work, which lacked an adequate phonetic foundation and in the worst cases tended toward a kind of Buchstabenschieberei or arbitrary paper phonology. In their neogrammarian “manifesto,” Osthoff and Brugmann (1878: iii) complained: “Languages were intensively investigated, but there was much too little investigation of the speaker” (⫽ Lehmann 1967: 198). Sweet (1888: xi), somewhat acerbically and perhaps not entirely objectively, describes in the preface to his History of English Sounds how he welcomed the “neo-philological” (that is, the neogrammarian) reformers, especially Paul and Sievers: “We set our faces against the ‘woodenness’ which then characterized German philology: its contempt for phonetics and living speech and reverence for the dead letter, its one-sidedly historical spirit, and disregard of analogy.” It is not surprising then that the pendulum swung wide, and the neogrammarians took precisely the opposite tack: the contemporary spoken dialects ⫺ and that of course at the time meant spoken, spontaneous language as opposed to standard or obsolescent languages ⫺ were elevated to the primary object of linguistic study. The commonsense logic behind this position is again laid out in Osthoff and Brugmann’s neogrammarian “manifesto” (see Lehmann 1976: 200⫺203): since the same physiological and psychological factors involved in language are at work in the present as in the past (the “uniformitarian principle,” Labov 1994: 21⫺23), the linguist’s skills should be honed first on languages that are directly accessible (that is, contemporary spoken dialects), and only then directed toward the understanding and reconstruction of obsolescent languages. The neogrammarian position is taken to its logical conclusion in Sievers’ (1901: 6) sage advice ⫺ still valid for present-day students of phonetics and phonology: “The linguist’s own dialect, spoken since childhood, should be the starting point of all phonetic investigations […]. Only once all the phonetic phenomena of the native dialect are clearly understood should the linguist gradually move on to the investigation of more distant dialects and languages” (similarly, Paul 1920: 30).

5. The neogrammarian tradition

3. Idiolects, dialects and languages Paul ([1880] 1920) serves as the classic neogrammarian “textbook,” which ⫺ in spite of its dense and sometimes opaque style ⫺ directly influenced generations of linguists. The fourth edition represents a mature rendering of the neogrammarian perspective informed by over three decades of vigorous scholarship and debate; see, for example, various references indicating the impact of non-neogrammarians, such as Hugo Schuchardt on “language mixture” (1920: 390) and the inadequacy of “tree theory” (1920: 43), and Mikołaj Kruszewski on sound laws (1920: 49). The neogrammarian perspective continued to be perpetuated in subsequent works such as Saussure (1916) and another classic textbook, Bloomfield (1933), both of which in their historical components did not deviate significantly from Paul. Weinreich, Labov and Herzog (1968), which remains a valuable, if somewhat uncharitable, historiographically-based discussion of Paul’s approach to language change and dialectology, argues that some of the most fundamental difficulties of modern linguistics can be traced back to Paul’s views (for general background on Paul, see Neumann 1996). In an approach not dissimilar to Chomsky’s (1986) distinction between E (externalized) and I (internalized) language, Paul considers terms such as dialect, language and language family to be entirely artificial and arbitrary constructs of “descriptive grammar,” which has the goal of describing the grammatical forms and conditions of use in a linguistic community at a given point in time. Since descriptive grammar does not deal with facts, but only with abstractions derived from observable facts, and since there are no causal connections between abstractions, but only between real objects and facts, descriptive grammar is far removed from a scientific investigation of language (1920: 24). For Paul it is the individual speaker’s mental and physical activities involved in the production and comprehension of language that constitute the true subject matter of linguistic science, so the language of an individual speaker ⫺ the idiolect ⫺ constitutes the only real and legitimate object of linguistic investigation (1920: 24). This position leads to an interesting consequence, which, as observed by Weinreich, Labov and Herzog (1968: 104), Paul embraces: “Since the individual psyche is seen as the locus of the associations and connections between language components, we realize why Paul isolates the individual as the primary carrier of a language, and brings the argument to its logical conclusion by asserting that ‘we must distinguish as many languages as there are individuals’.” Given this fundamental precept, Paul poses and attempts to answer some of the critical questions of the period relating to language change and dialects.

3.1. Why do languages change over time? For Paul, an individual’s linguistic capacity is the product of a highly complex mental object (Gebilde) operating at a subconscious level, which consists of numerous interconnected groups of images or representations (Vorstellungsgruppen) (1920: 25). This object, which Paul considers a psychological “organism,” can be viewed ⫺ perhaps somewhat anachronistically ⫺ as a mental grammar (see Weinreich, Labov and Herzog 1968: 105). It is self evident for Paul that the representations of the mental grammar are in a constant state of change, since they are a kind of subconscious reflection of the speaker’s own activities and what he or she is exposed to. With lack of reinforcement, established

73

74

II. Linguistic approaches to space representations weaken and even disappear, while new or altered representations are constantly arising through “every act of speaking, hearing, or thinking” (1920: 27), and, consequently, the associations among representations (or networks) are constantly shifting. Accordingly, the fact of “language” change is viewed as an inevitable consequence of the ever-changing nature of the idiolects that constitute language. (The individual’s language organism or grammar is the true carrier of change, and it is entirely misleading to speak of language change.) What really changes is the individual’s internalized grammar, although these changes may be reflected in the common language usage (Sprachusus) ⫺ an abstraction, which can be considered a kind of average deriving from a set of individual grammars ⫺ of a particular linguistic community (1920: 29). One of Paul’s primary concerns is to provide an explanation of regular (gradual) sound change. For Paul, a speaker’s production of a language sound depends on an idealized sound image or representation (Lautbild ), an abstract mental target that is formed on the basis of what the speaker both produces and hears. However, just as an archer who is shooting arrows can miss the mark, a speaker does not precisely hit the sound image target with every utterance, and Paul sees endless, random variation of production around this ideal target, which, due primarily to the balancing influence of verbal communication with other speakers of the same community (Verkehr, linguistic “intercourse”), takes place within relatively narrow limits. Paul argues that even though the variation occurs at a subconscious level, speakers have the ability to unconsciously modify sound images over time based on what they hear and produce (1920: 55), so sound change takes place without any awareness on the part of the speakers. In this regard, Weinreich, Labov and Herzog (1968: 125) would seem to overstate their case when they make reference to “the neogrammarian descriptivist conception of a homogeneous system as the sole legitimate object of analysis.” Beyond Paul’s assumption of significant idiolectal variation (which Weinreich, Labov and Herzog 1968: 104⫺ 119 themselves discuss), Sievers (1901: 272⫺273), for example, observes language specific differences in phonological variation; for example, k and x may be simple distributional variants in one language but contrastive in another. Noting that some “languages indiscriminately mix up phones that come across as entirely different to a more keen listener”, Sievers states the importance of taking such language specific factors into consideration when judging whether a particular sound change is completely regular or not. In addition, Sievers explicitly discusses the synchronic layering (coexistence) of old and new phonological material in idiolects that can result from sound change (1901: 268). He also sees no necessity of a preferred variant winning out, and even assumes that the gradual transition of a sound in a certain direction can be reversed (see Haas 1999: 131 for discussion of precisely such a case in Swiss German). At the same time, though, it is true ⫺ as Weinreich, Labov and Herzog (1968) discuss ⫺ that in Paul’s (and Sievers’) conception idiolectal variation is not encoded in any way; for example, in terms of social factors (see section 3.3 below).

3.2. How do dialects diverge and ultimately split? Given the intense research in the latter part of the nineteenth century on the history of the Indo-European languages and their development from a common source, the factors leading to dialect ⫺ and ultimately language ⫺ split were of great interest. As expected

5. The neogrammarian tradition from the neogrammarian perspective, for Paul the determination of dialects and their differentiation is based entirely on phonological factors. The direct transfer of other language properties among speakers, such as lexical or syntactic, can occur without difficulty, whereas in the case of phonology, sound material is never passed on exactly as it is received, so ultimately a much stronger differentiation can arise (1920: 47). Here Paul argues that his idiolectal perspective makes dialectal divergence another inevitable fact. “The endless variation of language as a whole and the incessant development of dialectal differences necessarily follow from the simple consideration of the endless variation and unique formation of each individual organism” (1920: 28). In fact, for Paul the question is turned on its head: the puzzle now is not why do dialects split, but rather why don’t individual idiolects drift away from each other to such a degree that extreme fragmentation results (1920: 39⫺40). The crucial balancing factor here is the verbal communication of linguistic intercourse, a very important factor for Paul since it is only within this context that an individual generates language (1920: 39). Paul compares the situation to the birth and biological development of an individual animal. By far the most important factor is the genetic inheritance stemming from the parents and, secondarily, the chance effects of things such as food, way of life, etc. Similarly, an individual is subject to the influences of his or her primary companions, whose idiolects serve as the “parents” of the particular individual’s idiolect. As individuals in verbal communication adapt their idiolects to one another, a certain leveling and compromise effect occurs, and no idiolect will ever radically differ from the given common language usage (Sprachusus) (1920: 38⫺48). Interestingly, in a somewhat controversial position at the time, Paul is explicit that a standard language ⫺ conceived of as an abstraction constituted of an ideal norm ⫺ can never arise naturally through such a leveling process: dialects are “sacrificed” to a standard language, which is, in fact, a “foreign idiom” (1920: 404). Paul also accepts what seems to be a necessary consequence of his premises; namely, that sound change is actually favored in isolated communities, since the compromising effects of leveling brought about by the broader linguistic intercourse found within a larger community is absent (1920: 60). Paul’s view here is precisely the opposite of most subsequent approaches (although see Andersen 1988: 70 for an apparently similar notion), even neogrammarianbased ones such as the Vienna School (see section 5 below). Again, Paul emphasizes the view that within any linguistic community there are as many dialects as there are individuals, so “dialect split means nothing more than the increase of individual differences beyond a certain degree” (1920: 38). At the same time, the sum effect of the individual changes ⫺ tempered by verbal communication ⫺ is manifested in changes in the general pattern of common language usage over time, and, depending on the degree of linguistic intercourse within and among groups of individuals, the language usage may change in different directions within different linguistic communities. However, complete absence of linguistic intercourse is not a necessary prerequisite for dialect split ⫺ although such absence can lead to stronger differentiation ⫺ and Paul discusses cases such as German, where, even though a dialect continuum is involved, mutual comprehensibility between idiolects at two given points can be lost (1920: 40⫺42). In this regard, he refers to the “shading off” (Abstufung) of sound changes over geographical distance being one important factor, with the High German shift being a classic example, and Georg Wenker’s Sprachatlas providing many more examples of the “extraordinary multiplicity” of the shading off effects (1920: 45).

75

76

II. Linguistic approaches to space It follows from Paul’s view that a clean separation of dialects and languages into groups and subgroups is never feasible. Accordingly, a genealogical tree or table might be useful for pedagogical purposes, but it cannot be considered an accurate reflection of reality. With a nod to wave theory, Paul notes that a sound change can spread through a linguistic area even after significant dialectal differentiation has taken place; for example, the case of open syllable lengthening of Middle High German forms such as le˘sen ‘read’, ge˘ben ‘give’, etc. in the Low and Middle German dialects. Accordingly, he warns against a simplistic assumption that would correlate the age of a sound change with greater geographic spread (1920: 42⫺43).

3.3. How do sound changes spread through a linguistic community? With a nod to Schuchardt, Paul states that “of all the questions that modern linguistics has to deal with none is more important than language mixture,” that is, the interaction of idiolects in contact (1920: 390). Especially from Paul’s perspective, one of the most difficult questions is to explain how an entire set of idiolects in linguistic intercourse can share a particular sound change (where we restrict our attention here to regular, gradual sound changes such as the voicing of intervocalic consonants, cluster assimilation, etc.). Paul’s theory is based on a consideration of both idiolect-internal and external factors. The internal factors involve a kind of markedness or preference theory, which Paul does not develop in any detail (1920: 32⫺33; see also Sievers 1901: 269). However, he clearly envisions an interpretation in which some particular sounds, sound sequences, etc. are more preferred or favored (bequemer “more convenient”) over others, where the primary basis for the preference is physiological; for example, an assimilated sequence atta is preferred over unassimilated akta. Such changes occur in tiny increments below the speaker’s threshold of awareness, and over time, both the sound image and the variants around the mental target move in tandem. The external factors involve linguistic intercourse and the propensity of speakers to adapt their speech to the speech of those they interact with. Of course, there is no necessity for speakers to spontaneously adjust their sound image in accordance with a particular preference, but given that some speakers have shifted their mental sound image (and their output in terms of the corresponding variants), other speakers can begin adjusting their sound images simply on the basis of exposure to the new variants. We would seem, then, to have the seeds of a rather straightforward theory of the spread of sound change through a linguistic community involving both innovation and adoption. The innovators spontaneously adjust their sound image (and the corresponding variants) in some preferred direction, while the adopters modify their sound image and variants due simply to exposure to the variants produced by the innovators. It is, however, a big leap from these basic assumptions to an account of regular sound change; that is, the explanation of how entire sets of idiolect speakers can shift their sound images in tandem, particularly given that regular sound changes typically carry on over generations. Part of Paul’s answer lies in child language acquisition, although in fact he seems to be somewhat divided on the role of acquisition in sound change. On the one hand, he makes no qualitative distinction between child language acquisition and the adult stage, which he considers more stable, but still subject to change (1920: 27). Similarly, Sievers

5. The neogrammarian tradition (1901: 267) explicitly attributes no special role to child language acquisition in language change, stating that it is likely immaterial whether an innovation begins in a particular generation or in the transmission from one generation to the next. On the other hand, Paul clearly views the language acquisition stage as a period of robust change and the primary locus for the explanation of changes (1920: 34, 62⫺63). Regardless, language acquisition takes on special importance in Paul’s theory in explaining how a sound change can ultimately take hold in an entire set of idiolects. For even if there is a minority of potential adopters who do not adjust their sound images in accordance with the majority, Paul assumes that the new generation will conform to the majority and the minority will simply die out. However, even with language acquisition added to the mix Paul still faces a deceptively simple question; namely, how does the majority develop in the first place? In fact, Paul finds himself in a kind of catch-22. On the one hand, he assumes that a critical mass within a dialectal group must be attained before adopters will begin to shift their sound images to conform to the shifted variants they are exposed to (although he is clear that it is not simply the number of speakers they are exposed to but the frequency of exposure that is crucial for the adopters). On the other hand, the innovators ⫺ in principle, every speaker within the given set of idiolects ⫺ are free to proceed along a variety of pathways, and the possibility of creating a critical mass seems remote. Here one might expect Paul to argue that a gradual shifting of a sound image might take hold spontaneously within a large subset of the idiolectal grouping precisely because it is preferred. Paul recognizes, however, that there is no possibility of explaining the development of a majority in this way, in part because there are simply too many conflicting preferences. He gives the example of a, which, in principle, can begin shifting in the direction of either i or u. In the end, Paul’s answer to the question is not particularly reassuring ⫺ he simply invokes chance: “In an area of especially intensive linguistic intercourse, it is easy for a particular tendency to become dominate simply through the play of chance” (1920: 61). In this case, Weinreich, Labov and Herzog (1968: 112) are justifiably critical in stating that “chance is here invoked illegitimately, since we are out to explain a specific, not random process.” Paul seems aware of the weaknesses of his approach and mentions two other relevant factors, although without extensive elaboration. First, he refers to the possibility that different motivations among speakers can lead to the same output (a theme taken up again in Evolutionary Phonology, Blevins 2004: 269): “the causes that push in a particular direction can be completely different for each individual” (1920: 61). Second, he realizes that his interpretation requires some further invisible hand mechanism, and here we find an incipiently structuralist approach. All languages have a certain harmony in their sound systems (Harmonie des Lautsystems) in which the movement of one sound in a particular direction can have impact on the other sounds in the system. Paul derives the systemic pressures from Sievers’ notion that every dialect has a particular neutral position (Indifferenzlage) of the speech organs (1920: 57). Regardless, in the end it is a fact that Paul cannot explain the spread of a sound change through a particular idiolectal grouping, and we must concur with Auer’s (2006) somewhat ironic conclusion; namely, although Paul “believes that variability is genuine and systematic, both in the individual and in the social group (1920: 55), and although he acknowledges the fact that it is through interaction between speakers that the language of an individual is ‘generated’ (1920: 37), his theory remains (psycho-)mechanistic and non-social.”

77

78

II. Linguistic approaches to space The non-social bias is also reflected in Paul’s somewhat counterintuitive insistence that sound change cannot spread from an individual. In fact, it is not difficult to misread Paul on this point, and that is precisely what one of Paul’s colleagues, William Wundt, apparently did when he refers to “the view, especially emphasized by H. Paul, that in language something can only become usual that was originally individual, that is, it comes from the individual.” Quoting the passage, Paul (1920: 62, note 1) is angrily indignant at this (allegedly intentional) misrepresentation of his ideas, and vehemently denies the possibility of interpreting his views in this way. However, Paul’s opposition is not a necessary consequence of his premises, and he does not give substantive reasons for rejecting this view. Sievers (1901: 367), for example, specifically states that a single individual can be the source of a sound change. In fact, Wundt’s statement seems only to be a kind of common sense consequence of Paul’s premises, which, interestingly, would put Paul in the same camp as one of the most renowned opponents of the neogrammarians, Hugo Schuchardt (see section 6 below). In this regard, Weinreich, Labov and Herzog (1968: 106) would again seem to overstate their case when they see “the creation of an irreconcilable opposition between the individual and society” in Paul’s conception. Indeed, although Paul himself did not conceptualize sound change in terms of social factors, this item was certainly on the neogrammarian research agenda (see next section).

4. A neogrammarian program or dialect description A neogrammarian program for dialect study is explicitly laid out in Wegener [1880] 1976 (“On German dialect research”), a funding and research proposal approved by Wilhelm Braune, Hermann Paul, Eduard Sievers and Jost Winteler. The primary goal of the document was to lobby the scholarly community in support of funding and resources from the Reichkanzler to carry out dialectology projects. As a proposal to secure government funding, it contains some masterful strokes justifying societal need. There is a sense of urgency based on the concern ⫺ intuitively plausible at the time but, as it turns out, exaggerated ⫺ that the German dialects were on the verge of extinction under the influence of the standard language (1976: 14). Accordingly, as a cultural legacy comparable to the marble pillars of Italy and the statues of ancient Greece, it was of the utmost importance to rescue the dialect information ⫺ the building blocks of the standard language ⫺ from oblivion, and in the process make a vital contribution to German unity (1976: 29). The funding and infrastructure requirements are clearly identified, including: the establishment of a government department on dialect study, the funding of dialect questionnaires to be administered by civil servants, the establishment of a German dialect library and publication subsidies (1976: 27⫺28). On this last point, it is surprising to note that Wegener’s call still resonates into the present, as important dialect dissertations, such as Bidner (1939), remain available even today in only limited copies of the original ⫺ impenetrably handwritten ⫺ manuscript. Wegener emphasizes the fact that the field of dialectology is still in its infancy, and that the generally negative perception of dialects and their study can only be overcome through a more scientific approach ⫺ that is, one that embraces the methodology of historical linguistics (1976: 3). In keeping with Paul’s treatment of the idiolect, broad

5. The neogrammarian tradition overviews of dialect regions are rejected in favor of detailed descriptions (grammars) of individual dialects, which are viewed as a necessary prerequisite to subsequent broader synthetic and comparative treatments. Also, as expected, there is a strong emphasis on phonetics a` la Sievers, and ⫺ particularly because of the requirement of an extremely detailed phonetic analysis ⫺ ideally the grammars should be written by a native speaker of the particular dialect based on self-analysis and fieldwork (1976: 26). These grammars came to be known as Ortsgrammatiken (village or local grammars, see section 5 below), and the already published Winteler (1876) served as a role model. Throughout the proposal, various theses describe the goals and structure of the Ortsgrammatik. The synchronic section is to provide a precise phonetic description of the particular dialect, including its segments, suprasegmentals, and what can be called phrasal or sentence phonology, including sandhi phenomena (1976: 12). The diachronic component describes the sound changes in the dialect starting with the “old German” stage, typically Middle High German but sometimes Old High German. Precise sound laws are to be provided that take into careful consideration the prosodic environment, for example, stressed versus unstressed syllable. All exceptions to the proposed sound laws are to be explicitly listed and categorized according to the interfering factor; for example, analogy within the dialect, borrowing from the standard language or a neighboring dialect, etc. (1976: 8). In the time honored tradition of Jacob Grimm, if factors behind an exception cannot be identified, it should be listed as unexplained. Given the role of paradigms in analogical leveling, there is also an emphasis on morphological description, and although phonetics and morphology are clearly the focus, there is acknowledgment of the importance of syntax and semantics. Wegener also makes a distinction between what later became known as the direct and indirect method of dialect investigation. Due to the sophisticated phonetic analysis required, the direct method (involving fieldwork, again ideally carried out by a native speaker) is appropriate, but for practical reasons the indirect method (that is, the elicitation of data by means of questionnaires) is a reasonable approach to morphology, syntax and semantics (1976: 27). Although the neogrammarians are often criticized for not considering the role of social factors in language description and change, Wegener touches on issues of linguistic variation due to age as well as gender factors. Social status is clearly conceived of as the prime external factor in dialect change and, indeed, Wegener could not be more explicit about its importance: “the question of language is essentially a question of social power” (1976: 16). Wegener conceptualizes the operation of social factors in terms of three concentric circles (1976: 15). The middle point is the dialect of the educated city elite (essentially a variant of the standard language), next is the dialect of the relatively uneducated city person, and then comes an outside circle consisting of the rural dialect (Bauernsprache). For Wegener movement is inexorably unidirectional toward the centre; in other words, the peripheral dialects will ultimately assimilate to the standard. Wegener promotes the view that such concentric circles should be investigated within dialectology, where “self-evidently” the starting point should be the rural dialect. As it is most isolated from the social effects of interaction with the standard language or the language of the urban centre, the rural dialect will best show the effects of the regular sound laws ⫺ tempered only by internal analogy ⫺ with very little dialect mixture.

79

80

II. Linguistic approaches to space

5. The Ortsgrammatik tradition Developed in the spirit of Wegener 1880, the Ortsgrammatik tradition has stood the test of time, producing a wealth of dialect material beginning with early classics such as Winteler (1876), Holthausen (1886), Heusler (1888), Schatz (1897) and Lessiak (1903). The general traits of the grammars are in remarkable conformance with Wegener’s proposal (see section 4 above). In the interests of avoiding issues of dialect mixture and observing the pure operation of sound laws and internal analogy, each grammar typically constitutes an in-depth description of one locality, or small geographical area considered to be relatively homogeneous. In keeping with the neogrammarian precept of detailed phonetic description, the grammars are usually written ⫺ often in the form of doctoral dissertations ⫺ by phonetically well-trained native speakers using a combination of selfanalysis and fieldwork (the direct method). Although most of the grammars have both a synchronic and diachronic component, they can differ in their relative emphasis. On this basis, Sonderegger (1968: 12) distinguishes three classic subtypes of neogrammarian dialect grammars ⫺ synchronic/phonetic (Winteler 1876), synchronic/historical (Heusler 1888) and historical (Kauffmann 1890). Already Winteler’s seminal Ortsgrammatik leaves no doubt about the death of the pre-neogrammarian Buchstabenschieberei and the corresponding shift in focus to the study of living languages, with particular emphasis on phonetics and phonology. Along with a general introduction to phonetics (17 pages), Winteler provides 66 pages of synchronic description of the dialect’s vowels and consonants, including quantity, along with 17 pages devoted to internal phonotactic and sandhi phenomena. An additional 53 pages cover the historical description of the sounds in terms of sound laws, and 41 pages are dedicated to a description of ⫺ primarily inflectional ⫺ morphology. As is typically the case in the Ortsgrammatik tradition, there is no syntactic or semantic description. As the science of phonetics was still in its early stages, a symbiotic relationship evolved between these detailed dialect descriptions and the field of Sieversian phonetics (Wiesinger 1976). The wealth of phonetic insights in the early works is striking, and the detailed work forms a necessary precursor to subsequent comparative and typological research. For example, Heusler’s (1888) investigation of Basel German, with its remarkably sophisticated discussion of fortis, lenis, geminate and ambisyllabic consonantal properties (see especially pp. 29⫺35) and description of the historically complex development of stressed syllable quantity (1888: 48⫺50), can still be read today with great benefit. Although ⫺ as expected ⫺ the search for regular sound laws and the explanation of exceptions in terms of dialect-internal analogy or dialect mixture is ever present in the Ortsgrammatiken, it is worth noting that the term sound law here continues to have a certain ambiguity characteristic of neogrammarian work: it can express either a diachronic formula, x > y, or a synchronic phonotactic constraint (for some discussion, see Weinreich, Labov and Herzog 1968: 115⫺116). For example, Pfalz’s (1913: 9) Law of Middle Bavarian is a purely synchronic phonotactic statement of prosodically determined segmental properties: “Following a short vowel or diphthong carrying an abruptly cut accent, only a fortis consonant can occur. Following a long vowel carrying a smoothly cut accent, only a lenis consonant can occur.” The first two and a half decades of the Ortsgrammatik tradition produced detailed synchronic descriptions of individual dialects on the basis of a sophisticated phonetic

5. The neogrammarian tradition foundation. In this regard, there is no doubt that Wegener’s (1880 [1976]: 4) concern that “one can never emphasize enough the difference between a letter and the sound” had been acknowledged and was being systematically dealt with. Further, the grammars clearly constitute crucial historical records of dialects that ⫺ it was assumed ⫺ were on the verge of extinction. However, there was still one important ingredient missing in this descriptive and analytic framework ⫺ actual sound recordings. This situation changed dramatically at the turn of the twentieth century. The Austrian Academy of Sciences Phonogrammarchiv, founded in 1899, proudly advertises itself as “the oldest audiovisual research archive in the world” (). Other recording centres were set up around Europe shortly thereafter; for example, at the University of Zurich, 1909, and the Lautabteilung der Preußischen Staatsbibliothek, Berlin, 1920. By 1901, the Phonogrammarchiv had developed the Wiener Archiv-Phonograph, which used Edison’s cylinder technique to make recordings on wax discs, initially limited to one to two minutes. Although immediately put into use throughout the world, it must be admitted that the earliest phonographs ⫺ weighing in at 45 kilograms ⫺ were not ideally suited for fieldwork, especially given the fact that even in Europe as late as the mid twentieth century the bicycle could serve as the primary mode of transport for the author of an Ortsgrammatik (see, for example, Kühebacher 1958: ii). Special travel phonographs were developed, and the last model in 1927 weighed “only” 9.5 kilograms. In spite of the obvious practical difficulties, the advent of new technologies was put to good use by the early dialectologists, with the first recording of a German dialect (Bavarian) dating to February 1901 (Oberösterreich, Attergau). For a brief historical overview, see Wagener (1988: 164⫺173). Although admittedly hampered by issues of sound quality, the early recordings of the Phonogrammarchiv, especially when combined with a detailed Ortsgrammatik, provide a golden opportunity for the direct study of sound change in real time. For example, Schatz’s (1897) classic description of the Imst dialect is complemented by two Phonogramm recordings, the first a 1904 recording of a worker, the second a 1909 recording of Schatz himself, who, following the Ortsgrammatik tradition, was a native speaker of the dialect. Revisiting the Imst dialect 76 years after Schatz’s original description, Hathaway (1973) demonstrates the value of the early material in her detailed analysis of the sound changes that occurred in Imst over the intervening period. Of the various approaches to dialect study that evolved into the twentieth century, the Vienna School ⫺ focused primarily on Bavarian ⫺ remained closest to neogrammarian principles, with a lineage that can be traced from Lessiak (1903), through Pfalz (1913) and Kranzmayer ([1926] 1981: 1956) to Wiesinger (1970) and beyond. One of the enduring contributions of this School is found in the many Ortsgrammatiken, often written by students as their doctoral dissertations (for an overview of the first six decades of this school, see Wiesinger 1976). There have also been impressive attempts to provide broader German dialect overviews (for example, Kranzmayer 1956; Wiesinger 1970). It is, however, one of the ironies of the German Ortsgrammatik tradition that the great wealth of material presented in the individual grammars does not lend itself easily to comparison and synthesis. In part, this is a natural consequence of the fact that we are dealing with a body of work that has evolved over decades through the efforts of many individual contributors. However, it must also be acknowledged that the difficulties in using the grammars for synthesis and comparative work are exacerbated by the lack of a standard, non-ambiguous transcription system, the absence of systemic, phono-

81

82

II. Linguistic approaches to space logical analyses and the tendency to provide theoretically ad hoc ⫺ and sometimes highly idiosyncratic ⫺ phonological interpretations (on this last point, see, for example, the mora-counting approach outlined in Wiesinger 1983). In spite of the difficulties, the Ortsgrammatiken provide enormous potential for detailed work by historical linguists and phonologists, particularly those who are interested in the study of sound change. Although much seminal analytic work has already been carried out, there is great need for the individual dialects and their interrelationships to be studied within broader well-grounded theoretical and typological frameworks. Given the enormity of data and interpretive difficulties, it is understandable that budding students of historical linguistics and phonology might approach the Ortsgrammatiken with great trepidation, viewing the consolidated works as some impenetrable monolith. In fact, though, this kind of situation is typical in historical linguistics ⫺ we are simply dealing with a manifestation of Roger Lass’s (1989: 75) statement, which can be considered a general law of history: “the closer to the present, the more confused the history […] becomes; this is nowhere truer than in phonology.” Lass suggests that the simplicity of the earlier periods is illusory, a product of the fact that the linguist is “deliciously unencumbered by hard evidence.” There is no doubt that the Ortsgrammatiken qualify as hard evidence, but even more impressively ⫺ scratches, flaws and even missing (syntactic) limbs notwithstanding ⫺ they do indeed constitute cultural gems of the first magnitude, as envisioned by Wegener and others more than 100 years ago.

6. Sound change and messy dichotomies William Labov’s presidential address to the Linguistic Society of America, entitled “Resolving the Neogrammarian Controversy” (⫽ Labov 1981; see also Labov 1994: 2001), reminds us of the recalcitrant nature of the issues that the neogrammarians faced in their attempt to develop a theory of sound change. Much of Labov’s own work has focused on some of the dichotomies we have inherited from the neogrammarian period. Sound change is classified as to whether it is gradual or abrupt, conscious or unconscious, socially or non-socially driven and whether it applies across-the-board (that is, simultaneously to all relevant lexical forms, determined only by phonological environment alone) or by lexical diffusion (word-by-word). There is no doubt that the assumption of dichotomies ⫺ even if they are over-idealized or downright false ⫺ can sometimes serve a useful heuristic function in our initial attempts to understand complex phenomena. Ultimately, however, artificial dichotomies only serve to obscure, and here it is important to remind ourselves that these inherited neogrammarian dichotomies are neither empirically nor theoretically well grounded. The neogrammarians themselves attempted to distinguish two types of sound change. As indicated in section 3 above, their primary focus was on gradual, across-the-board sound change (Lautwandel ), now often simply called neogrammarian change, but they were also aware, of course, that not all sound changes fit into this category. In addition, then, they recognized a class of abrupt changes, including, for example, dissimilation, and they made some attempt to determine non-arbitrary distinguishing properties of the two types. For example, Sievers (1901: 270) argues that whereas gradual regular change involves the creation of new phonological material, abrupt change does not but only involves the redistribution of existing phonological material (an echo of this type of

5. The neogrammarian tradition approach is found in later studies such as Kiparsky 1988). Pursuing a similar type of dichotomy, Labov (1994: 542⫺543). distinguishes: a) regular sound change or “change from below,” which involves the gradual transformation of a single phonetic feature of a phoneme in a continuous phonetic space” (for example, consonant changes in manner of articulation), and b) lexical diffusion or “change from above,” which is “the result of the abrupt substitution of one phoneme for another in words that contain that phoneme” (for example, consonant changes in place or articulation). In fact, it would seem that all such attempts to simplistically categorize sound changes into two types are doomed to failure (see Blevins 2004: 268⫺278 for discussion), and Sievers’ (1901: 272) hedge that it can, of course, be “difficult to clearly determine the border between abrupt and regular change” takes on a prophetic tone. First, abrupt changes that are regular and yield new phonological material have been identified (see discussion, for example, in Blevins 2004: 276). This type of change is not expected according to Labov’s neogrammarian-based classification. Indeed, from a strictly neogrammarian perspective, such changes cannot be categorized at all, because “true” sound change (that is, regular, law-governed change, Lautwandel ) must be ⫺ by definition ⫺ gradual (Paul 1901: 69). Second, there is also no doubt that gradual, regular change (neogrammarian change, senso stricto) can proceed by lexical diffusion (see Haas 1999; Bybee 2002). An interrelated dichotomy is found in treatments of dialect mixture. It became obvious to the neogrammarians that the complex phonological histories of dialects at border areas could not be explained in terms of sound laws alone. The working hypothesis developed that neogrammarian change (gradual, unconscious, regular and non-social) operates within a particular linguistic community, whereas in areas of linguistic intercourse between dialects a kind of borrowing could occur involving sound substitution (Lautersatz), which ⫺ proceeding by lexical diffusion (Wortverdrängung, “word displacement”) ⫺ is abrupt, conscious and socially driven (prestige). However, it was recognized very early, especially in the insightful work of Haag (1929⫺1930), that no clean categorization of the two types could be sustained. Haag (also Lessiak 1933: 6⫺7) recognized that, just like sound substitution, a regular sound change can be borrowed across dialects and be socially driven (see Seidelmann 1987; Auer 2006 for discussion). Our neogrammarian inheritance on these matters is clearly a mixed one, and we must acknowledge that it includes some unwanted baggage, including arbitrary dichotomies, which at this point in time would only seem to hinder progress. By contrast, it is appropriate to acknowledge that the fundamental insights of one of the neogrammarians’ main contemporary critics, Hugo Schuchardt ⫺ no supporter of the dichotomies discussed in this section ⫺ have stood the test of time. According to Schuchardt, a sound change spreads from individual words or groups of words through the lexicon by means of phonetic analogy. All change, whether gradual or abrupt, involves a certain degree of (more or less latent) awareness on the part of speaker. “In regard to the manner in which a sound change is transmitted from individual to individual, from community to community […] I do not see by any means the exclusive play of unconscious activity” (Schuchardt 1885 [1972]: 50). The determining factors in the spread of a sound change within and between linguistic communities are social, and there are no essential differences between the spread of a change from one idiolect to another or from one dialect to another (see Vennemann 1972 for a discussion of Schuchardt’s theory). The jury is still out on the role of child language acquisition in this mixture, and here, as mentioned

83

84

II. Linguistic approaches to space above in section 3.3, the neogrammarians themselves were not in agreement. One thing is for sure though: theories ⫺ neogrammarian, generative, or other ⫺ that ignore sociolinguistic factors cannot account for the directionality of sound change (drift).

7. Reerences Andersen, Henning 1988 Center and periphery: Adoption, diffusion, and spread. In: Jacek Fisiak (ed.), Historical Dialectology. Regional and Social, 39⫺83. Berlin: Mouton de Gruyter. Auer, Peter 2006 Phonological change. In: Ulrich Ammon, Norbert Dittmar, Klaus J. Mattheier and Peter Trudgill (eds.), Sociolinguistics, 1717⫺1727. Berlin: de Gruyter. Bidner, Hans 1939 Zur Lautlehre der Tuxer Mundart. PhD dissertation, University of Innsbruck. Blevins, Juliette 2004 Evolutionary Phonology. The Emergence of Sound Patterns. Cambridge: Cambridge University Press. Bloomfield, Leonard 1933 Language. New York: Holt. Bonfante, Giuliano 1947 The neolinguistic position. (A reply to Hall’s criticism of Neolinguistics.). Language 23: 344⫺375. Braune, Wilhelm 1880 [2004] Gotische Grammatik. Mit Lesestücken und Wörterverzeichnis. [20th rev. ed. by Frank Heidermanns. Tübingen: Niemeyer]. Bybee, Joan 2002 Lexical diffusion in regular sound change. In: David Restle and Dietmar Zaefferer (eds.), Sounds and Systems. Studies in Structure and Change. A Festschrift for Theo Vennemann, 59⫺74. Berlin: Mouton de Gruyter. Chomsky, Noam 1986 Knowledge of Language. Its Nature, Origin, and Use. New York: Praeger. Grimm, Jacob 1819 Deutsche Grammatik, 1st ed. Göttingen. Haag, Karl 1929⫺1930 Sprachwandel im Lichte der Mundartgrenzen. Teuthonista 6: 1⫺35. Haas, Walter 1999 Sprachwandel in apparent time und in real time. In: Wolfgang Schindler and Jürgen Untermann (eds.), Grippe, Kamm und Eulenspiegel, 125⫺144. Berlin: de Gruyter. Hathaway, Luise 1973 Der Mundartwandel in Imst in Tirol zwischen 1897 und 1973. Vienna: Braumüller. Heusler, Andreas [1888] 1970 Der alemannische Consonantismus in der Mundart von Baselstadt. In: Stefan Sonderegger (ed.), Schriften zum Alemannischen, I⫺131. Berlin: de Gruyter. Holthausen, Ferdinand 1886 Die Soester Mundart. Norden und Leipzig: Soltau. Jankowsky, Kurt R. 1972 The Neogrammarians. The Hague: Mouton. Kauffmann, Friedrich 1890 Geschichte der schwäbischen Mundart im Mittelalter und in der Neuzeit. Strasbourg: Karl J. Trübner.

5. The neogrammarian tradition Kiparsky, Paul 1988 Phonological change. In: Frederick J. Newmeyer (ed.), Linguistics. The Cambridge Survey, 363⫺415. Cambridge: Cambridge University Press. Koerner, E. F. K. 1981 [1989] The Neogrammarian doctrine: Breakthrough or extension of the Schleicherian paradigm: A problem in linguistic historiography. Folia Linguistica Historica 2: 157⫺178. [Reprinted as chapter 7 in: Practicing Linguistic Historiography, 79⫺100. Amsterdam: Benjamins.] Knoop, Ulrich 1982 Das Interesse an den Mundarten und die Grundlegung der Dialektologie. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke and Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 1, 1⫺23. Berlin: de Gruyter. Kranzmayer, Eberhard 1956 Historische Lautgeographie des gesamtbairischen Dialektraumes. Vienna: Böhlaus. Kranzmayer, Eberhard [1926] 1981 Laut- und Flexionslehre der deutschen zimbrischen Mundart. Edited by Maria Hornung. Vienna: VWGÖ. Kühebacher, Egon 1958 Dialektgeographie des oberen Pustertales. PhD dissertation, University of Innsbruck. Labov, William 1981 Resolving the Neogrammarian controversy. Language 57: 267⫺308. Labov, William 1994 Principles of Linguistic Change, vol. 1: Internal Factors. Oxford: Blackwell. Labov, William 2001 Principles of Linguistic Change, vol. 2: Social Factors. Oxford: Blackwell. Lass, Roger 1989 How early does English get “modern”? Or, what happens if you listen to orthoepists and not to historians. Diachronica 6: 75⫺110. Lehmann, Winfred P. 1967 A Reader in Nineteenth-Century Historical Indo-European Linguistics. Bloomington: Indiana University Press. Lessiak, Primus 1903 Die Mundart von Pernegg in Kärnten. Marburg: Elwert. Lessiak, Primus 1933 Beiträge zur Geschichte des deutschen Konsonantismus. Brünn: Rohrer. Neumann, Werner 1996 Hermann Paul. In: Harro Stammerjohann (ed.), Lexicon Grammaticorum: Who’s Who in the History of World Linguistics, 706⫺708. Tübingen: Niemeyer. Osthoff, Hermann and Karl Brugmann 1878 Morphologische Untersuchungen auf dem Gebiete der indogermanischen Sprachen, vol. 1. Leipzig. [Excerpt translated into English in: Lehmann 1967, 197⫺209.] Paul, Hermann [1886] 1920 Prinzipien der Sprachgeschichte, 5th ed. Halle/Saale: Max Niemeyer. Pfalz, Anton 1913 Die Mundart des Marchfeldes (Sitzungsberichte der Österr. Akademie der Wissenschaften 170/6.) Vienna: Hölder. Putschke, Wolfgang 1969 Zur forschungsgeschichtlichen Stellung der junggrammatischen Schule. Zeitschrift für Dialektologie und Linguistik 36: 19⫺48. Regel, Karl 1868 Die Ruhlaer Mundart. Weimar: Böhlau.

85

86

II. Linguistic approaches to space Saussure, Ferdinand 1916 Cours de linguistique ge´ne´rale. Edited by Charles Bally and Albert Sechehaye. Lausanne: Payot. Schatz, Joseph 1897 Die Mundart von Imst. Laut- und Flexionslehre. Strasbourg: Trübner. Schirmunsky, Viktor M. 1962 Deutsche Mundartkunde. Vergleichende Laut- und Formenlehre der deutschen Mundarten. Berlin: Akademie Verlag. Schleicher, August 1871 Compendium der vergleichenden Grammatik der indogermanischen Sprachen. Weimar: Böhlau. Schmeller, Johann Andreas 1821 Die Mundarten Bayerns grammatisch dargestellt. Munich: Thienemann. Schuchardt, Hugo 1885 [1972] Über die Lautgesetze. Gegen die Junggrammatiker. Berlin: Robert Oppenheim. [English translation in: Theo Venneman and Terence Wilbur (eds.), Schuchardt, the Neogrammarians, and the Transformational Theory of Phonological Change, 49⫺72. Frankfurt: Athenäum.] (Cited page numbers refer to English translation.) Seidelmann, Erich 1987 Über die Arten der Lautveränderung. In: Eugen Gabriel and Hans Stricker (eds.), Probleme der Dialektgeographie, 200⫺214. Bühl/Baden: Konkordia Verlag. Sievers, Eduard [1876] 1901 Grundzüge der Phonetik zur Einführung in das Studium der Lautlehre der indogermanischen Sprachen, 5th ed. Leipzig: Breitkopf & Härtel. [English translation of excerpt in: Lehmann 1967, 257⫺266.] Sonderegger, Stefan 1968 Alemannische Mundartforschung. In: Ludwig Erich Schmitt (ed.), Festschrift für Walther Mitzka zum 80. Geburtstag, 1⫺30. Wiesbaden: Steiner. Sweet, Henry 1888 A History of English Sounds from the Earliest Period. Oxford: Clarendon. Vennemann, Theo 1972 Hugo Schuchardt’s theory of phonological change. In: Theo Venneman and Terence Wilbur (eds.), Schuchardt, the Neogrammarians, and the Transformational Theory of Phonological Change, 117⫺179. Frankfurt: Athenäum. Wagener, Peter 1988 Untersuchungen zur Methodologie und Methodik der Dialektologie. Marburg: Elwert. Wegener, Philipp 1880 [1976] Über deutsche Dialectforschung. Zeitschrift für Deutsche Philologie 11: 450⫺480. [Reprinted in: Joachim Göschel, Norber Nail and Gaston van der Elst, Zur Theorie des Dialekts. Aufsätze aus 100 Jahren Forschung. Mit biographischen Anmerkungen zu den Autoren, 1⫺29. Wiesbaden: Steiner.] (Cited page numbers refer to reprint.) Weinhold, Karl 1853 Die Laut- und Wortbildung und die Formen der schlesischen Mundart. Wiesbaden: Dr. Martin Sändig. Weinreich, Uriel, William Labov and Marvin I. Herzog 1968 Empirical foundations for a theory of language change. In: Winfred P. Lehmann and Yakov Malkiel (eds.), Directions for Historical Linguistics, 97⫺195. Austin: University of Texas Press. Wiesinger, Peter 1970 Phonetisch-phonologische Untersuchungen zur Vokalentwicklung in den deutschen Dialekten, vol. 1, Die Langvokale im Hochdeutschen. Berlin: de Gruyter.

6. Traditional dialect geography

87

Wiesinger, Peter 1976 Die Wiener dialektologische Schule. In: Helmut Birkhan (ed.), Festgabe für Otto Höfler, 661⫺703. Vienna: Braumüller. Wiesinger, Peter 1983 Dehnung und Kürzung in den deutschen Dialekten. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke and Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 2, 1088⫺1101. Berlin: de Gruyter. Winteler, Jost 1876 Die Kerenzer Mundart des Kantons Glarus in ihren Grundzügen dargestellt. Leipzig and Heidelberg: C. F. Winter’sche Verlagshandlung.

Robert W. Murray, Calgary (Canada)

6. Language and space: Traditional dialect geography 1. 2. 3. 4. 5.

Introduction Adelbert von Keller (1812⫺1885): The initiator of traditional linguistic geography The Marburg School The Württemberg School References

1. Introduction This article examines the earliest “historical” approaches as they developed within traditional German dialectology: the Marburg and Württemberg Schools. The emergence of scholarly dialectology is usually associated with the beginnings of neogrammarian dialectology (cf. Murray in this volume) and with Georg Wenker’s founding of linguistic geography. Both of these approaches received important impetuses at the same time. In 1876, Jost Winteler’s neogrammarian monograph, Die Kerenzer Mundart des Kantons Glarus, which drew inspiration from Eduard Sievers’ Grundzüge der Lautphysiologie (likewise published in 1876), appeared; in the same year Georg Wenker distributed his questionnaire among the public school teachers of the Rhineland (see section 3.1). However, the roots of academic linguistic geography reach back much further than this. Johann Schmeller is generally acknowledged as the first scholarly dialectologist, but his studies were not followed up until much later. In contrast, the pioneering work of Adelbert von Keller (section 2) who, through his students Georg Wenker and Hermann Fischer, became the initiator of the rapidly developing field of linguistic geography (sections 3.1, 4.1) ⫺ is generally underrated. The earliest schools of linguistic geography concentrated exclusively on indirect data collection methods. The criticism of these methods raised by neogrammarians who oriented to the hard sciences’ ideal of exactitude, together with direct data collection on the Romance dialects (section 3.2), led to a systematic supplementing of indirect data with direct data in confined areas (section 4.2). Wrede, in decided opposition to the neogrammarian account of language change, went

88

II. Linguistic approaches to space on to develop his “social-linguistic” method, which attempted to discover correlations between highly localized linguistic differences and historical boundaries and to interpret linguistic boundaries as a result of old barriers to intercourse (Verkehrsgrenzen). In contrast, the representatives of the Württemberg School adopted a more intermediary position and remained more closely linked to the neogrammarian tradition.

2. Adelbert von Keller (18121885): The initiator o traditional dialect geography In 1833 Adelbert von Keller was appointed Chair of ‘Newer Philology’ at the University of Tübingen. Keller’s interest in spoken language, above all the empirical investigation of dialects, reset the focus in language history. He actively advocated the broad-range collection of dialectal materials, and in 1861 he sent out a Bitte um Mitwirkung zur Sammlung des schwäbischen Sprachschatzes ‘Request for cooperation in the collection of the treasures of the Swabian language’, together with extensive instructions, to school teachers in every village in Württemberg (Keller 1855). “All of the words used in Swabia that are not found in the written language or only with another meaning, as well as all […] words differing in inflection, gender or derivation, […] expressions which occur in documents, in the proper names of people, places, rivers, in singular idioms” (Keller 1855: 10; translation R. S.) were to be collected. On the basis of these linguistic data, Keller planned to draft a language map, which portrayed “not just the outward boundaries of the region”, but also “the internal dividing lines. […] The separation into larger groups, like Upper Swabian, Lower Swabian, Highlands, Lowlands, Black Forest and so forth remain[ed] a task for later determination” (Keller 1855: 21, translation R. S.). An essay based on Keller’s suggestion was printed and served as a prototype for 400 essays received from 320 Württemberg village schools (Ruoff 1964: 175, 1982: 128; see Baur 1978, map 5 on the locations). For the German language area, these papers represent the first collection of primary material from a closed dialect region for the analysis of phonological, morphological and lexical issues; they were the precursors of the later regional grammars (section 3.2). Keller’s contribution, Die Mundarten (1884), which resulted from these essays, is the first in a series of geographical dialect descriptions which appeared in Württemberg’s (later Baden-Württemberg’s) district descriptions from 1876 on.

3. The Marburg School 3.1. Georg Wenker (18521911): The indirect method Georg Wenker submitted his doctoral thesis in 1876 under the supervision of Adelbert von Keller on the language-historical topic Verschiebung des Stammsilben-Auslauts im Germanischen ‘The Shift of the Root Syllable Coda in Germanic’. Wenker was strongly influenced by Keller’s method of conducting empirical dialectological investigations. After an unsuccessful attempt to classify the dialects of the Rhineland on the basis of printed dialect samples, he turned to the teachers of the Prussian Rhine Province in 1876

6. Traditional dialect geography with the request that they use the normal alphabet to fill out a questionnaire with 42 sentences which were to be translated into the local dialect. The teachers were provided with diacritica for only a few vowels which could not be represented by the letters of the alphabet. Based on the 1,500 responses (out of 2,200 questionnaires distributed) Georg Wenker wrote a brief paper for the teachers, called Sprachkarte der Rheinprovinz nördlich der Mosel ‘Language Map of the Rhine Province North of the Moselle’ (Wenker 1877), which featured a division into the dialect formations of Westphalian, Low Franconian, Ripuarian and Moselle-Franconian that remains valid to this day (this was labeled by Wenker as Westphalian, Low Rhenish, Low Franconian/Northern Central Franconian and Central Franconian; cf. Schmidt and Herrgen, to appear). In 1877 Wenker sent out a new version of his questionnaire, now with 38 sentences, to the schools of Westphalia. The same year he was employed by the university library in Marburg. In 1878 ministries in Berlin approved funding for a questionnaire survey of all of Prussia. A few months later, a second analysis of the questionnaire was completed; this took the form of the first linguistic atlas in the world, the Sprach-Atlas der Rheinprovinz nördlich der Mosel sowie des Kreises Siegen, of which only a single hand-drawn exemplar with isoglosses in color on preprinted base maps (cf. Schmidt and Herrgen, to appear) was created. The original is kept at the Forschungszentrum Deutscher Sprachatlas in Marburg and has been published via the Internet (). Over the next two years, Wenker distributed the final, forty-sentence questionnaire. By 1881 he had already published two phonological and four morphological maps as the Sprachatlas von Nord- und Mitteldeutschland. Auf Grund von systematisch mit Hülfe der Volksschullehrer gesammeltem Material aus circa 30000 Orten. (For more on Wenker’s early work, see also Knoop, Putschke and Wiegand 1982: 6⫺68; Herrgen 2001: 1520⫺1525 offers a comprehensive overview of the creation of the Deutscher Sprachatlas). Work on these six maps made it clear to Wenker that he could not cope with the mapping of data from approx. 30,000 surveyed locations alongside his occupation as a librarian. He therefore sought financial support from the Office of the Imperial Chancellor, which he was granted on the condition that the survey be extended to cover the entire German Empire of that time. Thus, in 1887 and 1888, the 40 sentences were also sent out across the southern territory (Alsace, Lorraine, Baden, Württemberg and Bavaria). Also included in the survey were the German dialects of Luxemburg, Bohemia and Moravia, the Baltic governorates, Transylvania and the colonies along the Volga. Switzerland and Austria were later added between 1926 and 1933 under Wenker’s colleague and eventual successor, Ferdinand Wrede; in 1939 Wrede’s own successor, Walther Mitzka, added South Tyrol. With that, the entire German-speaking territory of Central Europe had contributed more than 44,000 sentences translated into dialect (König and Schrambke 1999: 15). The Sprachatlas des Deutschen Reichs that this survey resulted in exists as two full-color manuscripts in which 339 linguistic phenomena are displayed on 1,646 hand-drawn maps. The German-speaking territory is spread across three sheets (the northwest, the northeast and the southeast) at a scale of 1 : 1,000,000. Unfortunately, this detailed version of the Deutscher Sprachatlas was never conventionally published; it is, however, now viewable on the Internet as the Digital Wenker Atlas (DiWA). Wenker’s successor, Ferdinand Wrede, managed a partial publication, with greatly simplified maps, under the title of Deutscher Sprachatlas (DSA) which offered an overview of the distribution of 79 linguistic phenomena. A total of 23 installments were printed between 1927 and 1956. A reduced form of the DSA appeared from 1984 onward as the Kleiner

89

90

II. Linguistic approaches to space Deutscher Sprachatlas (KDSA) with maps of the entire DSA phonological geographical material, but with a significantly reduced number of mapped locations compared to the original (Veith and Putschke 1984⫺1999). Morphology was not covered by the KDSA. The first complete publication of the language atlas material was thus the digital version, DiWA (see above).

3.1.1. The 40 Wenker sentences The 40 sentences were designed to cover all of phonology and some morphology. Lexical issues played only a subordinate role; for instance, in sentence 1 (Im Winter fliegen die trocknen Blätter in der Luft herum ‘In winter the dry leaves fly around in the air’), the lexeme Winter serves to capture geographical variation in the root vowel in quantitative and qualitative terms; the lexeme Blätter is phonologically and morphologically interesting. But for some lexemes, Wenker found heteronyms, which cut across his phonological and morphological questions. This gave rise to lexical maps, e.g., for Pferd ‘horse’ (Map 8), Dienstag ‘Tuesday’ (Map 26) and Wiese ‘meadow’ (Map 41) as an unplanned byproduct.

3.1.2. Criticism o Wenkers data collection method In lectures and papers, Wenker repeatedly alluded to the “decline” of the dialects (e.g., Wenker 1886: 192). Nevertheless, in collecting his data he concentrated not on the older generation, but instead on the schoolchildren themselves. In his instructions to the teachers, he wrote: “Let a student or some suitable students make the translation; they know their dialect well enough and will find the work enjoyable. Only where the teacher was born where his school is located and has complete mastery of the dialect is it advisable to perform the translation oneself” (quoted in Stroh 1952: 423, translation R. S.). So Wenker did not rely exclusively on a homogeneous group, neither the schoolchildren nor the teachers, and thus neglected to control the parameters of age, occupation and origin by the selection of informants. This was one of the main points of criticism raised by Wenker’s neogrammarian opponents, especially his chief critic, Otto Bremer, a student of Eduard Sievers. In his Kritik von Wenkers Sprachatlas des Deutschen Reichs ‘Critique of Wenker’s Sprachatlas des Deutschen Reichs’, Bremer writes: If a certain region is hovering between two forms of a word, then naturally one teacher will list one form and another the other form; this is a matter of chance. If the teacher is an older man or if he has turned to older people as his source, then of course the older linguistic form will be listed in Wenker’s questionnaire; if he is perhaps younger or has let his pupils prepare the translation, then the younger linguistic form will certainly be filled out and Wenker will have to draw a line between the two villages. If, perchance, the two teachers had swapped roles, the map would have a completely different appearance. The Sprachatlas offers enough examples of such uncertain lines. (Bremer 1895: 8, translation R. S., cf. also 7, 11)

Wenker, however, was very aware of the problem of vertical variation. Indeed, he himself wrote that the “dialects, through blending with one another and with Standard German, are losing more and more of their purity and naturalness” (quoted from Martin 1934:

6. Traditional dialect geography 9, translation R. S.). Hence, it is also conceivable that Wenker deliberately took advantage of the teachers’ cooperation, assuming that they were familiar with the local base dialect and would intervene to correct uncertain or incorrect responses by the schoolchildren. Wenker’s immense research project could only be realized via written data collection. His full-coverage, phonetically inexact, pupil/teacher competence-based survey thus offers a counterpoint to the single-location, phonetically precise, investigator’s competence-based village grammars of neogrammarian researchers such as Jost Winteler. Criticism from the ranks of the neogrammarians therefore could not be avoided. Wenker met these attacks with the counterargument that the potential source of error of “linguistic variation” (due to the social heterogeneity of the informants) could be corrected by the sheer number of responses.

3.1.3. Results The widespread opinion that it was Wenker’s goal to use the maps of the Linguistic Atlas “to demonstrate the axiom of the exceptionlessness of the sound laws through living dialects” (Bach 1950: 40; translation R. S.) is no longer valid (cf. Wiegand and Harras 1971: 11⫺16; Herrgen 2001). The maps have at any rate disproved the neogrammarian postulate, since they show that “essentially, every single word and every single word form possesses its own zone of validity, its own boundaries in linguistic space; this cannot be captured solely through an acceptance of the effect of analogy and the occasional influence of neighboring dialects or written language admitted by the neogrammarians, if one insists on their understanding of the ‘sound laws’” (Bach 1950: 56⫺57; emphasis removed, translation R. S.). The diversity of isoglosses on the very first maps also disproved Wenker’s earlier hypothesis that “clear dialect boundaries” (Wenker 1886: 189) would emerge from his materials. This also falsified the theory that the boundaries of dialect and tribe coincide, as scholars prior to Wenker had postulated and Wenker too had hoped to prove. Wenker now sought for a new interpretation of dialect boundaries; he no longer saw the atlas as a research product but rather as a research instrument for the interpretation of linguistic geographic findings, for which he increasingly drew upon extralinguistic factors: “[The Sprachatlas] covers our present-day German dialects. Their gradual development into the current multiplicity is inseparably linked to history, the divisions, displacements, migrations, settlements and intermingling of the German tribes” (Wenker 1895a: 36⫺37; translation R. S.).

3.2. Ferdinand Wrede (18631934): Regional (landschatsbasierte) grammars From 1887 on, Wenker’s colleague Ferdinand Wrede continued his cartographic work on the linguistic atlas; as Map 56 of the Deutscher Sprachatlas he included the dialect classification map he had been continuously developing since 1903, originally for his lectures on the German dialects.

91

92

II. Linguistic approaches to space In 1911 he became Wenker’s successor (for an overview of Wrede’s work on the atlas project, see Mitzka 1952: 12). The third mapper was Emil Maurmann. Once the cartographic work was completed, Wrede took over the publication of the Sprachatlas. Under his leadership and together with his students, the Marburg dialectological school was formed, the great success of which was due in part to Wrede’s regular publications, including the reports in the Anzeiger für deutsches Alterthum und deutsche Litteratur (in issues 18, 1892 and 28: 1902), in which the results of the mapping process were made public for fellow scholars. Another significant factor for the school’s influence was the book series Wrede founded, Deutsche Dialektgeographie, which has been appearing since 1908 (see below). Wrede adopted the method used by Wenker in his later phase of interpreting the data on the basis of extralinguistic factors. He distinguished this method, which he labeled “social-linguistic”, from the individual-centered linguistic methods of sound physiology (Lautphysiologie) practiced in the nineteenth century. For Wrede, the term “social” implied “collective”; for him, “social linguistics” was based on “mass materials” from a “mass recording of German dialect resources”. Wrede used “mass” to evoke the “more than 44,000 dialect translations” of Wenker’s sentences (cf. Bellmann 1986: 32). The “social-linguistic factor”, according to Wrede, “includes all of the linguistic phenomena and changes that cannot be explained by reference to the individual, where only the interaction of many individuals comes into question, where manifold cultural influences and all manner of [social] intercourse and, above all, population mixtures are at work” (Wrede 1963: 310⫺311; translation R. S.). Thus, in applying the social-linguistic method to dialectology, the observation and interpretation of the individual dialect within the broadest possible dialect-geographical context and against the historical background of the region under study formed the focus of Wrede’s research interest. Of course, Wenker and Wrede were aware of the weaknesses of the Sprachatlas as described by Bremer. Wenker positioned himself by stressing that his goal was “[to obtain] a little from as many as possible rather than much from an inadequate number of locations” (Wenker 1881: VIII; translation R. S.). In his reply to Bremer’s criticism of the Sprachatlas’ indirect data collection method, he indicated that he was planning “individual studies as a follow-up to the atlas” (Wenker 1895b: 27; translation R. S.). Wrede viewed the Sprachatlas “as a first rough outline”, not a phonetic atlas, “but rather as the precursor to one” (Wrede 1908: IX; translation R. S.). These weaknesses were also manifest for the creators of the Sprachatlas thanks to the Atlas linguistique de la France (ALF); its founder Jules Gillie´ron had his student, Edmond Edmont, use direct methods to collect linguistic materials in order to extract as many phonetically precise data as possible from few locations. Gillie´ron selected 639 boroughs from the French-speaking territory in which Edmont then elicited almost 2,000 words, with great attention to the dialectal vocabulary. By 1903 the first 50 maps of the French atlas were already printed. In total the atlas included 1,920 maps with more than a million tokens (Gauger, Oesterreicher and Windisch 1981: 120⫺132). The advantages and disadvantages of the two very different linguistic atlases are obvious: the linguistic diversity documented by the French method, which included only two percent of all French settlements, lay far behind that of the German method, in which nearly one hundred percent of all localities were surveyed. On the other hand, the indirect data collection of the Germans necessitated subsequent interpretation whenever phonetic details were of interest.

6. Traditional dialect geography The need to supplement the Sprachatlas data with phonetic material was obvious. The neogrammarian monographs which emerged at the same time as the Deutscher Sprachatlas contained exact descriptions of the phonetic systems of their authors’ dialects. In the following years, the contents of the monographs were extended and systematized; in some volumes the phonetic part was followed by a short summary of the most important prosodic features, an overview of the dialect’s basic vocabulary sorted according to the phonemes of the Middle High German reference system and a selection of morphological special features of the dialect (for details see Schmidt and Herrgen, to appear). These local grammars provided an ideal basis on which to selectively verify the Sprachatlas material and check the variations in spelling in the questionnaires. Nonetheless, the theoretically and methodologically contradictory approaches of Wrede and the neogrammarians led Wrede, like Wenker before him, to ignore the neogrammarian monographs. Instead, Wrede encouraged investigations in smaller linguistic regions so as to create an accurate phonetic basis for the interpretation of the Sprachatlas maps. Wrede himself trained his students in phonetic transcription. In order to publish the results, Wrede founded the Deutsche Dialektgeographie (DDG) series, originally with the subtitle “Reports and Studies on G. Wenker’s Sprachatlas des Deutschen Reichs”, the first volume of which appeared in 1908. 100⫺200 locations were examined in each volume. For the series’ authors, the theoretical foundations and methods remained essentially the same for decades. They were described as follows by Gerhard W. Baur in his introduction to DDG volume 55a, Die Mundarten im nördlichen Schwarzwald: “The goal of dialect-geographical observation is to phonetically record the phonological, morphological and lexical inventory of the base dialect as precisely as possible, to make clear the spatial distribution of the linguistic phenomena and to explain their varying forms as far as possible” (Baur 1967: 4; translation R. S.). As a rule, the monographs also share the same structure. At the beginning one finds a comprehensive grammar of a survey location, described (in true neogrammarian tradition) against the backdrop of the Middle High German reference system. Then there is an overview of the geographic, political and religious history of the area under study. A third dialect-geographical section follows, the core of the regional grammar. At this point, Baur departs from the usual descriptive path with its listings of examples and the corresponding locations, instead concentrating on a cartographic depiction, dividing the grammar into a main, cartographic section with phonological, morphological and lexical maps and an explanatory subsection which complements the individual mapped phenomena. Finally, once again following DDG convention, Baur locates the phenomena within the context of the surrounding dialects, presents a linguistic classification of the entire region, compares the course of its internal and external boundaries with prevailing natural and historical conditions and notes linguistic movement. The fourth and last part concerns language change in the area over the past hundred years (Baur 1967: 5⫺10). An important element of the later prestructuralist DDG volumes was thus a historical investigation of dialectal change; for this purpose, Sprachatlas data was compared with more recent materials collected by the author; the questionnaires were also supplemented by phonological, morphological and lexical questions relevant to the region. Baur’s investigation is based on Keller’s 1855 survey, the 40 Wenker sentences and the 200 lemmas of the Deutscher Wortatlas (1951⫺1980; cf. Mitzka 1939). More recent DDG volumes also took into account the results of regional dialect atlases. For Baur this was a ques-

93

94

II. Linguistic approaches to space tionnaire developed by Fischer (1908), the linguistic atlas maps from the Geographie der schwäbischen Mundart (Fischer 1895) and the recordings of the German Sound Archive in Tübingen (cf. Reiffenstein 1982: 32⫺33; Ruoff 2004: 19⫺20; Frahm, Sauer and Schmid 2004: 55). The “linguistic dynamic” analysis in the DDG volumes increasingly expanded to encompass, as in Baur’s book, variables such as occupation, age and other sociolinguistic parameters. In this context, the “regional language” variety (for Baur: Verkehrs- oder Umgangssprache) increasingly came to be included in the investigation. Nearly all of the books in the DDG series included a combination map in which isoglosses and isogloss clusters were shown as boundaries of ranked significance. This mapping strategy can be traced back to Karl Haag (1898), who distinguished up to five levels of boundaries in his work on the Baar dialects (for more detail, see section 4.2.1); this method was adopted by Karl Bischoff (DDG 36), Hans Friebertshäuser (DDG 46), Erika Bauer (DDG 43), Peter Frebel (DDG 45) and Günter Bellmann (DDG 62). The method consisted of adding up all linguistic boundaries that separated two locations from one another. The disadvantage of the resulting “honeycomb maps” (polygonal maps) was that the number of records on which the classification was based varied from author to author (for the books named above, for instance, the threshold for a thirdlevel boundary lay between 29 and 70 records, starting from a base value that ranged between 0 and 18 records). Furthermore, the tokens were not weighted in qualitative terms, i. e. with reference to parameters such as frequency of use and degree of deviance from the standard (primary vs. secondary features); see below for more detail. Baur, however, did not create his combination map solely on the quantitative basis of the dialectal oppositions, but instead evaluated these in line with Haag’s guidelines, in general on phonological grounds. This method, in contrast to the purely additive maps, was rarely copied. Baur aside, only Rudolf Große (1955), Alfred Schirmer (1932) and Heinz Rosenkranz (1938) published evaluative combination maps. Wrede’s aim of being able to combine the dialect-geographical investigations of the DDG series into a contiguous network was attained for a large portion of the Germanspeaking territory. Maps 1 and 3 of the Bibliographie zur Grammatik der deutschen Dialekte (Wiesinger and Raffin 1982) offer an overview of the entire territory; for the southwest of Germany, see Maps 1 and 2 in Baur (2002).

3.3. The inluence o the Marburg School in central and northern Europe 3.3.1. The inluence o the Deutscher Sprachatlas (DSA) Aside from the neogrammarian ideas which were widespread in Europe, German dialectology, especially Wenker’s dialect geography, also exerted much influence, particularly in the north European countries.

3.3.1.1. Norway In Norway, the phonetician Johan Storm conducted numerous direct data collections from 1880 on. Influenced by Wenker, he then switched to indirect collection methods and drafted two questionnaires, initially a short one with 300 lemmas and later a more

6. Traditional dialect geography comprehensive one with 4,000 words, which he distributed to clergymen and teachers across the country (Storm 1882). This collection was to serve as the basis for a historical phonology of the Norwegian dialects. Two years later, Storm trimmed the word list back to 2,000 lemmas, including morpho-geographical questions (Storm 1884). That same year, under the title Norsk Lydskrift med Omrids af Fonetiken ‘Norwegian phonetic alphabet and sketch of Phonetics’, Storm also published suggestions for a phonetic transcription system based on the dialect material he had collected; he derived this from the Swedish transcription system published in the journal Svenska landsma˚l ‘Swedish dialect’ in 1879 by Johan A. Lundell. Storm’s treatise on Norwegian phonetics was published in the first issue of the journal Norvegia and was intended to assist the study of Norwegian dialects and folklore. In the introduction, Storm designs a grand program for dialect research. He argues for a systematic investigation of the dialects as a primary goal, so as to provide the basis for a classification of the Norwegian dialects. Both urban colloquial speech and the higher class Danish-Norwegian colloquial speech were also to be included in the study. Between 1905 and 1909, Hans Ross published a five-part account of the phonology and morphology of the Norwegian dialects with short monographs on the principal dialects of Norway and a dialect classification (Ross 1905⫺1909). He drew his material from a collection of more than 40,000 words that he published as Norsk Ordbog ‘Norwegian Dictionary’ (Ross [1895⫺1913] 1971). Amund B. Larsen is acknowledged as the founder of modern Norwegian dialect studies. Like his teacher, Storm, Larsen was a gifted phonetician indebted to both the neogrammarian tradition and linguistic geography. Like Storm, he also gained his thorough grounding in the Norwegian dialects through his own fieldwork, accompanied by his teacher in 1882. In 1892 he published the first linguistic-geographical map of the dialects of southern Norway with accompanying commentary; a second map followed in 1896 (Larsen 1892: 1896). By the end of the nineteenth century, he had completed his direct data collection for all of Norway and was able to publish a comprehensive overview of the Norwegian dialects in 1897 (Hoff 1968: 422⫺429; Bandle 1962: 299⫺301).

3.3.1.2. Sweden Around the turn of the century, both Johan A. Lundell (1880) and Adolf Nore´en (1903) published distribution maps of the Swedish dialects; both concentrated on the peculiarities of the Swedish marginal dialects. The Central Swedish dialects were given more attention in the classification published in 1905 by Bengt Hesselmann (Hesselmann 1905; also see Benson 1968: 359). Linguistic geography gained greater importance only in the 1930s with the linguists Natan Lindqvist and Delmar Olof Zetterholm at the University of Uppsala. They initiated a written survey, the results of which were published in the 1940s (Lindqvist 1947; Zetterholm 1940: 1953; see also Benson 1968: 364). Additionally, a number of studies which included linguistic-geographical maps emerged around Natan Lindqvist.

3.3.1.3. The Netherlands and Belgium (Flanders) In the Netherlands, the hypothesis ⫺ disproved by German dialectology ⫺ that dialectal and tribal boundaries (Stammesgrenzen) coincide also led to a search for fresh explanations for the emergence of dialect boundaries. Influenced by Wenker’s research methods, H. Kern, based in Leiden, collected written material in the northern Netherlands in 1876;

95

96

II. Linguistic approaches to space edited and supplemented by a survey by Jan te Winkel, this was published in two parts in 1895, after Kern’s death (te Winkel 1898). In the same year, P. Willems from Leuven distributed a questionnaire with 2,000 words across Belgium and the southern Netherlands; in 1914 J. J. Verbeeten, Jos Schrijnen and Jaques van Ginneken conducted indirect data collections in 170 boroughs in Limburg and East Brabant (Bach 1950: 45; Weijnen 1982: 192). In 1935, Dutch translations of the 40 Wenker sentences were sent out by the Dialect Committee of the Royal Dutch Academy of Sciences in Amsterdam as a third questionnaire (Vragenlijsten 1931⫺1958). The cooperation between Marburg and Amsterdam was the work of F. Wrede and P. J. Meertens, the secretary of the Dutch Dialect Committee. This survey was intended to enable the expansion of the boundaries of the Deutscher Sprachatlas into Dutch-speaking territory (Meertens 1936: 123⫺124).

3.3.2. The inluence o the Deutsche Dialektgeographie (DDG) series The book series Deutsche Dialektgeographie also attracted attention in neighboring countries. In Norway, Ingeborg Hoff (1946) published a study that included an extensive description of a village dialect as well as a linguistic-geographical overview of the dialects of the entire province; the latter was supplemented by 15 maps (Bandle 1962: 303⫺304). In the Netherlands and Flanders, over 50 village studies were published between 1884 and 1962 (for a list see Goossens 1968: 183⫺183, 186; Meertens and Wander 1958). In 1920 De Isoglossen van Ramisch in Nederland by Joseph Schrijnen appeared. In Flanders, the local monographs were supplemented by approx. 130 unpublished doctoral and graduate theses written primarily at the universities of Leuven and Ghent. Following the example of the DDG volumes from 1922 on, these included linguistic-geographical maps as well as a historical grammar. In Denmark in 1933, Karl Bock published a study about Niederdeutsch auf dänischem Substrat as DDG number 34. His investigation was based on responses to the 40 Wenker sentences from the Deutscher Sprachatlas augmented by his own survey and went on to present his results in a number of maps (Andersen 1968: 338).

4. The Württemberg School The power of the Marburg School extended to the regional research conducted in Württemberg, where dialectology was already firmly established. As shown above, a first connection between Tübingen and Marburg was furnished by Georg Wenker. But there were also dialectologists in Tübingen who remained loyal to the neogrammarian methodology. Outstanding researchers from the Württemberg School were Hermann Fischer, Karl Haag and Karl Bohnenberger.

4.1. Hermann Fischer (18511920): The indirect method Hermann Fischer was a student of Adelbert von Keller from 1870⫺1871. In 1888 he took over from Keller’s successor, Eduard Sievers, as a professor at Tübingen University.

6. Traditional dialect geography He assumed office with an inaugural lecture on the methods and goals of dialectology (Über Wege und Ziele der Dialektforschung). Keller gave Fischer access to the “conference papers” and 400,000 response slips that Keller had sorted as basis for a Schwäbisches Wörterbuch (Swabian Dictionary; Fischer 1904⫺1936), thus also providing a foundation for Fischer’s atlas, Geographie der schwäbischen Mundart (Fischer 1895). Fischer extended this collection by a written survey based on Wenker’s methods: in 1886 and 1887 he distributed a 200-question survey of dialectal phonology to all of the more than 3,000 parishes in Württemberg and surrounding areas. The questionnaire was returned by about half of the parishes, which Fischer used as a basis for his 28-map linguistic atlas. Fischer wanted the atlas to unburden his Schwäbisches Wörterbuch, the first installment of which appeared in 1901, of details about the geographical distribution of regular linguistic phenomena: “It seemed desirable to me to precede the Swabian Dictionary with a geographical-grammatical study so as to provide orientation about regular variation in the phonological form of individual words in advance” (Fischer 1895: III, translation R. S.). Fischer began his work on the atlas at almost exactly the same time as Wenker. His work was based on Wenker’s ideas, especially his Rheinischer Sprachatlas (1878) and the beginnings of the Deutscher Sprachatlas as described by Ferdinand Wrede in the Anzeiger für deutsches Alterthum und deutsche Litteratur. Although Fischer’s atlas is among the most significant dialect geographical contributions in the German-speaking world, it produced little reaction in comparison to the Deutscher Sprachatlas. Fischer, like Wenker and Wrede before him, was concerned with the most hotly discussed topics of his time ⫺ the regularity of phonological change and the correspondence of dialectal and other boundaries. Fischer had always been skeptical of the neogrammarian ideas; in the end he made his opposition public. Like Wrede and later Haag, he was convinced that the boundaries for individual words with the same vowel do not coincide regularly and concluded that “[a] sound law cannot apply with the necessity of a law of nature; otherwise it would always have had effect everywhere” (Fischer 1895: 83; translation R. S.). His main concern was the relationship between the old Germanic tribes and the current dialect boundaries, which had been discussed before and after Wenker up until the mid-twentieth century (cf. Nübling 1938; Moser 1951⫺1952, 1961). For Fischer, there is no Swabian dialect as such; instead, “every particular characteristic linguistic phenomenon has a closed territory and definite borders” (Fischer 1895: 82; translation R. S.). Every linguistic innovation, according to Fischer, arises in a certain location (or possibly several locations), from where it spreads out over a greater or smaller area. He thus aligns himself with Johann Schmidt’s wave theory, which he saw as “the only [theory] with which one succeeds at all in explaining the history of language” (Fischer 1895: 83; translation R. S.). These linguistic innovations stop at natural or political boundaries. From his maps, Fischer ⫺ again in agreement with Wenker and Wrede ⫺ was able to detect neither a coincidence of tribal and dialect boundaries nor telltale features which could characterize Swabian as an old tribal dialect. Although there were clusters of boundary lines on Fischer’s combination map (for instance in the northeast, where Swabian Alemannish borders on East Frankish, and in the east, where it adjoins Bavarian), Fischer drew no further conclusions from this. In the foreword to his atlas he wrote: “Combining the boundary lines from my first 25 maps on a single map reveals a picture

97

98

II. Linguistic approaches to space of extreme lawlessness. The lines run in every conceivable direction and there are only a few localities on the map with no boundary of some kind running through them” (Fischer 1895: 82; translation R. S.).

4.2. Small-scale linguistic geography in the Württemberg School: Karl Haag and Karl Bohnenberger 4.2.1. Karl Haag (1860-1946) Keller’s method of conducting small-scale dialect geographical research on an empirical basis was taken up by Karl Haag. He abandoned the “murky atmosphere of armchair hypotheses” (Haag 1900: 139, translation R. S.) to undertake a “collection of recordings from over 200 locations, which form a closed area of almost 60 square miles” using direct methods which he supplemented by an indirect collection of data with the help of teachers in order to fill “the unfortunately constantly recurring gaping omissions” (Haag 1898: 3⫺4; translation R. S.). He thus achieved the ideal combination of the research methods applied by the Deutscher Sprachatlas and the Atlas linguistique de la France. Just how much he followed Marburg research in doing so is indicated by the fact that he intervened in the dispute between Bremer, Wenker and Wrede. Haag’s combined material forms the basis for his monograph, Die Mundarten des oberen Neckar- und Donaulandes, a seminal work for the entire field of dialect geography ⫺ not in the least thanks to Haag’s new cartographic procedures, which also occupied an essential place in all of his subsequent dialect-geographical papers. In his foreword to this first large dialect-geographical study he writes: The accumulated phonological geographical material could only be made clear with a map; this became increasingly central to the work, especially when, in searching for an explanation for the lines that emerged, I made the happy discovery that there was a surprisingly extensive agreement with old political boundaries. This led to a division of the work at hand into four: a monograph on [Schwenningen], a comparative account of the entire area, an investigation of the nature of the boundaries revealed by the map, and a modest collection of samples. (Haag 1898: 4; translation R. S.)

Haag was the first dialectologist to insist that the features used to delimit a dialect region must be weighted: “The significance with which we credit individual boundaries is to be judged in three ways: the number of forms affected by the sound change, the frequency of use of these forms, and the degree to which they have changed” (Haag 1898: 93, translation R. S.). These three criteria are narrowed down by Haag. On point 1, “The question of the number can only be answered by the most comprehensive collection of material. This effort would facilitate a statistical analysis of sounds and allow the sorting of the MHG word and form inventory into phonetically meaningful units: initial sounds, final sounds with short vowels, [and final sounds] with long vowels would need to be posited” (Haag 1898: 93, translation R. S.). On point 2: Phonological change affecting the core vocabulary of a language ⫺ pronouns, particles, modal verbs ⫺ is of especially great consequence for the interpretation of the boundaries;

6. Traditional dialect geography thus the group of aˆn: haˆn, staˆn, gaˆn, laˆn is of more weight than that of ht, although it has five times as many forms. Statistics on the rate of occurrence of frequently used words would also be a precondition for an accurate assessment here. (Haag 1898: 93, translation R. S.)

On point 3, “The auditive salience [Ohrfälligkeit] of the phonological change, that which beyond the boundary is felt to be particularly odd, is an important, subjective dimension of the change. This reflection is, however, precisely due to its subjective character, only permitted little room on the map” (Haag 1898: 93, translation R. S.). Haag concludes: “Considerations of this kind […] lead to a numerical ranking of the boundaries, as drawn on the map. Five border strengths are distinguished: the border lines are 4, 3, 2, 1, ½ mm in breadth; the number and frequency of the forms they circumscribe increase almost in an exponential relationship: (more than) 200, 100, 51, 21, 10” (Haag 1898: 93⫺94; translation R. S.). Haag’s polygon map of the Baar dialects reveals 78 boundaries for individual dialectal features (Haag 1898: Appendix). Haag calculated the relative strength of every boundary by counting and weighting the common features of each enclosed segment. This method of display produced a central “core area” in every “linguistic landscape” surrounded by a transition zone, a “fringe area” or, as Haag later termed it, “a vibration area”. Haag was convinced that the boundaries of the linguistic landscape were connected with the political traffic (intercourse) controls of the late medieval territories. Since, according to Haag, former political boundaries do not persist in language for longer than 300 years (Haag 1900: 139), there is no connection between old tribal borders and dialect boundaries unless an old tribal boundary coincides with a later political border (Haag 1946: 3⫺4). Geographical barriers are less relevant for the establishment of boundaries than political ones. Whether these boundaries are linguistically reinforced depends in the final instance on the “intercourse” that effects the spread of a linguistic innovation; if it halts at an obstacle, the linguistic adjustment is also interrupted. For Haag, the second important factor for the spread of a linguistic innovation is the higher prestige of a form. Haag agrees with Fischer that the extent of the spread is in line with Schmidt’s wave principle. Between 1925 and 1930, Haag expanded his linguistic geographical research to cover all of Württemberg. In his Die Grenzen des Schwäbischen in Württemberg (1946) he attempted a demarcation of the individual linguistic regions: We determine them according to the strength of the strands of the phonological boundaries that traverse the state, and in so doing keep an eye out for the boundaries of the erstwhile political areas to which they are oriented. The Roman Empire of the German Nation, in all its remarkable variegation with its princes, churches, knightly and civil governances, will then provide the names for labeling the linguistic regions. (Haag 1946: 10; translation R. S.)

Haag distinguishes fifteen linguistic regions, but concedes that “only the urgent need to clamber from the confusion of diversity to a vantage point [is] redressed” (Haag 1946: 10; translation R. S.). Fischer’s account of the “lawlessness” of the boundaries was thus disproved; in fact, Fischer later agreed with Haag’s linguistic geographic approach. Haag’s significant findings on the exceptionlessness of the sound laws were a result of his empirical data collection methods. According to Haag, there are two kinds of sound change: “current” and “older”. Current change is found where a sound law is still in

99

100

II. Linguistic approaches to space effect. It includes all words belonging to the affected cohort without exception; this sound change is triggered by minor articulatory shifts of which the speaker remains unaware. “Relaxing lip closure makes every b into a w; fronting the stricture makes of every ix an ic¸; opening makes of every ir an e˛r, every ¯ın becomes an e¯n; closing the nose denasalizes all vowels” (Haag 1929⫺1930: 28; translation R. S.). For Haag, these results can only be derived from directly collected linguistic data: “The transmission from individual to individual, from dialect to dialect resembles unconscious infection; neighboring dialects display only subtle differences of degree, like a slow increase or decrease in a phenomenon; their geographic spread is therefore only rarely precisely determinable; this present-day sound change thus has largely fluid boundaries” (Haag 1898: 88; translation R. S.). Haag sees this sound change as extremely important, because many words are affected. “Older sound change”, in contrast, is sound replacement in which only single words belonging to a cohort are borrowed from a neighboring dialect via “conscious importation” (Haag 1898: 89; translation R. S.). Haag’s suggestion, never enacted, was to conduct direct recordings across all of Germany. The Deutscher Sprachatlas was to be divided into sectors, the results coordinated and depicted on a single map (Haag 1928: 167).

4.2.2. Karl Bohnenberger (18631951) The phonetic and grammatical lectures of his neogrammarian teacher, Eduard Sievers, provided decisive impetuses for Karl Bohnenberger. Yet while Sievers had no interest in the linguistic geographic material of his predecessor, Adelbert von Keller, on which neogrammarian axioms could have been tested, linguistic geography was a central topic of research for Bohnenberger. Like Karl Haag, he was an empiricist; the two dialectologists from Tübingen collected linguistic data in the field at the same time. His reservations about the Deutscher Sprachatlas, which he never shook off (Bohnenberger 1902: 329), were a result of (among other factors) the differing approaches to data collection. Steadfast in the neogrammarian tradition, he argued against the collection of written data, which in his opinion obscured the causes of language change, an issue that concerned him as much as it did Karl Haag. In his Über die ostgrenze des alemannischen he emphasized in “Digression 3, On the extent of sound change and the causes of linguistic boundaries” (1928b: 63, translation R. S.) that statements about the phenomena of language change are only possible when based on direct data collection: “Here again in many things a clear distinction arises between researchers who proceed principally from written sources and those who principally proceed from spoken and heard speech” (translation R. S.). He also criticized not just Fischer’s questionnaire methods “with all due respect to the services that this atlas has provided and continues to provide” (Bohnenberger 1928b: 54⫺55; translation R. S.), but also “the Marburg pronouncements”, especially Kurt Wagner’s results published in 1927 in his work Deutsche Sprachlandschaften (DDG 23): “Thus it appears to me incomprehensibly daring, when K. Wagner […] indulges in opinions about Alemannish or Bavarian” (Bohnenberger 1928b: 55, footnote 1; translation R. S.). He praised all the more the direct recordings initiated by Wrede: “Obviously, the excellent and very useful direct dialect recordings from individual districts in central and northern Germany, commissioned by the Marburg Center and in-

6. Traditional dialect geography cluded in Wrede’s familiar collection, must be considered field-based research.” (Bohnenberger 1928b: 56, footnote 1). Bohnenberger posits two kinds of linguistic change: on the one hand, indigenous, inherited (überkommen) phonological change and on the other, constant and sporadic change. Most linguistic boundaries can be attributed to inherited sound change and the effects of analogy. Again in agreement with Karl Haag, Bohnenberger sees the triggers of language change in the higher prestige of a linguistic form; this is dependent on the political and economic primacy of certain regions and the standing of the speakers, especially with regard to their occupation (Bohnenberger 1953: 259⫺260). Unlike Wenker, Fischer or Haag, Bohnenberger saw reflections of former tribal borders in contemporary linguistic boundaries; later research did in fact vindicate this view, at least in some instances (cf. Moser 1951⫺1952, 1961; Nübling 1938: 241⫺242). Whether or not these tribal boundaries were maintained depended on the communicative intercourse (Verkehr; cf. Bach 1950: 66; Bohnenberger 1953: 225; Engel 1964: 225); thus Bohnenberger, like Haag, assigned communication a decisive role in the establishment and maintenance of dialect boundaries. Bohnenberger analyzed his material in many papers; they provided the basis for the internal and external classification in Die Mundarten Württembergs and the comprehensive account of Die alemannische Mundart (Bohnenberger 1928a, 1928b: 1953) . It was above all in Über die ostgrenze des alemannischen (1928b) that he intensively considered a “classificatory scheme” for dialects, especially for the fixing of border lines. Although he distinguished full Alemannic from a band of contiguous transitional dialects, he maintained that a sharp delineation was possible here, too: “Pre-boundary lines separate full dialects from transitional dialects” and “total boundary lines separate complete dialects” (Bohnenberger 1928b: 30; translation R. S.).

5. Reerences Andersen, Poul 1968 Dänische Mundartforschung. In: Ludwig Erich Schmitt (ed.), Germanische Dialektologie. Festschrift für Walther Mitzka zum 80. Geburtstag, vol. 2, 319⫺347. (Zeitschrift für Mundartforschung, Beihefte NF 6.) Wiesbaden: Steiner. Bach, Adolf 1950 Deutsche Mundartforschung. Ihre Wege, Ergebnisse und Aufgaben. 2nd ed. Heidelberg: Winter. Bandle, Oskar 1962 Die norwegische Dialektforschung. Ein Überblick über ihre Geschichte und ihre heutige Situation. Zeitschrift für Mundartforschung 29: 289⫺312. Bauer, Erika 1957 Dialektgeographie im südlichen Odenwald und Ried. (Deutsche Dialektgeographie 43.) Marburg: Elwert. Baur, Gerhard W. 1967 Die Mundarten im nördlichen Schwarzwald. (Deutsche Dialektgeographie 55a, b.) Marburg: Elwert. Baur, Gerhard W. 1978 Bibliographie zur Mundartforschung in Baden-Württemberg, Vorarlberg und Liechtenstein. (Idiomatica 7.) Tübingen: Niemeyer.

101

102

II. Linguistic approaches to space Baur, Gerhard W. 2002 Bibliographie zur Mundartforschung in Baden-Württemberg, Vorarlberg und Liechtenstein von den Anfängen bis zum Jahr 2000. 2nd ed. (Idiomatica 7.) Tübingen: Niemeyer. Bellmann, Günter 1961 Mundart und Umgangssprache in der Oberlausitz. Sprachgeographische und sprachgeschichtliche Untersuchungen zwischen Schwarzwasser und Lausitzer Neiße. (Deutsche Dialektgeographie 62.) Marburg: Elwert. Bellmann, Günter 1986 Zweidimensionale Dialektologie. In: Günter Bellmann (ed.), Beiträge zur Dialektologie am Mittelrhein, 1⫺55. (Mainzer Studien zur Sprach- und Volksforschung 10.) Stuttgart: Steiner. Benson, Sven 1968 Die Erforschung der schwedischen Mundarten. In: Ludwig Erich Schmitt (ed.), Germanische Dialektologie. Festschrift für Walther Mitzka zum 80. Geburtstag, vol. 2, 348⫺368. (Zeitschrift für Mundartforschung, Beihefte NF 6.) Wiesbaden: Steiner. Bischoff, Karl 1935 Studien zur Dialektgeographie des Elbe-Saale-Gebietes in den Kreisen Calbe und Zerbst. (Deutsche Dialektgeographie 36.) Marburg: Elwert. Bock, Karl Nielsen [1933] 1974 Niederdeutsch auf dänischem Substrat. Studien zur Dialektgeographie Südostschleswigs. (Deutsche Dialektgeographie 34.) Walluf: Sändig. Bohnenberger, Karl 1902 Sprachgeschichte und politische Geschichte. Zeitschrift für Hochdeutsche Mundarten III: 321⫺326. Bohnenberger, Karl 1928a Die Mundarten Württembergs. Eine heimatkundliche Sprachlehre. (Schwäbische Volkskunde 4.) Stuttgart: Silberburg. Bohnenberger, Karl 1928b Über die ostgrenze des alemannischen. Tatsächliches und grundsätzliches. Mit 1 Karte. [Reprint from Beiträge zur Geschichte der deutschen Sprache und Literatur, vol. 52.] Halle, Saale: Niemeyer. Bohnenberger, Karl 1953 Die alemannische Mundart. Umgrenzung, Innengliederung und Kennzeichnung. Tübingen: Mohr. Bremer, Otto 1895 Beiträge zur Geographie der deutschen Mundarten, in Form einer Kritik von Wenkers Sprachatlas des Deutschen Reichs. (Sammlung kurzer Grammatiken deutscher Mundarten 3.) Leipzig: Breitkopf & Härtel. Deutscher Wortatlas 1951⫺1980 (Based on the Sprachatlas des Deutschen Reichs by Georg Wenker, initiated in simplified version by Ferdinand Wrede, continued by Walther Mitzka and Bernhard Martin.) Gießen: Schmitz. Deutscher Sprachatlas 1927⫺1956 (Based on the Sprachatlas des Deutschen Reichs by Georg Wenker, initiated by Ferdinand Wrede, contined by Walther Mitzka and Bernhard Martin.) Marburg: Elwert. Engel, Ulrich 1964 Karl Bohnenberger 1863⫺1951. In: Zur Geschichte von Volkskunde und Mundartforschung in Württemberg, 210⫺242. (Volksleben 5.) Tübingen: Tübinger Vereinigung für Volkskunde. Fischer, Hermann 1895 Geographie der schwäbischen Mundart. Tübingen: Laupp. Fischer, Hermann 1904⫺1936 Schwäbisches Wörterbuch (continued by Wilhelm Pfleiderer). Tübingen: Laupp.

6. Traditional dialect geography Fischer, Hermann 1908 Einige Winke über Forschungen über Schwäbische Mundarten. Korrespondenzblatt für höhere Schulen Württembergs 15: 81⫺98. Frahm, Eckart, Eckhard Sauer and Sigrid Schmid 2004 Digitalisierung der historischen Tonbandaufnahmen. Technische Angaben und Erfahrungsberichte. In: Dialekt und Alltagssprache. Arbeitsstelle Sprache in Südwestdeutschland, 53⫺64. (Tübinger Korrespondenzblatt 57.) Tübingen: Gulde. Frebel, Peter 1957 Die Mundarten des westlichen Sauerlandes zwischen Ebbegebirge und Arnsberger Wald. (Deutsche Dialektgeographie 45.). Marburg: Elwert. Friebertshäuser, Hans 1961 Sprache und Geschichte des nordwestlichen Althessen. (Deutsche Dialektgeographie 46.) Marburg: Elwert. Gauger, Hans-Martin, Wulf Oesterreicher and Rudolf Windisch 1981 Einführung in die romanische Sprachwissenschaft. Darmstadt: Wissenschaftliche Buchgesellschaft. Goossens, Jan 1968 Zur Geschichte der niederländischen Dialektologie. In: L. E. Schmidt (ed.), Germanische Dialektologie (⫽ FS W. Mitzka), 180⫺208. Wiesbaden: Steiner. Große, Rudolf 1955 Die meißnische Sprachlandschaft. Dialektgeographische Untersuchungen zur obersächsischen Sprach- und Siedlungsgeschichte. (Mitteldeutsche Studien 15.) Halle, Saale: Niemeyer. Haag, Karl 1898 Die Mundarten des oberen Neckar- und Donaulandes (Schwäbisch-alemannisches Grenzgebiet: Baarmundarten). Reutlingen: Hutzler. Haag, Karl 1900 7 Sätze über Sprachbewegung. Zeitschrift für Hochdeutsche Mundarten I: 138⫺141. Haag, Karl 1928 Drei Wege der Mundartforschung. Germanisch-Romanische Monatsschrift XVI: 165⫺167. Haag, Karl 1929⫺1930 Sprachwandel im Lichte der Mundartgrenzen. Teuthonista 6: 1⫺35. Haag, Karl 1946 Die Grenzen des Schwäbischen in Württemberg (mit 3 Karten). (Schwäbische Volkskunde NF 8.) Stuttgart: Wissenschaftliche Verlagsgesellschaft. Herrgen, Joachim 2001 Die Dialektologie des Deutschen. In: Sylvain Auroux, E. F. K. Koerner, Hans-Josef Niederehe and Kees Versteegh (eds.), Geschichte der Sprachwissenschaften. Ein internationales Handbuch zur Entwicklung der Sprachforschung von den Anfängen bis zur Gegenwart, vol. 2, 1513⫺1535. (⫽ Handbücher zur Sprach- und Kommunikationswissenschaft 18.2.) Berlin/New York: de Gruyter. Hesselmann, Bengt 1905 Sveama˚len och de svenska dialekternas indelning [Svea-Dialects and the Classification of the Swedish Dialects]. Uppsala: Appelberg. Hoff, Ingeborg 1946 Skjetvema˚let : utsyn over lydvoksteren i ma˚let i Skiptvet i Østfold i jamføring med andre østfoldske ma˚l [Skjetve-Dialect: Overview of the Phonemic Inventory of the Skiptvet Dialect in Østfold in Comparison to other Dialects in Østfold]. (Skrifter / Det Norske videnskaps-akademi i Oslo 2. Historisk-filosofisk klasse.) Oslo: Dybwad. Hoff, Ingeborg 1968 Norwegische Mundartforschung. In: Ludwig Erich Schmitt (ed.), Germanische Dialektologie. Festschrift für Walther Mitzka zum 80. Geburtstag, vol. 2, 398⫺458. (Zeitschrift für Mundartforschung, Beihefte NF 6.) Wiesbaden: Steiner.

103

104

II. Linguistic approaches to space Keller, Adelbert von 1855 Anleitung zur Sammlung des schwäbischen Sprachschatzes. In: Einladungsschrift der Universität Tübingen zum 27. September 1855. Tübingen: Fues. Keller, Adelbert von 1884 Die Mundarten. In: Das Königreich Württemberg. Eine Beschreibung von Land, Volk und Staat, Part II, vol. 1, 166⫺177. Stuttgart: Kohlhammer. Knoop, Ulrich, Wolfgang Putschke and Herbert Ernst Wiegand 1982 Die Marburger Schule: Entstehung und frühe Entwicklung der Dialektgeographie. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke, Herbert Ernst Wiegand (ed.): Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 1, 38⫺92. (⫽ Handbücher zur Sprach- und Kommunikationswissenschaft 1.1.) Berlin/New York: de Gruyter. König, Werner and Renate Schrambke 1999 Die Sprachatlanten des schwäbisch-alemannischen Raumes. Bühl, Baden: Konkordia. Larsen, Amund Bredesen 1892 Oversigtskart over visse dialektfænomeners udbredelse i Kristiansands stift [Overview Map of the Spread of Several Dialectal Phenomena in the Bishopric of Kristiansand]. (Forhandlinger. Det Norske videnskaps-akademi 9.) Christiania: Dybwad. Larsen, Amund Bredesen 1896 Om de norske dialekters forhold til nabosprogene [On the Norwegian Dialects’ Relation to Neighboring Languages]. In: Sproglig-historiske studie. Tilegnede Professor C. R. Unger, 1⫺11. Kristiana: Aschehoug. Larsen, Amund Bredesen 1897 Oversigt over de norske bygdema˚l [Overview of Norwegian Rural Dialects]. Kristiania: Aschehoug. Lindqvist, Natan 1947 Sydväst-Sverige i spra˚kgeografisk belysning [Southwest Sweden from a Linguistic-Geographical Perspective]. (Skrifter utgivna genom Landsma˚lsarkivet i Lund 2.) Lund: Carl Blom. Lundell, Johan August 1880 De svenska folkma˚lens frändskaper och etnologiska betydelse [The Swedish Folk Dialects and their Relatedness and Ethnological Significance]. Antropologiska sektionens tidskrift 1(5). Martin, Bernhard 1934 Georg Wenkers Kampf um seinen Sprachatlas (1875⫺1881). In: Bernhard Martin and Ferdinand Wrede, Von Wenker zu Wrede, 1⫺37 2nd ed. (Deutsche Dialektgeographie 21.) Marburg: Elwert. Meertens, Pieter J. 1936 Niederländische Mundartforschung. Zeitschrift für Mundartforschung 12: 125⫺127. Meertens, Pieter J. and Boy Wander 1958 Bibliografie der dialecten van Nederland. 1800⫺1950 [Bibliography of the Netherlands Dialects. 1800⫺1950]. Amsterdam: Noord-Hollandsche uitgevers maatschappij. Mitzka, Walther 1939 Der Fragebogen zum deutschen Wortatlas. Zeitschrift für Mundartforschung 15: 105⫺ 111. Mitzka, Walther 1952 Handbuch zum deutschen Sprachatlas. Marburg: Elwert. Moser, Hugo 1951⫺1952 Stamm und Mundart (mit 3 Karten). Zeitschrift für Mundartforschung 20: 129⫺ 145. Moser, Hugo 1961 Noch einmal: Stamm und Mundart. Zeitschrift für Mundartforschung 28, 32⫺42.

6. Traditional dialect geography Nore´en, Adolf 1903 Va˚rt spra˚k: nysvensk grammatik i utförlig framställning I [Our Language: New Swedish Grammar Presented in Detail I]. Lund: Gleerup. Nübling, Eduard Friedrich 1938 Die “Dreistammesecke” in Bayern (Schwäbisch-Bairisch-Fränkisch) in sprachlicher und geschichtlicher Betrachtung. PhD thesis, University of Augsburg. [Reprinted from the Jahrbuch des Historischen Vereins für Schwaben und Neuburg 53, 185⫺300.] Reiffenstein, Ingo 1982 Das phonetische Beschreibungsprinzip als Ergebnis junggrammatischer und dialektologischer Forschungsarbeiten. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke and Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 1, 23⫺38. (⫽ Handbücher zur Sprach- und Kommunikationswissenschaft 1.1.) Berlin/New York: de Gruyter. Rosenkranz, Heinz 1938 Mundart und Siedlung im Gebiet der oberen Saale und des nördlichen Frankenwaldes. Jena: Fischer. Ross, Hans [1895⫺1913] 1971 Norsk Ordbog [Norwegian Dictionary]. Oslo: Univ. Forl. Ross, Hans 1905⫺1909 Norske Bygdemaal [Norwegian Rural Dialects]. (VSS II, 1905, no. 2; 1906, no. 3; 1907, no. 5; 1908, no. 4; 1909, no. 3.) Kristiania: Dybwad. Ruoff, Arno 1964 Hermann Fischer. In: Zur Geschichte von Volkskunde und Mundartforschung in Württemberg, 171⫺192. Tübingen: Tübinger Vereinigung für Volkskunde. Ruoff, Arno 1982 Die Forschungstätigkeit der Württembergischen Schule als Beispiel regionaler Dialektologie. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke, Herbert Ernst Wiegand (ed.): Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 1, 127⫺ 144. (⫽ Handbücher zur Sprach- und Kommunikationswissenschaft 1.1.) Berlin/New York: de Gruyter. Ruoff, Arno 2004 Die Tübinger Arbeitsstelle “Sprache in Südwestdeutschland” 1955⫺1995. In: Dialekt und Alltagssprache. Arbeitsstelle Sprache in Südwestdeutschland, 19⫺52. (Tübinger Korrespondenzblatt 57.) Tübingen: Gulde. Schirmer, Alfred 1932 Beiträge zur nordthüringischen Dialektgeographie. (Deutsche Dialektgeographie 26.) Marburg: Elwert. Schmidt, Jürgen Erich and Joachim Herrgen to appear Sprachdynamik. Eine Einführung in die moderne Regionalsprachenforschung. Berlin: Erich Schmidt. Schrijnen, Joseph 1920 De Isoglossen van Ramisch in Nederland [The Isoglosses of Ramisch in the Netherlands]. (Wetenschappelijk onderzoek der Zuid-Oostelijke dialekten 1.) Bussum: Brand. Sievers, Eduard 1876 Grundzüge der Lautphysiologie: Zur Einführung in das Studium der Lautlehre der indogermanischen Sprachen. Leipzig: Breitkopf und Härtel. Storm, Johan 1882 Norsk Ordliste til Lydlæren [Norwegian Word List of Phonetics]. Kristiania: Foreningen for norske Dialekter og Folketraditioner. Storm, Johan 1884 Kortere Ordliste med Forklaring af Lydskriften [Shorter Word List with Explanation of Phonetic Transcription]. Publisher and place unknown.

105

106

II. Linguistic approaches to space Stroh, Friedrich 1952 Handbuch der germanischen Philologie. Berlin: de Gruyter. Veith, Werner H. and Putschke, Wolfgang 1983⫺1999 Kleiner Deutscher Sprachatlas. Tübingen: Niemeyer. Vragenlijsten met register 1931⫺1958 Bidragen en Mededelingen der Dialectenkommissie van de Koninklijke Nederlandse Akademie van Wetenschappen te Amsterdam XXII [Contributions and Communications of the Dialect Comission of the Royal Dutch Academy of Sciences in Amsterdam XXII]. Amsterdam: Noord-Holland. Wagner, Kurt 1927 Deutsche Sprachlandschaften. (Deutsche Dialektgeographie 23.) Marburg: Elwert. Weijnen, Antonius A. 1982 Deutsche Dialektologie und europäische Dialektforschung: wechselseitige Wirkungen. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke, Herbert Ernst Wiegand (ed.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 1, 190⫺202. (⫽ Handbücher zur Sprach- und Kommunikationswissenschaft 1.1.) Berlin: de Gruyter. Wenker, Georg 1877 [1915] Das rheinische Platt. Den Lehrern des Rheinlandes gewidmet von Georg Wenker. Düsseldorf. [Reprinted in: Ferdinand Wrede (ed.), Das rheinische Platt (Deutsche Dialektgeographie 8.) Marburg: Elwert.] Wenker, Georg 1878 Sprachatlas der Rheinprovinz nördlich der Mosel sowie des Kreises Siegen. Handwritten document. Wenker, Georg 1881 Sprach-Atlas von Nord- und Mitteldeutschland. Auf Grund von systematisch mit Hülfe der Volksschullehrer gesammeltem Material aus ca. 30 000 Orten. 1st delivery. Straßburg: Trübner. Wenker, Georg 1886 Über das Sprachatlasunternehmen. In: Verhandlungen der 38. Versammlung deutscher Philologen und Schulmänner in Gießen vom 30. September bis 3. Oktober 1885, 187⫺194. Leipzig: Teubner. Wenker, Georg 1895a Über den Sprachatlas des Deutschen Reichs. In: Verhandlungen der 43. Versammlung deutscher Philologen und Schulmänner, 34⫺43. Leipzig: Teubner. Wenker, Georg 1895b Herrn Bremers Kritik des Sprachatlas. In: Georg Wenker, Ferdinand Wrede. Der Sprachatlas des Deutschen Reichs. Dichtung und Wahrheit, 3⫺30. Marburg: Elwert. Wiegand, Herbert Ernst and Gisela Harras 1971 Zur wissenschaftshistorischen Einordnung und linguistischen Beurteilung des Deutschen Wortatlas. With the assistance of Gisela Harras. Germanistische Linguistik 1⫺2: 1⫺204. Wiesinger, Peter and Elisabeth Raffin 1982 Bibliographie zur Grammatik der deutschen Dialekte. Laut-, Formen-, Wortbildungs- und Satzlehre 1800 bis 1980. Frankfurt a. M.: Lang. te Winkel, Jan 1898 De Noordnederlandsche tongvallen. Atlas van taalkaarten met tekst [The North Netherlands Dialects. Linguistic Atlas with Text]. Leiden: Brill. Winteler, Jost 1876 Die Kerenzer Mundart des Kantons Glarus in ihren Grundzügen dargestellt. Leipzig/Heidelberg: Winter. Wrede, Ferdinand 1892 Berichte über G. Wenkers Sprachatlas des deutschen Reichs I⫺III. Anzeiger für Deutsches Alterthum und Deutsche Litteratur 18, 300⫺300, 405⫺413.

7. The kulturmorphologische Ansatz in dialectology Wrede, Ferdinand 1902 Berichte über G. Wenkers Sprachatlas des deutschen Reichs XIX. Anzeiger für Deutsches Alterthum und Deutsche Litteratur 28, 160⫺174. Wrede, Ferdinand 1908 An Georg Wenker. In: Jacob Ramisch. Studien zur niederrheinischen Dialektgeographie. Ferdinand Wrede. Die Diminutiva im Deutschen. VII⫺XIII. (Deutsche Dialektgeographie 1.) Marburg: Elwert. Wrede, Ferdinand 1963 Kleine Schriften. Edited by Luise Berthold, Bernhard Martin, Walther Mitzka. (Deutsche Dialektgeographie 60.) Marburg: Elwert. Zetterholm, Delmar Olof 1940 Dialektgeografiska undersökningar [Dialect-geographical Investigations], vol. 1, 2. (Skrifter: Ser. A / Landsma˚ls- och Folkminnesarkivet i Uppsala 1.) Uppsala: Almqvist & Wiksell. Zetterholm, Delmar Olof 1953 Dialektgeografiska undersökningar [Dialect-geographical Investigations], vol. 3, 4. (Skrifter: Ser. A / Landsma˚ls- och Folkminnesarkivet i Uppsala 10.) Uppsala: Almqvist & Wiksell.

Renate Schrambke, Freiburg (Germany)

7. Language and Space: The kulturmorphologische Ansatz in dialectology and the German language space ideology, 19201960 1. 2. 3. 4. 5. 6. 7. 8.

Aim and scope German dialectology after World War I Dialect, standard language and their respective cultural impact The discourse environment of cultural morphology Metaphors of space (and how they change) Stamm, Raum, Volk ⫺ competing key terms of the dialectological debates in the 1930s Conclusion References

1. Aim and scope Among the schools of German dialectology, the kulturmorphologische Schule (lit. ‘cultural morphology approach’) is special in many ways. Established in the 1920s in the context of the Institute for Rheinische Landesgeschichte (‘Regional history of the Rhineland’) in Bonn, it soon became legendary. For several decades, cultural morphology was considered to be a most promising perspective for dialect geography. Mitzka even refers to it as “das Hochziel der Dialektgeographie” [‘the highest aim of dialect geography’]

107

108

II. Linguistic approaches to space (Mitzka 1943: 17). The founding document outlining the principles of this historicalcultural approach to dialectology (Aubin, Frings and Müller 1926) puts the vast cartographic material collected by Wenker and Wrede to a completely new use. What was originally designed to represent synchronic traits of German dialects at the time of data collection (from 1876 onwards), now was reinterpreted as the frozen result of a long cultural-historical process during which linguistic traits had spread along routes of trade and communication (Verkehr) and medieval territorial boundaries, and were structured and bundled in and around cities. Surely, Wrede had already begun to transform the aims of his dialect atlas towards a dynamic view of Soziallinguistik ‘social linguistics’, but what really seemed revolutionary to his contemporaries was the fact that, within the framework of cultural morphology, the spreading of dialectal traits was shown to be a prototypical, model case for the spreading of culture in space. Reevaluating the cartographic material of the Sprachatlas within the framework of historical-cultural theory solved, as will be shown below, a major problem faced by the field of dialectology at the time: once a well-respected field of linguistics, dialectology had faded into the background in terms of relevance and popularity. Upon being seen as a model case of cultural historiography, the reputation of practical relevance was once again restored to the field of dialect research. At the time when the kulturmorphologische Schule was established, the dominant paradigm was that of the neogrammarians (cf. Murray in this volume). Dialectology was primarily associated with sound laws, questions of speech physiology and at best folklore or the boundaries of old tribal territories. Nothing dialectologists used to concern themselves with attracted much public interest, nor did dialectology have any impact or influence on historiography and social science. As pointed out by Auer (2004), dialectology in Germany was, from the very beginning, part of a nation state ideology. During the nineteenth century, dialects were connected to the standard variety by ethno-romantic constructions from which the ancient Stämme were derived (the tribes such as the Swabians or the Bavarians who had given their names to the German dialects). The Stämme were in a sense “imagined communities” (Anderson 1983) living and interacting in their dialects, which in turn had laid the grounds for the German standard language. Although this romantic view had been seriously challenged by the findings of the dialect atlas from the very beginning, modified versions of it survived much longer, as will be shown. Cultural morphology provided a promising perspective for the reentry of dialectology into a modernized interdisciplinary national project, characterized by key terms such as Volk, Kultur and Verkehr. Thus, from today’s point of view, cultural morphology can be viewed in two ways. On the one hand, it is an integral part of the aggressively expansionist ethno-science (Volksforschung) that evolved in the 1920s and supplied Nazism with expertise for their plans to push the boundaries of the Reich far into the west (Westforschung, cf. Dietz, Gabel and Tiedau 2003) and into the east (Ostforschung, cf., e.g., Oberkrome 1993). On the other hand, cultural morphology is part of the history of a sociolinguistic perspective on language change, language contact and language expansion in space. Lately, modern sociolinguists have had a tendency to list cultural morphology among the ‘extralinguistic’ approaches to variation and do not seem willing to trouble themselves further with this aspect of their tradition (cf. Barbour and Stevenson 1998: 73⫺74). For instance, in Herrgen’s (2001: 1524) historical account of German dialectology, there is only one single sentence which reminds the reader of the fact that cultural morphology once existed: the

7. The kulturmorphologische Ansatz in dialectology sentence refers to the fact that the atlas maps were interpreted historically and correlated with extralinguistic data. Prior to that time, however, cultural morphology had been considered to be a most prominent part of the dialectological enterprise (cf. GroberGlück 1982). Historiographers praised cultural morphology for being the leading school of German dialect geography from the moment it entered the field: In der zweiten Hälfte der zwanziger und zu Beginn der dreißiger Jahre entwickelte sich die Dialektgeographie zur führenden Forschungsrichtung innerhalb der Dialektologie. Ausschlaggebend für den nachhaltigen Einfluß, den diese Disziplin auf die Dialektologie und die Sprachwissenschaft insgesamt ausübte, war die Kulturraumforschung. ‘In the second half of the 1920s and early 1930s dialect geography developed into the leading approach within dialectology. This lasting impact of the discipline on dialectology and linguistics in general was due to research on cultural spaces.’ (Wilking 2003: 66)

It is clear that the kulturmorphologische Ansatz cannot be equated with historical-cultural approaches to dialectology as a whole, but it still is emblematic of the antagonist forces that influenced German dialectology in the first half of the twentieth century. With cultural morphology, dialectology became a part of the broader enterprise termed Kulturraumforschung by historiographers. Emerging in the 1920s, this was (as Ditt 2003: 929 notes) a modern, innovative and interdisciplinary branch of historiography, oriented toward cartographic representations of social and cultural everyday phenomena, and serving the political need to reinvestigate the basic traits of the German Volk and its Raum (which was considered to be the central aim of the Geisteswissenschaften).

2. German dialectology ater World War I It is symptomatic for the situation of German linguistics (and for the humanities in general) after World War I that a sense of a deep crisis prevailed (Knobloch 2005: 193⫺ 208). The days of the unquestioned hegemony of German linguistics were thought to have come to an end. French sociology of language (Antoine Meillet, Jules Gillie´ron, Charles Bally) rapidly gained popularity and attractiveness throughout Europe. The leading figures in German Geisteswissenschaft were convinced that only Geist would be able to get the country back to the top where, in their minds, it definitely belonged. Geist, Volk, Kultur and Raum were the key terms designed to counter western positivism and sociology and to regain ground for Germany and its tradition in the sciences and humanities. Neogrammarian method and theory were dismissed as “positivistic”. Linguistic nationalism rapidly gained ground in the 1920s, within the linguistic profession as well as other areas. Leo Weisgerber, who like Theodor Frings and Adolf Bach was closely associated to the Institut für geschichtliche Landeskunde der Rheinlande in Bonn (founded in 1920; now known as IGL), provided a linguistic manifesto in his book Muttersprache und Geistesbildung ‘Mother Tongue and Education of the Mind’ (Weisgerber 1929), which would become exceptionally influential in the decades to follow. The political manifesto of linguistic nationalism was first articulated by Georg Schmidt-Rohr in Muttersprache (1933). Both books converge in the belief that it is the native language that actively shapes the forces and qualities of the Volk. In the context of ethnic radicalization, these elements of Herder’s and Humboldt’s thinking competed with race for the

109

110

II. Linguistic approaches to space status of a new key term. Hutton (1999) has referred to this segment of the evolving Nazi ideology as “mother-tongue fascism”. Questioning and de-legitimizing the political borders fixed by the Treaty of Versailles became an obsession of public discourse, and the humanities tried to prove their newly acquired relevance by contributing to it in the best way they could. The common denominator of all branches of culturalist Volksforschung was correcting political history and political boundaries that assigned regions and territories to states whose language and culture had not shaped the space in question. Raum (and in particular cultural space) was semantically opposed to political government, and it was semantically coordinated with Volk. At that time, dialectology differed significantly from other fields of linguistics. Research on dialects was well-organized and centralized (in Marburg), as well as rooted in the regions and supported by a large body of collaborating school teachers and lay persons. Dialect had always been seen as the language of the people (Volkssprache), so aligning themselves with a new ethnic science (Volkswissenschaft) was relatively easy for dialectologists. Therefore, restructured humanities saw dialectology among the winners (cf. Ehlers to appear), while the remaining fields of linguistics had a hard time adjusting to the new standards of ethnic thinking. Dialectology had been Gemeinschaftsforschung long before the word turned programmatic after 1933. In the IGL in Bonn, the models and methods of cultural morphology evolved through close cooperation between dialectologists, historicists and other social scientists in the context of Westforschung. For some time it had been questioned whether the current border between Germany and France should mirror the old tribal boundaries between the settlements of Germanen (by which the speakers of Germanic languages were meant) and (romanized) Franken. Research within the IGL, inspired by Wrede’s Sprachatlas methods, focused on patterns of everyday culture that spread together and shaped cultural space: ways of settlement and house building, naming traditions (toponomastics), dialect features and customs. The patterning of these features was considered to shape a cultural space in which the lines of different cultural elements could be shown to converge. The result of IGL research was that a large portion of northern France, Belgium, and the Netherlands had originally been germanischer Kulturboden ‘Germanic cultural space’, while the contemporary linguistic (and political) borders reflected Ausgleichslinien ‘lines of compromise’ fixed by the patterns of cultural contact and conflict between Germanen and Romanen (cf. Petri 1937; Steinbach and Petri 1939). It has been argued by Ditt (2003: 931), that the IGL-findings could serve opposing political options: either to underline common (European) ground in the Western border regions or to support (German) claims to legitimate possession of these territories. It was the second option that was favored by the political contexts until 1945, and it was the first option that ensured the survival of cultural morphology after 1945 (Rusinek 2003). In retrospect it was, of course, easy to emphasize that it was a “European” program all along.

3. Dialect, standard language and their respective cultural impact In volk-oriented German linguistics until 1945 there was a deep divide over the respective cultural impact of dialect and standard language. Was it the (mainly rural) dialects of the settlers that rooted German culture in space, or was it the cultural force of the

7. The kulturmorphologische Ansatz in dialectology educated standard speakers that kept the ethnic Germans abroad in close ties with the Germans in the Reich? Walther Mitzka, who from 1934 was the head of the most powerful institutions of German dialectology (namely the Sprachatlas, the Kartell der Mundartwörterbücher ‘The Association of Dialect Dictionaries’, and the section Volkssprache of the Reichsgemeinschaft für deutsche Volksforschung ‘The Association for German Folk Research’), advocated the superiority of dialect. Mitzka himself was in fact critical of cultural morphology and favored a modernized and dynamic version of the Stämmemodel (see Wilking 2003 for more detail). Contrary to Verkehr ‘trade and communication’, another key term, which Theodor Frings had borrowed from Hermann Paul and the Neogrammarians, Mitzka (1941, 1943) maintained/contended that it was prestige and socio-cultural evaluation that conditioned the expansion of linguistic features in space. The lasting results of the colonization in eastern regions were interpreted as having proven that rural dialects (Bauernsprache) indeed attached themselves to occupied ground and that urban centers of trade and commerce were less important (at least in the eastern regions) than the IGL-linguists had thought. The question of the cultural impact of dialect and standard language was far from academic since it had very practical consequences in key matters of the politics towards the Grenz- und Auslandsdeutschtum, i. e., the Germans living outside the German territory (see Knobloch 2005: 111⫺136 for the role of Auslandsdeutschtum in German linguistics). While the proponents of dialectal superiority advocated strengthening the ties between peripheral Germans and the Reich by cultivating their dialects, their opponents tried to overcome the cultural limitations inherent in the dialect and to establish the standard variety as a reference in school, administration and media. The conflict is clearly demonstrated in quotes such as the following: Die Mundart verbindet uns stärker als die Schriftsprache mit unsern völkischen Ursprüngen. Volksfremde Denkweise, wie sie die Politik überwunden hat, wird von ihr sowieso übersprungen. Daher ihr volksbiologischer Wert. ‘Dialect links us more strongly to our origins as a Volk than written language can. Thinking that is alien to our Volk, which has now been left behind in politics, is fully disregarded in dialect anyway; thus its innate folk-biological value.’ (Mitzka 1943: 5)

In Mitzka’s ethno-romantic view, dialect is believed to have the strength and vitality to resist the corrupting forces of modernity. This position was, however, far from uncontroversial. It was challenged by many of the leading figures in linguistics and language policy. Georg Schmidt-Rohr writes on the linguistic situation in the Alsace: Die alemannische, die fränkische Mundart ist schon als Mundart grundsätzlich schwächer, weniger lockend und werbekräftig als die französische Hochsprache. In Familien, in denen die deutsche Mundart nur Haussprache ist, während französische Schulen besucht, französische Zeitungen gelesen werden, muß notwendigerweise die Sprache des größeren Kreises (die Schriftsprache) allmählich die Sprache des engeren Kreises (die Mundart) verdrängen. ‘The Alemannic and the Franconian dialect are as such weaker, less attractive and less alluring than the French standard language. In families where the German dialect is only spoken at home, while French schools are attended and French newspapers are read, the language of greater linguistic reach and influence (the written language) will gradually and inevitably encroach upon and marginalize the language of lesser linguistic reach (the dialect).’ (Schmidt-Rohr 1933: 346)

111

112

II. Linguistic approaches to space For Schmidt-Rohr, standard and written varieties symbolize ethnic unity, while dialects stand for diversity, isolation, lack of communicative reach and lack of social recognition in modern societies. It is significant that the rhetoric on space served both the modernist and the conservative factions. Raum can be conceived as ground (Boden), as the soil for growing in the fields (Scholle), but also as dynamic space mediating trade, communication and exchange. The connotations of Raum spread in opposite directions, and this fact evidently contributed to the success of spatial metaphors after World War I (Köster 2002: 101⫺118; see section 5). In this controversial field, the strategy of cultural morphology involved the use of Sprachatlas maps as a resource for a new approach to the history of the German written standard variety (Frings 1956: 1⫺21), by opting for a dynamic interpretation of maps, and for a diachronic reading of the synchronic distribution of linguistic traits. Like Hermann Aubin (who left the IGL to teach in Breslau), Theodor Frings took the IGLviews and the Bonn IGL-network with him to the eastern part of Germany. From 1927 onwards, he taught at Leipzig University and sought to apply the methods developed in his Rheinische Landesgeschichte to the linguistic data of eastern Germany. The construction of the emergence of the written German standard language in the fourteenth century that Frings offers has all elements of a unification myth, and it goes as follows: In the west, the modern borderline between the French and German speaking areas is an Ausgleichslinie ‘line of compromise’ between romanized Franks and those Germanic tribes who had maintained their ethnic and linguistic traditions. A standard language could not develop here. On the contrary, the river Rhine attracted waves of trade and traffic between the north and the south which led to maximal diversity and relic areas on both sides of that main axis of communication. The romanized western Franks took the lead in developing a central state and a unified written standard: Die Sprachgrenze ist nicht, wie man geglaubt hat, ein Ergebnis der Völkerwanderungszeit, sondern eine späte Ausgleichslinie zwischen romanischen und deutschen Franken. Jenseits der Linie, im Westreich, entstehen Einheitsstaat, Schrift- und Hochsprache, diesseits, im Ostreich ⫺ und das ist der tragische Augenblick im germanisch-deutschen Sprachgeschehen ⫺ erlahmte die Kraft. Die Franken schenkten Frankreich die Grundlage der sprachlichen Einigung, nicht Deutschland. ‘The language border is not the result of the Great Migration as once believed, but a recent line of compromise between romanized and German Franks. Beyond that line, in the western empire, a unified state, written language and a standard language emerge, while in the eastern empire ⫺ and this is the tragic moment in the Germanic-German language history ⫺ the forces become weak. The Franks gave the foundations of linguistic unification to France, not to Germany.’ (Frings 1956: 12)

If it is not the west that has brought linguistic unity to the Germans, where should we look for it then? It cannot come from the east either, because civilization spreads from west to east and never the other way around. The Marburg-maps according to Frings show that the main features of the standard language converge along the river Main between Mainz, Würzburg and Bamberg: Das ist gewiß eine Herzlage und eine Sammelstelle, auf die wir schon einmal stießen, aber ohne staatliche Bedeutung, ohne Stoßkraft, in Eigenleben befangen wie die anderen Räume. […] Dem deutschen Altland war die Kraft der sprachlichen Einigung entschwunden.

7. The kulturmorphologische Ansatz in dialectology ‘This is indeed a core territory and point of convergence, which we have encountered once before, but without state relevance, without outward energy, preoccupied by a life of its own as is true of the other spaces […]. The old German land has lost the power of linguistic unification.’ (Frings 1956: 15)

Yet this region of the Altland is known to be a starting point of the German settlement in the east, and so the central features of the (now) written standard move east with the Mainfranken, mix with other varieties and finally settle by forming an Ausgleichssprache ‘koine’ that is backed by the state of Wettin and by Leipzig as an urban centre. It is through the dynamics of the Ostkolonisation that linguistic unity is brought about. Neither educated humanists nor the chancelleries are credited with the “invention” of the standard variety. This proves to be incompatible with the myth of the creativity of the Volk (yet compatible with the myth of the educated middle-class citizen or the administration of the state). It is the settlers moving east that build linguistic unity, not great individuals or a central state (like in France): Das neue Deutsch war im Munde der Ostsiedler vorgeformt und wurde gesprochen lange bevor es seit dem 13. Jahrhundert in die Schreibstube einzog. Es ist ein Gewächs des neudeutschen Volksbodens, eine Schöpfung des Volkes, nicht des Papiers und des Humanismus. ‘The new German had existed in the minds and mouths of the settlers in the east and was spoken long before it entered the clerks’ offices. It grew on new German soil, and was a creation of the people, not of paper and humanism.’ (Frings 1956: 16)

Interestingly enough, this mythic conception of linguistic unification was not only supported by the humanities of the pre-National Socialistic and National Socialistic eras, but also by communist post-war Germany (where Frings continued to hold his chair at Leipzig in the German Democratic Republic). All in all, cultural morphology held a middle course in the conflict on the cultural impact of dialect or standard language, as did Mitzka, who in his 1941 contribution to the infamous volume Kriegseinsatz der Geisteswissenschaften ‘War Deployment of the Humanities’ attributed the formation of German to the impact of Bauernsprache ‘farmers’ language’ and Bürgersprache ‘citizens’ language’ (cf. Mitzka 1941: 87). When dialectologists at the time (like Martin 1939) described the role of the dialects in the development of German, they use metaphors of force and vitality: Urkraft, der breite alte Strom der Mundarten, Quelle deutschen Volkstums, Grundlage unseres Sprachlebens [‘original power’, ‘the broad, ancient pulsing current of the dialects’, ‘source of the German folk’, ‘foundation of our language life’] (Martin 1939: 4⫺5). Metaphors for the written standard tended to be taken from culture and order instead: Die Schriftsprache ist strengen Regeln unterworfen im Wortschatz, Formenbildung, Satzbau und Ausdruck (sic); sie führt ein Leben in Zurückhaltung, in steter Anpassung an das geistige Leben der Nation. In ihr erklingen die Werke der Dichter, Gelehrten; sie muß in Zeitung und Rundfunk angewandt werden, in ihr wünschen wir die großen Reden unserer Volksführer zu hören. […] Sie ist das Sinnbild der deutschen Einheit. ‘The written language is subject to strict rules of vocabulary, morphology, syntax and expression; it leads a life of restraint, continually adapting to the mentality of the nation. It echoes the works of poets and scientists; it must be printed in the newspaper and heard on the radio, as it is the written language that we wish to hear used in the great speeches of our people’s leaders. […] It is the symbol of German unity.’ (Martin 1939: 5)

113

114

II. Linguistic approaches to space While dialectology attempted to cash in on the spiritual warmth of Heimat ‘homeland’ connotations, standard varieties seemed to lack this feeling. Simple minds like Bernhard Martin even emphasize that the written variety preserves the best of the various dialectal achievements, but still has no Heimat: Während die Schriftsprache das ihr Gemäße auswählt und festhält aus dem sprachlichen Gute aller Stämme und Landschaften, dadurch nirgends und überall daheim ist, steht die Mundart mit festen Füßen auf der heimischen Erde, ist ein Ausdruck dieser Heimat selbst. '‘While the written language selects that which is appropriate, retaining the linguistic good of all tribes and regions, it is thus nowhere and everywhere at home, whereas the dialect is firmly grounded on the soil of the homeland; it is an expression of this homeland in itself.’ (Martin 1939: 9)

4. The discourse environment o cultural morphology In the dialectological writings of the 1930s, 1940s and 1950s, cultural morphology is held in high esteem. Frings and his followers (such as Friedrich Maurer, Adolf Bach, Ernst Schwarz) are said to have overcome the isolated existence of dialect research and to have promoted the status of this branch of the humanities to a ground-breaking and integral partner of German cultural history (see Schwarz 1950: 193; Martin 1939: 92). It is repeated almost ritually that cultural morphology has shaped our perspective on the large scale dynamics of cultural traits in space. The term cultural morphology was selected, as it could easily be associated with Spengler’s popular philosophy of emerging and declining cultures, well-known among the educated people of the 1920s (Der Untergang des Abendlandes, printed first in 1918⫺ 1922). In Spengler’s Kulturkreislehre, the vitality of cultures in history is assessed in terms of competition with rival cultures. By calling dialect geography kulturmorphologisch, it was implied that dialectological maps could serve as evidence for the description of these cultural dynamics in space, elevating Spengler’s philosophy of culture to the level of sound empirical research. Dialectological cultural morphology was by far not the only attempt to benefit from Spengler’s immense popularity. In 1936, the philosopher Eduard Spranger published a short volume entitled Probleme der Kulturmorphologie, claiming to help “die Aufgaben verstehen, die die kommenden Kulturbegegnungen uns auferlegen werden” [‘to understand the tasks which coming cultural encounters will impose on us’] (Spranger 1936: 40). What happens when dynamic and expanding cultures (either by settling outside their territories, or in colonization, or in the remote contacts of trade and communication) interact with other cultures in space? The settings in question mirror exactly the factors assumed to be relevant for dialect dynamics: settlements abroad, colonization, trade and commerce. Spranger suggests that the results of cultural contact differ widely according to the circumstances of the contact: settlement and colonization confront cultures in the same space, so traits have to either mix and mingle, or else the superior one destroys and replaces the others. Trade, communication, and political exchange do not take place on the same territory, and they are said to result in a leveling (Ausgleich, Spranger 1936: 7⫺8). The examples discussed by Spranger include the spread of Christianity, the exchange between Romanen and Germanen in the West, and the German colonies lost

7. The kulturmorphologische Ansatz in dialectology during World War I. In reading Spranger’s text, we also understand that the new humanities (among them dialectology) lay claim to political relevance: Die tiefgehende nationale Selbstbesinnung, zu der das deutsche Volk seit kurzem wieder erwacht ist, legt die Frage nach dem Eigenen und Ursprünglichen der deutschen kulturformenden Kräfte nahe. ‘The deep-rooted national self-reflection to which the German people re-awoke a short while ago brings the question of what is proper and original in the forces which shape German culture into the foreground.’ (Spranger 1936: 9)

Returning to dialectology proper, we find that towards the end of the National Socialist era, there were two competing models claiming to account for the spread of dialectal features in space, and both, in a broad sense, had accepted the innovations brought about by Theodor Frings and his followers in cultural morphology. The model advocated by Frings himself was based on the idea of sequences of “waves”, piling up and overlapping with other waves, pushing features continually forward, surrounding and isolating obstacles. The populations in this space could either show reluctance or eagerness in adopting new features. This model gives prominence and strength to urban centers of trade, communication and political power. It is these factors that send the waves of change towards the periphery, and the more distant they become, the less impact they have. The competing model, advocated by Walther Mitzka, included primarily rural, loosely-knit networks of isolated settlements (Streusiedlungen), that combined into larger networks and finally dominated the territory in question completely, thereby isolating the old features and serving as models to adopt the new ones. There was no privilege for urban centers in this model. The practitioners of the Sprachatlas (such as Martin 1939: 106⫺107) had no qualms accepting either model. They were convinced that the two models were compatible with one another and that they both had areas to which they could be profitably applied. It is remarkable though that adherence to one or the other of the positions correlates highly with the commitment to the cultural superiority of either dialect or written standard. The followers of Frings’ model (many of whom were outside of dialectology proper, like Leo Weisgerber or Hennig Brinkmann) argued that the unity and strength of the Volk depend primarily on the cultural reach of written standard, while dialects are seen as a source of fragmentation. The followers of Mitzka, on the other hand, underlined the stabilizing forces of dialect. However, the basic tenets of cultural morphology, reshaping the dynamic interrelations of folklore (Volkstum) and the land (Landschaft), can be considered common ground in the early 1930s. Adolf Bach, a close follower of Theodor Frings, also makes reference to cultural spaces, “die auf deutschem Volksboden von den Mächten einer schicksalhaften Geschichte ausgeformt worden sind” [‘which where formed on the German space by the powers of a fateful history’] (Bach 1934: 168). At the same time he recommends that cultural spaces and the dynamics investigated by cultural morphology be made the fundament of an “organic” restructuring of the provinces of the Reich. He continues: Dies Ziel wird dort von besonderer Bedeutung sein, wo wir hinausgehen über die gegenwärtigen Grenzen des Deutschen Reiches. Den jenseits von ihnen wohnenden Volksgenossen (und denen, die Gewalt über sie haben) kann hier aufs nachdrücklichste zum Bewußtsein gebracht

115

116

II. Linguistic approaches to space werden, daß politische Grenzen dort sinnlos gezogen sind, wo sie keine Rücksicht nehmen auf das Volkstum einer Landschaft. ‘This aim will become particularly relevant once moving beyond the present borders of the German Reich. It will most emphatically enter the consciousness of the compatriots living beyond the borders (and the consciousness of those who have power over them) that political borders there have been thoughtlessly drawn and fail to take the Volkstum of a region into consideration.’ (Bach 1934: 168⫺169)

Cultural morphology established itself as the interdisciplinary study of the dynamic relations between Volkstum and Landschaft (‘region’ is only an approximate translation) with the help of dialectology. It was widely accepted even by competitors in the same field.

5. Metaphors o space (and how they change) It is a linguistic truism that political as well as scientific key terms can only be fixed in relation to their antonyms. Programmatic expressions should not be understood in terms of what they stand for, but rather in terms of what they stand against. The changing semantics of Raum ‘space’ in the German history of ideas have been described in detail by Köster (2002). The public career of Raum compounds in Germany before 1945 is evidently related to imperialism, where the semantics of Raum were meant to cover national territories together with spheres of (colonial) interest outside the state borders. From the beginning, Raum was opposed to official state borders. This line of opposition was of course enforced by the territorial losses of Germany after the war in the Treaty of Versailles. As of this point in time, there were a number of regions which could be said to be part of the German space, but did not belong to the German state territory. There is, however, a second line of reasoning, a second semantic opposition which is less evident. We tend to think of Raum as static, but it can be shown that in the 1920s popular compounds underline its dynamic meaning. Raum is seen as a scene of potential action, a place where forces meet and compete, as a dynamic scene for competing actors. Köster (2002: 12) cites contemporary sources according to which Raum was interpreted as a metaphor for the chance to use ones own forces. It is essentially by this dynamic connotation that the Raum-semantics could be used simultaneously to entail territorializing and de-territorializing ambitions: the forces that shaped any given cultural space would at the same time be the forces transcending any spatial limits. From this point of view, languages that resisted territorialization (like Yiddish) were particularly fascinating (Auer 2004: 150). In the late 1920s, pioneers of sociolinguistics in Germany (such as Heinz Kloss, Franz Thierfelder; cf. Hutton 1999: 200⫺205) described Yiddish as a language that could help to expand the influence of German in Eastern Europe. The contrast between dialect and standard could easily be paralleled with the contrast between static, spatially bounded varieties on one side and expanding (space-transcending) varieties on the other. This is why Yiddish could not be considered a German dialect: it had no territorial boundaries. It is evident that early sociolinguists (Georg Schmidt-Rohr as well as those mentioned above) were convinced that the cultural superiority of written standard languages depended exactly on their forces as “raumüberwindende Mächte” ‘space-transcending powers’ (cf. Schmidt-Rohr 1934), while dialects were considered to

7. The kulturmorphologische Ansatz in dialectology be static. German language politics abroad (developed by institutions such as the Deutsche Akademie in Munich) trusted in the ability of the German standard language (with its tradition in culture, science, philosophy) to gain new ground abroad and eventually overcome spatial boundaries. This interpretation of Raum as a dynamic scene where forces meet and compete was especially compatible with the psychological twist that Georg Simmel, Kantianist philosopher and sociologist, gave to spatial semantics (Simmel 1992: 687⫺687, Köster 2002: 81⫺88). According to Simmel, space is a mental construct; it is the way we perceive the scenes for all forms of social interaction. Borders are not real in Euclidean space, but are the results of social constructions, which evolve from interaction between individuals; we perceive them as features of the territory: Nicht der Raum, sondern die von der Seele her erfolgende Gliederung und Zusammenfassung seiner Teile hat gesellschaftliche Bedeutung. ‘Not space, but the mental structuring and unification of its parts have social meaning.’ (Simmel 1992: 688)

It can easily be shown that this ambiguous mix of naturalist and territorializing conceptions of space on one hand, and psychological and interactive conceptions on the other, was also relevant in the above-mentioned controversies regarding cultural morphology. Walther Mitzka’s critical attitude towards cultural morphology is related to his psychological concept of space which attributes cultural superiority to the individuals’ positive attitudes towards their dialect, while the model of cultural morphology is much more dependent on the results of physical barriers, medieval political territories and routes of trade and communication. But, of course, this type of conflict is quite common for hegemonic key terms. They can be considered key terms because every person in the discursive field has to clarify his own position and perspective on its implications, not because there is consensus on these implications. All word compounds ending in Raum though ⫺ such as Lebensraum (life-space), Volksraum (ethnic space), Sprachraum (language space), Wirtschaftsraum (economic space), Kulturraum (cultural space) ⫺ which flooded discourse during the 1930s, related the social activities of the Volk to the realm of political power and geography. Operating on common (ethno-Darwinist) ground, the proponents of spatial concepts disagreed widely on the importance of language, economy, culture, race for space. Linguists, of course, tended to believe that it was language that created the Volk. In an attempt to sort out the various concepts, Schmidt-Rohr (1936) writes: Der Volksraum, der als geopolitische Größe oft übersehen wurde, ist der geographische Raum, insofern er von einer wesenseinheitlichen Bevölkerung bewohnt wird. Diese Gebiete sind wesentlich Einheiten infolge eines einheitlichen geistigen Verkehrs. Die Grenzen sind wesentlich da gezogen, wo Verschiedenheit der Mittel des geistigen Verkehrs die Möglichkeit dieses Verkehrs aufhebt, wo verschiedene Sprachen aneinander stoßen. ‘The Volksraum, which is often overlooked as a geopolitical factor, is the geographical space inasmuch as it is inhabited by a population similar in kind. These territories are relevant units because of a unified, shared mentality. The borders are by and largely drawn where the differences in the means of communication render discourse impossible, in other words, where different languages meet.’ (Schmidt-Rohr 1936: 242)

117

118

II. Linguistic approaches to space And while Schmidt-Rohr, a critic of dialectal fragmentation in Germany, predicted that in contact situations the winning language will be the one that is structurally superior in culture, literature, press, dialectologists like Mitzka claimed the superiority of the dialect by stressing its closer psychological ties to the speakers and their local identities. Spaces of language contact are determined by the dynamic value differences (“dynamische Wertverschiedenheit”, Schmidt-Rohr 1936: 243) and by the cultural reach of the languages involved. Bilingual settings do not last, they are considered to be unnatural. One of the languages will have to retreat eventually. The power of (written standard) languages to overcome distances in space is amplified by technical means, as SchmidtRohr states (1934: 203). There are no limits governing the extent to which languages create and enlarge intellectual spaces for their speakers. Proto-sociologists of language such as Schmidt-Rohr (1936: 243) or Heinz Kloss (e.g., 1929) tend to identify the cultural scope and value of a language with its political force. By this equation, strengthening the cultural and intellectual community of one’s mother tongue becomes a political program. Emigrants from Germany are thought to be part of the Volk as long as they still command their mother tongue ⫺ hence the political wish to reinforce their loyalty by strengthening their linguistic ties to the Reich. In border regions and linguistic enclaves (Sprachinseln; see Kuhn 1934), people have to be protected against the temptations and seductions of attractive contact languages they are exposed to. And abroad, in the territory of the enemy, minority languages (like the Celtic languages in France, Ireland, Wales) have to be encouraged in their struggle against the linguistic centralism of the state. The basic principle of linguistic nationalism was that political borders should mirror linguistic borders, which tends to render political power illegitimate for linguistic minorities in any state. These in turn find their protector (Schutzmacht) in states, where their mother tongue is the official language.

6. Stamm, Raum, Volk  competing key terms o the dialectological debates in the 1930s From this point of view, it was a severe disadvantage of the traditional key term of dialectology, the Stamm ‘tribe’, that it had no dynamic connotations; instead, it presupposed stable territorial grounding and particularism with respect to the standard language. In the nineteenth century, dialect borders had been explained by the settlement borders of the old tribal groups going back to the age of the Great Migration. Their dialects were neither written nor vehicles of social advancement, nor fit to be propagated by modern means of communication. The tribes were thought to be leftovers of a world that had been overcome by the rise of nation-state. In the eyes of modernists (like Bach 1934), the rhetoric of the Stamm suggested that dialects (and dialectologists) were living in the past. While England and France looked back on centuries of political and linguistic centralism, German Stämme (and their linguistic correlates, the traditional dialects) served as reminders that behind the short history of the German central state there lurked the remains of primitive social bodies. The archaism behind tribal rhetoric, however, does not give the whole story. In the Weimar Republic, the German Stämme were part of the political agenda and even present in the text of the constitution, which had “das deutsche Volk einig in seinen

7. The kulturmorphologische Ansatz in dialectology Stämmen” [‘the German people united in its tribes”] (cf. Oberkrome 2004: 15). Among the most popular works of literary history was Josef Nadler’s Literaturgeschichte der deutschen Stämme und Landschaften. The proceedings of the Siebenter deutscher Soziologentag in Berlin 1930 provide evidence that the most prominent speakers of German humanities seriously debated the problem of Stämme. The reasons for this untimely boom are not easy to grasp. Firstly, the appeal of the Stämme reflects the high priority that was given to the Grenz- und Auslandsdeutschtum as a field of political relevance in the new humanities. Linguistic and cultural identities of Germans abroad were thought to be heavily impregnated by their tribal loyalties. Secondly, it was the Stämme that marked the semantic opposition towards the waves of western popular culture (Stammeskultur versus Massenkultur) that proved to be so threatening for the German conservative middle class identity. Strengthening regional and cultural ties seemed to be an effective means of protection against the temptations of western popular culture. Thirdly, being communities of descent placed the tribes closer to racial categories. Therefore, within German dialectology, there remained influential groups of researchers who relied on a modernized Stämme terminology, among them Walther Mitzka. In order to modernize tribal rhetoric in dialectology, it had to be separated from the old, static concept of space and adjusted to the new, dynamic one propagated by cultural morphology. Ironically, it was of all people Hermann Aubin, co-founder of cultural morphology, who set out to modernize this relationship between Stamm and space, the same person who denied any credit to the archaic and static model of the German tribes with their invariable nature-like properties, their common descent and stable settlement places. Aubin redefined Stämme as historical and cultural subjects, shaped by the ever changing influences of geography, politics, religion, ideology, commerce and contact with other groups. In short, the formation of tribes was considered to be an ongoing affair, not an accomplished fact of history (Aubin 1931; Oberkrome 2004: 16⫺17). Max Hildebert Boehm, a famous spokesman of the conservative branches of the humanities in 1930, subtly dissected the key terms of his time in his programmatic sketch of the humanities as Volkswissenschaften ‘ethnic sciences’: Wir brauchen aber dem Abstammungsmoment nur eine kleine Wendung zu geben, um die Vorstellung vom Stamm als natürlich bedingter Art auf eine bezeichnende Weise zu verändern. […] Wir leben als Stammesgenossen von heute in der Vorstellung, nicht bloße Anlagen, sondern eine bestimmende Art und einen besonderen Geist geerbt zu haben. Damit aber verschmilzt die naturhafte Artprägung des Stammes mit einem partikularen geistigen Überlieferungszusammenhang und einem darauf gegründeten Gemeinschaftsgefühl. Der Stamm fühlt sich älter als das Volk. Er hat sich als gefestigte Volkspersönlichkeit aufgegeben, aber seine besondere Volksart nicht nur leiblich, sondern auch geistig bewahrt. Und dies sein verdämmerndes, gestalt- und willenlos gewordenes Sondersein erscheint ihm auch auf der geistigen Ebene als ein Stück Natur im Vergleich zu der formenden, fordernden, normsetzenden Kultur des Volkes, in die der Mensch erst durch Erziehung und Lehre völlig hineinwächst. […] Aber das Stammestum erhält sich im Gefühl des Menschen gleichsam als ein umhegter Bereich des Mütterlichsten. […] Und dieser lebendige Stamm ist es, der sich mit Behelfen aus dem Bestande der naiven Popularwissenschaft seiner Zeit den Glauben an Artgleichheit, gemeinsame Abstammung, Taten der Väter, umgreifbares Gemeinschaftsgut und dergleichen mehr auferbaut. Wie trügerisch dieser Glaube sein kann, hat beispielsweise die jüngste deutsche Mundartforschung gezeigt.

119

120

II. Linguistic approaches to space ‘But we only need to give a small twist to the notion of descent in order to change the view of the Stamm as a naturally conditioned kind in a characteristic way. […] Today, we people of the same Stamm believe to have inherited not only predispositions, but also a certain way of being and a special Geist. In this way, the natural formation of the Stamm becomes amalgamated with a particular mental tradition and a feeling of togetherness which is based on it. The Stamm thinks of itself (sic) as being older than the people. It has given up itself as a solid Volk personality but has maintained his special ethnic way of being not only physically, but also spiritually. And this gloomy specialness of being that has lost shape and will, now appears to it as a piece of nature on the spiritual level, when compared to the forming, demanding and norm-setting culture of the Volk, into which man only grows completely by education and teaching. […] But in man’s emotions the Stamm is in a sense preserved like a cherished area of the most motherly. […] And it is this lively Stamm which construes the belief in similarity of the kind, in common descent, deeds of the fathers, embraceable shared values and so on with props from the repertoire of naı¨ve popular science. Recent German dialectology, for instance, has shown how this belief can be deceiving.’ (Boehm 1932: 107⫺108)

Boehm was definitely hostile towards the Stämme. From his perspective, the use of tribal discourse implies that a nation is incapable of doing away with its origins. Such a nation will dissolve and fall apart and eventually be nothing more than a collection of tribes, which in his opinion is a dim and miserable fate, exemplified by Luxemburg (Boehm 1932: 110). The allusion to late German dialectology in the end of the citation is aimed at cultural morphology. Some of the proponents of cultural morphology (in particular Adolf Bach; see Bach 1934: 46) did not even use the word Stämme without distancing quotation marks ⫺ this shows that the discourse of the Stamm in its pure and unrefined form was incompatible with cultural morphology as well as with the ideology of the standard language. It was also Walther Mitzka who re-implemented modernized Stämme in the heart of German dialectology, the Sprachatlas. Whereas dialectology in the traditions of the nineteenth century had only known the Stämme, Mitzka introduced a two-way terminology, consisting of Altstämme and Neustämme (paralleled with Altland and Neuland). Cultural-historical dynamics, introduced into the debate by Aubin, were seen mainly on the side of Neustämme which were thought to have been formed in colonization and settlement movements of German origin in Eastern Europe. It is important to keep in mind that the role of Stämme was still debated in German dialectology as late as the 1960s (cf. Moser 1961). Cultural morphology on the whole, however, pursued Ferdinand Wrede’s position on the Stämme, which amounted to the statement that they were widely irrelevant (Wrede 1902). As the only evidence available for tribal boundaries was dialectal evidence, they were suspected to be circular. For Wrede, only if later medieval territories reflected and continued tribal territories, the latter were considered to be relevant at all. Even Mitzka (1943: 40) states explicitly that the relevance of old Stamm borderlines cannot be proven, if later territorial boundaries are drawn along the same lines: the dialectal line in that case could also be of later origin. It should also be noted that vague, romantic notions such as Stamm had no place in sober neogrammarian dialectology which was already fairly sociological in the 1890s (e.g., Wegener 1891). Cultural morphology adds the idea that many traits of local dialects cannot be explained by local tradition alone. In addition, the radiations of cultural centers (such as cities, administrative and religious centers) must be taken into account (Martin 1939: 101).

7. The kulturmorphologische Ansatz in dialectology Interlocking dialectal regions with urban centers were seen as a result of trade and communication in cultural morphology. It is important to note that Mitzka’s construction, though different in argument, serves very similar purposes. For him, the compact dialectal unity of the Altstämme is lost in their expansion towards the East. German speaking settlements, partly urban and partly rural, are pushed forward into territories with different dominating languages. These outposts, with their developing ties and networks, transform primary dialects into Verkehrsmundarten ‘regional dialects’, varieties with a broader reach (Mitzka 1943: 87). Mitzka and the proponents of cultural morphology shared a certain type of linguistic frontier myth: while the majority of the other European countries had well-defined territories and long standing centralized states, the Germans had been continually battling at their western and eastern frontiers and formed their cultural and linguistic identities in these battles. Due to (as Frings 1956: 5 writes) the German expansion of territory for two millennia, it is not surprising to find that they accomplished their linguistic unity at a meaningful moment of their colonial history. For Mitzka, the most dynamic regions of the whole process are the border regions and not the cultural and economic centers. He claims that in the Prussian Northeast, linguistic space was formed by the rural dialect alone, without any urban influence (Mitzka 1943: 16). Evidently, Mitzka preferred a version of the frontier myth that credited the persistence of the German Volk mainly to the peasants with their attachment to the soil, while Frings and the rest of cultural morphology tended towards a more modernist version of the same myth. A final point is of interest. There had been a rather paralyzing tendency towards increasingly fine grained localism in German dialectology until the 1920s. For the popularity of cultural morphology it was certainly helpful that Theodor Frings and his followers promised to refocus on the “global” tendencies and dynamics of dialects in Germany. In a programmatic sketch of cultural morphology, Frings and Tille (1925⫺1926: 1) write: Aus der Kleinarbeit, die F. Wredes Schüler in ihren Dissertationen zum SA geleistet haben, ist allmählich ein neuer verheißungsvoller Gesichtspunkt entstanden: das Studium der Großbewegungen, die das alte deutsche Stammland von den Alpen bis zur Nordsee durch die Jahrhunderte mit gleichem, immer wiederkehrendem Rhythmus, in immer wieder erneutem Südnordstrom durchflutet haben. ‘Out of the detailed work completed by the students of Wrede in their dissertations for the Sprachatlas, a new, promising point of view has finally emerged: the study of the large movements that have flooded the old German land from the Alps to the North Sea through the centuries in the same, ever-repeating rhythm and ever-renewing stream from the south to the north.’

When we compare to this the theories and models of Mitzka and the Sprachatlas of the 1930s, it becomes apparent that they too were trying to reconnect dialectal localism with the global accomplishments of German cultural history, although it was their belief that it was the colonization of Eastern Europe that kept the Volk together. In the end, it is under the unifying roof of Kulturraumforschung ‘research on cultural space’ that the different schools of German dialectology meet. Again ironically, the political context after 1945 supported a renaissance of cultural morphology (with its inclination to the West) as well as a renaissance of Stämme dialectology that had become part of Ostforschung. While the former recoded their western frontier framework in terms of the common history of Europe and the occident, the latter turned from eastern coloniza-

121

122

II. Linguistic approaches to space tion to the fate of the German refugees (Vertriebene). Both schools had in fact mixed and mingled in the 1920s and 30s. Hermann Aubin (Breslau) and Theodor Frings (Leipzig) had moved to the east. Walther Mitzka (Marburg) had been an early activist for eastward oriented Grenz- und Auslandsdeutschtum before he became head of the Marburg Sprachatlas. Adolf Bach spent a few years at the German Reichsuniversität in Strassburg before he had to return to Bonn and the IGL. Ernst Schwarz, head of the Sudetendeutsche Mundartforschung in Prague, went to Erlangen after the war.

7. Conclusion Cultural morphology was shaped in strategic competition with the rising French sociology of language and dialect. Even the name cultural morphology is reminiscent of Emile Durkheim’s morphologie sociale (1897), a branch of sociology that was to investigate the elementary forms of social life, such as housing, distribution of population, communication. In German Geisteswissenschaften, the empirical fields of French sociology were usually replaced by compounds with either Kultur or Geist. At the same time, the name cultural morphology also related to Oswald Spenglers immensely popular and alarmist Der Untergang des Abendlandes where the great cultures of world history are shown to follow timeless laws of rise and decline. Cultural morphology was a promising approach for dialectologists. It lead linguists out of their bemoaned isolation from the other humanities, and instead secured methodological leadership for them: nothing was equal to the Sprachatlas and its dynamic cartography of culture. The Atlas der deutschen Volkskunde ‘Atlas of German Folklore’ (founded by the Deutsche Forschungsgemeinschaft/Notgemeinschaft in 1928 as a national flagship project of ethnic science) followed the methods that had been tested in dialectology by the Sprachatlas and that had been refined by cultural morphology. Proponents of cultural morphology initially had a strong position in this atlas, and dialectologists tried to establish the Wortatlas within it (which was later successfully established by Mitzka in Marburg). Ethnography and historiography both felt indebted to the dialectologists. In 1936, Hermann Aubin, one of the leading proponents of cultural morphology, wrote: Modern linguistics has taught us to recognize transformations and assimilations in regional cultural provinces. Words and forms spread from original localities, the varieties of which, for some reason or other, are particularly highly valued and therefore imitated as superior. These forms compete for dominance with the native ones they encounter; they often conquer wide areas and spaces. The same goes for other cultural traits. (Aubin 1936: 12; translation C. K.)

Reframed as cultural morphology, dialectology turned modern, dynamic and sociological. As a model approach within Kulturraumforschung, it managed to gain prominence and high reputation among the Geisteswissenschaften, as well as generous funding by the Deutsche Forschungsgemeinschaft, the German Research Council (Ehlers to appear). While neogrammarian dialectologists were ridiculed as unworldly Lautschieber (sound shifters), cultural morphology became the most prominent branch of applied linguistics in Germany and remained in a hegemonic position for at least 40 years. With cultural

7. The kulturmorphologische Ansatz in dialectology morphology, dialectology turned “operative”, for there was something it had to offer to (ethno)politics. While the impact of cultural morphology is currently down-played in textbooks of dialectology, one must acknowledge that many social dialectologists and sociolinguists who are considered modern today were inspired by cultural morphology. Reference to Aubin, Frings and Müller (1926) can be found in Uriel Weinreich’s Languages in Contact (Weinreich 1977), and the first generation of “applied” dialectologists and sociolinguists devoted to “ethnopolitics” (e.g., Walter Kuhn and Heinz Kloss), is still present in the text books of our day.

8. Reerences Anderson, Benedict 1983 Imagined Communities: Reflections on the Origin and Spread of Nationalism. London: Verso. Aubin, Hermann 1931 Die geschichtlichen Grundlagen der deutschen Stämme. In: Deutsche Gesellschaft für Soziologie (ed.), Verhandlungen des siebenten deutschen Soziologentages vom 28. September bis 1. Oktober 1930 in Berlin, 257⫺268. Tübingen: Mohr. Aubin, Hermann 1936 Schlesische Siedlung beiderseits der Sudeten. Schlesisches Jahrbuch 7: 9⫺28. Aubin, Hermann, Theodor Frings and Josef Müller 1926 Kulturströmungen und Kulturprovinzen in den Rheinlanden. Bonn: Röhrscheid. Auer, Peter 2004 Sprache, Grenze, Raum. Zeitschrift für Sprachwissenschaft 23(2): 149⫺180. Bach, Adolf 1934 Deutsche Mundartforschung. Ihre Wege, Ergebnisse und Aufgaben. Eine Einführung. Heidelberg: Winter. Barbour, Stephen and Patrick Stevenson 1998 Variation im Deutschen. Soziolinguistische Perspektiven. Berlin/New York: de Gruyter. Boehm, Max Hildebert 1932 Das eigenständige Volk. Volkstheoretische Grundlagen der Ethnopolitik und Geisteswissenschaften. Göttingen: Vandenhoeck & Ruprecht. Dietz, Burkhard, Helmut Gabel and Ulrich Tiedau (eds.) 2003 Griff nach dem Westen. Die “Westforschung” der völkisch-nationalen Wissenschaften zum nordwesteuropäischen Raum (1919⫺1960). 2 vols. Münster: Waxmann. Ditt, Karl 2003 Die Politisierung der Kulturraumforschung in Dritten Reich. Das Beispiel Franz Petri. In: Dietz, Gabel and Tiedau, vol. 2, 927⫺944. Durkheim, Emile 1897⫺1898 Morphologie sociale. L’Anne´e sociologique 2: 520⫺521. Ehlers, Klaas-Hinrich to appear Staatlich geförderte Sprachforschung von 1920 bis 1970. Wiesbaden: Steiner. Frings, Theodor 1956 Sprache und Geschichte II. (Mitteldeutsche Studien 17.) Halle: Niemeyer. Frings, Theodor and Edda Tille 1925⫺1926 Kulturmorphologie. Teuthonista 2: 1⫺18. Grober-Glück, Gerda 1982 Die Leistungen der kulturmorphologischen Betrachtungsweise im Rahmen dialektgeographischer Interpretationsverfahren. In: Werner Besch, Ulrich Knoop, Wolfgang

123

124

II. Linguistic approaches to space Putschke and Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 1, 92⫺113. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.1.) Berlin/New York: de Gruyter. Herrgen, Joachim 2001 Die Dialektologie des Deutschen. In: Sylvain Auroux et al. (ed.), Geschichte der Sprachwissenschaften. Ein internationales Handbuch zur Entwicklung der Sprachforschung von den Anfängen bis zur Gegenwart, vol. 2, 1513⫺1535. (Handbücher zur Sprach- und Kommunikationswissenschaft 18.2.) Berlin/New York: de Gruyter. Hutton, Christopher M. 1999 Linguistics and the Third Reich: Race, mother-tongue fascism and the science of language. London: Routledge. Kloss, Heinz 1929 Nebensprachen. Eine sprachpolitische Studie über die Beziehungen eng verwandter Sprachgemeinschaften. Wien: Braunmüller. Knobloch, Clemens 2005 “Volkhafte Sprachforschung”. Studien zum Umbau der Sprachwissenschaft in Deutschland zwischen 1918 und 1945. Tübingen: Niemeyer. Köster, Werner 2002 Die Rede über den “Raum”. Zur semantischen Karriere eines deutschen Konzeptes. Heidelberg: Synchron. Kuhn, Walter 1934 Deutsche Sprachinsel-Forschung. Geschichte, Aufgaben, Verfahren. Leipzig: Hirzel. Martin, Bernhard 1939 Die deutschen Mundarten. Leipzig: Quelle and Meyer. Martin, Bernhard 1959 Die deutschen Mundarten. 2. ed. Marburg: Elwert. Mitzka, Walther 1941 Bauern- und Bürgersprache im Ausbau des deutschen Volksbodens. In: Franz Koch et al. (eds.), Von deutscher Art in Sprache und Dichtung, vol. 1. 67⫺96. Berlin: Kohlhammer. Mitzka, Walther 1943 Deutsche Mundarten. Heidelberg: Winter. Moser, Hugo 1961 Noch einmal: Stamm und Mundart. Zeitschrift für Mundartforschung 28: 32⫺43. Oberkrome, Willi 1993 Volksgeschichte. Methodische Innovation und völkische Ideologisierung in der deutschen Geschichtswissenschaft 1918⫺1945. Göttingen: VR. Oberkrome, Willi 2004 Die Kultur der deutschen Stämme. Räumlich-tribalistische Volkstumstheorien in den deutschen Geisteswissenschaften 1880⫺1960. In: Georg Bollenbeck and Clemens Knobloch (eds.), Resonanzkonstellationen. Die illusionäre Autonomie der Kulturwissenschaften, 15⫺32. Heidelberg: Synchron. Petri, Franz 1937 Germanisches Volkserbe in Wallonien und Nordfrankreich. Die fränkische Landnahme in Frankreich und den Niederlanden und die Bildung der westlichen Sprachgrenze. Bonn: Röhrscheid. Rusinek, Bernd 2003 “Westforschungs”-Traditionen nach 1945. Ein Versuch über Kontinuität. In: Dietz, Gabel and Tiedau, vol. 2: 1141⫺1204. Schmidt-Rohr, Georg 1933 Mutter Sprache. Vom Amt der Sprache bei der Volkwerdung. Jena: Diederichs. Schmidt-Rohr, Georg 1934 Die Sprache als raumüberwindende Macht. In: Karl Haushofer (ed.), Raumüberwindende Mächte, 202⫺230. Leipzig: Teubner.

8. Structuralist and generative approaches

125

Schmidt-Rohr, Georg 1936 Mensch und Raum. Die badische Schule 3: 242⫺244. Schwarz, Ernst 1950 Die deutschen Mundarten. Göttingen: Vandenhoeck & Ruprecht. Simmel, Georg 1992 [1908] Soziologie. Untersuchungen über die Formen der Vergesellschaftung. Vol. 11 of the complete edition, edited by Otthein Rammstedt. Frankfurt am Main: Suhrkamp. Spranger, Eduard 1936 Probleme der Kulturmorphologie. Berlin: ADW. Steinbach, Franz and Franz Petri 1939 Zur Grundlegung der europäischen Einheit durch die Franken. Leipzig: Hirzel. Wegener, Philipp 1891 Die Bearbeitung der lebenden Mundarten. In: Hermann Paul (ed.), Grundriss der germanischen Philologie, vol. 1, 931⫺944. Strassburg: Trübner. Weinreich, Uriel 1977 [1953] Sprachen in Kontakt. München: Beck. [Originally published in English.] Weisgerber, Leo 1929 Muttersprache und Geistesbildung. Göttingen: Vandenhoeck. Wilking, Stefan 2003 Der deutsche Sprachatlas im Nationalsozialismus: Studien zur Dialektologie und Sprachwissenschaft zwischen 1933 und 1945. Germanistische Linguistik, Vol. 173⫺174. [PhD dissertation, Heidelberg 1998.] Wrede, Ferdinand 1902 Ethnographie und Dialektwissenschaft. Historische Zeitschrift 88: 22⫺43.

Clemens Knobloch, Siegen (Germany)

8. Language and space: Structuralist and generative approaches 1. 2. 3. 4. 5.

Introduction Structuralist approaches Generative approaches Conclusion References

1. Introduction In their strictest versions, structuralist and generative approaches to language research are not concerned with language external determinants of linguistic variation such as geographical space. This becomes clear when the goals of these theoretical approaches are compared with the goals of traditional dialectology.

126

II. Linguistic approaches to space The goal of traditional dialectology is to describe and compare individual dialect systems and to show the geographic distribution of their linguistic properties. The focus is on variation in the lexical, phonetic and morphological domain. Geographic maps are the primary output of this research. These maps show dialect areas and (bundles of) isoglosses, the borders between different dialect areas defined by one or more linguistic variables. On the basis of these geographic maps, the interaction between isoglosses and external factors are studied, such as the influence of natural communicative boundaries (e.g., swamps, rivers, mountains), political boundaries etc. Geographic maps are also thought to represent and help study historical change and the way linguistic properties propagate across a geographical area. Under the assumption that synchronic geographical variation reflects the diachronic development of a linguistic property, dialectologists seek to order different synchronic variants in terms of stages of a language, using the oldest stage as a point of reference. In structuralist approaches to language variation, an individual language variety is taken to be a synchronically closed system. The goal of structuralist research is to uncover the structure of this system, which is thought to be the sum of the relations between elements of this system. Linguistic elements cannot be studied in isolation, but only in their synchronic relations with elements within the same system. Since a linguistic element is defined by the set of relations it entertains with other elements within the same system, an element of language variety A that appears in a shape identical to that of an element of language variety B need not be structurally identical with it as it may be part of different sets of oppositions in the two systems. As a result, structuralist research of dialects has often been restricted to a description of the phoneme inventory of a language variety (or a subsection of it) by analysis of the linguistic distribution of these phonemes within this variety and to an analysis of phonemes in terms of distinctive features. Structuralist research on dialects at the level of the lexicon, morphology or syntax has been much less common. The structuralist perspective does not immediately invite comparison of diachronically or synchronically related language varieties and consequently does not yield as many geographical maps as traditional dialectology. Yet comparative structuralist dialect research does exist, starting with the influential work of Weinreich (1954) (cf. section 2). Jongen (1982) provides a good overview of structuralist dialectology and points out that structuralist dialect geography has made a major contribution to our understanding of language system internal conditions for innovation and language change, in addition to the language external conditions that are central in traditional dialectology. Important work in this domain includes Moulton (1961), Goossens (1969) and Wiesinger (1970). If language varieties are synchronically closed systems and should be studied individually, the delimitation of individual language varieties is important. Since what laymen call a language or dialect is in reality an aggregate of language varieties with partially overlapping linguistic properties, this is a serious problem. The most extreme solution to this problem, in particular in American structuralism, is to break down the object of description to the idiolectal level, i. e., the total set of speech habits of a single individual at a given time. However, this is only an apparent solution, as the resulting object usually will still be heterogeneous: e.g., every speaker masters different styles, and many dialect speakers also speak the regional and standard varieties related to their dialect. Generative linguistics can be seen as the successor of structuralism and it shares some central assumptions with it. In this approach, the description and analysis of language systems is only an intermediate step towards the eventual goal, which is to characterize

8. Structuralist and generative approaches the innate human capacity to acquire a language and to explain the process of language acquisition. This innate human capacity is taken to be a set of abstract linguistic structuration principles that all natural languages have in common and which define the limits of language variation. Generative linguistics, in particular in its early stages (i. e. the 1960s and 1970s), explicitly abstracts away from external factors and from much of the variation. Like structuralism, in its strictest sense, generative grammar studies idealized idiolects as these would most closely represent language competence. In theory, then, it would be sufficient to study the language variety of one speaker in great depth to reach these goals. In practice, however, the comparative method has played a major role in generative linguistics. Initially this involved the comparison of genetically less closely related languages. From the 1990s on, comparison of closely related dialects has been gaining ground rapidly as well. Some examples are Beninca` (1987), Kayne (2000), Poletto (2000), Zwart (1997), Barbiers et al. (2005, 2008) for syntax and Antilla (2006) for phonology. To the extent that such research produces geographic maps, these maps are not the goal of the research but only a helpful tool. Analysis of the external distribution of linguistic features has a very low priority. While the structuralist approach is largely restricted to the (morpho-)phonological system, in the generative approach both (morpho-)phonology, (morpho-)syntax and semantics are central research domains. The study of the lexicon and thus lexical variation is less prominent and the same holds for phonetic variation. Finally, the generative approach differs from structuralism and dialectology methodologically. The latter approaches can be characterized as primarily descriptive. Generative linguistics is more theoretically oriented and uses the hypothethicodeductive method for data collection and analysis. In view of its structuralist roots and its increasingly micro- and macrocomparative orientation current generative linguistics can be seen as an eclectic combination of dialectological, structural, typological and formal-theoretical approaches.

2. Structuralist approaches The structuralist approach to language variation developed out of the Vienna tradition of the Neogrammarians (e.g., Pfalz’s 1918 Reihenschritttheorie; cf. Murray, this volume, and Schmidt, this volume, for the relation between the neogrammarian and the structuralist approach), and is usually taken to start with the publication of de Saussure’s Cours de Linguistique ge´ne´rale in 1916. De Saussure (1949: 20) defines as the task of general linguistics to describe all the linguistic systems of the world. Weinreich (1954: 390) argues that if this means that idiolects in the sense described above are the object of study, the task of structuralist linguistics would be inexhaustible, and its results would be trivial and hardly worth the effort. Comparison of language varieties therefore requires a more flexible version of structuralism. To make a structuralist approach to spatially determined variation possible, Weinreich proposes that so called diasystems should be the object of structuralist research. The linguistic analyst can construct a diasystem out of two or more systems that have partial similarities. Diasystems are not just the linguists’ constructions but have psychological reality for bilingual (or: bidialectal) speakers. Structural dialectology is the investigation of the problems arising out of the joint treatment of different systems which

127

128

II. Linguistic approaches to space show partial similarity. In its focus on the language internal properties of linguistic systems, structuralist dialectology can fruitfully abstract away from extra-structural factors such as geography, ethnography, political and cultural history. In fact, the study of diasystems had already been an important part of traditional dialectology before the advent of structuralism. However, the structuralist approach to diasystems yields questions and results that are considerably different from the results of dialectology. This holds both at the level of the analysis of the system and at the level of the geographical description (mapping) of linguistic variables. The main difference between the diasystems constructed by traditional dialectologists and those constructed by structuralists is that traditional dialectologists tend to ignore the structures of the constituent varieties. Weinreich (1954: 392) notes that dialectology compares linguistic elements belonging to different systems without sufficiently stressing their intimate membership in those systems. This amounts to a non-phonemic (or phonetic) approach to variation in the domain of sounds. An example from Weinreich (1954: 392⫺393) serves to illustrate the different analytical results of the dialectological and structural approaches. Suppose four speakers of a language pronounce the word man in the following ways: (1) [man], (2) [man], (3) [ma˚n] and (4) [ma˚n]. On the basis of these pronunciations alone, a non-structural analysis would conclude that (1) and (2) should be grouped together and distinguished from the group formed by speakers (3) and (4). Suppose now the following hypothetical system. Informant (1) speaks a variety in which vowel length is a distinctive feature, such that a short vowel in a word cannot be replaced by a long vowel without changing the meaning of that word. The form of informant (1) is then phonemically /ma˘n/. Informant (2) does not distinguish vowel length and has /man/. In the system of informant (3), the pronunciation of [a˚] in [ma˚n] is a positional, non-distinctive rounded variant of [a] showing up between /m/ and /n/. Phonemically, informant (3) has /man/. The fourth variety does not have this positional variation. Informant (4) has /mon/. Thus, according to this structural analysis, system (2) and (3) are the same and group together, while systems (1) and (4) are different. The dialectological analysis of this hypothetical system yields a map of the /a/ area that is split into two parts by the isogloss between [a] and [a˚]. According to the structural analysis, however, this isogloss is less important than the one between [ma˚n] ⫽ /man/ and [ma˚n] ⫽ /mon/. The latter isogloss is absent on the dialectological map as it does not correspond to any difference in pronunciation. Thus, structural dialectology can make its own important contribution to our insights about the geographical distribution of diasystems and that of the individual language varieties within such systems, despite the fact that it is primarily concerned with the system-internal properties of language varieties. This is relevant for traditional dialectology, too, with its focus on the external distribution of linguistic variables. At the same time, structural dialectology with its central hypothesis of language varieties as synchronically closed systems needs the methods of traditional and modern dialectology to be able to delimit the language varieties and diasystems of which it wants to investigate the structural properties. This gives rise to advanced statistical analysis of the correlations between linguistic variables within one linguistic level (e.g., phonology), across levels (e.g., correlations between phonological and syntactic variables) and between linguistic variables and language-external factors. Jongen (1982) provides a more extensive introduction to structural dialectology including many illustrative case studies.

8. Structuralist and generative approaches

3. Generative approaches Generative linguistics investigates I-language, the internal mental representation of a language variety as a specific instantiation of Universal Grammar, the hypothesized set of innate abstract linguistic building principles that define the limits of linguistic variation. E-language, the external properties and distribution of linguistic variables in social and geographical space and time, does not belong to its research domain. Like structuralism, generative linguistics studies language competence, not language use. As a result, generative linguistics initially abstracted away from a large part of variation, including dialectal variation. This does not mean that dialects cannot be the subject of generative research, since every dialect is just another instantiation of Universal Grammar. However, until the early 1990s, dialects played a very minor role in generative research, mainly for methodological reasons. Since native speakers’ intuitions about linguistic constructs are the primary source of data in generative linguistics and since the intuitions of the researcher and his or her circle, often speakers of a standard language, are the most directly accessible ones, there were hardly any attempts to shift the attention to dialects. At the same time, it was clear from the start of the generative enterprise that comparing different language varieties is necessary to discover universal properties, and it was also clear that any linguistic theory based on the Universal Grammar hypothesis, should be able to explain the attested variation. In practice then, until the 1990s, generative grammarians on the one hand investigated the (morpho-)phonological, (morpho-)syntactic and semantic properties of usually standard or national languages, and developed explicit formal grammars that were designed to generate all and only the well-formed structures of a language. Obviously, these always were partial grammars, since the tremendous complexity of languages and the delimitation problem made it impossible to formulate grammars for entire language varieties. On the other hand, by comparing the grammars of different languages, a theory was developed to account for the attested variation between these languages (macrovariation). However, a theoretical and methodological apparatus for investigating intra-language variation (microvariation) was lacking. If we assume the generative period to start with Chomsky (1955, 1957), the period from 1955 until today can be roughly divided into two: a rule-based part and a constraint-based part. In (morpho-)syntax, primarily rule-based approaches gradually developed into constraint-based approaches such as the Principles and Parameters model (cf. Chomsky 1981) and the Minimalist Program (Chomsky 1992: 1995). In phonology, the switch to a constraint-based model coincides with the beginnings of Optimality Theory (McCarthy and Prince 1993; Prince and Smolensky 1993).

3.1. Phonological microvariation 3.1.1. Rule-based approaches to phonology Generative phonology starts with The Sound Pattern of English (SPE; Chomsky and Halle 1968). It presents a comprehensive view of the phonology of English, and is a landmark in the field of phonology. Like Chomsky’s syntactic work, SPE is based on

129

130

II. Linguistic approaches to space the assumption that the mental grammar has a modular organization. The lexical, phonological and syntactic components of this grammar are relatively independent but interacting modules. In this view, phonology can be studied as a linguistic subsystem, separate from other components of the grammar. The phonetic form uttered by a speaker is the result of applying context-sensitive derivational rules to an underlying phonemic sequence. As in structuralism, phonemes are not primitive units but bundles of plus-or-minus valued features based on articulatory and perceptual properties of sounds such as vocalic, high, back, anterior, nasal, sonorant, etc. The oppositions between these features define the inventory of phonemes of a language variety. The phonological component of each lexical entry is considered to be a linear sequence of these feature bundles. Derivatives of SPE have made modifications by changing the inventory of segmental features, considering some to be absent rather than having a positive or negative value, or adding complexity to the linear, segmental structure assumed by Chomsky and Halle. Two major successor theories are Lexical Phonology and Morphology, and Autosegmental Phonology. Lexical Phonology and Morphology (Kiparsky 1982) involves a theory of the organization of grammar. It deals with the relationship between phonology, morphology and the lexicon and has as its basic claim that all morphological processes, and many phonological ones, are located in the lexicon. Phonological rules fall into two classes, lexical rules, which may interact with morphological rules (see below), and postlexical rules, which may not interact with morphological rules. Morphology and phonology apply in tandem. After every word-formation rule, lexical phonological rules re-apply. The effect of this is that lexical phonological rules apply cyclically. Autosegmental Phonology was developed by Goldsmith (1990 [1976]) based on ideas of Bloch (1948), Firth (1948) and Hockett (1955). According to this theory, phonological representations consist of more than one linear sequence of segments. Each linear sequence constitutes a separate tier. The co-registration of elements (or autosegments) on one tier with those on another is represented by association lines. Each phonemic feature in a language appears on exactly one tier. A large part of phonological generalizations are interpreted as a restructuring or reorganization of the autosegments in a representation. For example, two vowels that are separated by a consonant in a linear sequence can be adjacent on their own tier, which explains why assimilation processes between linearly non-adjacent vowels are sometimes possible. Some important examples of the usefulness of autosegmental analysis include the study of vowel and nasal harmony systems (Clements 1976) and the vocalism and consonantism of Arabic (McCarthy 1981).

3.1.2. Microvariation in rule-based phonology In a rule-based theory, the potential loci of (for instance, dialectological) variation within a language include: (i) variation in the inventory of distinctive features that are active in a language variety, and thus variation in the phoneme inventory (cf. Weinreich’s hypothetical example discussed above involving the absence of the distinctive feature vowel length); (ii) variation in the transforming rules that are active, including the optional or obligatory character of their application; (iii) variation in the linguistic context in which a rule applies; (iv) variation in the order in which different rules apply (cf. Bailey 1973);

8. Structuralist and generative approaches

131

(v) variation in the levels at which different rules apply (e.g., root, stem, word, postlexical). The rule-based approach to microphonological variation, in particular the effect of rule ordering and optionality of rule application, can be illustrated with the following Finnish case taken from Antilla (2006). The relevant rules of the phonological system of Finnish are given in (1), each with an example. The degemination rule applies to all geminate stops in the given context. (1)

Vowel deletion

a J Ø / _i

Assibilation (optional)

t J s / _i

Degemination

tt J t / _VC

Apocope (optional)

i

J Ø

huuta-i J ‘shout-past’ huut-i ~ ‘shout-past’ ott-i-n J ‘take-past’ huus-i ~ ‘shout-past’

huuti huusi otin huus

These rules must apply in the order given in (1) to derive the facts: (2) a. Vowel deletion feeds Assibilation /huuta-i / J huuti J huusi b. Degemination counterfeeds Assibilation /otta-i-n / J ottin J otin (*osin) c. Apocope counterfeeds Degemination /hakkat-i / J hakkasi J hakkas (*hakas) d. Vowel deletion feeds Degimination /otta-i-n / J ottin J otin e. Apocope counterbleeds Assibilation /huuta-i / J huuti J huusi J huus

‘shout-past’ ‘take-past-1p.sg’ ‘beat-past’ ‘take-past-1p.sg’ ‘shout-past’

This ordering of rules predicts a four-way variation, given that assibilation and apocope apply optionally: (3) Vowel Deletion Assibilation (optional) Degemination Apocope (optional)

/lentä-i/ lenti ⫺ ⫺ ⫺ [lenti] ~

/lentä-i/ lenti lensi ⫺ ⫺ [lensi] ~

/lentä-i/ lenti ⫺ ⫺ lent [lent] ~

/lentä-i/ lenti lensi ⫺ lens [lens]

Southeastern Finnish is an example of a dialect in which exactly these four variants occur. The extrinsic ordering of rules is crucial. If Assibilation were to follow Apocope, a three-way variation was predicted: lenti ~ lensi ~ lent, but no *lens, since the relevant environment for Assibilation is lost after Apocope of /i/ from [lenti]. There are no Finnish dialects with this three-way variation. Optionality in the phonological system is often not completely free: the application of a rule is frequently determined by language-external (sociolinguistic) factors. Labov

132

II. Linguistic approaches to space (1969) and Cedergren and Sankoff (1974) developed a statistical technique to estimate the relative weight of internal and external factors based on corpus frequencies. They proposed to associate the resulting weights with the structural descriptions of SPE-style rules, giving rise to so called variable rules.

3.1.3. Constraint-based approaches to phonology The constraint-based approach to phonology currently known as Optimality Theory (OT) started with Prince and Smolensky (1993 [2004]) and McCarthy and Prince (1993) and has now almost completely replaced the derivational rule model of SPE and its successors. In the OT paradigm there are no derivational rules. The central idea of OT is that the observed forms of language arise from the interaction between conflicting constraints. An OT grammar has three basic components (cf. Archangeli and Langendoen 1997; Kager 1999; Antilla 2003 for good introductions): (i)

The lexicon, which provides the set of underlying forms. A crucial hypothesis of OT is that there are no language-specific constraints on the input (the so called richness of the base): Every grammar is able to deal with every input. For example, the grammar of a language that does not allow complex consonant clusters should tell us what happens with the underlying form /flask/. Two options are vowel epenthesis (e.g., falasak) and consonant deletion (e.g., las). (ii) The generator (GEN) generates the list of possible outputs, or candidates, on the basis of the input from the lexicon. Lexical items, i. e. linear phoneme sequences, are the input to the generator. GEN is maximally permissive, i. e. any output candidate is permitted for any input candidate within the limits of structural wellformedness. Thus, in the example in (i), vowel epenthesis and consonant deletion are not rules or operations, but terms that describe differences between the input and the output. (iii) The evaluator (EVAL) contains a universal constraint set (CON). There are two types of constraints. Faithfulness constraints prefer the output that most closely resembles the input. Markedness constraints prefer the phonologically least marked output. All constraints are active in all languages, but languages differ in the ranking of the constraints. On the basis of this constraint ranking, the evaluator selects the least offensive candidate from among the output candidates. This is the optimal candidate and the actual output.

3.1.4. Microvariation in constraint-based phonology Here is an example from Antilla (2003: 214⫺215) to illustrate how this theory works. The relevant constraints are given in (4). We assume the ranking in (5) for a hypothetical dialect of English, where X >> Y means that X outranks Y. (4)

Onset *CxCod Max(C)

Onsets are required No coda clusters No consonant deletion

(markedness) (markedness) (faithfulness)

8. Structuralist and generative approaches (5)

Onset

>>

*CxCod

>>

133 Max(C)

Suppose the input form is /m=st/ and two relevant output forms delivered by the Generator are [m=st] and [m=s]. Both variants have an onset and thus satisfy Onset. The variant [m=st] has a coda cluster [st] and therefore violates the constraint *CxCod. The variant [m=s] has no coda cluster but violates the constraint Max(C) because the final consonant [t] is deleted. According to the ranking in (5), the identity of the input to the output (faithfulness) imposed by Max(C) is less important in this dialect than the markedness of a coda cluster. As a result, [m=s] is the optimal candidate even though it violates Max(C). In this hypothetical dialect, then, final [t] is invariably deleted to avoid coda clusters. If the ranking of *CxCod >> Max(C) were the reverse such that it would be more important to not delete the consonant than to have a complex coda, we would get a dialect in which the final [t] is invariably retained. Thus, according to OT, language varieties do not differ in the constraints that are active, but in the rankings of universal constraints. The language particular rankings of constraints are the primary locus of cross-linguistic variation in OT. In a diasystem, we expect minimal differences in the ranking of constraints to capture the minimal differences between the language varieties, and we expect more rigorous rerankings to yield different diasystems. Constraint ranking is not the only locus of variation in OT. Antilla (2003) mentions the following additional loci: tied violations, pseudo-optionality, multiple grammars, stratified grammars and continuously ranking grammars. We have a tied violation if two or more candidates incur exactly the same violations with respect to all the constraints of a grammar. Since the grammar in such a case defines two equivalent outputs, this gives rise to intradialectal optionality. Pseudo-optionality arises when the lexicon is the locus of variation. There is a lexical choice between two minimally different lexical items, while the constraint ranking is constant. For example, a language variety has two items in its lexicon which both mean ‘cost’: /cost/ and /cos/. These two variants will not be compared in the competition between constraints, they each survive their own competition. An argument in favor of this approach is that different (socio-)linguistic information may be associated with the lexical items, e.g., that /cos/ is typical for working class colloquial speech. Pseudo-optionality can account for variation both within and across language varieties. The central hypothesis of the Multiple Grammars Model (Kroch 1989: 1994) for (morpho-)syntactic variation is that an individual can simultaneously possess several grammars. This hypothesis is necessary to account for multilingualism. Applying it to microvariation implies that every speaker is considered to be multilingual. In OT, the Multiple Grammars Model means that a single speaker commands a set of distinct total rankings of constraints. This model makes interesting quantitative predictions, as the following example from Antilla (2003: 219) shows. Suppose a speaker commands three grammars with the three rankings in (6): (6)

Input a. /m=st/ b. /m=st/ c. /m=st/

Ranking Onset >> *CxCod >> Max(C) Onset >> Max(C) >> *CxCod Max(C) >> *CxCod >> Onset

Output /m=s/ /m=st/ /m=st/

134

II. Linguistic approaches to space If the three grammars are equivalent, this means that there is 1 : 3 chance that the output will be /m=s/ and a 2 : 3 chance that it will be /m=st/. This is expected to correspond with usage frequency. The speaker is predicted to use /m=s/ 33 percent of the time and /m=st/ 67 percent. Special cases of the Multiple Grammar Model include Stratified Grammar (Antilla 2006), which consists of internally unranked strata of constraints which are externally ranked, Floating Constraints (Reynolds 1994) and Partially Ordered Grammars (Antilla and Cho 1998). Finally, a variant of OT that deserves mentioning is the Continuous Ranking Model (Zubritskaya 1997; Boersma 1998; Boersma and Hayes 2001). According to this model, ranking of constraints is not discrete but continuous. Each constraint defines a range that may (temporarily) overlap with the range of another constraint, giving rise to variation. Boersma and Hayes (2001) developed the Gradual Learning Algorithm that is able to learn continuously ranked grammars. They also show that Continuously Ranked Grammars predict the empirical frequency of different variants of a form quite precisely.

3.1.5. Rule-based vs. constraint-based approaches There are several advantages of OT over the derivational rule model (Roca 1997): (i) Constraints play an important role in the derivational model too, but in an unprincipled way and at various levels of the derivation. The derivational model has the disadvantage that ungrammaticality has two sources: the rules themselves and the constraints. In OT, only the constraints are responsible for filtering out output structures. (ii) OT provides a direct formal framework for representing universality and markedness, because both GEN and the constraints are assumed to be universal and related to markedness. (iii) OT provides a fertile ground for explaining the learnability of the system, because GEN and all constraints are included in UG, such that the learnability burden reduces to constraint ranking (and establishment of lexical representations, in common with traditional generative theory). (iv) OT dispenses with the power inherent in rule interaction. (v) OT allows parallel operation of the constraints, and thus a declarative (as against procedural) formulation of them. An important problem for OT approaches is Opacity, which becomes clear when we try to model counterfeeding phenomena of the type discussed in section 3.1.2 in terms of constraint ranking. The logic of the problem is as follows (adapted from Roca 1997): Suppose we have a rule A J B /A__ that derives AB from the input AA, and a second rule B J C /__ ] that changes string final B into C. The first rule feeds the second rule; together, they turn AA into AC. Suppose we give these two rules the converse order. Rule B J C /__ ] now does not apply because the string AA does not contain B. Rule A J B /A__ does apply and this time the output is AB. By the external ordering of rules, rule A J B /A__ cannot feed Rule B J C /__ ], a counterfeeding relationship. The resulting output, [AB] is opaque in the context of this grammar, because it meets the structural description of one of the rules of this grammar (Rule B J C /__ ] ) to which nonetheless it has not been subjected. In an OT approach, the two rules can be translated into the constraints *AA and *B], disfavoring a sequence of two A’s and a string final B respectively. If these constraints work on the input AA, the result will always be AC, regardless of constraint

8. Structuralist and generative approaches ordering. The output string AA violates *AA, the output AB violates *B], whereas the output AC violates neither. Consequently, the string AB which can be derived by rule ordering in a derivational rule model cannot be accounted for in a constrained-based model. One approach to solve this problem is Stratal Optimality as proposed in Kiparsky (2003) and Antilla (2006). As noted above, in Stratal Optimality constraints belong to different strata depending on whether they apply in stem phonology, word phonology or post-lexical phonology.

3.2. Syntactic microvariation 3.2.1. Rule-based approaches to syntax The development from rule-based to constraint-based approaches discussed for phonology in the previous sections has also taken place in syntax. Here as well, the derivational rule model dominant from the 1950s onwards (Chomsky 1955: 1957) has developed into the Minimalist Program (Chomsky 1992: 1995), a model with a very impoverished generative module and a major role for constraints. However, the role of constraints in Minimalism is less clear than in Optimality Theory. In most Minimalist work, constraints are not ranked and there is no claim to their universality. There is an Optimality theoretic branch of generative syntax starting with Pesetsky (1997), but this is much less influential than it is in phonology. There are various other branches of generative grammar, including Relational Grammar (cf. Perlmutter 1980; Blake 1990), Lexical Functional Grammar (cf. Bresnan 2001), Categorial Grammar (cf. Steedman 1996), Generalized Phrase Structure Grammar (Gazdar et al. 1985), Head-driven Phrase Structure Grammar (Pollard and Sag 1994). But since the Chomskyan version of generative grammar, most recently the Principles and Parameters model (Chomsky 1981) and the Minimalist Program (Chomsky 1992: 1995), is by far the most dominant approach to syntactic variation, the discussion will be restricted to this version. A rule-based syntax should characterize the syntactic competence of the speaker, i. e. the knowledge that is necessary to produce and interpret all and only the grammatical sentences of a language. A rule-based generative syntax has the following components: (i)

Lexicon. This specifies the idiosyncratic (i. e., unpredictable) phonological, morphological, syntactic and semantic properties of a lexical item. Morphology is usually taken to be part of the Lexicon, too. (ii) Rewriting rules. A set of context free rules operating on syntactic categories. Example: PP J (AdvP) P NP states that a prepositional phrase consists of an (optional) adverbial phrase, a preposition and a noun phrase (e.g., right on my head). Phrases expand to sequences of categories. These categories are then substituted by lexical items of the relevant category. (iii) Transformations. A set of operations that change the structures that are the output of the rewriting component. Examples of transformations are movement of a constituent (e.g., fronting of a wh-word) and deletion. (iv) Semantic component. A set of rules to interpret the output of the transformational component semantically. This component interacts with other, non-linguistic cognitive modules.

135

136

II. Linguistic approaches to space (v) Phonological component. A set of rules that interprets the output of the transformational component phonologically. This component is the input for phonetics. Each module is restricted by universal and language specific constraints.

3.2.2. Microvariation in rule-based syntax All of the grammar components can be loci of cross- and intralinguistic syntactic variation. An example of lexical variation is the particle verb bijhebben, lit. ‘by-have, have with’. In Brabantish, sentence (7a) is grammatical due to the fact that the combination bij ‘by’ and hebben ‘have’ is lexicalized and expresses a reflexive verb. The same sentence is impossible in Standard Dutch (7b), in which bijhebben is not lexicalized and bij requires an overt reflexive (7c). The absence of zich in (7a) cannot be explained by assuming deletion or a silent version of the reflexive in (7a), as the reflexive normally cannot be left out in Brabantish when combined with bij and a different verb (cf. Barbiers and Bennis 2003; Barbiers et al. 2005): (7) a.

De timmerman heeft geen spijkers bij. the carpenter has no nails by ‘The carpenter has no nails with him.’

Brabantish

b. *De timmerman heeft geen spijkers bij. the carpenter has no nails by refl

Standard Dutch

De timmerman heeft geen spijkers bij zich. the carpenter has no nails by refl

Standard Dutch

c.

The rewriting rules can be a locus of syntactic (micro-)variation, too. For example, at the macrolevel, prepositional (e.g., English) and postpositional languages (e.g., Japanese) can be distinguished by the rules PP J P NP and PP J NP P. An example at the microlevel is the expansion of noun phrases. A possessive NP with a third person masculine possessor is expressed as (8a) in Standard Dutch. Colloquial varieties of Dutch have (8b). The difference can be captured in the rewriting rules in (8c) (irrelevant details and expansions of the rule are omitted): (8) a. zijn boek his book

Standard Dutch

b. hem zijn boek him his book c. Standard Dutch: NP J PossPron Colloquial Dutch: NP J Pron PossPron

Colloquial Dutch N N

This example also shows that it is not a priori clear which component of the grammar is the locus of a particular case of variation. An alternative analysis of the contrast between (8a) and (8b) would be to assume that both Standard and Colloquial Dutch have the rewrite rule NP J Pron PossPron N, but that only Standard Dutch has a

8. Structuralist and generative approaches deletion operation (i. e., a transformation in syntax or phonology) that removes the personal pronoun. A possible example of macrovariation in the transformational component is the distinction between wh-moving languages (e.g., English) and wh-in-situ languages (e.g., Chinese). On such an account, the wh-movement transformation would be absent in languages like Chinese. Alternatively, languages have been taken to differ in the level at which a transformation applies, overtly in syntax as in English or covertly at the level of semantic interpretation (LF: Logical Form) (Huang 1982). The example of microvariation given in (9) can also be interpreted as a difference in the transformational component, with the relevant transformational rule in Standard Dutch being blocked (but see below): (9) Brabantish/Dutch a. e´e´n zo’n raar kind one such a strange child

Brabantish/*Dutch b. zo’n raar kind e´e´n such a strange child one

c. Transformation: Move [NP zo’n raar kind ] across [e´e´n].

3.2.3. Microvariation in constraint-based syntax A central hypothesis of the currently most dominant version of generative grammar, the Minimalist Program (Chomsky 1995), is that there is no variation in the rewriting component and the transformational component and that there is no optionality in syntax. All syntactic variation is supposed to be reducible to the lexicon, more specifically to morphosyntactic feature specification of functional elements, and to the level of phonological form (PF), where the output of the syntactic module is turned into a phonological structure. It is often assumed that morphology is not part of the Lexicon, as in the preminimalist times, but part of PF, as in Distributed Morphology (Halle and Marantz 1993). There are various reasons for this shift in the analysis of linguistic variation. Both the rewrite component and the transformational component had been impoverished to a minimum before Minimalism started. Thus, the rewriting component, containing language and category specific and partially optional rules was reduced to a general X-bar theory (Jackendoff 1977) generalizing over all categories and capturing the hypothesis that at an abstract level the architecture of phrases is the same for all categories. Later reductions of this component include Chomsky’s (1995) Bare Phrase Structure that does away with categories and projection labels and leaves Merge (the merger of two elements or constituents) as the only structure building principle, and Kayne’s (1994) Antisymmetry, according to which all syntactic structures in all languages are uniformly binary and right-branching. Similarly, the transformational component was reduced to one operation, Move, which is the operation Merge applied to a constituent that is already part of the constituent that it is remerged with. The structures built by Merge are subject to various universal conditions, such as structural conditions on the operation Move (e.g., agreement, locality) and conditions of semantic interpretation (e.g., a verb should have all the arguments that it is specified for in the lexicon). An advantage of locating variability in the Lexicon and at PF is that it makes the task of language acquisition easier:

137

138

II. Linguistic approaches to space lexical items have to be learnt anyway and variation at PF is directly observable. Syntactic principles are innate and universal and don’t have to be learnt. This framework can be called constraint-based because it has a very general structure generating and manipulating component and a set of conditions or constraints that filter out ungrammatical structures. The constraint-based era in generative syntax in fact already started with Chomsky’s (1981) Principles and Parameters (P&P). But other than in the Minimalist Program, in P&P syntactic variation still can be located in the syntactic module of the grammar, where building principles are parameterized, the value of parameters being set in the process of language acquisition on the basis of the input. The idea of parameters was abandoned in the 1990s as too powerful a mechanism leading to an explosion of merely descriptive parameters. In the Minimalist framework, word order variation as exemplified in (9) cannot be explained in terms of variation in the syntactic component. Rather, movement of the constituent [NP zo’n raar kind ] across [e´e´n] is universally allowed provided that there is morphosyntactic agreement between these two elements. In Brabantish, there is agreement as the numeral [e´e´n] is overtly inflected for gender and agrees with the NP that is lexically specified for gender. In Standard Dutch, there is no agreement because there are no gender distinctions on [e´e´n], while the NP is lexically specified for gender. Thus, word order variation is here the result of the interaction between Move, the universal condition on movement Agree and variation in the morphosyntactic feature specification of [e´e´n]. This effectively reduces this case of syntactic variation to the lexicon (cf. Barbiers 2005). A possible case of reduction to phonological form can be made on the basis of the data in (10), four variants that occur in varieties of Dutch. (10) Ik weet niet (‘I don’t know’) a. wie of dat er komt who if that there comes ‘who comes’ b. wie of er komt c. wie dat er komt d. wie er komt We assume that the syntactic structure of the left periphery of the embedded clauses in (10) is identical, consisting of three positions, a wh-word position, a position for interrogative complementizers and a position for unmarked (declarative) complementizers. At PF, some of these positions can be silent up to recoverability. For example, the unmarked complementizer dat [finite, subordinate] in Dutch can not be left out except when it cooccurs with the interrogative complementizer of that has a superset of the features of dat: [finite, subordinate, interrogative]. An important difference with Optimality Theory is that the constraints in Minimalism are not ranked and language variation is not modeled as variation in the ranking of constraints. An Optimality branch of generative syntax does exist, however, starting with Pesetsky (1997). A recent example is Broekhuis (2008), a framework in which the universal properties of language are part of the generator and the variable properties part of an OT type evaluator. Broekhuis (2008) also provides a more extensive comparison of the various stages of generative grammar with OT.

8. Structuralist and generative approaches

4. Conclusion In conclusion, spatially determined phonological and syntactic microvariation play an important role in current generative theory, in the search for a theory capturing universal and varying properties of natural language. The main motivation is to enlarge the empirical basis of generative research. Another important motivation is that in Minimalism, syntactic variation is hypothesized to be the result of minor lexically specified morphosyntactic differences. Research into closely related language varieties is an excellent testing ground for this hypothesis, as the effects of minimal morphosyntactic differences on syntactic structure can be directly observed. The advantage of micro- and macrocomparison of large numbers of language varieties is also that potential correlations can be tested statistically (cf. Spruit 2008). Geographical maps can be helpful in visualizing such correlations. Eventually, an overall theory of linguistic variation should explain both the internal and the external factors determining such variation, a goal that is still far away.

5. Reerences Antilla, Arto 2003 Variation and phonological theory. In: J. K. Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.), The Handbook of Language Variation and Change, 206⫺243. Malden: Blackwell. Antilla, Arto 2006 Variation and opacity. Natural Language and Linguistic Theory 24: 893⫺944. Antilla, Arto and Young-mee Yu Cho 1998 Variation and change in Optimality Theory. Lingua 104, 31⫺56. [Special Issue on Optimality Theory.] Archangeli, Diana and D. Terence Langendoen (eds.) 1997 Optimality Theory: An Overview. Malden: Blackwell. Bailey, Charles-James 1973 Variation resulting from different rule orderings in English phonology. In: Charles-James Bailey and Roger W. Shuy (eds.), New Ways of Analyzing Variation in English, 211⫺252. Washington DC: Georgetown University Press. Barbiers, Sjef 2005 Variation in the morphosyntax of ONE. Journal of Comparative Germanic Linguistics 8(3): 159⫺183. Barbiers, Sjef and Hans Bennis 2003 Reflexives in dialects of Dutch. In: Jan Koster and Henk van Riemsdijk (eds.), Germania et Alia. A Linguistic Webschrift for Hans den Besten. Available at . Barbiers, Sjef, Hans Bennis, Gunter De Vogelaer, Magda Devos and Margreet van der Ham 2005 Syntactic Atlas of the Dutch Dialects, vol. 1. Amsterdam: Amsterdam University Press. Barbiers, Sjef, Johan van der Auwera, Hans Bennis, Eefje Boef, Gunter De Vogelaer and Margreet van der Ham 2008 Syntactic Atlas of the Dutch Dialects, vol. 2. Amsterdam: Amsterdam University Press. Beninca`, Paola 1987 Dialect Variation in the Theory of Grammar. Dordrecht: Foris. Blake, Barry J. 1990 Relational Grammar. London: Croom Helm.

139

140

II. Linguistic approaches to space Bloch, Bernard 1948 A set of postulates for phonemic analysis. Language 24: 3⫺47. Boersma, Paul 1998 Functional Phonology. The Hague: Holland Academic Graphics. [PhD dissertation, University of Amsterdam. Available from .] Boersma, Paul and Bruce Hayes 2001 Empirical tests of the gradual learning algorithm. Linguistic Inquiry 32(1): 33⫺40. Bresnan, Joan 2001 Lexical Functional Syntax. Malden: Blackwell. Broekhuis, Hans 2008 Derivations and Evaluations: Object Shift in the Germanic Languages. Berlin/New York: Mouton de Gruyter. Cedergren, Henrietta J. and David Sankoff 1974 Variable rules: Performance as a statistical reflection of competence. Language 50: 333⫺355. Chomsky, Noam 1955 [1975] Logical Structure of Linguistic Theory. New York/London: Plenum. Chomsky, Noam 1957 Syntactic Structures. The Hague: Mouton. Chomsky, Noam 1981 Lectures on Government and Binding. Dordrecht: Foris. Chomsky, Noam 1992 A Minimalist Program for Linguistic Theory. (MIT Working Papers in Linguistics.) Cambridge, MA: MIT. Chomsky, Noam 1995 The Minimalist Program. Cambridge, MA: MIT Press. Chomsky, Noam and Morris Halle 1968 The Sound Pattern of English. New York: Harper & Row. Clements, George N. 1976 [1980] Vowel Harmony in Nonlinear Generative Phonology: An Autosegmental Model. Bloomington: Indiana University Linguistics Club. Firth, John R. 1948 Sounds and prosodies. Transactions of the Philological Society 47(1): 127⫺152. Gazdar, Gerald, Ewan Klein, Geoffrey Pullum and Ivan Sag 1985 Generalized Phrase Structure Grammar. Oxford/Cambridge MA: Blackwell/Harvard University Press. Goldsmith, John 1990 Autosegmental and Metrical Phonology. Malden: Blackwell. Goossens, Jan 1969 Strukturelle Sprachgeographie: Eine Einführung in Methodik und Ergebnisse. Heidelberg: Winter. Hockett, Charles F. 1955 A Manual of Phonology. (Indiana University Publications in Anthropology and Linguistics 11.) Baltimore: Waverly Press. Halle, Morris and Alec Marantz 1993 Distributed Morphology and the Pieces of Inflection. In: Kenneth Hale and Samuel J. Keyser (eds.), The View from Building 20, 111⫺176. Cambridge, MA: MIT Press. Huang, C.-T. James 1982 Move wh in a language without wh-movement. The Linguistic Review 1: 369⫺416. Jackendoff, Ray 1977 X-bar-Syntax: A Study of Phrase Structure. (Linguistic Inquiry Monograph 2.) Cambridge, MA: MIT Press.

8. Structuralist and generative approaches Jongen, Rene´ 1982 Theoriebildung der strukturellen Dialektologie. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke, Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 1, 248⫺276. (⫽ Handbücher zur Sprach- und Kommunikationswissenschaft 1.1.) Berlin: de Gruyter. Kager, Rene´ 1999 Optimality Theory. Cambridge: Cambridge University Press. Kayne, Richard S. 1994 The Antisymmetry of Syntax. Cambridge, MA: MIT Press. Kayne, Richard S. 2000 Parameters and Universals. New York: Oxford University Press. Kiparsky, Paul 1982 Lexical phonology and morphology. In: The Linguistic Society of Korea (ed.), Linguistics in the Morning Calm, vol. 2, 3⫺91. Seoul: Hanshin. Kiparsky, Paul 2003 Finnish noun inflection. In: Diane Nelson and Satu Manninen (eds.), Generative Approaches to Finnic and Saami Linguistics, 109⫺161. Stanford: CSLI. Kroch, Anthony 1989 Reflexes of grammar in patterns of language change. Journal of Language Variation and Change 1(3): 199⫺244. Kroch, Anthony 1994 Morphosyntactic variation. In: K. Beals et al. (eds.), CLS 30, vol. 2, The Parasession on Variation in Linguistic Theory, 180⫺201. Chicago: Chicago Linguistic Society. Labov, William 1969 Contraction, deletion and inherent variability of the English copula. Language 45: 715⫺762. McCarthy, John J. 1981 A prosodic theory of non-concatenative morphology. Linguistic Inquiry 12(3): 373⫺418. McCarthy, John J. and Alan Prince 1993 Prosodic Morphology I: Constraint Interaction and Satisfaction. (Report no. RuCCS-TR-3, Rutgers University Center for Cognitive Science.) Available from . Moulton, William G. 1961 Lautwandel durch innere Kausalität: Die ostschweizerische Vokalspaltung. Zeitschrift für Mundartforschung 28(3): 227⫺251. Pesetsky, David 1997 Optimality theory and syntax: Movement and pronunciation. In: Archangeli and Langendoen: 134⫺170. Perlmutter, David M. 1980 Relational grammar. In: Edith A. Moravcsik and Jessica R. Wirth (eds.), Syntax and Semantics, 195⫺229. (Current Approaches to Syntax 13.) New York: Academic Press. Pfalz, Anton 1918 Reihenschritte im Vokalismus. In: Beiträge zur Kunde der bayerisch-österreichischen Mundarten, 22⫺42. (Kais. Akademie der Wissenschaften in Wien. Philosophisch.-historische Klasse. Sitzungsberichte 190/2.) Wien: Hölder. Pollard, Carl and Ivan A. Sag 1994 Head-Driven Phrase Structure Grammar. Chicago: University of Chicago Press. Poletto, Cecilia 2000 The Higher Functional Field in the Northern Italian Dialects. New York: Oxford University Press. Prince, Alan and Paul Smolensky 1993 [2004] Optimality Theory: Constraint Interaction in Generative Grammar. Available from . Malden, MA: Blackwell.

141

142

II. Linguistic approaches to space Reynolds, William T. 1994 Variation and phonological theory. Unpublished PhD dissertation, University of Pennsylvania. Available from . Roca, Iggy 1997 Derivations and Constraints in Phonology. Oxford: Clarendon. Saussure, Ferdinand de 1949 Course de linguistique ge´ne´rale. Paris: Payot. Spruit, Marco Rene´ 2008 Quantitative Perspectives on Syntactic Variation in Dutch Dialects. Utrecht: LOT. Available from . Steedman, Mark 1996 Surface Structure and Interpretation. Cambridge, MA: MIT Press. Weinreich, Uriel 1954 Is a structural dialectology possible? Word 14: 388⫺400. Wiesinger, Peter 1970 Phonetisch-phonologische Untersuchungen zur Vokalentwicklung in den deutschen Dialekten. 2 vols. Berlin: de Gruyter. Zubritskaya, Katya 1997 Mechanisms of sound change in OT. Language Variation and Change 9(1) 121⫺148. Zwart, Jan-Wouter 1997 Morphosyntax of Verb Movement. A Minimalist Approach to the Syntax of Dutch. Dordrecht: Kluwer.

Sjef Barbiers, Amsterdam (The Netherlands)

9. Language and space: The variationist approach 1. 2. 3. 4. 5. 6.

Introduction The sidelining of the spatial in early variationism Early geolinguistics Subsequent geolinguistic work Concluding remarks References

1. Introduction In this chapter I discuss the extent to which variationist sociolinguistics has (and has not) engaged with concepts of geographical space and spatiality over the past half century. Three clear issues will become apparent. Firstly, variationism has at worst largely ignored spatiality and at best treated it quite distinctly and separately from other social factors until relatively recently. Secondly, when it has engaged with space, it has tended to be a socially devoid, Euclidean, distance-is-all type of space, rather than a socially rich spatiality, which recognizes that “the fact that social processes take place over space

9. The variationist approach and in a geographically-differentiated world affects their operation” (Massey 1985: 16), again until relatively recently. And thirdly, space has not, yet again until relatively recently, seen the sort of critique in sociolinguistics that has been witnessed by concepts such as style (Labov 1966; Bell 1984; Eckert and Rickford 2001; Coupland 2007), gender (Eckert 1990; Meyerhoff 1996), age (Eckert 1997) and so on. Below, I discuss how space has been integrated into variationist sociolinguistics. Unsurprisingly, the conceptualization of space in modern studies of language variation and change has tracked the development of the same in human geography, but a generation or so later (Britain 1991: 2002). Perhaps more surprising, however, is the fact that a lot of early variationist work drew inspiration from economic rather than human geography, and this heritage survives until the present-day, in, for example, the way that gravity models still seem central to much work on innovation diffusion (see the critiques of the application of gravity models to dialectology in Britain 2002, 2004, 2005).

2. The sidelining o the spatial in early variationism It must have been particularly exciting to be a sociolinguist in the 1960s. A new discipline, radically different both from the emerging giant of generative linguistics and from traditional dialectology, but drawing inspiration from both (from the former at least in the very early days, with the concept of variable rules, surviving today in the statistical package, Varbrul), was busy addressing, amongst other things, the many pressing questions encapsulated in Weinreich, Labov and Herzog (1968). The social embedding of change, captured through the perspective of quantitative variation within the apparenttime model, was one of the most important breakthroughs in these early heady days, and one which continues to drive sociolinguists in their research. As the discipline matured, so many of the tools of the trade were refined, many social categories unpicked and explored, but the fundamental question of how change was embedded in social practice endured. That social embedding, however, did not contemplate a place for space. Space was, instead, treated in a number of ways. Sometimes it was quite simply ignored. Since most very early sociolinguistic research was conducted in cities (with one extremely important exception, to be discussed below), and other social factors seemed more pressing than geography at the time ⫺ class, ethnicity, gender ⫺ and, since random sampling was used (so there was a nagging pressure to hold all other factors constant to enable the careful control and analysis of the chosen variable social constraints), space didn’t enter the picture. Labov’s (1966) study of “New York City”, for example, was a study of just the Lower East Side of NYC and, even then, only a part of it (Labov 2006: 104). So space was carefully controlled out of the study, and spatial variation within the neighborhood (let alone within the city) itself not examined. Trudgill’s (1974a) study of Norwich, however, random-sampled from five different parts of the city, since doing so “opens up the possibility of investigating geographical variation” (Trudgill 1974a: 22). He addressed geographical variation within the city specifically in the analysis of (o¯) (the reflexes of ME o¯ in words such as boat and goat), showing that in the suburb with areas of newest housing ⫺ Lakenham ⫺ peripheral to the urban centre, more diphthongs were found than in other suburbs, when social class was held constant. But these were rare forays into geographical variation in the new urban sociolinguistics. It was seen as dis-

143

144

II. Linguistic approaches to space tinct from the crucial social factors of the day and, if mentioned at all in more theoretical work, was conceptualized largely as a surrogate for time. Here is Weinreich et al. (1968: 155): “the problem of accounting for the geographical transition of dialects across a territory […] appears to be symmetrical with the problem of accounting for the transition of dialects through time in one community”. It was typical of the time in the social sciences to see geographical patterns as fossilizations or slices of time, betraying a historicism much chided by geographers. As Soja has reiterated: An essentially historical epistemology continues to pervade the critical consciousness of modern social theory. It still comprehends the world primarily through the dynamics arising from the emplacement of social being and becoming in the interpretive contexts of time […]. This historicism […] has tended to occlude a comparable critical sensibility to the spatiality of social life, a practical theoretical consciousness that sees the life-world of being creatively located not only in the making of history, but also in the construction of human geographies, the social production of space and the restless formation and reformation of geographical landscapes. (Soja 1989: 10⫺11)

With the apparent-time model at the very heart of the variationist enterprise, it is perhaps not too surprising that, in the early days at least, this historicism also pervaded studies of language variation and change. There are understandable reasons why space was downplayed in early variationist studies. Partly it could have been a reaction to the largely rural focus of traditional dialectology. I have argued before (Britain 2002) that, beyond the wonderful maps of the dialect atlases, there was actually very little that can be considered truly geographical, let alone spatially sensitive, in the work of the traditional dialectologists who had a historicist agenda every bit as keen as the later variationists. The role of space was largely reduced to that of a canvas onto which dialectological findings could be painted. A combination of the rejection of the rural focus of a largely asocial “map-heavy” traditional dialectology, on the one hand, and the view, on the other, that cities represented the best places to find the most fluid, heterogeneous, complex communities, facilitating investigations of the social embedding of change “where it’s all happening”, led variationist sociolinguistics to throw the rural baby out with the traditional dialectological bathwater. In the preamble to the Social Stratification of English in New York City ⫺ one of the key texts of early variationism ⫺ Labov (2006 [1966]) contrasted his earlier work on largely rural Martha’s Vineyard (Labov 1972) with the current research on the Lower East Side, making it clear that the latter represented “a much more complex society” (Labov 2006: 3). Nonetheless NYC was ultimately distilled down to the variables of age, class, ethnicity and gender, factors, which, as is made very clear (Labov 1972: 4⫺6), are also some (but not all) of the key pivots of social diversity in Martha’s Vineyard. There, groups of speakers of Portuguese, Native American and other miscellaneous ethnicities made up half if not more of the population (Labov 1972: 6), even before we take into consideration a small resident population coming originally from the Mainland and the large numbers of tourists. In addition, these populations were not distributed geographically evenly across the island, and were, naturally, engaged in a range of economic activities. As the results of Labov’s analysis made it clear, the community showed considerable sociolinguistic diversity with respect to age, location, occupation, ethnicity, orientation towards the island and desire to stay or leave (1972: 22, 25, 26, 30, 32, 39). In terms of social and linguistic structure, Martha’s Vineyard hardly fitted the rural

9. The variationist approach stereotype of quiet and sleepy pastoralism, or of traditional dialectological NORMs, as Labov showed. By contrasting a highly rural area with a highly urban one, Labov demonstrated that there are large-scale social(-linguistic) processes which are perhaps most obviously and vividly expressed in cities but are not confined politically, sociologically or epistemologically to an urban context. However, this perspective was not picked up and the resulting urban turn in sociolinguistics had significant consequences on the direction that variationism took. Urbanism still pervades much of the discipline ⫺ the rural is still portrayed as the insular, the conservative, the backward, the isolated, the static, as an idyll of peace and tranquility rather than as composed of heterogeneous communities, of contact, of change and progress and conflict. (See, for example, in the human geographic literature, Cloke 1999, 2005; Cloke and Little 1997; Macnaghten and Urry 1998; Shucksmith 2000; Short 2006; for further discussion of the irrelevance of the urban-rural divide, see Britain to appear b). Indeed variationism is sometimes even defined in terms of its urban focus: “urban sociolinguistics” or “urban dialectology” are terms used to describe variationism in general by scholars such as Chambers (2003: 1), Chambers and Trudgill (1980, 1998), Coupland (2007), Walters (1988: 119). To engage in a variationist analysis of an isolated rural area is, if we are to follow this nomenclature, to engage in urban dialectology. Others are careful to make the point that diversity ⫺ social and linguistic ⫺ can be found everywhere: “No matter how small and seemingly homogeneous the community, social status differences play an essential role in shaping dialect differences and can never be entirely discounted” (Wolfram and Schilling-Estes 1998: 32), and there are (just) a few early and significant examples of variationist studies of rural areas (e.g., Wolfram and Christian’s (1974) analysis of Appalachian English). The fetishism of the urban which struck sociolinguistics at the time was not, of course, confined to this discipline but was prevalent right across the social sciences in the post-war period (Britain 2002), and came under sustained critique from social theorists such as Harvey (1973) and Castells (1977) from the 1970s onwards. As Johnston makes clear from that point “urban areas changed their roles within geography: rather than being the focus of attention per se, they became the contexts within which cultural, economic, social and political processes and conflicts were played out” (Johnston 2000: 878).

3. Early geolinguistics During the burgeoning period of sociolinguistics in the 1960s, there was relatively little variationist geolinguistics. Indeed a review of what had gone on in the field in 1980 (Chambers and Trudgill 1980) listed little more than the work of the authors themselves. In the 1970s and early 1980s it was Chambers and Trudgill who did most to consider applying geographical perspectives to variationist research, Chambers by quantifying and performing spatial analyses of data from the Survey of English Dialects (e.g., Chambers 1982; Chambers and Trudgill 1998: 106⫺118), and Trudgill by adopting and adapting work from time geographers such as Torsten Hägerstrand, and introducing quantitative gravity models from economic geography to variationist sociolinguistics to shed light on innovation diffusion (e.g., Trudgill 1974b, 1975: 1983). Trudgill (1974b) began by discussing how cartographical methods for displaying behavioral variability (in this case the uptake of motorcars in Southern Sweden, from Hägerstrand 1952) could be

145

146

II. Linguistic approaches to space applied to dialect data by placing a grid of equally sized “cells” across the landscape under investigation, sampling from each cell and calculating the proportion that different variants of a variable are used, on average, in that cell, enabling both contrasts to be made with other cells and, consequently, the proportions of variants used in different geographical areas to be compared and explained. Trudgill argued, like Hägerstrand, that we should attempt to “ascertain the spatial diffusion of ratios” (Trudgill 1974b: 222, his emphasis) and that the work could be applied to examine the geographical dissemination of linguistic innovations. He proposed that such approaches would enable us to compare different time periods, using the apparent time model, as well as consider physical and social barriers to uptake of innovations (as Hägerstrand had done). Then, through an analysis of phonological variation and change on the Brunlanes peninsula near Larvik on the south coast of Norway, Trudgill was able to show the geographical distribution of a change in progress ⫺ the lowering and backing of /æ/ from [e] to [aœ≈] ⫺ and point to higher levels of innovation in the urban centers of the peninsula than in the intervening rural areas. His work was an early example of the urban hierarchy model of linguistic innovation diffusion, whereby innovations descend down an urban hierarchy of large city to city to large town, to town, village and country. The usual explanation for this model is that whilst the distance between the diffusing centre and the receiving location plays some role, interaction between urban centers in modern societies is likely to be greater, and therefore a more frequent and effective channel for the transmission of new forms than between urban and rural, an account which owes much to Christaller’s (1933) famous central place theory. Transportation networks tend to link urban with urban, economic and consumption infrastructures tend to be based in and oriented towards urban centers, with the ensuing consequences for employment, commuting and leisure patterns, and these feed the hierarchical nature of diffusion (Britain 2002, 2005, 2006). Trudgill then went further, adopting gravity models from economic geography to see whether it is possible to predict which places will come under a greater influence from a vigorous innovation than others. He examined the extent to which three East Anglian urban centers in England were likely to be affected by the deletion of /h/, assumed to be spreading from London. In order to estimate the likely influence (including linguistic influence) that one place would have over another, the gravity models employed took into consideration two factors: the distance between the two places and the populations of the two places: Mij ⫽

Pi Pj (dij)2

where M ⫽ the interaction between the two places i and j, P ⫽ population and d ⫽ distance. Trudgill added a third factor, a weighting to take into account how similar the accents of the different interacting urban centers already were. The model predicted that the city of Norwich would be affected by the innovation most and the smaller town of King’s Lynn the least, and subsequent studies of h-dropping in these places confirmed these results. Callary’s (1975) research on the raising of /æ/ in Northern Illinois, USA supported the urban hierarchy model more or less ⫺ towns with a population of over 10,000 had greater levels of raising than smaller settlements, and a subsequent gravity model analysis by Chambers and Trudgill (1998: 185) of the Illinois data demonstrated a good, though admittedly not perfect fit with the data.

9. The variationist approach

4. Subsequent geolinguistic work The 1980s saw little rapprochement between “urban” sociolinguistic variationism and the rather sporadic engagement with geolinguistics. A similar period of disengagement had taken place in the social sciences too, but somewhat earlier. Massey notes that “in terms of the relation between the social and the spatial, this was the period of perhaps the greatest conceptual separation […]. For their part, the other [non-geographic; DB] disciplines forgot about space altogether” (Massey 1984: 3) and “continued to function, by and large, as though the world operated, and society existed, on the head of a pin, in a spaceless, geographically undifferentiated world” (Massey 1984: 4). Geographers were expected to “take up a position at the end of the transmission belt of the social sciences, dutifully mapping the outcomes of processes which it was the role of others to study” (Massey 1985: 12). Spatial patterns, at that time, were seen merely as the outcomes of processes of non-geographical kinds. So it was none too surprising that there was little interaction between sociolinguists and geolinguists either. Let’s make it clear, the two sides were not in dispute, but simply they proceeded as if neither needed the other ⫺ there was little geographical input to “urban” variationism and the geolinguistics of the time was largely asocial. Both camps agreed both that more had to be done and that the geographical dimension of variationism was underexplored. Chambers and Trudgill (1980), in the first edition of Dialectology, for instance, having presented the limited amount of work that had been carried out in this framework, presented geolinguistics as a potential future avenue for sociolinguistic investigation. Labov, in his summary paper of the first twenty years of variationism, firmly separated “spatial” contributions to language change from the “social”, and treated the study of linguistic heterogeneity in space, society and time as a “natural alliance” (Labov 1982: 20), but of separate disciplines. He (Labov 1982: 42) went on to state that “the study of heterogeneity in space has not advanced at the same tempo as research in single communities”. The division implies that heterogeneity in both time and society are somehow not in space, and that spatiality has not shaped the evolution of social variation in the communities under investigation. But this view was typical of the time (see further Britain 2002, to appear b). Human geography began to move beyond this theoretical and methodological separation of space and society in the late 1970s and 1980s finding a space for itself again, remembering the importance of place and that “the fact that social processes take place over space and in a geographically-differentiated world affects their operation” (Massey 1985: 16). From this point on, human geographers focused on conceptualizing spatiality as a contingent effect contributing to the contextual conditions mediating the operation of causal powers under the banner of “the difference that space makes” (Massey 1984; Sayer 1985; Cochrane 1987; Cooke 1989a, 1989b; Duncan 1989; Savage et al. 1987; Johnston 1991). This rediscovery of the relevance of, for example, place and region led to a recognition that these are dynamic entities, always in a state of evolution, in which social and economic processes are constantly being played out through a geographically differentiated filter. “Space and spatiality in general is socially constructed […]. That process of construction is constantly evolving. The spatialities of our lives are the product of continual ‘negotiation’, the outcome of the articulation of differentially powerful social relations” (Allen et al. 1998: 138). This rediscovery also triggered a concerted effort to develop critical perspectives on the geographies of socially marginalized groups,

147

148

II. Linguistic approaches to space such as those of gender, childhood, sexuality, ethnicity, ill-health and disability and “otherness”. Variationist sociolinguistics is slowly beginning to come to terms with this turn in human geographic thought and so more recently, there has been a raised interest in how a richer, socially sensitive approach to space within the discipline can shed light on variation and change. I examine some of the most salient, which fall broadly within three categories: innovation diffusion, dialect supralocalization and the increasing adoption of identity and practice based approaches in variationist sociolinguistics.

4.1. Innovation diusion The literature to date on the spread of new linguistic forms has proposed a number of different models or patterns: ⫺ wave or contagion diffusion (Trudgill 1986; Bailey et al. 1993; Britain 2002, 2004), whereby innovations, like the ripple effect caused when a pebble is dropped into a puddle, radiate over time out from a central focal area, reaching physically nearby locations before those at ever greater distances. Bailey et al. (1993: 379⫺380) suggest that contagion diffusion is operating to spread lax nuclei of /i/ in words such as field across Oklahoma, in the US, and Labov (2003: 17) suggests that contagion diffusion was at work in the spread of the term hoagie to represent a long filled bread roll from Philadelphia across the state of Pennsylvania; ⫺ urban hierarchical or cascade diffusion (Trudgill 1974b, 1983, 1986; Callary 1975; Gerritsen 1988; Gerritsen and Jansen 1980; Bailey et al. 1993; Herna´ndez Campoy 2003, Britain 2002, 2004, 2005, 2006; Labov 2003) whereby innovations descend down a hierarchy of large city to city to large town, to town, village and country. In Britain (2004), for example, I show how the diffusion of l-vocalization into the East Anglian Fens appears to be reaching small towns before the surrounding countryside, and Kerswill’s (2003: 235⫺236) investigations of the diffusion of labiodental variants of /θ Î/ in Britain show that [f v], having diffused, he claims, from London, have reached the cities of Bristol, Derby and Norwich before the nearer but smaller towns of Romsey and Wisbech. It has been from within the urban hierarchical model that gravity models, adopted from economic geography, have been applied to dialectological data (e.g., Trudgill 1983; Chambers and Trudgill 1998; Herna´ndez Campoy 2003 etc.); ⫺ cultural hearth diffusion (Horvath and Horvath 1997, 2001, 2002) whereby the innovation gains a foothold in both town and country in one particular region before diffusing to other parts of the country; Horvath and Horvath (1997) use the model to show how /l/-vocalization has taken hold in the area around Adelaide and coastal South Australia, before diffusing further afield in the country, and; ⫺ contra-hierarchical diffusion (Trudgill 1986; Bailey et al. 1993), whereby innovations diffuse against the urban hierarchy, arising in rural areas and spreading to urban ones (also cf. Vandekerckhove, Section 4.1. in this handbook). Trudgill (1986) demonstrates that smoothing processes ⫺ reducing triphthongs to diphthongs or long vowels ⫺ appear to be spreading southwards from rural north Norfolk in Eastern England into rural and urban areas of the rest of the county and into neighboring Suffolk, and Piercy (2007) shows that the diffusion of long back vowels in the BATH lexical

9. The variationist approach set is spreading against the urban hierarchy in Dorset in the South-West of England, despite the attrition of rhoticity operating hierarchically. The type of diffusion found appears very much to be context-dependent, however, and there are examples of the same variable diffusing differently in different places. Labov (2003), for example, reminds us that while urban hierarchy diffusion seems to be operating in the diffusion of the cot-caught merger in Oklahoma (Bailey et al. 1993), this merger is not diffusing in the same way across the border from Canada into the United States (Boberg 2000). The adoption of models of geographical diffusion in the variationist literature owes a lot to the work of the Swedish time-geographer Torsten Hägerstrand who pioneered the modeling of the spatial spread of innovations (e.g., Hägerstrand 1952). It was his work, as was mentioned earlier, that inspired Peter Trudgill’s application of the method to dialectological data most notably in southern Norway and in Eastern England (Trudgill 1974b: 1983) . The evolution of human geographic thought has not, however, been particularly kind to Hägerstrand’s approach. Both Blaikie (1978) and Gregory (1985) made cutting critiques of this approach to diffusion, highlighting its lack of sensitivity to social context and its reliance on outdated application of spatial laws and spatial processes (see also Britain 2009, to appear b). Problematically, Gregory (1985) argued, for example, the investigation of the spatial diffusion of innovations has proceeded a) on the assumption that the diffused form, in spreading spatially, totally levels away the formerly used traditional forms in its path, whereas, for example, dialectological research has shown that often diffusing innovations “mutate” en route. Contact between innovations and local traditional forms occasionally results in hybrid forms characteristic of neither the original innovation nor the conservative dialect (see, for example, Trudgill 1986; Britain 2005), and it is not unknown at all for the social evaluation of an innovation to change as the innovation spreads too. Glottal stops, most characteristic, perhaps, of informal working class speech in the South-East of England (though today they are diffusing across the country and spreading to middle class and more formal speech) have been found more among middle class women than working class speakers in both Cardiff (Mees and Collins 1999) and Newcastle (Watt and Milroy 1999). In Cardiff, Mees and Collins (1999: 201) argued that a shift to glottal stop use represented a shift to “more sophisticated and fashionable speech”; b) on the assumption that non-adoption of the innovation is a “passive state where the friction of distance applies a brake to innovation […] rather than ‘an active state arising out of the structural arrangements of society’” (Gregory 1985: 319). So rather than considering very slow moving or fossilized innovations (such as the diffusion of /a:/ in the BATH lexical set and /v/ in the STRUT lexical set in southern British English) as having simply “run out of steam”, we should expect to find active lines of resistance to those innovations. Indeed, as Wells (1982: 354) suggested, even middle class people in the north of England “would feel it to be a denial of their identity as northerners to say BATH words with anything other than short [a]”. As Gregory (2000: 176) made clear, resistance to an innovation “connotes a process of sustained struggle: considered and collective action on the part of people whose evaluation of the available information may be strikingly different to that of the ‘potential adopters’”; c) paying little attention to the structural contradictions on innovation adoption (Gregory 1985: 323) ⫺ in other words, to mention a dialectological example sometimes the

149

150

II. Linguistic approaches to space linguistic system of an accent in the path of a diffusing innovation may be incompatible with that innovation and consequently slow down or block its adoption. An example of the combination of this factor and the previous one is the widespread resistance in the English of Liverpool to the adoption of glottal stops for /t/ reported by Watson (2006). He found that despite the prevalence of glottal stops across a large swathe of contemporary England, including the North, not one glottal stop was found in prepausal position in his data from Liverpudlian adolescents (Watson 2006: 59). In addition, the use of [h] for /t/ reported in earlier work on Liverpool English by Knowles (1973: 234) only following short vowels in small monosyllabic function words (e.g., at and what) and not found at all in the surrounding areas of the North-West, or indeed elsewhere in England, not only remained robust in those contexts in the speech of the Liverpudlian adolescents, but had also spread to polysyllabic non-function words where the /t/ was preceded by an unstressed vowel (e.g., biscuit, bucket, chocolate) (Watson 2006: 59). Liverpool is also known for spirantization of oral stops in intervocalic and word-final position (Watson 2007; Honeybone 2001; Sangster 2001). A shift from a fricated form to the use of glottal stops consequently would entail a more complex change than in varieties without the spirantized forms, and Watson comments that the glottal stop is extremely restricted in Liverpool English to contexts preceding /l/ and other syllabic consonants such as /n/. But Liverpool also has an extremely distinctive local identity and a distinctive demographic history, and its avoidance of glottal stops could be seen as part structural contradiction, part social rejection and local loyalty; d) paying little attention to social constraints on innovation adoption (Gregory 1985: 323). Gravity models used in the dialectology of diffusion, for example, assume that everyone who uses the innovation has an equal chance of passing it on and that everyone in the geographical path of the innovation has an equal chance of adopting it. However, we live in a socially differentiated world, where access to the resources enabling mobility and contact are unevenly and unequally distributed. Gravity models do not take the complexity of social structure into account, leading Gregory (1985: 328) to argue that such diffusion models “failed to cut through the connective tissue of the world in such a way that its fundamental integrities are retained” and that there is a need “for a more sensitive intellectual surgery”. He argues that gravity models and diffusion theories have operated through “the detachment of ‘potential adopters’ from their social moorings and the displacement of subjects from social struggles” (1985: 328). Few attempts have been made to enrich linguistic diffusion models with social information about the speakers. A few sociolinguists have conducted comparative analyses of diffusion across the sorts of social categories typical of early variationist studies (e.g., contrasting ages, sexes, etc.; see, for example, Trudgill 1983; Britain 1991; Horvath and Horvath 1997, 2001, 2002), and Boberg (2000) demonstrates the effects of a national boundary in undermining predictions based on gravity models. Most notable, perhaps, however, is the important work of Wolfram and Schilling-Estes (see for example, Wolfram 1997, 2002; Wolfram and SchillingEstes 1995; Schilling-Estes and Wolfram 1999) contrasting patterns of change in two seemingly similar communities: Ocracoke off the East Coast of North Carolina and Smith Island in the Chesaspeake Bay of Maryland in the United States. Both were once relatively isolated but today more and more speakers from both islands are coming into contact with people from the mainland and are moving away to seek

9. The variationist approach better employment prospects. Both share some distinctive dialect characteristics: a back and raised nucleus of /ai/: [v= ] and a front gliding realization of /au/: [æ= ]. However, on the one hand, Ocracokers appear to be losing these distinctive features, while on the other, Smith Islanders are increasing their use of them and continuing to diverge from neighboring dialects (see Wolfram 2002: 770). The important sociodemographic distinction between the two communities is that while Ocracoke is becoming a popular destination for short and long term residence by non-islanders, few people are moving onto Smith Island and many are leaving, resulting in a concentration of the dialect in the mouths of the few that remain. Such communities are, of course, becoming increasingly rare, but the important point here is that the types of change underway on the islands are indicative of differences in their socio-spatial “becoming” (to use a term from Pred 1985: 338); e) paying little attention, furthermore, to the structural and social consequences of the adoption of the innovation (Gregory 1985: 304). If the innovation is adopted, does this adoption serve to dismantle social and linguistic divisions? If it is not adopted, does this rejection serve to bolster local identities as being distinct from those of the promoters of the innovation? In fact, Gregory (2000: 177), fifteen years after his original critique of diffusion theory, concludes rather pessimistically about the current state of this approach, suggesting that there had been little development in the search for a socially and contextually rich account of spreading innovations. Even in one of the more well-trodden topics of enquiry in the human geographic diffusionist literature, the spread of HIV/AIDS infection, he claims that much research had been “subordinated to the objectivist logics of spatial science” (see Brown 1995 for a noteworthy exception).

4.2. Supralocalization o dialects Diffusion, as became clear in the previous section, should be conceptualized as a form of varietal contact. Consequently, and not surprisingly, where breaks in contact frequency are found, we also find that linguistic breaks ⫺ isoglosses or dialect transitions ⫺ occur (Chambers and Trudgill 1998; Britain 1991, 2001: 2002). These breaks sometimes arise because of physical barriers to inter-regional communication. They are also shaped, however, by breaks in routinized social practice within speech communities (Britain 1997, 2002, 2005, 2009, to appear a, b). Giddens has argued that routines form “the material grounding of […] the recursive nature of social life” (Giddens 1984: xxiii), and channel everyday human behavior into a set of self-perpetuating socio-geographical “grooves”. Dialect boundaries and transitions are often found in breaks between the socio-spatial networks of these “grooves”. The geographies and histories of our social networks and those of the social, economic and political institutions which guide our daily lives in the West are played out, routinized and reproduced within functional zones usually centered around (or in the sometimes distant shadow of) one or a number of urban areas. Subsequently, the sociogeographical trajectories of speakers and their institutions are often strongly guided by past practices, by attitudinal considerations and by physical factors, and hence regions are formed (Britain 2002). As Johnston states,

151

152

II. Linguistic approaches to space places differ culturally, in terms of […] the ‘collective memory’. For a variety of reasons, some associated with the local physical environment, people’s responses to the problems of surviving collectively vary from place to place, at a whole range of scales. How they respond becomes part of the local culture, the store of knowledge on which they draw […]. That store […] becomes the inheritance of those who succeed, being transmitted intergenerationally to others who will modify it as they in turn tackle problems old and new. (Johnston 1991: 50)

Intra-regional mobility, whilst breaking down networks and routines at the very local level, reinforces supra-local structure, and with, for example, improvements in transportation routes, the shift from primary and secondary to tertiary sector employment as the backbone of the economy, higher levels of university attendance (at sites often well away from the local speech community), the normalization of long-distance commuting, labor market flexibility and the consequent geographical elasticity of family ties and other social network links, these supralocal functional zones are probably larger than ever before (see Britain to 2009; Sayers 2009). The previously mentioned social and geographical mobility within these supralocal zones has led to dialect contact between the varieties spoken within them. The result has been the emergence over time of regional koines ⫺ leveled supralocal varieties which are replacing some of the linguistic diversity that once reigned within individual regions. We now have a good deal of variationist evidence that intra-regional mobility is breaking down local varieties (though there are of course exceptions ⫺ see the Liverpool example earlier) in favor of larger supra-local ones, creating a smaller number of geographically expansive regiolects (see, among many, Milroy; Milroy and Hartley 1994; Milroy et al. 1994; Milroy 1999; Watt and Milroy 1999; Watt 2002; Watts 2006; also cf. Berruto and Herrgen in this volume). Milroy, Milroy and Hartley (1994), for example, demonstrated that in Newcastle-upon-Tyne, the highly local [?t] variant of /t/ is losing out, particularly among younger female speakers, in favor not of the standard [t] variant, but, as we saw earlier, the regionally more widespread but non-standard glottal form [?]. And Watt (2002) found that the regional Northern non-standard variants [e:] and [o:] of the FACE and GOAT variables respectively were taking over in Newcastle from the much more locally current [=e] and [we] variants (also discussed earlier), with the national standard forms serving as insignificant minority variants and used only among middle class speakers. Similarly, research by both Mathisen (1999) and Watts (2006) has suggested that [ng] forms of (n) may well be “rejuvenating” to become a supra-local form of the Midlands and parts of the North-West. Watts examined dialect contact and diversity in Wilmslow, a town situated some 20 km south of Manchester. Wilmslow is a predominantly middle class town, where many of Manchester’s wealthier business people, white-collar workers and football stars live. In the 1970s, a large local authority-run housing estate ⫺ Colshaw ⫺ was built on the side of Wilmslow to rehouse people who had previously been living in substandard housing in Manchester. Watts (2006: 147, 294, 170) found that while the middle class residents of Wilmslow used extremely low levels of many widespread non-standard forms (e.g., less than 14 percent /=n/ in unstressed words (94 percent in Colshaw estate), less than three percent th-fronting in words such as think and mother (> 43 percent in Colshaw), non-standard [ng] forms of stressed (n) (in words such as sing and ring) were used in more than 50 percent of tokens (and over 80 percent among the 11⫺18 year olds). Mathisen (1999: 120) found a very similar pattern for

9. The variationist approach speakers in Sandwell, near Birmingham. A number of other studies suggest its retention among young speakers across the Midlands and parts of the North-West, e.g., Redditch (Ryfa to appear a), the Wirral (Grainger 2005) and Liverpool (Watson 2007). Such supra-local koineization is also underway, as Ellis (1889) noted over a century ago, in south-eastern England, much to the excitement of journalists and politicians, the emerging regiolect having been named “Estuary English” (for discussions see Maidment 1994; Parsons 1998; Peys 2001; Przedlacka 2001; Ryfa to appear b, etc.). Torgersen and Kerswill (2004), for example, have demonstrated that the somewhat different vowel systems of the towns of Reading and Ashford in south-east England are converging and Britain (2005) has highlighted the variable acquisition of supralocal south-eastern forms in the Fens. There is, of course, variation within this regiolect, and local varieties still exist. Regional supralocalization is still underway, only affecting certain features and not others, and distinct local dialects form part of the mix that has engendered the regiolect in different places ⫺ the East Anglian version of the regiolect is distinct from, say, the Sussex version, since, obviously, East Anglian dialects helped shape its very emergence. Allen, Massey and Cochrane (1998) investigated the growth of the “South-East” as the focus of Thatcher’s neo-liberal project of the 1980s, and, using a number of measures such as income levels, wage increases, house price inflation, hi-tech employment growth, government public spending increases, revisited the “region” as a geographical concept. They emphasized that the South-East isn’t “out there”, well defined and demarcated, waiting to be documented, but is created, shaped and reworked by practice, both individual and institutional, within it. The region is both fluid and diverse. Dialectological evidence from this region highlights that at some levels it shows considerable linguistic homogenization, but at others provides examples of considerable linguistic diversity. Recent work on relative pronoun marking in the South-East of England emphasizes that there is still considerable intra-regional differentiation. Cheshire, Fox and Britain (2007), comparing London and the Fens (150 km apart, but both in the “South-East”), show radically distinct choices of pronoun marking. As Figure 9.1 shows, the two varieties

Fig. 9.1: Distribution of relative markers in subject and object function in (inner) London and the Fens (from Cheshire, Fox and Britain 2007)

153

154

II. Linguistic approaches to space agree in only one respect ⫺ the avoidance of which as a relative pronoun. Regions are still complex, diverse and fluid entities and the dialect patterns of the south-east of England exemplify that supralocal homogeneity is far far from complete. Mentioned earlier was the fact that both physical barriers and breaks in routinized communication networks can be the cause of dialect boundaries. Important, though, also is the recognition that these physical and social factors, along with attitudinal ones, together can serve to reinforce and recreate boundaries. The Fens in Eastern England is the site of a large bundle of isoglosses separating East Anglian dialects from East Midland ones as well as “Northern” ones from “Southern” ones (see Britain 1991: 2001). But social, physical as well as attitudinal considerations need to be taken into consideration to account for why this area is host to so many dialect transitions. Before the seventeenth century, the area was largely boggy marshland with sparsely distributed communities settled on small patches of higher ground which were often subjected themselves to regular flooding; because of their impenetrability, the area served as a regional and military frontier. Both the place and the people were stereotyped negatively by nonFenlanders. Darby (1931: 61) claims that there arose “a mythical fear of a land inhabited by demons and dragons, ogres and werewolves”, and he quotes both Felix who claimed the Fens were “especially obscure, which ofttimes many men had attempted to inhabit, but no man could endure it on account of manifold horrors and fears and the loneliness of the wide wilderness ⫺ so that no man could endure it, but everyone on this account had fled from it” as well as the famous diarist Samuel Pepys who in an entry on 18 September 1663, describes his travels “over most sad fenns, all the way observing the sad life which the people of the place do live, sometimes rowing from one spot to another and then wading”. These physical and attitudinal and political reinforcements of the Fens as a boundary led to social routines respecting this boundary too, and thereby (to recite Giddens), a set of self-perpetuating socio-geographical “grooves” developed which reinforced the view of the “Fens-as-barrier”. Nowadays, connections across the Fens are much easier, but the traces of the historical frontier can be still be seen in the geographical patterning of all but the very most recent linguistic innovations arriving in the area. Britain (1991, 1997: 2001) provides linguistic evidence of the north-east/south-west dialect boundary in the Fens which mirrored the path of the historical “frontier”, and Britain (2002: 612⫺ 616) shows how at a more local level, public transportation networks, drainage channel systems, relative population densities, political boundaries, local rivalries and competition between two urban centers with competing service, employment and entertainment provision reinforce local (and highly salient) dialect boundaries. Very much more recent innovations from southern England, especially non-salient, supra-regional ones, such as the vocalization of /l/, have ignored these local contextual conditions and created northsouth differences in the Fens, rather than east-west differences. These differential geographical patterns can, of course, only be fully understood through an appreciation of the contextual historical, social, political, attitudinal, economic and geographical development of the area.

4.3. Practice-based approaches The examples above have highlighted how social practice helps create and reinforce spatial distributions of linguistic features at the regional and subregional level. Research at the very local level, at the level of the local neighborhood, has also been fruitful in

9. The variationist approach demonstrating how our variable social geographies as speakers can serve to create and reinforce variable linguistic geographies. Two recent research projects are particularly notable. Firstly, at the very local level, is Eckert’s (2000) investigation of the relationship between dialect, adolescent communities of practice in Detroit and the appropriation of space in and outside school by those young people (also cf. Eckert and Quist in this volume). Eckert shows, for instance, how engagement in out-of-school urban and suburban sites, hanging-out in parks and cruising along the streets connecting the suburbs with more central parts of Detroit are all intertwined with language change. Strong correlations between the use of advanced variants of changes in progress and engagement as a cruiser or a frequenter of the urban parks show that those adolescents who brave the freedom, excitement and danger of these activities are those most likely to be leading many of the changes in Eckert’s study (cf. Eckert in this volume). For example, the backing of /v/ to [c], a change led clearly by “Burnouts”, shows clear positive correlations with cruising activity, while /v/ fronting to [e], on the other hand, a “Jock”-led change, correlates negatively with cruising (Eckert 2000: 151⫺152). Eckert concludes that “[t]he use of urban variables […] is a resource for the entire population in the construction of linguistic styles related to engagement in urban practice” (Eckert 2000: 153). Secondly we can look to the work of Barbara Johnstone and her colleagues examining how dialect features arise, become associated with and stereotyped to a particular place, and how speakers then synthesize local attitudes, media representations and other discourses concerning this stereotype to “trademark” or “enregister” the dialect of the place as a commodity, highlighting that a multitude of sources can together create a context for, shape and renew local patterns of language variation. Johnstone, Andrus and Danielson (2006), drawing data from corpora of newspaper reports about the Pittsburgh dialect, from interviews with local city residents and from extensive ethnographic observation, show how the monophthongization of /au/ to [a:] has taken the journey from indicator to marker to stereotype (in Labov’s (1972) parlance), both in terms of sociolinguistic patterning in the speech community and in terms of media and other commercial representations of the dialect, and in so doing has joined a small number of local dialect features which are clearly above the level of speaker consciousness. These features have become “legitimized” through ever more serious exposure in the media, and speakers use them as a resource for marking identity, “performance” and other discourse strategies, signaling, importantly for our discussion here, a link between language and place. Johnstone et al. (2006) make it clear that /au/ > [a:] originally signaled “working class” and “male” but not necessarily “Pittsburgh”, but that the evolution of the form in the speech community, its representation in the local Pittsburgh media, and its subsequent commodification in the city (in the form of T-shirts and mugs with dialect forms printed on them) have recycled or reallocated what were class and gender associations into place-based ones. Besides these more socially sensitive approaches to geographical language variation, another branch of geolinguistics which has seen somewhat of a revival in recent years is dialect cartography (cf. Nerbonne and Heeringa and Lameli in this volume), a revival sparked both by improvements in cartographical technology and software, and by a desire, following Labov’s (1994) important work theorizing patterns of phonological merger, split and vowel chain shifts, to investigate the geographical dispersion of those important language changes affecting North America. Telsur, a large-scale project col-

155

156

II. Linguistic approaches to space lecting and analyzing recordings of telephone interviews with over 750 speakers from across the US and parts of Canada culminated in the publication of the Atlas of North American English (Labov, Ash and Boberg 2006) showing the geographical distribution of a wide range of the changes addressed in Labov (1994). Labov et al.’s examination of the maps that resulted from their analyses of the Telsur data led to similar conclusions as those based on the English Fens data and the evidence of supralocalization discussed above: regional dialect boundaries from the past (for example from much earlier dialect atlas work, often concentrating on lexical items rather than phonological variation) had often persisted to the present day and where the modern boundaries were different this was often a result of the most recent changes in the speech community. Labov, Ash and Boberg argued that “regional dialects are becoming increasingly differentiated from each other […] within most of the regional boundaries, linguistic changes in progress have the effect of solidifying and developing the regional pattern. Many local dialects are indeed disappearing, but they are assimilating to larger regional patterns rather than to a national or international model” (2006: 119). The cartographic work carried out by Labov and his associates is exemplary, but it is so because it is grounded in a thorough theoretical understanding of the nature of phonological change in the US based on decades of careful empirical investigation and on sophisticated acoustic analysis of large corpora of carefully collected data.

5. Concluding remarks All in all, variationism has not fully or adequately engaged yet with socially rich theorizations of geographical space. It has tended both to fetishize some types of space (e.g., urban spaces as opposed to rural ones) as well as, often, to divorce space from social structure and language from the geographical “becoming” of the speech community. Interestingly, differences in academic focus have meant that it has been hard to contrast, for example, whether the many parts of the English speaking world are all undergoing similar types of geolinguistic change in the early twenty-first century. The recent focus in British variationism on consonantal change has led to an academic focus on diffusion, leveling and convergence, whereas the North American focus on vowels has led to a broader discussion of differentiation and divergence. Yet, as we have seen, both appear to be undergoing supralocalization and thereby reinforcing regional dialect areas. The examination of mergers, splits and chain shifts in North America, on the other hand, has not been particularly successfully applied to changes currently afoot in British English, where in the South, researchers have found contact-based approaches more fruitful to explain recent phonological developments (see, for example, Torgersen and Kerswill 2004; Fox 2007) and where in the North, changes underway have been explicitly excluded from Labov’s (1994) typological focus on vocalic chain shifts. But a welcome critical consciousness about the role of space is beginning to enthuse scholars of language variation and change, and this volume, as well as others (e.g., Lameli, Kehrein and Rabanus to appear) are symptomatic of a blossoming of interest in the language-space interface that extends well beyond variationist sociolinguistics.

9. The variationist approach

6. Reerences Allen, John, Doreen Massey and Allan Cochrane 1998 Rethinking the Region. London: Routledge. Bailey, Guy, Tom Wikle, Jan Tillery and Lori Sand 1993 Some patterns of linguistic diffusion. Language Variation and Change 3: 359⫺390. Bell, Allan 1984 Language style as audience design. Language in Society 13: 145⫺204. Blaikie, Piers 1978 The theory of the spatial diffusion of innovations: A spacious cul-de-sac. Progress in Human Geography 2: 268⫺295. Boberg, Charles 2000 Geolinguistic diffusion and the U.S.-Canada border. Language Variation and Change 12: 1⫺24. Britain, David 1991 Dialect and space: A geolinguistic study of speech variables in the Fens. Unpublished PhD dissertation, Department of Language and Linguistics, University of Essex. Britain, David 1997 Dialect contact and phonological reallocation: “Canadian Raising” in the English Fens. Language in Society 26: 15⫺46. Britain, David 2001 Welcome to East Anglia! Two major dialect “boundaries” in the Fens. In: Jacek Fisiak and Peter Trudgill (eds.), East Anglian English, 217⫺242. Woodbridge: Boydell and Brewer. Britain, David 2002 Space and spatial diffusion. In: Chambers, Trudgill and Schilling-Estes (eds.), 603⫺637. Britain, David 2004 Geolinguistics and linguistic diffusion. In: Ulrich Ammon, Norbert Dittmar, Klaus Mattheier and Peter Trudgill (eds.), Sociolinguistics: International Handbook of the Science of Language and Society, 34⫺8. (⫽ Handbücher zur Sprach- und Kommunikationswissenschaft 3.1.) Berlin: Mouton de Gruyter. Britain, David 2005 Innovation diffusion, “Estuary English” and local dialect differentiation: The survival of Fenland Englishes. Linguistics 43: 995⫺1022. Britain, David 2006 Language/Dialect contact. In: Keith Brown (ed.), Encyclopaedia of Language and Linguistics, vol. 6, 651⫺656. 2nd ed. Oxford: Elsevier. Britain, David 2009 One foot in the grave? Dialect death, dialect contact and dialect birth in England. International Journal of the Sociology of Language 196/197: 121⫺155. Britain, David to appear a “Big bright lights” versus “green and pleasant land”? The unhelpful dichotomy of ‘urban’ versus ‘rural’ in dialectology. In: Enam Al-Wer and Rudolf De Jong (eds.), Arabic Dialectology. Leiden: Brill. Britain, David to appear b Conceptualizations of geographic space in linguistics. In: Alfred Lameli, Roland Kehrein and Stefan Rabanus (eds.), Language and Space: An International Handbook of Linguistic Variation, vol. 2: Language Mapping. Berlin: Mouton de Gruyter. Brown, Michael 1995 Ironies of distance: An ongoing critique of the geographies of AIDS. Environment and Planning D: Society and Space 13: 159⫺183.

157

158

II. Linguistic approaches to space Callary, Robert 1975 Phonological change and the development of an urban dialect in Illinois. Language in Society 4: 155⫺170. Castells, Manuel 1977 The Urban Question: A Marxist Approach. London: Arnold. Chambers, Jack K. 1982 Geolinguistics of a variable rule. Discussion Papers in Geolinguistics 5: 1⫺18. Chambers, Jack K. 2003 Sociolinguistic Theory. 2nd ed. Oxford: Blackwell. Chambers, Jack K. and Peter Trudgill 1980 Dialectology. 1st ed. Cambridge: Cambridge University Press. Chambers, Jack K. and Peter Trudgill 1998 Dialectology. 2nd ed. Cambridge: Cambridge University Press. Chambers, Jack K., Peter Trudgill and Natalie Schilling-Estes (eds.) 2002 The Handbook of Language Variation and Change. Oxford: Blackwell. Cheshire, Jenny 2002 Sex and gender in variationist research. In: Chambers, Trudgill and Schilling-Estes (eds.), 423⫺443. Cheshire, Jenny, Sue Fox and David Britain 2007 Relatives from the South. Paper presented at UK Language and Variation Change 6, Lancaster University, September 2007. Christaller, Walter 1933 Die zentralen Orte in Süddeutschland. Jena: Gustav Fischer. Cloke, Paul 1999 The country. In: Paul Cloke, Philip Crang and Marc Goodwin (eds.), Introducing Human Geographies, 256⫺267. London: Arnold. Cloke, Paul 2005 Conceptualising rurality. In: Paul Cloke, Terry Marsden and Patrick Mooney (eds.), Handbook of Rural Studies, 18⫺29. London: Sage. Cloke, Paul and Jo Little (eds.) 1997 Contested Countryside Cultures. London: Routledge. Cochrane, Allan 1987 What a difference the place makes: The new structuralism of locality. Antipode 19: 354⫺363. Cooke, Philip 1989a The contested terrain of locality studies. Tijdschrift voor economische en sociale geografie 80: 14⫺29. Cooke, Philip 1989b Localities: The Changing Face of Urban Britain. London: Unwin Hyman. Coupland, Nikolas 2007 Style: Language Variation and Identity. Cambridge: Cambridge University Press. Darby, Henry 1931 The role of the Fenland in English history. Unpublished PhD dissertation, Department of Geography, St Catherine’s College, Cambridge University. Duncan, Simon 1989 Uneven development and the difference that space makes. Geoforum 20: 131⫺139. Eckert, Penelope 1990 The whole woman: Sex and gender differences in variation. Language Variation and Change 1: 245⫺267. Eckert, Penelope 1997 Age as a sociolinguistic variable. In: Florian Coulmas (ed.), The Handbook of Sociolinguistics, 151⫺167. Oxford: Blackwell.

9. The variationist approach Eckert, Penelope 2000 Linguistic Variation as Social Practice. Oxford: Blackwell. Eckert, Penelope and John Rickford (eds.) 2001 Style and Sociolinguistic Variation. Cambridge: Cambridge University Press. Ellis, Alexander 1889 On Early English Pronunciation, vol. 5. London: Truebner and Co. Foulkes, Paul and Gerard Docherty (eds.) 1999 Urban Voices: Accent Studies in the British Isles. London: Arnold. Fox, Sue 2007 The demise of Cockneys? Language change in London’s “traditional” East End. Unpublished PhD dissertation, Department of Language and Linguistics, University of Essex. Gerritsen, Marinel 1988 Sociolinguistic developments as a diffusion process. In: Ulrich Ammon, Norbert Dittmar, Klaus Mattheier and Peter Trudgill (eds.), Sociolinguistics: An International Handbook of the Science of Language and Society, vol. 3.2., 1574⫺1591. (⫽ Handbücher zur Sprachund Kommunikationswissenschaft 3.2.) Berlin: de Gruyter. Gerritsen, Marinel and Frank Jansen 1980 The interplay of dialectology and historical linguistics: Some refinements of Trudgill’s formula. In: Peter Maher, Allan R. Bomhard and Ernst F. K. Koerner (eds.), Papers from the Third International Conference on Historical Linguistics, 11⫺38. Amsterdam: Benjamins. Giddens, Anthony 1984 The Constitution of Society: Outline of the Theory of Structuration. Cambridge: Polity Press. Grainger, Heather 2005 Is the Wirral really Scouse? The use of local and supralocal forms in East Wirral. Unpublished MA dissertation, Department of Language and Linguistics, University of Essex. Gregory, Derek 1985 Suspended animation: The stasis of diffusion theory. In: Gregory and Urry (eds.), 269⫺336. Gregory, Derek 2000 Diffusion. In: Ron Johnston, Derek Gregory, Geraldine Pratt and Michael Watts (eds.), The Dictionary of Human Geography, 175⫺178. 4th ed. Oxford: Blackwell. Gregory, Derek and John Urry (eds.) 1985 Social Relations and Spatial Structures. London: Macmillan. Hägerstrand, Torsten 1952 The Propagation of Innovation Waves. Lund: Gleerup. Harvey, David 1973 Social Justice and the City. London: Arnold. Herna´ndez Campoy, Juan Manuel 2003 Exposure to contact and the geographical adoption of standard features: Two complementary approaches. Language in Society 32: 227⫺255. Honeybone, Patrick 2001 Lenition inhibition in Liverpool English. English Language and Linguistics 5: 213⫺249. Horvath, Barbara and Ronald J. Horvath 1997 The geolinguistics of a sound change in progress: /l/ vocalisation in Australia. University of Pennsylvania Working Papers in Linguistics 4(1): 109⫺124. [Special issue: Charles Boberg, Miriam Meyerhoff and Stephanie Strassel (eds.), A Selection of Papers from NWAVE 25.] Horvath, Barbara and Ronald J. Horvath 2001 A multilocality study of a sound change in progress: The case of /l/ vocalization in New Zealand and Australian English. Language Variation and Change 13: 37⫺57.

159

160

II. Linguistic approaches to space Horvath, Barbara and Ronald J. Horvath 2002 The geolinguistics of /l/ vocalisation in Australia and New Zealand. Journal of Sociolinguistics 6: 319⫺346. Johnston, Ron 1991 A Question of Place: Exploring the Practice of Human Geography. Oxford: Blackwell. Johnston, Ron 2000 Urban geography. In: Ron Johnston, Derek Gregory, Geraldine Pratt and Michael Watts (eds.), The Dictionary of Human Geography, 875⫺878. 4th ed. Oxford: Blackwell. Johnstone, Barbara, Jennifer Andrus and Andrew Danielson 2006 Mobility, indexicality, and the enregisterment of “Pittsburghese”. Journal of English Linguistics 34: 77⫺104. Kerswill, Paul 2003 Dialect levelling and geographical diffusion in British English. In: David Britain and Jenny Cheshire (eds.), Social Dialectology: In Honour of Peter Trudgill, 223⫺244. Amsterdam: Benjamins. Knowles, Gerry 1973 Scouse: The urban dialect of Liverpool. Unpublished PhD dissertation, Department of Linguistics, University of Leeds. Labov, William 1966 The Social Stratification of English in New York City. Washington, DC: Center for Applied Linguistics. Labov, William 1972 Sociolinguistic Patterns. Oxford: Blackwell. Labov, William 1982 Building on empirical foundations. In: Winfred Lehmann and Yakov Malkiel (eds.), Perspectives in Historical Linguistics, 79⫺92. Amsterdam: Benjamins. Labov, William 1994 Principles of Linguistic Change, vol. 1: Internal factors. Oxford: Blackwell. Labov, William 2003 Pursuing the cascade model. In: David Britain and Jenny Cheshire (eds.), Social Dialectology: In Honour of Peter Trudgill, 9⫺22. Amsterdam: Benjamins. Labov, William 2006 The Social Stratification of English in New York City. 2nd ed. Cambridge: Cambridge University Press. Labov, William, Sharon Ash and Charles Boberg 2006 Atlas of North American English: Phonetics, Phonology, and Sound Change. New York: Mouton de Gruyter. Lameli, Alfred, Roland Kehrein and Stefan Rabanus (eds.) to appear Language and Space: An International Handbook of Linguistic Variation, vol. 2: Language Mapping (HSK 30.2). Berlin: Mouton de Gruyter. Macnaghten, Phil and John Urry 1998 Contested Natures. London: Sage. Maidment, John 1994 Estuary English: Hybrid or Hype? Paper presented at the 4th New Zealand Conference on Language and Society, Lincoln University, Christchurch, New Zealand. Massey, Doreen 1984 Introduction: Geography matters. In: Doreen Massey and John Allen (eds.), Geography matters!, 1⫺11. Cambridge: Cambridge University Press. Massey, Doreen 1985 New directions in space. In: Gregory and Urry (eds.), 9⫺19. Mathisen, Anne Grethe 1999 Sandwell, West Midlands: Ambiguous perspectives on gender patterns and models of change. In: Foulkes and Docherty (eds.), 107⫺123.

9. The variationist approach Mees, Inger and Beverley Collins 1999 Cardiff: A real-time study of glottalisation. In: Foulkes and Docherty (eds.), 185⫺202. Meyerhoff, Miriam 1996 Dealing with gender identity as a sociolinguistic variable. In: Victoria Bergvall, Janet Bing and Alice Freed (eds.), Language and Gender Research: Rethinking Theory and Practice, 225⫺252. London: Longman. Milroy, James, Lesley Milroy and Sue Hartley 1994 Local and supra-local change in British English: The case of glottalisation. English WorldWide 15(1): 1⫺33. Milroy, James, Lesley Milroy, Sue Hartley and David Walshaw 1994 Glottal stops and Tyneside glottalisation: Competing patterns of variation and change in British English. Language Variation and Change 6: 327⫺357. Milroy, Lesley 1999 Women as innovators and norm-creators: The sociolinguistics of dialect levelling in a northern English city. In: Suzanne Wertheim, Ashlee C. Bailey and Monica CorstonOliver (eds.), Engendering Communication: Proceedings of the 5th Berkeley Women and Language Conference, 361⫺376. Berkeley: BWLG Publications. Parsons, Gudrun 1998 From “RP” to “Estuary English”: The concept “received” and the debate about British pronunciation standards. Unpublished MA dissertation, Department of Linguistics, University of Hamburg. Peys, Ste´phanie 2001 In pursuit of Cockney speak. Unpublished MA dissertation. Department of Linguistics, Universite´ de Pau et des pays de l’Adour. Piercy, Caroline 2007 “We don’t say oo /ar/”: Diffusion of Phonological Change in Dorset English. Research on languages and linguistics at Sussex (ROLLS) seminar series, University of Sussex, 12 November 2007. Pred, Allan 1985 The social becomes the spatial, the spatial becomes the social: Enclosures, social change and the becoming of places in the Swedish province of Ska˚ne. In: Gregory and Urry (eds.), 337⫺365. Przedlacka, Joanna 2001 Estuary English and RP: Some recent findings. Studia Anglica Posnaniensia 36: 35⫺50. Przedlacka, Joanna 2002 Estuary English? A Sociophonetic Study of Teenage Speech in the Home Counties. Bern: Peter Lang. Ryfa, Joanna to appear a A sociolinguistic investigation of the New Town dialect of Redditch. Unpublished PhD dissertation, Department of Language and Linguistics, University of Essex. Ryfa, Joanna to appear b Englishes of London and the South East of England. In: Tometro Hopkins, John McKenny and Kendall Decker (eds.), World Englishes. London: Continuum. Sangster, Catherine 2001 Lenition of alveolar stops in Liverpool English. Journal of Sociolinguistics 5: 401⫺412. Savage, Mike, James Barlow, Simon Duncan and Peter Saunders 1987 “Locality Research”: The Sussex programme on economic restructuring, social change and the locality. Quarterly Journal of Social Affairs 3: 27⫺51. Sayer, Andrew 1985 The difference that space makes. In: Gregory and Urry (eds.), 49⫺66. Sayers, Dave 2009 Reversing Babel: Declining linguistic diversity in late modern societies. Unpublished PhD dissertation, Department of Language and Linguistics, University of Essex.

161

162

II. Linguistic approaches to space Schilling-Estes, Natalie and Walt Wolfram 1999 Alternative models of dialect death: Dissipation vs. concentration. Language 75: 486⫺ 521. Short, Brian 2006 Idyllic ruralities. In: Paul Cloke, Terry Marsden and Patrick Mooney (eds.), Handbook of Rural Studies, 133⫺148. London: Sage. Shucksmith, Mark 2000 Exclusive countryside? Social Inclusion and Regeneration in Rural Britain. York: Joseph Rowntree Foundation. Soja, Edward W. 1989 Postmodern Geographies: The Reassertion of Space in Critical Social Theory. New York: Verso. Torgersen, Eivind and Paul Kerswill 2004 Internal and external motivation in phonetic change: Dialect levelling outcomes for an English vowel shift. Journal of Sociolinguistics 8: 24⫺53. Trudgill, Peter 1974a The Social Differentiation of English in Norwich. Cambridge: Cambridge University Press. Trudgill, Peter 1974b Linguistic change and diffusion: Description and explanation in sociolinguistic dialect geography. Language in Society 3: 215⫺246. Trudgill, Peter 1975 Linguistic geography and geographical linguistics. In: Christopher Board, Richard J. Chorley, Peter Haggett and David R. Stoddart (eds.), Progress in Geography, 227⫺252. London: Arnold. Trudgill, Peter 1983 On Dialect: Social and Geographical Perspectives. Oxford: Blackwell. Trudgill, Peter 1986 Dialects in Contact. Oxford: Blackwell. Walters, Keith 1988 Dialectology. In: Frederick J. Newmeyer (ed.), Language: The Socio-cultural Context, 119⫺139. (Linguistics: The Cambridge Survey 4.) Cambridge: Cambridge University Press. Watson, Kevin 2006 Phonological resistance and innovation in the North-West of England. English Today 22(2): 55⫺61. Watson, Kevin 2007 Liverpool English. Journal of the International Phonetic Association 37: 351⫺360. Watt, Dominic 2002 “I don’t speak with a Geordie accent, I speak, like, the Northern accent”: Contact-induced levelling in the Tyneside vowel system. Journal of Sociolinguistics 6: 44⫺63. Watt, Dominic and Lesley Milroy 1999 Patterns of variation and change in three Tyneside vowels: Is this dialect levelling? In: Foulkes and Docherty (eds.), 25⫺46. Watts, Emma 2006 Mobility-induced dialect contact: A sociolinguistic investigation of speech variation in Wilmslow, Cheshire. Unpublished PhD dissertation, Department of Language and Lingustics, University of Essex. Weinreich, Uriel, William Labov and Marvin Herzog 1968 Empirical foundations for a theory of language change. In: Winfried Lehmann and Yakov Malkiel (eds.), Directions for Historical Linguistics, 97⫺195. Austin: University of Texas Press.

10. Social anthropology and interactional sociolinguistics Wells, John Christopher 1982 Accents of English: The British Isles, vol. 2. Cambridge: Cambridge University Press. Wolfram, Walt 1997 Issues in dialect obsolescence: An introduction. American Speech 73: 1⫺12. Wolfram, Walt 2002 Language death and dying. In: Chambers, Trudgill and Schilling-Estes (eds.), 764⫺787. Wolfram, Walt and Donna Christian 1974 Appalachian Speech. Washington, DC.: Center for Applied Linguistics. Wolfram, Walt and Natalie Schilling-Estes 1995 Moribund dialects and the language endangerment canon: The case of the Ocracoke Brogue. Language 71: 696⫺721. Wolfram, Walt and Natalie Schilling-Estes 1998 American English. Oxford: Blackwell.

David Britain, Essex (Great Britain)

10. Whos there? Language and space in social anthropology and interactional sociolinguistics 1. 2. 3. 4. 5. 6.

Introduction The place of ethnography in the study of variation The nature of meaning in variation Space and the construction of meaning Conclusion References

1. Introduction A theory of variation has to include a theory of meaning, for whatever its role in linguistic change and whatever it can tell us about grammatical structure, variation constitutes a robust social indexical system. Space is central to the study of social indexicality, because space is imbued with social meaning, and the distribution of linguistic forms across space is key to the construction of meaning in variation. Social practice is located in space, it constructs space, and the social itself is a metaphorical space. The meaning of sociolinguistic variation invokes geography at all levels from the neighborhood hangout to the international and the transnational, and social space from the family to imagined communities. Understanding the meaning of variation, therefore, requires that we engage with space in all these senses. Focusing on the ethnographic study of variation, the following pages will emphasize that dialectological, survey and ethnographic approaches to the study of variation provide complementary information about the social life of sociolinguistic variation, and that we can only understand the social value of

163

164

II. Linguistic approaches to space variation ⫺ its meaning ⫺ if we combine all three. But it is only at the ethnographic level that we can gain access to meaning in variation as it is constructed in day-to-day social practice, in other words, as it functions in speakers’ sociolinguistic competence.

2. The place o ethnography in the study o variation An ethnographic study of variation allows us to focus on quite specific people, observing how they deploy variables in practice across situations, over time, and in groups whose dynamics we understand. It also allows us to talk to speakers repeatedly and to be present when they talk with others. We can learn where they go and whom they interact with, whom they would like and not like to interact with, how they view the world around them and who they see as occupying that world. We can learn how people view language in the world from their local perspective ⫺ what kinds of social distinctions they make and how they connect those distinctions to linguistic differences. And while an ethnographic study might be located in a community that we define in advance, these people will eventually lead us beyond the boundaries that we’ve attributed to it. In this process, we learn that social space is never circumscribed by a fixed physical space, but extends into adjacent, distant and even imagined spaces, and the meaning of variation is constructed across them all. The study of variation has ranged from a broad survey approach examining distributions of variables over large populations, to ethnographic studies focusing on the use of variation in face-to-face communities. The survey approach examines the differential use of variables among demographic categories (particularly, class, age, gender, ethnicity), which provide a kind of abstract map of social space (cf. Mæhlum in this volume). Ethnographic studies, on the other hand, examine the local categories and practices that underlie the demographic patterns. Class, age, gender and ethnicity are broad categories that structure societies, but they are experienced on the ground in more local ways. When we find, for example, that women overall use more standard grammar than men, we mean that there is a general, repeated, and replicable statistical result by which women, in the aggregate, use standard grammar more than men. This says nothing about why this pattern exists, what kinds of behaviors and ideologies underlie these patterns, what kind of meaning people attach to standard forms, who does and doesn’t fit the pattern and why. It says nothing about the relation between gender as a social phenomenon and language use, and it says nothing about why the same generalization applies to class stratification ⫺ not only women, but middle class people, use more standard grammar than men and working class people. “Be a man” and “sit like a lady” invoke the global gender binary of male and female, but they do not have the same significance everywhere. Being a man or sitting like a lady is part of an individual and local history of admonishments, invoking not gender globally, but a local construction of gender (and class and age and ethnicity), just as “hold your fork properly” is part of a local construction of class (and gender and age and ethnicity). In other words, these demographic categories are abstractions, and they interact intensely, so that while the global patterns of correlation of linguistic variables with these categories provide a social map, they do not provide explanations for the shape of that map. Explanations for global patterns lie in the relations among the myriad local practices that add up to produce them. How is

10. Social anthropology and interactional sociolinguistics gender or class constructed in a particular place at a particular time? How do these combine to create the global pattern ⫺ and what of those who do not conform to this pattern? The large-scale distributions of survey studies are the outcome of differential behavior of speakers living their daily lives in very different places in society, and those daily lives, in turn, are structured by their location in the larger political economy. Thus global patterns and local practices are not simply complementary, but are in a mutually reproductive relation (Bourdieu 1977a). Class, gender, age and ethnicity are categories that structure social life on the ground. They constrain the kinds of communities of practice (Eckert and McConnell-Ginet 1992) that people participate in, and the practices in these communities in turn shape the local construction of the categories. Gendered divisions of labor, childcare practices and leisure activities are constrained by the material conditions that constitute class. This may involve the formation of particular types of social networks, particular kinds of communities of practice, particular ways of doing things, particular needs, desires and emotional lives. And these in turn reproduce people’s place in the social order and the social order itself. The construction of meaning in variation is part of people’s participation in these networks and communities of practice, hence of the process of social reproduction. Variation is part of speakers’ active participation in, and construction of, the social world and of the construction of selves in that world. Much of current theorizing about variation is still based on the demographic categories of survey studies, with some serious consequences. The demographic categories used in these studies are so embedded in social science that they are commonly taken as preordained ⫺ as pre-existing the variation that correlates with them. The view of variation as reflecting a pre-existing and relatively static social order has a variety of implications that require examination. First, it yields a view of variation as a reflection or an expression of social address, and leads to a view of the acquisition of variable competence as a simple learning of the (adult) social structure and of the variables that go with that structure. It views change as spreading across the social map, and social analysis has focused on who the change reaches when, or who adopts or resists the change. The ethnographic view reminds us that changes in progress constitute a very small subset of the forms that make up the system of social variation. And most variables ⫺ even changes in progress ⫺ are always present in the community, to be taken up or not, by all speakers. Social stratification of the variables, then, it is due to selective use, not to a spread of accessibility through the socioeconomic hierarchy. Finally, it views variation as resulting from, and marking, demographic categories. The move from correlation to the notion of marking skirts the issue of meaning as linguists label certain variables as “local” or “working class” or “vernacular” or “female-led changes”. But clearly, people ⫺ even women ⫺ are not saying “I’m a woman” when they use a “female-led” change. Rather, that variable means something that is in turn related to gender. The very fact that the same variables may stratify regularly with multiple categories ⫺ e.g., gender, ethnicity and class ⫺ indicates that their meanings are not directly related to these categories, but to something that’s related to all of them (Ochs 1991). The demographic patterns must be understood in a broader perspective ⫺ one that encompasses the social agency (not to be confused with intentionality) of speakers as they use variation in day to day practice. The reactive view of variation ignores the role of variation in the actual production of social categories. Social categories can exist only inasmuch as they can be distinguished, and linguistic differentiation is a fundamental

165

166

II. Linguistic approaches to space way of creating distinctiveness. Variation, then, emerges at the nexus of structure and change, as integral to social practice rather than as a simple output of structure. The explanation for much of what we see in variation lies in the meanings that are associated with variables ⫺ in how those meanings come about and how they change.

3. The nature o meaning in variation We can assume that the meaning of variation is indexical ⫺ that is, variables gain their value contextually, specifically through association with the categories of speakers who use them. In its simplest sense, indeed the sense emanating from the survey tradition, the use of a variable “points to”, or indexes, a category of speaker. But this indexicality is not fixed; on the contrary, much of its value lies in its indeterminacy. The association of a variable with a category of speakers may be taken, for example, to pick out some characteristic of those speakers, yielding an additional indexical value in a growing set of associations ⫺ an indexical order (Silverstein 2003). Thus, for example, if working class men are viewed as tough, a variable that is primarily used by working class men (first order index) may later become associated with toughness (second order index). This variable, then, could become available to other categories of people to index toughness. It is crucial to emphasize that at this point, the original association loses its primacy in the indexical system, so the status of first order index is purely historical. Indeed, we have no way of knowing in most cases how the many values that an index may have came about. But a variable can take on a variety of meanings in this way, yielding what one might call an indexical field. Which of the potential meanings is invoked by the use of a variable depends on the context and on the style in which it is embedded (I will return to this below). A linguistic form (in this case, a variable) becomes an index when speakers come to distinguish it from other forms and to associate it with some social distinction. The social distinction is reinforced ⫺ perhaps created ⫺ in this association. A social distinction does not have to be just a category of person, but can be a mood, a stance, which may or may not be associated with a particular category of person. Negative concord (I didn’t do nothing) is socially stratified throughout the English speaking world. It is associated with working class status and with lack of education, and one cannot say that one of these meanings is primary. Indeed, class and education are inseparable, along with differences in occupations, leisure activities and consumption patterns. And negative concord is inseparable from other patterns deemed “vernacular” which may or may not differ across the English speaking world. But in every place, the linguistic and the social distinction go hand in hand. The production of distinction has been described in detail by Susan Gal and Judith Irvine (Irvine and Gal 2000), as involving three semiotic processes: erasure, recursivity and iconicity. The distinction between a category of people seen as working class and as middle class involves (1) a selective ignoring of similarities between these two groups of people and of differences within the groups ⫺ both social and linguistic (erasure); (2) an embedding of multiple distinctions in this opposition ⫺ other linguistic differences as well as, for example, consumption patterns, dress, education (recursivity); and (3) the forging of an apparently natural connection between the two, for instance viewing both negative concord and the working class mind as illogical

10. Social anthropology and interactional sociolinguistics (iconicity). This process is central to the workings of variation, and is a process that speakers engage in as part of moving around in the world. Presumably this process begins very early in development ⫺ infants differ their voice pitch depending on their interlocutor, and children certainly notice early on the linguistic cues for parental affection, reprimand, anger and joy. And many of the cues to these emotions function in broader patterns of sociolinguistic variation ⫺ sounds of adult power in the home are heard from the mouths of teachers in school, thus moving from the personal to the more institutional and public. Development is, among other things, a gradual expansion from the most intimate to the most impersonal and public. And this expansion involves a continual interpretation of the world around, of the types of people that inhabit that world, and of the ways of speaking that go with different places and people. It makes sense to assume that the analytic abilities that allow the small child to construct a grammar from the language he or she hears are at work throughout life as speakers interpret the relation between linguistic variation and social characteristics. This interpretation does not take place at the level of the individual person or the individual variable, however, but at the level of style. Style has traditionally been treated in variation as a continuum of formality ⫺ a link that emerged from a methodological focus on the effects of speakers’ monitoring their own speech. Labov (1972) reasoned that a speaker’s most ingrained and automatic patterns emerge when they are paying little attention to their speech, and that attention to speech leads prestige norms to interrupt these patterns. He defined style, therefore, as a continuum of formality-informality ⫺ a continuum that to some extent replicates the socioeconomic stratification of features within the speech of individuals. This approach to style is uniquely related to Labov’s desire to elicit automatic speech (the vernacular). In every other field that deals with style, style is a socially distinctive system. In all areas of art, style is what constitutes schools, periods and individuals. Style has this function in everyday language use as well, and this kind of style, which is orthogonal to the formality continuum, is the best level for approaching the meaning of variation. The view of style as audience design (Bell 1984), based in studies of speech accommodation, indeed focuses on the socially distinctive view of style. In this case, though, the focus is on individuals’ style shifting as they speak for different (present or imagined (Bell 2001)) audiences. It’s one step further to extend our view of style beyond accommodation to speakers as defined by dialects, to a view of style as constituting a far more complex persona ⫺ a persona that is not only from Boston, female, middle class, white ⫺ but that is sexy, flirtatious, politically conservative and ⫺ at this particular moment ⫺ angry at her kid. For not all variables are related to geographic space or social address, but they can be related to things as fleeting as attitude, mood, emotions, sexual desire. These in turn, used more regularly, can become personal qualities, along with such things as political orientation, religiosity, sluttiness, independence, comicalness. Stylistic practice, in other words, is the production of persona. And in stylistic practice we identify, segment, produce and reproduce personae.

4. Space and the construction o meaning As discussed at length in this volume and elsewhere (Johnstone 2004), physical space is imbued with social meaning. My use of the term physical space here only distinguishes it from the purely metaphorical term social space when referring to “place” in the social

167

168

II. Linguistic approaches to space order. Residential patterns, local territories, rural and urban spaces, regions and nations all involve the physical placement of distinctive populations and linguistic varieties. And this placement in turn plays a role in reproducing distinctiveness. Social differences in a single locality can invoke space at all levels ⫺ from very local territories to the transnational, as people move about in, and orient differently to, the world around them. As they do so, they encounter and adopt (or not) new linguistic features. Spaces also serve as soundscapes (Inoue 2003). People circulate in spaces with quite distinctive sounds and come to associate those sounds with their understanding of what those spaces are about. Richard Bauman, in a study of market calls in Mexico (Bauman 2001), points out that children walking through the market will hear stylistic innovations that vendors use to elaborate their calls, and may associate these innovations with their impressions about the market scene. A child growing up on the upper west side of New York City will hear different things all day long than a child living even just across town. Peasants in the French Pyrenees used to frequent one of two restaurants in town when they went to market. The noisiness of this “peasant” restaurant stood in stark contrast to the other, quiet, bourgeois restaurant. This quiet, peasants told me, was highly suspicious ⫺ a sign that the occupants had something to hide. The noisiness of the peasant restaurant, meanwhile, connects to life back in the village, where one hears loud voices continually, as people call to their animals and to each other as they pass in open spaces, or as they gather in the village cafe´. It is also related to the public nature of life in the intimate social milieu of the village, where secrets cannot last. Here, the doors of homes are open for the visitor to step in and shout out, as opposed to the closed (and locked) doors of bourgeois as they talk quietly in their homes in town. But speakers are not just passive hearers and potential recipients of spatially distributed styles ⫺ they are also purveyors (or hoarders) of style. Pierre Bourdieu (1977b, 1982) defines linguistic competence as not simply the ability to produce grammatical utterances, but the ability to have those utterances heard and the ideas in them taken up. If we view space from this perspective, we can see the clear relation between language, space and power. Speakers who dominate central visible spaces, including particularly the media, are in a position to have their utterances heard. And the central venue lends weight to those utterances and to the variety in which they are made. Since a linguistic variety is closely related to social groups, the construction of a distinctive variety highlights that group, thus a group wishing to achieve visibility and ascendance will also want to develop and display the variety. (Concomitantly, a group that seeks invisibility may want to enhance a variety for purposes of their own solidarity, but keep it out of the public eye.) The opportunity that dominating a soundscape gives for making one’s own language heard emerges clearly in schools. An ethnographic study of two elementary schools in Northern California (Eckert 1996) followed an age cohort as it approached adolescence. Space emerged as an important commodity as the cohort transformed itself into a peerbased social order. This process, which takes place in elementary schools across the US (and no doubt beyond) requires the appropriation and distribution of adult power into the age group. Transcendence of the teacher-dominated classroom and control of central school space is fundamental to this process, and the establishment of a peer hierarchy is closely tied to the reorganization of space. A heterosocial “popular crowd” emerges in late primary school, as certain small friendship groups enter into alliances, creating a significant mass of collaborating boys

10. Social anthropology and interactional sociolinguistics and girls. It is the crowd’s express purpose to become a social force and to gain power and status by leading their age cohort into adolescence. Central to their joint practice is the construction of a heterosexual market, as the crowd collaborates in pairing up individual boys and girls as “couples”. The crowd’s socially “advanced” activities make them highly visible, and they enhance this visibility by controlling central spaces as stages for their activities. Their games, negotiations, dramas and stylistic performances take up the center of the playground while their peers can watch from the periphery. As their alliances bring together kids from multiple classrooms, they transcend the space of the classroom and the adult domination symbolized by that space. Their behavior in their own classrooms underlines their spatial dominance as they are scattered throughout the class, dominating space with their looks, gestures, note-passing and coordinated performances. This visibility allows them to achieve symbolic dominance, as their public performances highlight stylistic innovations. These kids lead their cohort in the use of flamboyant styles, in which the California Vowel Shift is accelerated. Their language use defines “cool” ⫺ from their use of new speech activities and speech acts (e.g., friendship and heterosexual drama and compliments) to their use of phonological variables.

4.1. Communities and local identity People give meaning to space, from the back yard to gang territories to meeting places, malls, towns, regions, continents and countries as part of making sense of themselves, each other and the world. The meaning is based on their perceptions of what goes on in these spaces and of the people that they see as inhabiting and claiming this landscape. The defined speech communities of variation studies are typically analytic strategies ⫺ “convenient fictions” that allow the linguist to delimit a population. These communities are also, one hopes, selected as sufficiently salient to the population to explain the variation of the speech within. The very notion of speech community seems to be designed to define an inward-looking linguistic unit (Pratt 1988). This is not to say that variation studies study speech communities in complete isolation from the outside, but that they generally treat the outside as undifferentiated. While variation studies have located communities with respect to regions and conurbations, the focus is generally on variation within the community rather than the relation between this variation and the outside and commonly the identity of this community is taken for granted. The notion of “local identity” is often invoked, but only vaguely with respect to what constitutes the local. Is it the neighborhood? The town? A particular group of people living in some specific place? The very notion of local can exist only in a discursive space that constructs the “other”, and social practice in any community constructs the borders and the outside. How people situate themselves in their own communities, as individuals and as groups, is inseparable from how they situate themselves with respect to the rest of the world. A study of social variation in space is best done, therefore, by suspending the boundaries of local communities, and focusing on how different speakers use space and how they view space ⫺ from the closest to the most remote. The first quantitative ethnographic study of variation, William Labov study of Martha’s Vineyard (1963), was very much about local identity. This study showed variation to be a resource for the construction of quite specifically local meaning, but meaning that was located in turn within the broader sociogeographic context. At the time of

169

170

II. Linguistic approaches to space Labov’s study, Martha’s Vineyard was facing a growing incursion from the mainland US as the tourist economy and the purchase of summer and even year-round homes by mainlanders threatened many islanders’ sense of history and autonomy. Labov related the pronunciation of /ay/ to differing ideologies with respect to mainland incursion in the island economy. A specifically island feature, the centralization of the nucleus of /ay/ appeared most frequently in the speech of categories of people who were associated with the long-standing local English fishing community and resisting the growing tourist economy. It was not “island” identity that was associated with centralization, but a specific kind of island identity. And this specific kind emerged in a local contestation of the meaning and the future of the island in opposition to local people who welcomed the opportunities that come with the tourist industry. In this way, a person’s place in island society was inseparable from his or her orientation to the mainland, and to mainland people and interests on the island. Local identity is not a simple identification with a delimited space, but with a particular, and often contested, construction of that space.

4.2. Moving through local space Another aspect of the local is the way social practice unfolds in local physical space. Local identity has been more widely associated with the working class and the lower middle class ⫺ people that Labov (2000) has labeled “centrally located groups”. The relation between class and the local has been central to the study of social constraints in variation. The repeated socioeconomic patterning of variation shows an inverse relation between class level and the local distinctiveness of speech. Thus working class speakers have the most locally distinct speech while the upper middle class have a speech pattern that transcends the local and even the regional. This corresponds to a variety of ways in which working class people engage with the local, while upper middle class speakers disengage in favor of networks and institutions that transcend the local. Lesley Milroy’s work on social networks (1980) is based on the fact that working class networks are more locally based than middle class networks. Patterns of resource management (jobs, homes, etc.) lead to locally-based networks, whose greater density and multiplexity, Milroy argue, leads to the concentration of local features as speakers encounter local norms across situations and communities of practice. Labov (2000) has argued that leaders in sound change are women with many network ties both in their neighborhood and outside the neighborhood ⫺ in other words, people who move around locally; not people who are locked in some static social and geographic unit. In addition, though, the working class and lower middle class are the people whose work is locally-based. These are the people who are responsible for the local community, whether by maintaining the local infrastructure (e.g., people in the building trades, storekeepers) or by protecting the local community (e.g., police officers, fire fighters). A sense of local ownership, or of cultural citizenship, therefore, is fundamental to working class and lower middle class ideology. The local community, though, is not the neighborhood or the town, but encompasses the surrounding area that provides resources to the immediate locality. This pattern begins well before adulthood. The relation between class and geographic circulation emerged as central in the ethnographic study of adolescents in Detroit suburban high schools (Eckert 1989: 2000). Detroit suburban adolescents operate in a space that is defined above all by race and class, as the predominantly African American pop-

10. Social anthropology and interactional sociolinguistics ulation of the city of Detroit gives way to a predominantly White suburban population. And as one moves outward from the city, the socioeconomic level rises gradually as the area becomes decreasingly urban. Each high school has a socioeconomic character that is related to its place in the suburban continuum, and class merges with urban culture as kids ⫺ and their schools ⫺ closer to the city are identified as poorer, more urban, hence more streetwise, more self-sufficient, tougher. Within each high school, class differences are realized in opposed social categories ⫺ students who organize around alienation from the school (burnouts) and constitute a working class culture, and students who organize around engagement with the school institution (jocks) and constitute a middle class culture. While family socioeconomic status does not completely predict affiliation in one category or the other, the burnouts and jocks come primarily from the lower and upper ranges, respectively, of the local socioeconomic hierarchy. The daily practices of jocks and burnouts bring together orientation to school with an orientation to the conurbation in which the school is located. Set within the gradual difference in socioeconomic status across the suburban area, this yields the kind of recursivity described by Irvine and Gal (2000). An additional recursivity is within the approximately two thirds of the student body who do not identify as jocks or as burnouts, but refer to themselves as in-betweens, and who designate their place in the social order on a continuum between the two. In schools throughout the Detroit conurbation, burnouts gather in marginal spaces ⫺ such as loading docks and parking lots. Jocks, on the other hand, gravitate to, and control, the spaces that are central to school life ⫺ eating and athletic spaces, areas around lockers, front halls and activities offices. This territorial specialization gives each category the space to consolidate, and the proximity of territories enhances the visibility of each category to each other and to the rest of the school population. This mutual exposure facilitates the construction of opposed styles. The burnouts’ marginal territory puts them closer to the outside, making it easier to avoid the adult gaze and to leave school surreptitiously, while the jocks’ territory makes it easier for them to control school resources. The burnouts’ orientation to the outside is not simply a negative orientation to school, but a positive orientation to the community around the school ⫺ and to the entire conurbation. The social networks that make up the burnout community are continuations of age-heterogeneous neighborhood networks in childhood, so that spatial orientation has a deep history imbued with the particular kinds of ties and preoccupations of these networks. And school, with its strict age grading, interrupts and stigmatizes the burnouts’ neighborhood-base networks. Burnouts value continuity and loyalty to childhood friends, while jocks value relationships based in institutional participation. Jocks come often from families in which parents maintain more control over their friendships, arranging play dates in childhood rather than letting them loose in the neighborhood. School offers the first opportunity for the jocks to make their own friends, and their friendship groups remain school-based, and they focus their activities and their identities on the school itself. Their neighborhoods do not function as significant social spaces, and they carry out their social activities in the school and in suburban public spaces, limiting their use of urban space to institutions ⫺ museums, stadiums and malls. The class-based orientations to space, then, are not superficial, but deeply rooted in kids’ ideologies and emotional lives. This fundamental difference also structures their use of the space defined by the wider conurbation. The jocks may establish collegial relations with jocks from other schools if they come in contact through institutional

171

172

II. Linguistic approaches to space activities. But their friendship groups remain located in their own school and are tied to school loyalty. The burnouts’ local orientation extends beyond the neighborhood and town towards the urban center. They claim neighborhood school yards as hangouts, but they also frequent parks closer to the urban center. Burnouts from schools across the suburbs follow set cruising roots ⫺ streets that lead into the city, where they converge in the evening on a few popular strips. Burnouts strive to extend their networks towards the city, looking to the wider conurbation for resources, excitement and opportunities. They view kids coming from the urban end of the conurbation as strong and streetwise ⫺ qualities that they admire. The burnouts’ local and urban orientation is reflected in their significantly greater use of sound changes that are moving outwards from the urban center. The backing of /uh/ and the raising of the nucleus of /ay/ are both robustly more advanced closer to Detroit, and in every school examined in the suburbs, the burnouts use more of both of these variants than the jocks. Furthermore, among the in-betweens, those who place themselves closer to the burnouts ⫺ kids who cruise Detroit ⫺ also use more of these variants than those who do not. The use of urban variables, in other words, indexes qualities associated with urban life. While the burnouts expand into urban space, jocks dominate institutional space. This gives them access to the formal venues that confer legitimacy. Their words are heard by larger audiences, and imbued with institutional gravitas. For the jocks, institutional status is primary, and the more engaged students are with the school ⫺ even within the jock category ⫺ the less they use non-standard grammar and urban phonology. So, for example, the most standard, least urban speakers among the jocks are those that dominate student government as opposed to those who participate primarily in sports.

4.3. The meaning o that place The survey studies that characterized the “first wave” of variation studies (Eckert 2005) were generally studies of regional dialects, as spoken in a major urban center: e.g., New York (Labov 1966), Detroit (Shuy et al. 1967), Norwich (Trudgill 1974), Tehran (Modaressi 1978). In these studies the place of the community in geographic space was relevant only insofar as it located the speech community within a dialect area, thus defining the linguistic features to be studied. The social analysis did not take local identity or the sociogeographic surroundings into consideration, but focused on demographic categories that abstract away from the local inasmuch as they can be found in any community. The speech community itself, then, was incidental ⫺ simply a container for the dialect and for the social hierarchies that structure its use. In this sense, the theory viewed local features, whatever the locality, as equivalent. But there can be little question that to the extent that linguistic features are associated with localities or regions, the reputation of that locality or region must affect their indexical value both at home and elsewhere. The study of the local construction of “Pittsburghese” (Johnstone, Andrus and Danielson 2006) makes it clear that the character of cities has great local salience, which must play a role in the indexical value of particular local variables much as the historic fishing culture was linked to /ay/ centralization on Martha’s Vineyard. Qing Zhang’s work on Beijing Mandarin (2005) offers a rich view of phonological variables that are tightly linked to place through the relation between social types and local meaning. Zhang traces ⫺ both through literature and through speaker reports ⫺ a connection

10. Social anthropology and interactional sociolinguistics between some phonetic features of Beijing dialect and classic Beijing social personae. Rhotacization (including the rhotacization of finals and the lenition of obstruent initials) is emblematic of Beijing, and is commonly said to contribute an “oily” “smooth” sound to Beijing Mandarin. This oiliness and smoothness is further iconically linked to a fabled male Beijing character ⫺ a type who appears in novels as a social operator who always manages to come through anything by virtue of his “smoothness.” The double iconicity ⫺ of rhotacization with oiliness, and oiliness in turn with a particular character type ⫺ makes this feature iconic of Beijing. The other Beijing type that Zhang discusses is the “Alley Saunterer” ⫺ a feckless individual who wanders around the back streets of the city waiting for something to happen. This type is associated with another Beijing feature, the interdental pronunciation of /z/. Once again, this link is to be found in popular literature, and emerged as well in Zhang’s conversations with Beijing speakers. Place, then, takes on character through the construction of local types ⫺ types that are in turn part of the construction of Beijing’s particular urban character. These features play a complex role in the indexical landscape in Beijing. Zhang compared the speech of managers in state-owned businesses with managers of equal educational and professional status in the foreign-owned financial sector. While the stateowned businesses maintain a local base, the new financial sector is a new and fast-paced transnational milieu, and a cosmopolitan persona is essential to success in this sector. The people working in this sector are the new Beijing “yuppies” ⫺ an emerging social category that is carving out an entirely new highly consumptive life style. And integral to this style is the construction of a cosmopolitan linguistic style. Zhang found that while managers in state-owned businesses used the Beijing variables described above quite frequently, managers in the financial sector did not. Further, while male financial managers made some use of the “smooth operator” variable, which is not completely at odds with desirable business characteristics, they did not use the “alley saunterer” variable which is at odds with the dynamic business persona. Yuppie women, on the other hand, completely avoided both of these variables. At the same time, both male and female yuppies reach out beyond Beijing and beyond mainland China into the transnational Mandarin sphere, to enhance the cosmopolitan sound of their style. The yuppies, and the yuppies alone, adopt a tone feature that is characteristic of varieties of non-mainland Mandarin associated with capitalism. In Beijing Mandarin, unstressed syllables are sufficiently reduced that they lose distinctive tone, assimilating to the tone of the preceding syllable. In non-mainland varieties, however, these syllables are less reduced and maintain their distinctive tone. The managers in the financial sector often use this full tone, giving their speech a cosmopolitan flavor, and, by virtue of the very different rhythm this yields, a “crisp” sound which itself has iconic value in a fast-moving capitalist style. Thus reputations of cities, regions and countries enter into the meaning of linguistic forms associated with them, and ultimately make these forms potentially available for bricolage elsewhere. New York and Jewishness are closely tied in the American imagination, as New York is known for its early large, powerful and intellectual Jewish community. The late comedian Lenny Bruce’s famous quote pretty much wraps up the relation between place and type in this case: “Even if you’re Catholic, if you live in New York you’re Jewish. If you live in Butte, Montana, you’re going to be goyish even if you’re Jewish.” This Jewish imaginary appears to play a role outside of New York. In a study of an affluent community in Michigan, Rebecca Knack (1991) found that Jewish speakers were less likely than their Gentile neighbors to front /oh/ as part of the Northern Cities Shift. She found,

173

174

II. Linguistic approaches to space further, that the extent to which female speakers resisted this change was related both to the extent to which they were integrated into the local Jewish community, and to the extent that their social networks included New Yorkers. The diphthongization and raising of /oh/ is arguably the most salient feature of New York City area speech, since it is the only one not shared in some way with other North American dialects. In Michigan, this amounts to a reversal of the Northern Cities change. Whether the non-fronters of /oh/ hear that change as “goyish” or if they disidentify more generally with the Midwest as a goyish region, or whether they hear the New York backing as sophisticated and/or Jewish is unclear. This is a point at which experimental methods could build profitably on the ethnography. Variability above the local level ⫺ regional and national dialects ⫺ is the most loaded with indexical meaning. Variables can be less adaptable across dialects, and it is probable that only variables that are not part of an incompatible systemic shift will move across regions. The north-south divide is a powerful linguistic and ideological line in the US (and in many other countries) and plays an important role in variation. Work in perceptual dialectology shows that Americans are acutely conscious of a north-south dialect boundary and that they have strong ideologies about these two regions and their dialects. In some respects regional stereotypes are shared on both sides of the boundary. The social evaluation of southerners and of their speech as uneducated has been robustly shown in perceptual dialectology studies (Preston 1989 and in this handbook), suggesting that the indexical value of southern features is somewhat uniform across the US. It is part of a wider set of values commonly found between standard and vernacular varieties (Lambert et al. 1960), in which northern speakers are heard as intelligent and educated while southern speakers are heard as casual, friendly and down-to-earth. (ING) varies everywhere in the English-speaking world between a velar [-=n] and an apical [-en] variant. This is not generally thought of as a regional feature ⫺ rather, in all communities examined, the apical variant correlates inversely with socioeconomic status and with formality of style. In an experimental study of this variable in the US, Kathryn Campbell-Kibler (Campbell-Kibler 2005) found that speakers overwhelmingly associate it with educational level. But this association interacts further with an association with region ⫺ people from both north and south associate the apical variant with southern speech. And they also associate southern speech with lack of education. At a higher level of geographic abstraction, the opposition between American and British speech is extremely salient. The dynamic between Britain and the US involves shared stereotypes of Americans as crass and uncultured and the British as cultured and refined. Many aspects of British English, therefore, carry connotations from these stereotypes ⫺ most particularly the non-raising of /ae/ and /t/ release. Lowered /ae/ simply sounds pretentious in most dialects of American English (this is changing in California, where the lowering of /ae/ appears to be a change in progress). But the /t/ release is an American variable with an apparently complex relation to British /t/ release. In casual American speech, intervocalic /t/ is flapped, and word-final /t/ is either flapped or not released at all. Studies have shown the use of released stops in these two positions to be associated with a range of groups that would appear to have nothing in common. Mary Bucholtz’s study (Bucholtz, 1996) of a Northern California high school included a group of girls who lay claim to an intellectual identity by styling themselves as “nerds”. They shared an “intellectual” linguistic style that included the use of /t/ release. In her study of an Orthodox Jewish community, Sarah Benor (2002) found that /t/ release was

10. Social anthropology and interactional sociolinguistics more common among boys than among girls and that boys who had been to Yeshiva led everyone in the use of /t/ release. The link between /t/ release and talmudic learning was supported by the observation that this feature emerged in conversation when boys were making important points. /t/ release has also been associated with gay speech (Podesva, Roberts and Campbell-Kibler 2002). Rob Podesva’s study (2004) of a gay doctor as he moves from situation to situation particularly underlines the significance of this variable. Podesva found that the doctor made significantly greater use of /t/ release in the clinic with patients than in informal interactions. But further, the performance of gay diva that he often adopted in informal situations with his friends might have had fewer /t/ releases, but those /t/s that were released had significantly longer bursts than in the clinic. Clearly, stop release can have a general meaning associated with education and articulateness in the general American population, and there are no doubt several sources contributing to this indexical value. It is reasonable to assume that at least part of the indexical force of stop release comes from the contrast between British and American speech and the popular association of British English with education, refinement and high culture. Indeed, an over-use of flapping is a common device for constructing a “loutish” American style. This international difference coincides with more local ⫺ even iconic ⫺ uses of /t/ release. It is a feature of clear speech and of emphatic speech. Children may hear parents releasing their /t/s when giving pointed instructions, when they are scolding, when they are angry. Teachers may release their /t/s when they are being explicit ⫺ as well as when they’re impatient and scolding. At the same time, both parents and teachers may scold a child for not speaking clearly, setting an example for clear speech that is rife with hyperarticulation. This sets up an indexical field for /t/ release that encompasses clarity and emphasis, but extends as well into negative feeling, power, education and refinement (Eckert 2008). Furthermore, this indexical field leaves room for exaggeration through lengthening the burst, yielding an impression of prissiness in a gay diva performance. Other languages are also powerful contributors to the repertoire of variables. The use of foreign accents in Disney films (Lippi-Green 1997) exploits and, in the process reproduces, a powerful set of linguistic stereotypes. Similarly, the use of mock Spanish (Hill 1993) across the US ⫺ by using Spanish phrases badly ⫺ serves as a force of pejoration. A true variable has emerged from substratum effects in bilingual communities. The interdental fricatives /th, dh/ are most commonly pronounced as alveolar stops by first generation English speakers. Their easy adaptability into English phonology as /t, d/ is no doubt responsible for the common survival of the stop pronunciations in the ensuing generation of native speakers of English. It has been found in Polish, Cajun, German and Mexican communities, and it very quickly takes on indexical value based on the nature of the community. The frequency of the stop variant has been found to correlate with the strength of the speaker’s ethnic neighborhood ties in the Polish neighborhood of Hamtramck in Detroit (Edwards and Krakow, 1985). Dubois and Horvath (1998) also found that it has been commodified along with ethnic identity in the Cajun tourist industry. In a Wisconsin farming town, Mary Rose (2006) has found it to be associated with the family farming and hard work ethic characteristic of the local German population, but not with a particular speaker’s German-ness. Norma Mendoza-Denton (2008) found this variable to correlate in a Chicano community with gang affiliation, and to be used almost categorically in expressions that invoke common community knowledge (and everything).

175

176

II. Linguistic approaches to space

5. Conclusion The role of linguistic variation in weaving a level of indexical meaning into day-to-day speech is fundamental to linguistic and social practice. Speakers live in a vast and continually changing sociogeographic landscape, and throughout life are engaged in constant sense-making in that landscape. They come to associate types of people with places, and to associate those people’s speech styles with social types. This involves a complex social and linguistic analysis as they assess what is socially distinctive about the types, and then which bits of their linguistic patterns relate to those distinctions. This continual analysis allows speakers to identify resources, and to appropriate them, incorporating them into their own repertoires, recombining them with other resources to construct new meanings in an ongoing process of bricolage (Hebdige 1984). People tweak their styles to take stances and to present themselves as particular kinds of people. These particular kinds are not defined simply by age, class, gender or ethnicity, but are more specific types located at, but not defined by, the intersections of these categories. And it is their linguistic practice at these locations that creates the large-scale patterns that emerge in survey studies. This act of tweaking thus affects not only the person’s way of speaking, but yields a tiny change in the stylistic landscape, the system of social meaning, the social system that style represents, and the meaning of the variable itself. The study of variation has so far only scratched the surface of the variable features that speakers use to make social meaning, as the focus has been on variables that define regional, class, or ethnic dialects. If we move away from our focus on dialectal features and changes in progress, we come upon a congeries of resources that make up the system of variation ⫺ emotive features, features of “baby talk”, prosodic, morphosyntactic, discourse and lexical features are only the beginning of a vast list that are constantly in use. But none these resources occur in isolation. Their meaning is dependent on ⫺ and contributes to ⫺ the style in which they occur. For it is ultimately in styles and their relation to personae and social types that the indexical value of variation is grounded in the social.

6. Reerences Bauman, Richard 2001 The ethnography of genre in a Mexican market: Form, function, variation. In: Penelope Eckert and John Rickford (eds.), Stylistic Variation in Language, 57⫺77. Cambridge: Cambridge University Press. Bell, Alan 1984 Language style as audience design. Language in Society 13: 145⫺204. Bell, Allan 2001 Back in style: Reworking audience design. In: Penelope Eckert and John Rickford (eds.), Style and Sociolinguistic Variation, 139⫺169. Cambridge: Cambridge University Press. Benor, Sarah 2002 Sounding learned: The gendered use of /t/ in Orthodox Jewish English. Penn working papers in linguistics: Selected papers from NWAV 2000. Bourdieu, Pierre 1977a Outline of a Theory of Practice. Cambridge: Cambridge University Press. Bourdieu, Pierre 1977b The economics of linguistic exchanges. Social Science Information 16, 645⫺668.

10. Social anthropology and interactional sociolinguistics Bourdieu, Pierre 1982 Ce que parler veut dire. Paris: Fayard. Bucholtz, Mary 1996 Geek the girl: Language, femininity and female nerds. In: Natasha Warner et al. (eds.), Gender and Belief Systems, 119⫺131. Berkeley: Berkeley Women and Language Group. Campbell-Kibler, Kathryn 2005 Listener perceptions of sociolinguistic variables: The case of (ING). Unpublished PhD dissertation, Department of Linguistics, Stanford University. Dubois, Sylvie and Barbara Horvath 1998 From accent to marker in Cajun English: A study of dialect formation in progress. English World Wide 19: 161⫺188. Eckert, Penelope 1989 Jocks and Burnouts: Social Categories and Identity in the High School. New York: Teachers College Press. Eckert, Penelope 1996 Vowels and nailpolish: The emergence of linguistic style in the preadolescent heterosexual marketplace. In: Jocelyn Ahlers et al. (eds.), Gender and Belief Systems: Proceedings of the 1996 Berkeley Women and Language Conference, 183⫺190. Berkeley: Berkeley Women and Language Group, University of California. Eckert, Penelope 2000 Linguistic Variation as Social Practice. Oxford: Blackwell. Eckert, Penelope 2005 Variation, convention, and social meaning. Paper presented at the Annual Meeting of the Linguistic Society of America. San Francisco. Eckert, Penelope 2008 Variation and the indexical field. Journal of Sociolinguistics 12(3): 453⫺476. Eckert, Penelope and Sally McConnell-Ginet 1992 Communities of practice: Where language, gender and power all live. In: Kira Hall, Mary Buchholtz and Birch Moonwomon (eds.), Locating Power: Proceedings of the Second Berkeley Women and Language Conference, 89⫺99. Berkeley: Berkeley Women and Language Group, University of California. Edwards, Walter and Cheryl Krakow 1985 Polish-American English in Hamtramck: A sociolinguistic study, Paper delivered at Conference on New Ways of Analyzing Variation in English, Georgetown University. Hebdige, Dick 1984 Subculture: The Meaning of Style. New York: Methuen. Hill, Jane H. 1993 Hasta la vista, baby: Anglo Spanish in the American Southwest. Critique of Anthropology 13: 145⫺176. Inoue, Miyako 2003 The listening subject of Japanese modernity and his auditory double: Citing, sighting, and siting the modern Japanese woman. Cultural Anthropology 18: 156⫺193. Irvine, Judith T. and Susan Gal 2000 Language ideology and linguistic differentiation. In: Paul V. Kroskrity (ed.), Regimes of Language: Ideologies, Politics, and Identities, 35⫺83. Santa Fe NM: SAR Press. Johnstone, Barbara 2004 Place, globalization, and linguistic variation. In: Carmen Fought (ed.), Sociolinguistic Variation: Critical Reflections, 65⫺83. Oxford/New York: Oxford University Press. Johnstone, Barbara, Jennifer Andrus and Andrew Danielson 2006 Mobility, indexicality, and the enregisterment of “Pittsburghese”. Journal of English Linguistics 34: 77⫺104. Knack, Rebecca 1991 Ethnic boundaries in linguistic variation. In: Penelope Eckert (ed.), New Ways of Analyzing Sound Change, 252⫺272. New York: Academic Press.

177

178

II. Linguistic approaches to space Labov, William 1963 The social motivation of a sound change. Word 18: 1⫺42. Labov, William 1966 The Social Stratification of English in New York City. Washington DC: Center for Applied Linguistics. Labov, William 1972 Some principles of linguistic methodology. Language in Society 1: 97⫺120. Labov, William 2000 Principles of Linguistic Change: Social Factors. Cambridge: Blackwell. Lambert, Wallace, R. Hodgson, R. Gardner and S. Fillenbaum 1960 Evaluative reactions to spoken language. Journal of Abnormal and Social Psychology 60: 44⫺51. Lippi-Green, Rosina 1997 English with an Accent: Language, Ideology, and Discrimination in the United States. New York: Routledge. Mendoza-Denton, Norma 2008 Home Girls. Cambridge/New York: Blackwell. Milroy, Lesley 1980 Language and Social Networks. Oxford: Blackwell. Modaressi, Yahyah 1978 A sociolinguistic analysis of modern Persian. Unpublished PhD dissertation, Linguistics Department, University of Kansas. Ochs, Elinor 1991 Indexing gender. In: Alessandro Duranti and Charles Goodwin (eds.), Rethinking Context, 325⫺358. Cambridge: Cambridge University Press. Podesva, Robert 2004 On constructing social meaning with stop release bursts. Paper presented at Sociolinguistics Symposium 15, Newcastle upon Tyne. Podesva, Robert J., Sarah J. Roberts and Kathryn Campbell-Kibler 2002 Sharing resources and indexing meanings in the production of gay styles. In: Kathryn Campbell.Kibler, Robert J. Podesva, Sarah J. Roberts and Andrew Wong (eds.), Language and Sexuality: Contesting Meaning in Theory and Practice, 175⫺190. Stanford: CSLI Press. Pratt, Mary Louise 1988 Linguistic utopias. In: Nigel Fabb, Derek Attridge, Alan Durant and Colin MacCabe (eds.), The Linguistics of Writing: Arguments between Language and Literature, 48⫺66. New York: Methuen. Preston, Dennis 1989 Perceptual Dialectology. Dordrecht: Foris. Rose, Mary 2006 Social meaning in rural Wisconsin. Unpublished PhD dissertation, Department of Linguistics, Stanford University. Shuy, Roger W., Walter A. Wolfram and William K. Riley 1967 Linguistic correlates of social stratification in Detroit speech: Final Report, Research Project No. MH 15048-01, National Institute of Mental Health. Silverstein, Michael 2003 Indexical order and the dialectics of sociolinguistic life. Language and Communication 23: 193⫺229. Trudgill, Peter 1974 The Social Differentiation of English in Norwich. Cambridge: Cambridge University Press. Zhang, Qing 2005 A Chinese yuppie in Beijing: Phonological variation and the construction of a new professional identity. Language in Society 34: 431⫺466.

Penelope Eckert, Stanford (USA)

11. Language, space and the folk

179

11. Language, space and the olk 1. 2. 3. 4. 5. 6. 7. 8.

Introduction Similarities and differences Hand-drawn maps Dialect maps ⫺ maps that aren’t quite maps Where are you from? Conclusion Appendix References

“Space ⫺ the final frontier” (Captains Kirk and Picard, from the American science fiction television program “Star Trek”)

1. Introduction When people ask me where I am from, I hesitate. I was born in the United States in southern Illinois but grew up in southern Indiana (see the Appendix for a map of the US with state names). From a linguistic point of view, that move was probably uninteresting; they are both in what US dialectologists would call the “South Midlands,” and I still have many of its characteristics ⫺ conflation of the vowels in such pairs as pin and pen; second person plural you all (not the more southern y’all); such double modals as ought to could and might could (but not the more southern may can); partial monophthongization of /ay/; the same vowels in off and on; [z], not [s], in the verb grease and the adjective greasy (but not in the noun grease), and so on. I hesitate partly because many people want to know where one is from now, and, at the time of writing, I lived in the Lower Peninsula of Michigan, where none of the above speech characteristics exist, except among older immigrants like me, and some of their descendents, although Evans (2001) has shown that the third generation of South Midlanders in Michigan have little or no trace of their native dialect. Like many people worldwide, my birthplace, my childhood home, and my current residence are different, and such differences may have linguistic repercussions. If I say I am from Michigan, people sometimes even overtly observe that I don’t sound like I am from there. What lies behind nonlinguists’ observations of areal differences? This article deals with folk or nonspecialist notions of how a language is distributed in space ⫺ a perceptual dialectology. As in professional dialectology, the primary folk notion of language and space is that of region or area, and the defining characteristic of area is boundary. Nonlinguists know that groups of people form bounded zones of linguistic similarity, and the principal task in a perceptual dialectology is to find out where they believe those boundaries are. How can these folk notions be determined?

2. Similarities and dierences The oldest cognitive mapping method asked respondents to rate surrounding areas on the basis of their similarity to or difference from their own speech. The construction of

180

II. Linguistic approaches to space maps from these ratings caused the first controversy in the field of perceptual dialectology: the Dutch scholar Antonius Weijnen believed that Japanese scholars did not find good correspondences between perception and traditional or production dialectology because they asked the wrong question. The Japanese work focused on where respondents found dialects to be different while, for Weijnen, the proper question was to ask where a respondent felt surrounding varieties were similar, since, in his opinion, respondents in places like the Netherlands and Japan would always find some little difference between their own speech and that of even very nearby and linguistically similar villages (Weijnen 1968 [1999: 132]). Whether seeking similarity or difference (or both), a number of techniques were developed in the Netherlands and Japan to infer folk dialect boundaries from such respondent characterizations. In the Netherlands, the prevailing method was known as the Pfeilchenmethode (‘little-arrow method’), in which an arrow was drawn from one area to another when a respondent in the first said that the second area’s speech was the same as his or her own. This technique was developed in the late nineteenth century by Pieter Willems, who died before his work was finished (1886), but a detailed account of this earliest work on perceptions can be found in Goeman (1989 [1999: 138⫺139]). Map 11.1 shows an application of this technique. The dark, thick lines are the traditional dialect divisions, and the perceptual areas are determined by clusters of sites connected by arrows. That is, one may outline perceptual areas by drawing boundaries around all areas where there is a gap between little arrow connections. I have outlined these latter with lighter gray lines (not done in the original). In the northwest (i. e., upper left) section of Map 11.1, for example, the respondent from W (Willemstad) indicates that no nearby community sounds like W by drawing no arrow towards any, and since no surrounding communities have identified W as similar, no arrows are drawn towards it; therefore, the gray circle around W indicates that it is a perceptual isolate. In contrast, the respondent from D (Dinteloord) believes that the variety in F (Fijnaart) is the same as D, and the respondent from F returns the favor; hence, an arrow from D to F and one from F to D. The F respondent also identifies K (Klundert) as the same, but this perception is not reciprocal. Again, a gray boundary line indicates this perceptual region, and, although it is easy to see that W is an isolate, one must consult the arrows in detail to see the different relationships between D and F (reciprocal), F and K (one-way, F to K), and D and K (mediated by an intervening site). A more complex relationship exists in the area just to the east. There, Z (Zewenbergen) identifies M (Moerdijk, just to the north) as the same, although reciprocal identification is not given, and Z itself is identified as the same by a respondent from one site to its southwest. In both these cases, however, the production boundary just to the east is not crossed. A respondent from Z, however, asserts the similarity to Z of both ZH (Zevenbergschen Hoek) and L (Langeweg) across the production boundary, although the respondent at neither ZH nor L identifies Z as similar. This one crossing, however, creates a very large perceptual unity, one that runs across the entire east-west area of the territory. A better representation of the perceptual areas might be made by circling only the most densely clustered areas of arrow linkages, but there do not appear to have been any early proposals about how that might be done quantitatively (but see Goeman 2002). Rensink (1955) provided the first general map of Dutch-speaking areas based on the perceptions, in which lines are drawn around the bundles of little arrows, as I have done with thick, gray lines in Map 11.1.

11. Language, space and the folk

Map 11.1: The westernmost section of the North Brabant, showing production boundaries (dark thick lines), little arrows of respondent similarity perceptions, and perceptual areas (gray thick lines) (enlarged and adapted from Weijnen 1946)

In general, the principal motivation in this early Dutch research seemed to have been a desire to give production dialect boundaries greater or lesser weight by establishing their folk validity, combining folk and scientific notions of linguistic space. This is thoroughly discussed by Weijnen (1968) and is explicitly realized in Daan (1969) who, in an ambitious study of contiguous Dutch-speaking areas, provides a map based on both perception and production data. Kremer (1984) is another interesting little-arrow study of the perception of varieties by German and Dutch speakers within and across national boundaries. All these studies, however, make use of the technique of drawing perceptual boundaries around areas not crossed by any little arrows, i. e., judgments of dialect similarity by folk respondents.

181

182

II. Linguistic approaches to space In Japan different methods were developed to realize maps from respondent judgments, and I will illustrate this with only one study, that of Mase (1964) from alpine Japan. He first mapped responses to two questions ⫺ which sites sound the same and which sites sound a little different from the home site of the respondent. Map 11.2 shows how he combines these results in constructing his map.

Map 11.2: The subjective dialect boundaries indicated by a respondent at site #57 (Mase 1964)

The respondent at #57 in Map 11.2 has identified #58 and #59 as the same. He has also noted that #62, #63, #56, #55 and several sites in Nagawa are a little different. Speakers from #58 and #59 agree, not only that their own regions are similar to one another and #57 but also that the same sites in Nagawa are a little different. Finally, not shown in Map 11.2, respondents from surrounding areas classify #57, #58, and #59 together in their evaluations. In short, the perceptual dialect area made up of these three sites is based on reciprocal perceptions of similarity, on similar perceptions of minor degrees of difference, and on the perception by surrounding areas of their similarity to one another. Mase also calculates places that respondents refer to as “quite different,” “totally different,” and “incomprehensibly different” and assigns higher scores to higher degrees of difference. Map 11.3 shows a map for the entire alpine area studied, combining all respondents’ ratings for all degrees of difference and similarity; thicker lines indicate greater differences, and dotted lines indicate areas regarded as only slightly different. Note in the bottom left of the map that #57, #58, and #59, shown as similar in Map 11.2, are, in fact, rated considerably different from all surrounding areas when all these similarity and difference rankings are taken into consideration. This method, however, reduces our ability to distinguish which regions are identified on the basis of similarities (exclusively or predominantly) and which are identified on the basis of differences. Nevertheless, Mase’s treatment of boundaries is quantitatively sophisticated in comparison with the little arrow technique, in which only one connection (i. e., one similarity judgment) appears to cause a site to be included in a perceptual area. Other Japanese mappings of degrees of similarity and difference include Grootaers (1964), Nomoto (1963) and Sibata (1959). The mapping of similarity and difference has reappeared in several newer studies of perceptual dialectology, in which the respondents are asked to determine differences between areas such as states or other regional or political zones pre-selected by the

11. Language, space and the folk

Map 11.3: A combined perception map of an alpine area of Japan (Mase 1964)

researcher (Montgomery 2007). For example, Preston (1996: 317⫺20) compares the ratings for “degree-of-difference” on a four-point scale (1 ⫽ same, 2 ⫽ slightly different, 3 ⫽ different, and 4 ⫽ unintelligibly different) for all fifty US states, New York City, and Washington, D.C. from the point of view of three geographically distinct sets of respondents in the US. Since such studies predetermine the areas asked about and use a scale, means scores and area shadings (and more sophisticated statistical techniques) can be used to represent the degree of difference or similarity perceived by a group of respondents about surrounding areas (and, of course, their own). In Preston (1996), for example, respondents from southeastern Michigan found a much larger area of similarity to their own speech (six bordering or nearby states) than did those from southernmost Indiana (only two bordering states). This style of degree-of-difference study has been used in other areas of the US and in France (Kuiper 1999), Turkey (Demirci and Kleiner 1999), French-speaking Canada (Evans 2002), Switzerland (L’Eplattenier-Saugy 2002), and Spain (Moreno Ferna´ndez and Moreno Ferna´ndez 2002). In several of these studies, demographic subdivisions within the group studied have been elaborated on statistically. Moreno Ferna´ndez and Moreno Ferna´ndez (2002), for example, find interesting differences in Madrid perceptions of regional Spanish according to sex, age, and education of the respondents. Tamasi (2003) provides an interesting option to all these respondent-centered techniques of identifying similarity and difference. In a technique borrowed from cognitive anthropology called pile sort, she asks respondents to sort into as many piles as they like the regions whose speech practices are the same. She then provides maps of these pile sorts based on cluster analyses that allow specification of clusters at different levels of agreement. Map 11.4 shows a .25 agreement level for her respondents from Georgia. Map 11.4 makes it clear that several states are either viewed as single dialect areas or that there is insufficient agreement to align them with other areas (California, Hawai’i, Alaska, Florida and Texas). The remainder of the country was sorted into six zones at this level of agreement. Cognitive or folk maps of dialects, however, can be derived even more directly than from respondent views of similarity and difference.

183

184

II. Linguistic approaches to space

Map 11.4: A pile-sort mapping at a .25 level from a cluster analysis of the fifty states by Georgia respondents (adapted from Tamasi 2003: 66)

3. Hand-drawn maps The next method to be used in mapping folk dialect spaces followed work done in the general area of mental maps (e.g., Gould and White 1974) and has relied on the most straightforward way to determine what the folk think about the limits of language areas ⫺ have them draw a map. The earliest collection of hand-drawn maps of respondent ideas about language distribution is apparently Preston (1981, 1982). Map 11.5 is a typical hand drawn map of a speech area, in this case by a respondent from Hawai’i, who was asked to outline and label on a US map containing only state lines the areas of the US where people spoke differently. Map 11.5 shows that respondents do not make exclusive use of state lines or other guides given them for orientation. For example, “Texas drawl” does not include all of the state of Texas but does include all or almost all of three other states (Oklahoma, Arkansas and Louisiana) and sizeable portions of two others (Kansas and Missouri). This map may be the cognitive or mental map of US regional speech areas for this respondent, but such ethnographically interesting but perhaps idiosyncratic maps have been combined to reveal speech community rather than individual mental maps of dialect areas. These combinations were first done by tracing all respondent boundaries for a single area onto a map and tallying the correspondences. Map 11.6 shows such a map

11. Language, space and the folk

Map 11.5: A Hawai’i respondent’s hand-drawn map of US regional speech areas (from the author’s personal collection)

of the outlines of the Tohoku dialect area (in the northernmost part of Honshu, the main island of Japan) as drawn by 60 respondents from Aichi Prefecture (also in Honshu, but quite south of Tohoku, between Tokyo and Osaka). Map 11.7 shows the results of combining the lines that occurred most frequently in the same place. For example, all sixty respondents agreed that Tohoku dialect was spoken in the area in solid black farthest north, but in the area just to the south of this onehundred percent agreement, the east is regarded as more Tohoku-like (50 to 59 respondents) as opposed to the west (only 40 to 49 respondents). Such distinctions allow researchers to look for both linguistic and other cultural notions that might support such a heavier concentration of the Tohoku dialect identity in the east. This technique of combining numerous hand drawn lines to determine a perceived speech area boundary has been used in a number of studies, including Preston (1981, 1986, 1988: 1989), Dailey-O’Cain (1999), Kuiper (1999), Evans (2002), Benson (2003), and Montgomery (2007). A slightly different approach to generalizing lines drawn by respondents is taken by Inoue (1996), Hartley (1999, 2005), Lance (1999), Demirci (2002), and Fought (2002). Inoue, for example, in determining the folk dialect boundaries of England, determines which counties have been included by respondents in their line-drawing task and then prepares a mental map on that basis. Lance uses US states and Demirci areas surrounding principal Turkish cities in much the same way. If political or other zones within the area to be studied are small enough or can be reasonably subdivided into subzones, this approach may be a viable alternative to tracing the precise pathways of individual lines and determining where they cluster. Interestingly, in an attempt to follow the clusters of lines using the first approach to the determination of general boundaries, Long and Yim (2002), in their study of mental maps of Korean dialect division, found this second

185

186

II. Linguistic approaches to space

Map 11.6: A map showing 60 Aichi respondents’ boundaries for the dialect area of Tohoku (Long 1999: 180)

approach preferable since nearly all their respondents drew lines along prefectural boundaries. In still another technique developed by Preston and Howe (1987), outlines of areas were traced onto a digitizing pad and percentages of respondent agreement about boundaries were calculated. Map 11.8, for example, shows the results of such a computational procedure for 147 hand drawn maps of US dialect regions provided by respondents from southeastern Michigan when a fifty percent agreement criterion is used. This computational approach has been used, with adaptations to local territories, by Long (1999), Long and Yim (2002), and Montgomery 2007. Whether done by deriving boundary lines by simply counting co-occurrences of hand drawn lines, by combining political or other districts contained within hand-drawn lines into dialect areas, or by deriving boundaries computationally from hand drawn lines, this approach to cognitive dialect boundaries asks respondents to represent directly on a map their notion of the number and extent of distinct speech areas. In this style of research, respondents are not asked to use their own speech as a measure of difference, as in the little arrow, Japanese, and degree-of-difference methods outlined above, al-

11. Language, space and the folk

Map 11.7: A composite map of 60 Aichi respondents’ hand-drawn maps of the Tohoku dialect area (Long 1999: 183)

Map 11.8: A computationally generalized map of US dialect areas derived from 147 hand-drawn maps (at a fifty-percent level of agreement) provided by southeastern Michigan respondents (Preston 1996: 305)

187

188

II. Linguistic approaches to space though that might indeed be a criterion used by the respondents in constructing their maps. For that reason, post-task interviews that try to uncover such respondent strategies are highly recommended (e.g., Preston 1999: xxxiv). The work carried out within newer mental map traditions is more focused on the agreements about boundaries of speech differences from the perspective of a single speech community and, like much more recent sociolinguistic work, focuses on drawings or judgments from a large number of respondents from a single area. This allows testing of typical sociolinguistic demographic differences within a group of respondents. Demirci (2002), for example, contrasts hand-drawn maps of Turkish dialect areas by gender. Not surprisingly, the older efforts in the Netherlands and Japan focused on single respondents from a number of sites, just as traditional dialectology often chose a single, exemplary speaker from an area. Perhaps more importantly, the newer mental maps tradition has focused equally on nearby and distant places for judgment; the earlier tradition asked respondents to evaluate nearby places only, or, at least, did not insist on or imply that judgments of distant places were required. It is not surprising, therefore, that research that has focused on local areas will find more local details in the respondent representations and that work that focuses more globally will find less local detail. When Benson (2003) asked respondents to draw maps of linguistic differences in Ohio, she found two, three, and even four-part divisions of the state. In contrast, in earlier (unpublished) work with sixteen Ohio respondents who were asked to map the entire country, I found that eleven did not divide Ohio at all and only five divided it into two areas ⫺ northern and southern. Only three of Benson’s eleven respondents did not subdivide the state while four found two areas, three outlined three, and one found four. Other work on mental maps of dialects has derived boundaries of linguistic space by referring to concepts quite different from those usually considered by dialectologists.

4. Dialect maps  maps that arent quite maps The Japanese dialectologist Fumio Inoue was the first to develop a strategy for mapping dialects based on concepts other than the respondent’s ability to rate similarities or differences in surrounding regions or to outline or identify speech regions on the basis of their identities. Inoue was so convinced (e.g., 1996) that nonlinguistic geographic knowledge (e.g., political boundaries) guided respondent answers to requests to draw boundaries that he devised a method for determining what he called the dialect image (Inoue 1995) of an area, a method in which the mental position of a dialect was derived from the attitudinal factors associated with places and speakers. In short, Inoue wants to know how different areas are related to one another in the conceptual space allotted to them by different nonlinguistic factors. To accomplish this he borrows heavily from the semantic differential technique adapted to language attitude studies (e.g., Shuy and Fasold 1973). Inoue first elicits from native respondents a number of evaluative words that describe the dialects of the region he is interested in. A multidimensional analysis of these items shows that they can be grouped and that the grouping can be shown along axes in

11. Language, space and the folk

Fig. 11.1: Multidimensional distribution along two axes of evaluative words provided by British university students (Inoue 1995 [1999: 153])

multidimensional space. Figure 11.1 shows the two-axes groupings of sixteen evaluative words for English accents given by British university students (Inoue 1995 [1999: 153]). The two dimensions (horizontal and vertical) are clearly polarized, and Inoue labels the right side of the horizontal dimension “STANDARDNESS” and the left side “ACCENTEDNESS.” He calls the vertical dimension “PASTORAL (RURAL)” at the top and “URBANITY” at the bottom. He then assigns these ratings to varieties and produces the multidimensional plot shown in Figure 11.2. This is indeed a mental map that does not correspond to geographical space at all, but it clearly shows that British university students have much stronger ideas about nearby varieties; the extremes of the axes are Liverpool and Cambridge University on the “Class Difference” or “STANDARDNESS” horizontal axis and London and Villages in Norfolk on the “Urbanization” vertical one.

189

190

II. Linguistic approaches to space

Fig. 11.2: Multidimensional distribution along two axes for selected English varieties provided by British university students (Inoue 1995 [1999: 154])

Inoue has also applied this multidimensional scaling technique to hand-drawn map results using only the frequency of combination of counties and has shown a strong horizontal axis for a north-south dimension (Geordie to Cockney) in English dialects but a strong vertical dimension for an east-west distinction for dialects in only the south of England (“Eastern Norfolk” to “Western”) (Inoue 1996: 153). Other applications of this multidimensional scaling technique have sought to determine dialect images of areas for other factors. For example, Kuiper (1999) provides multidimensional plots (with embedded cluster analyses) of French respondents’ ratings of French dialect areas for degree-of-difference, correctness, and “pleasantness.” Hartley does the same for ratings given by residents of the state of Oregon in the US (1999) and for residents of Boston, Massachusetts (2005), both for the entire US, and Evans (2002) does the same for Montreal French-speaking students’ ratings of French varieties worldwide. It is clear that clustering and multidimensional scaling techniques add to our understandings of the mental images respondents have of language distribution, even if the distributions are in created mental spaces rather than predetermined geographic ones. Such work as Inoue’s and others’, if fully considered, might make this article twice as long. In addition to geographic space and the demographic features that clearly interact with the folk respondent’s cognitive construction of it, the folk also map language in social space. This, however, makes useful but metaphorical sense of the notions space and map. We could map (perceptually) an individual’s and a speech community’s perceptions of all the sociolinguistic commonplaces ⫺ sex, gender, age, ethnicity, community of practice, etc. There is no doubt of the folk respondent’s awareness of these categories. To take only ethnicity as an example, folk respondents have been shown to be sensitive to this feature in language varieties on the basis of global or overall speech styles (e.g., Giles and Bourhis 1976) and on the basis of specific linguistic features (e.g., Graff, Labov and Harris 1986; Purnell, Idsardi and Baugh 1999). This article deals only with those cases in which such demographic factors clearly interact with the folk respondent’s construction of space in its more ordinary geographic sense.

11. Language, space and the folk

Map 11.9: The sites in the US from which nine well-educated, middle-aged male voice samples were taken (Preston 1996: 322)

5. Where are you rom? Finally we ask how people relate language and space when they hear actual speech rather than prompts based on regional labels or evaluative words. An early example of this in my own work was the submission of nine well-educated, middle aged, male voice samples placed at fairly even intervals along a north-south continuum in the middle of the US. The samples contained no lexical or grammatical clues that would help identify their provenience. The respondents discussed here, all from southeastern Michigan, were shown Map 11.9 and asked to identify which of these sites each voice, played in a scrambled order, was from. Since these respondents were those from whom the composite mental map of US dialects was derived based on their hand-drawn representations, one might expect the boundaries of voice perception to be in roughly the same place as those shown in that cognitive map (i. e., Map 11.8) and with the same degree of salience (i. e., drawn by a greater number of respondents) indicated there. If that is so, then one would expect the following placements: (1) Nashville TN and Florence and Dothan AL would be most consistently recognized as “South” and provide the strongest boundary (94 percent in Map 11.8). (2) Saginaw and Coldwater MI and South Bend and Muncie IN would be consistently recognized as “North” and provide the second strongest boundary (61 percent in Map 11.8). (3) New Albany IN and Bowling Green KY would be consistently recognized as “Inner South” and provide the third strongest boundary (30 percent in Map 11.8). (4) Nashville TN would be recognized as “Inner South” rather than “Southern” and provide the fourth strongest boundary (30 percent in Map 11.8).

191

192

II. Linguistic approaches to space

Map 11.10: Results of an ANOVA test on the site identifications of the nine voice samples; areas colored the same were not statistically different

(5) Muncie, IN is so close to the southern edge of area #2 and New Albany IN is so close to the northern edge of area #6 that they might be grouped together to form a minor (fifth place) boundary or each form a separate minor boundary. An ANOVA analysis of the assignment of voices to sites was performed, and the results are shown in Map 11.10. Prediction (1) was confirmed. The three southernmost areas form a region; prediction (2) was partially confirmed; three of the four northern areas form a region, although Coldwater MI was excluded from this group and was combined instead with New Albany IN and Bowling Green KY. Prediction (3) was confirmed; New Albany IN and Bowling Green KY formed an area, although, as noted above, Coldwater MI was surprisingly identified with them rather than with the other northern areas. None of the less salient predictions suggested in (4) and (5) were confirmed. As Map 11.8 shows, the most salient perceptual regions along this continuum of voices (areas #1, #2, and #6) had sample voices correctly positioned in them with the exception of one northern one, and the evidence given the respondents was only that of regional pronunciation. In other regional speech identification work, Diercks (1988) in a study in northern Germany used sample voices and asked respondents to determine how far from their home site such voices are, usually with accurate results. Williams, Garrett and Coupland (1999) asked Welsh teachers and students to identify fourteen regional voices (two each from six sites in Wales and two English “RP” voices, all male adolescents). Although teachers made more correct assignments than students, overall correct placement was low, around 50 percent accuracy for the teachers compared to about 30 percent for students (351), and there was a great deal of variation even in the recognition of two voices from the same region. For example, adolescent judges from Cardiff correctly identified Cardiff voice #2 100 percent of the time but Cardiff voice #1 at only a 43.5 percent rate (352). An innovative alternative procedure for relating voice sample recognition to perceptual areas is provided in Montgomery (2007). He asked respondents, all from the north

11. Language, space and the folk

Fig. 11.3: Starburst pattern of Carlisle respondent placements of a voice sample from Carlisle (Montgomery 2007: 295)

of England, to put an “X” on a blank map of the country to indicate where they thought a voice sample was from. He then calculated the distance between the respondent placement and the actual voice site. When lines are drawn from the actual site to the respondent placement, the result is a starburst. Figure 11.3, for example, shows the starburst pattern of the placement of a voice from Carlisle (near the English-Scottish border in the northwest of the country) by respondents from Carlisle. Note that some respondents actually place Carlisle off the 175 mile limit of this diagram but that all such displacements are to the southeast. Since Carlisle is near the coast, they could not place it to the west, but nothing would prevent them from putting it farther north except the understanding that it could not be Scottish, and nothing would prevent it from being placed farther east except the understanding that it could not be a Newcastle (“Geordie”) variety. These results, and the results of Carlisle respondents’ placements of six other regional voices, are superimposed on an actual map in Map 11.11. Interestingly, the Carlisle respondents’ placements of the local voice are the worst (the purple lines from the “starburst” in Figure 11.3), but, in general Montgomery shows that most respondents from Carlisle (and other sites he studied) actually placed sample voices within the perceptual territory he determined for that respondent group from a hand-drawn map and boundary task previously submitted. In addition to these general voice placements, perceptual dialectology also hopes to discover not only what factors from the nonlinguistic world make people think about language and space the way they do but also how sensitive speakers are to various linguistic forms as they construct language and space relationships. Plichta, Rakerd and Preston (2005) selected a well-known southern US speech stereotype (/ay/ monophthongization) and resynthesized a sample of the word guide so that it increased in monophthongization in seven regular steps from a fully diphthongal form to a fully monophthongal one. The seven voice samples (one male and one female) were played three times

193

194

II. Linguistic approaches to space

Map 11.11: Carlisle respondents’ placements of seven regional voices, with the starburst patterns superimposed on a map (Montgomery 2007: 298)

at each of the seven steps for a total of forty-two judgments. In each case, the respondent was to assign the word to one of the nine sites shown in Map 11.9. The sites were numbered one through nine (Saginaw to Dothan) so that a numeric base could be used to ascertain if degree of monophthongization was perceived by the respondents (from all over the US) as an increasingly southern feature. Table 11.1 shows the results. An ANOVA post-hoc test shows that each of these mean scores is significantly different from every other one, revealing considerable sensitivity to very minor changes in degree of diphthongization and very clearly showing a strong association between monophthongization and the respondents’ perception of its southernness. At every one of these seven steps, however, there was also a significant difference (based on independent t-tests) between the male and female voice scores. The female voice, at the same degree of monophthongization as the male one, always has a lower

11. Language, space and the folk

195

Tab. 11.1: Mean scores based on regional values assigned each step of the increasingly monophthongized versions of /ay/ Step

Mean

1 2 3 4 5 6 7

2.85 3.17 3.87 4.89 5.99 6.58 7.02

Region 1. 2. 3. 4. 5. 6. 7. 8. 9.

Saginaw Coldwater South Bend Muncie New Albany Bowling Green Nashville Florence Dothan

(i. e., “more northern”) score. This finding surely tells us that mental maps are based on cultural stereotypes already well known to sociolinguists. Women are regarded in nearly all the sociolinguistic literature as more likely users of standard speech (e.g., Trudgill 1972), but this finding is apparently also a folk stereotype, for our respondents regard a female voice with no distinctive characteristics other than sex identity as “more northern” (a feature characteristically regarded as “more standard” in US folk linguistics, e.g., Preston 1996). Mental maps will have to be cognizant of such sociocultural stereotypes as formative in the construction of linguistic mental spaces. In other work on specific features or types of features, Clopper and Pisoni (2004) conducted a study that asked respondents to place characteristically regional voices from across the US within broadly defined dialect regions. A regression analysis revealed that the respondents were especially sensitive to only four (out of a potential seven) robust phonetic indicators of region, and, although the ratings were better than chance, they were accurate overall at only about a 30 percent rate. Van Bezooijen and Gooskens (1999) played sample voices from several English and Dutch sites with various types of filtering, allowing a focus on segments, voice quality, and prosody. In general, they found that respondents were best at identifying regional speech on the basis of segmental evidence, although the degree of success was considerably different across languages. Fridland, Bartlett and Kreuz (2004) showed that respondents were better at identifying some resynthesized vowels as “southern US” if they were ones represented in their own production and if they were ones more uniquely related to the area.

6. Conclusion A continuing issue in perceptual dialectology is, with any of these techniques, whether respondents judge areal linguistic differences on the basis of their experiences with and knowledge of language and linguistic details or on the basis of other factors that have to do with regional delineation. Dann ([1969] 1999: 21⫺22) is suspicious that many folk speech areas may be constructed on the basis of nonlinguistic features (e.g., religious differences), and a part of an early argument in the study of folk dialectology mentioned above had to do with just this theme. Researchers in Japan (e.g., Grootaers 1964 and

196

II. Linguistic approaches to space Sibata 1959) concluded that the “consciousness” of dialects was based on the folk knowledge of such matters as geographical barriers, political boundaries, and even school districts (but see Mase 1964 and Nomoto 1963). Inoue (1996) came to a similar conclusion after preparing his perceptual map of British dialects: “Stereotypical images of a place or of people living there help form dialect images” (158). While all researchers recognize that some such boundaries as school districts and the like might actually correspond to production boundaries, in general, the Japanese researchers were less likely than the Dutch to note extensive correlations between production and perception boundaries. Recall that Weijnen (1968), however, found that lack of correlation to be the result of a methodological flaw in not observing the different responses one might get by asking respondents where the speech of others was “different” rather than by asking where it was “the same.” It is very doubtful that many respondents form their general notions of space on the basis of language differences, although that awareness may play a bigger role where there are linguistic correlates to tribal or clan divisions, particularly in areas where exogamous marriage or some other cultural practice is based on linguistic differences (e.g., Stanford 2007). Since we know from work in cultural geography that different respondents have very different notions of even the physical spaces around them, it has perhaps been inappropriate to provide respondents with even minimal outlines of areas on which to draw their mental maps of linguistic differences. Perhaps one ought first to determine the general, nonlinguistic mental map of respondents and only then allow them to go on to characterize the speech regions of it. On the other hand, the mental map a respondent has of an area in general might have a very different outline from that same respondent’s map of linguistic differences in the same area. Perhaps we will have to content ourselves with investigating the folk mental maps of space with regard to language, comparing them, on the one hand, with maps of actual (production) language differences and, on the other, with other folk mental maps of constructs such as physical space, desirability of residence, political climate, and the like. However we conduct this perceptual research, it is very unlikely that we will find the mental map of an individual or a group. The various studies outlined above show variation among the same respondents when given different tasks, and that is surely the most likely outcome when we seek a cognitively based but learned rather than innate characteristic of human behavior. When we foreground sex in some aspect of the task, then sex will probably play a role in our respondents’ positioning of language in space; when we highlight a linguistic feature or a status related linguistic factor, those elements will figure prominently in the folk notion of how language is distributed in space. We expect this sort of variability in linguistic performance; why would it not appear in linguistic regard and, in this case, in the folk realizations of language as it is distributed in space?

11. Language, space and the folk

7. Appendix

197

198

II. Linguistic approaches to space

8. Reerences Benson, Erica J. 2003 Folk linguistic perceptions and the mapping of dialect boundaries. American Speech 78(3): 307⫺330. Clopper, Cynthia G. and David B. Pisoni 2004 Some acoustic cues for the perceptual categorization of American English regional dialects. Journal of Phonetics 32: 111⫺140. Daan, Jo 1969 [1999] Dialekten. In: Jo Daan and D. P. Blok, Von Randstad tot Landrand, 7⫺43. (Bijdragen en Mededelingen der Dialecten Commissie van de Koninklijke Nederlandse Akademie van Wetenschappen te Amsterdam 37.) Amsterdam: Noord-Hollandsche U.M. [English translation: Dialects. In: Preston (ed.), 9⫺30.] Dailey-O’Cain, Jennifer 1999 The perception of post-unification German regional speech. In: Preston (ed.): 227⫺242. Demirci, Mahide 2002 Gender differences in the perception of Turkish regional dialects. In: Long and Preston (eds.), 41⫺50. Demirci, Mahide and Brian Kleiner 1999 The perception of Turkish dialects. In: Preston (ed.): 263⫺281. Diercks, Willy 1988 [2002] Mental maps. Linguistisch-geographische Konzepte. Zeitschrift für Dialektologie und Linguistik 55(3): 280⫺305. [English translation: Mental maps: Linguistic-geographic concepts. In: Long and Preston (ed.), 51⫺70.] Evans, Betsy E. 2001 Dialect contact and the northern cities shift in Ypsilanti, Michigan. Unpublished PhD dissertation, Michigan State University. Evans, Betsy E. 2002 Attitudes of Montreal students to varieties of French. In: Long and Preston (eds.), 71⫺93. Fought, Carmen 2002 California students’ perceptions of, you know, regions and dialects? In: Long and Preston (eds.), 113⫺134. Fridland, Valerie, Kathryn Bartlett and Roger Kreuz 2004 Do you hear what I hear? Experimental measurement of the perceptual salience of acoustically manipulated vowel variants by Southern speakers in Memphis, TN. Language Variation and Change 16(1): 1⫺15. Giles, Howard and Richard Y. Bourhis 1976 Voice and racial categorization in Britain. Communication Monographs 43: 108⫺114. Goeman, A. C. M. 1989 [1999] Dialectes et jugements subjectifs des locuteurs. Quelques remarques de me´thode a propos d’une controverse. Grenoble: ELLUG, Universite´ Stendhal ⫺ Grenoble 3. [English translation: Dialects and the subjective judgments of speakers: Remarks on controversial methods. In: Preston (ed.), 135⫺144.] Goeman, A. C. M. 2002 Perception of dialect difference: Standard and dialect in relation to new data on Dutch varieties. In: Long and Preston (eds.), 135⫺149. Gould, Peter and Rodney White 1974 Mental Maps. Harmondsworth: Penguin. Graff, David, William Labov and Wendell Harris 1986 Testing listeners’ reactions to phonological markers of ethnic identity. In: David Sankoff (ed.), Diversity and Diachrony, 45⫺58. (Current Issues in Linguistic Theory 53.) Amsterdam: Benjamins.

11. Language, space and the folk Grootaers, Willem 1964 [1999] La discussion autour des frontie`res dialectales subjectives. Orbis 13, 380⫺398. [English translation: The discussion surrounding the subjective boundaries of dialects. In: Preston (ed.), 115⫺129.] Hartley, Laura C. 1999 A view from the west: Perceptions of U.S. dialects by Oregon residents. In: Preston (ed.), 315⫺332. Hartley, Laura C. 2005 The consequences of conflicting stereotypes: Bostonian perceptions of U.S. dialects. American Speech 80(4): 388⫺405. Inoue, Fumio 1995 [1999] Classification of dialects by image ⫺ English and Japanese. In: Wolfgang Viereck (ed.), Verhandlungen des Internationalen Dialektologenkongresses, Bamberg, 29. 7.⫺4. 8. 1990, 75⫺77 and 355⫺368. Stuttgart: Steiner. [Reprinted in Preston (ed.), 147⫺159.] Inoue, Fumio 1996 [1999] Subjective dialect division in Great Britain. American Speech 71(2): 142⫺161. [Reprinted in Preston (ed.), 161⫺176.] Kremer, Ludger 1984 [1999] Die niederländisch-deutsche Staatgrenze als subjektive Dialektgrenze. In: Grenzen en Grensproblemen. Een Bundel Studies Uitgegeven door het Nedersaksisch Instituut van de R. U. Groningen ter Gelegenheid van zijn 30-jarig Bestaan, 76⫺83. (Nedersaksische Studies 7 / Driemaandelikse Bladen 36.) [English translation: The Netherlands-German border as a subjective dialect boundary. In: Preston (ed.), 31⫺36.] Kuiper, Lawrence 1999 Variation and the norm: Parisian perceptions of regional French. In: Preston (ed.): 243⫺262. Lance, Donald M. 1999 Regional variation in subjective dialect divisions in the United States. In: Preston (ed.): 283⫺314. L’Eplattenier-Saugy, Caroline 2002 A perceptual dialect study of French in Switzerland. In: Long and Preston (eds.), 351⫺ 365. Long, Daniel 1999 Geographical perceptions of Japanese dialect regions. In: Preston (ed.): 177⫺198. Long, Daniel and Dennis R. Preston (eds.) 2002 Handbook of Perceptual Dialectology, vol. 2. Amsterdam: Benjamins. Long, Daniel and Young-Cheol Yim 2002 Regional differences in the perception of Korean dialects. In: Long and Preston (eds.), 249⫺275. Mase, Yoshio 1964 [1999] Hoˆgen ishiki to hoˆgen kukaku. In: Misao To¯jo¯ (ed.), Nihon hoˆgen kenkyuˆkai, 270⫺ 302. Tokyo: Tokyodo. [English translation: Dialect consciousness and dialect divisions. In: Preston (ed.), 71⫺99.] Montgomery, Christopher 2007 Northern English dialects: A perceptual approach. Unpublished PhD dissertation, University of Sheffield. Moreno Ferna´ndez, Juliana and Francisco Moreno Ferna´ndez 2002 Madrid perceptions of regional varieties in Spain. In: Long and Preston (eds.), 295⫺320. Nomoto, Kikuo 1963 [1999] Kotoba no ishiki no kyoˆkai to jissai no kyoˆkai. Jinruikagaku [Anthropological Sciences] 15, 271⫺281. [English translation: Consciousness of linguistic boundaries and actual linguistic boundaries. In: Preston (ed.), 63⫺69.]

199

200

II. Linguistic approaches to space Plichta, Bartłomiej, Brad Rakerd and Dennis R. Preston 2005 The /ay/s have it. In: Tore Kristiansen, Nikolas Coupland and Peter Garrett (eds.), Acta Linguistica Hafniensia 37(9): 107⫺130. Preston, Dennis R. 1981 Perceptual dialectology: Mental maps of United States dialects from a Hawai’ian perspective (summary). In: Henry Warkentyne (ed.), Methods IV / Me´thodes IV, 192⫺198. (Papers from the 4th International Conference on Methods in Dialectology.) Victoria, BC: University of Victoria. Preston, Dennis R. 1982 Perceptual dialectology: Mental maps of United States dialects from a Hawai’ian perspective. In: Dennis R. Preston (ed.), Working Papers in Linguistics 14(2): 5⫺49. [Department of Linguistics, University of Hawai’i.] Preston, Dennis R. 1986 Five visions of America. Language in Society 15(2): 221⫺240. Preston, Dennis R. 1988 Methods in the study of dialect perception. In: Alan Thomas (ed.), Methods in Dialectology, 373⫺395. Clevedon/Avon/Philadelphia: Multilingual Matters. Preston, Dennis R. 1989 Perceptual Dialectology. Dordrecht: Foris. Preston, Dennis R. 1996 Where the worst English is spoken. In: Edgar Schneider (ed.), Focus on the USA, 297⫺ 360. Amsterdam: Benjamins. Preston, Dennis R. (ed.) 1999 A Handbook of Perceptual Dialectology, vol.1. Amsterdam: Benjamins. Preston, Dennis R. and George M. Howe 1987 Computerized generalizations of mental dialect maps. In: Keith M. Denning, Sharon Inkelas, Faye C. McNair-Knox and John R. Rickford (eds.), Variation in Language: NWAV-XV at Stanford, 361⫺378. Stanford: Department of Linguistics, Stanford University. Purnell, Thomas, William Idsardi and John Baugh 1999 Perceptual and phonetic experiments on American English dialect identification. Journal of Language and Social Psychology 18(1): 10⫺30. Rensink, W. G. 1955 [1999] Dialectindeling naar opgaven van medewerkers. Mededelingen der Centrale Commissie voor Onderzoek van het Nederlandse Volkseigen 7: 20⫺23. [English translation: Informant classification of dialects. In: Preston (ed.), 3⫺7.] Shuy, Roger W. and Ralph W. Fasold 1973 Language Attitudes: Current Trends and Prospects. Washington, DC: Georgetown University Press. Sibata, Takesi 1959 [1999] Hoˆgen kyoˆkai no ishiki. Gengo Kenkyuˆ 36, 1⫺30. [English translation: Consciousness of dialect boundaries. In: Preston (ed.), 39⫺62.] Stanford, James 2007 Dialect contact and identity: A case study of exogamous Sui clans. Unpublished PhD dissertation, East Lansing, MI: Michigan State University. Tamasi, Susan L. 2003 Cognitive patterns of linguistic perceptions. Unpublished PhD dissertation, University of Georgia. Trudgill, Peter 1972 Sex, covert prestige and linguistic change in the urban British English of Norwich. Language in Society 1(2): 179⫺195.

12. Language and space: The linguistic dynamics approach Van Bezooijen, Rene´e and Charlotte Gooskens 1999 Identification of language varieties: The contribution of different linguistic levels. Journal of Language and Social Psychology 18: 31⫺48. Weijnen, Antonius A. 1946 De grenzen tussen de Oost-Noordbrabantse dialecten onderling [The borders between the dialects of eastern North Brabant]. In: Antonius A. Weijnen, J. M. Renders, and Jac van Ginneken (eds.), Oost-Noordbrabantse Dialectproblemen [Eastern North Brabant Dialect Problems], 1⫺15. (Bijdragen en Mededelingen der Dialectencommissie van de Koninklijke Nederlandse Akademie van Wettenschappen te Amsterdam 8.) Amsterdam: Noord Hollandsche Uitgevers. Weijnen, Antonius A. 1968 [1999] Zum Wert subjektiver Dialektgrenzen, Lingua 21: 594⫺596. [English translation: On the value of subjective dialect boundaries. In: Preston (ed.), 131⫺133.] Williams, Angie, Peter Garrett, and Nikolas Coupland 1999 Dialect recognition. In: Preston (ed.), 345⫺358. Amsterdam: Benjamins. Willems, Pieter 1886 De enqueˆte werd gehouden in 1886, de antwoorden zijn het eigendom van de Koninklikke Vlaamsche Academie voor Taal- en Letterkunde, en worden daar bewaard [The inquiry was done in 1886; the responses are the property of the Royal Flemish Academy of Languages and Literature in Ghent where they are preserved]. Microcopies are held at the Institute of Dialectology and Phonetics in Leuven, the Catholic University Nijmegen and the P. J. Meertens Institute Amsterdam.

Dennis R. Preston, Oklahoma State (USA)

12. Language and space: The linguistic dynamics approach 1. 2. 3. 4. 5.

Introduction Spatially distributed language as a research laboratory for a theory of language (change) Theoretical consequences A rudimentary explanation of twentieth-century dialect dynamics References

1. Introduction The research goal of the linguistic dynamics approach can best be presented in contradistinction to that of classical dialectology. Traditional dialectology was shaped by an axiom of dissolution: the dialect change observable from the outset was interpreted as a process of rapid disappearance as a result of the increasing prevalence of standard varieties. Research focused on the reconstruction of dialects, assumed to have been stable for centuries before the dissolution process set in. Accordingly, the research goal was the

201

202

II. Linguistic approaches to space description of the spatial distribution of the oldest and least standard forms of language in isolation and its explanation in terms of historically remote processes. In contrast, the primary research goal of the linguistic dynamics approach is the precise analysis of the dynamic processes at work within complex language systems and their explanation in terms of individual cognitive and interactive-cum-communicative factors. The data basis consists of directly observable empirical processes of stabilization or modification. For a (linguistic) dynamic dialectology, this means that the continuing development dialects have shown since they began to be studied is not a source of interference in the reconstruction of a putative “initial state”, but rather an ideal source of detailed insight into the processes acting on data samples distributed across space (both groups of speakers and varieties) and time. The same is true for variation within a dialect at a particular point in time. Concurrent variants do not obscure an idealized homogeneous variety; they need instead to be analyzed as an integral aspect of the constant varietal flux. Shifting the research goal necessitates a modification of the observational basis in line with Labov’s (1994) requirement that apparent-time and real-time analyses be linked. Within the linguistic dynamics approach, a research tool has been developed to this end: the dynamic language atlas, which allows the accurate tracking of selected data on the linguistic competence and performance of various groups of speakers across space and time. This new “research laboratory for a theory of language (change)” is introduced in section 2. It can be located in a tradition reaching back to neogrammarian linguistics, in which extensive dialectological studies are recruited to empirically test and further develop theorems of language change. Section 2.1 presents the new research possibilities opened up by such broad-scale real-time analyses and the opportunities for validation that arise when the data presented in dialect atlases are systematically related to more comprehensive surveys and direct performance records (village and regional grammars, sound recordings) and modern language variation studies. Section 2.2 aims to make visible the theoretical challenges raised by this new research tool. To this end, four surprisingly clear-cut preliminary findings from the to-date most advanced dynamic language atlas (the Digital Wenker Atlas) are presented. They reveal completely unexpected developments which require a fundamental readjustment of the explanatory schema. Section 3 describes how the theoretical consequences of these findings have been elaborated under the rubric of the “linguistic dynamics approach” by the Marburg research team (cf. Schmidt 2005a, 2005b, 2005c, 2008; Schmidt and Herrgen, to appear; Herrgen 2006; Lameli 2004; Lenz 2003; Rabanus 2008; Kehrein 2006, 2008). The theoretical core is the concept of synchronization (section 3.1), developed in response to the Saussurean idea of synchronie in light of Hermann Paul’s model of dynamic individual competence. The root cause of linguistic dynamics is seen as the constant calibration of an individual’s language competence through linguistic interaction (i. e., language performance), leading to either a stabilization or modification of the participants’ linguistic competence. Mesosynchronizations, in which groups of speakers in direct contact with one another construct common linguistic knowledge over extended periods of time, are the decisive factor in the dynamic of regional varieties. The second pillar of the approach is the linguistic dynamic understanding of variety (section 3.2). This concept distinguishes between elements of linguistic knowledge that an individual can easily overrule within a specific synchronization act and those elements modifiable only through phases of active learning, i. e., usually only in the course of generational change. The latter

12. Language and space: The linguistic dynamics approach include the sign-generating (phonological, prosodic) and combining (morphosyntactic) rules that form the foundations of linguistic competence. They are also the basis of full varieties (Vollvarietäten) like the dialects and regiolects. The theory’s third stanchion is the concept of a modern regional language (section 3.3). It is impossible to study the development of a dialect in isolation without ignoring the fact that dialects have long been an integral part of a dual variety formation and that all twentieth-century and contemporary dialect speakers from countries where a well-developed school system provides instruction in a standard variety have bivarietal competence (in dialect and regiolect). Alongside the well-documented horizontal dialect formations (of coexistent base dialects; Dialektverbände), sketchily studied vertical regional language formations have emerged in the past 200 years or so. Their historical origins and varietal structure are explored here. In a closing section it will be shown how the unexpected findings from the new “research laboratory” can be explained with the help of these elements of the linguistic dynamics theory.

2. Spatially distributed language as a research laboratory or a theory o language (change) One hundred and thirty years ago, the neogrammarians recognized that the meticulous examination of an individual dialect (as in the village monographs, cf. Winteler 1876) represented an outstanding resource for the development and validation of theories of language change and hence of language itself. At roughly the same time, the initial findings of dialect geography (Wenker 1878) were rapidly proving a theoretical challenge par excellence to the contemporary account of language change (cf. the highly informative overview in Auer 2005). Early dialect geography challenged the neogrammarian attempt to explain change in isolated varieties in terms of internal (sound laws, chain shifts) and cognitive factors (“analogy”; cf., respectively, Paul 1886: 46⫺65, 85⫺98; Pfalz 1918; the overview in Murray in this volume) with its finding that “every word has a history of its own” (Malkiel 1967; cf. Wenker 1889: 22⫺23) and developed an alternative attempt to account for the palpably observable distribution of linguistic phenomena in space through external (sociodemographic/interactive) factors. In so doing, correlations between (bundles of) isoglosses and geographical, historicopolitical and denominational borders were explained as the ultimate effects of former barriers to communication or “intercourse” (Verkehrsgrenzen; cf. Wrede 1903: 30⫺35; Frings 1924: 8). To this day, the central questions which arose continue, in modified forms, to dominate theories of language change. This is true of the old conundrum of whether a sound change equally affects all the words of a variety in the same phonological context (“Does sound change proceed one word at a time. Or does it change phonemes as a whole?” asks Labov [1994: 502]). But it is equally true of the far more basic issue of the nature of the relationship between internal and external factors. Can a simple allocation of these to the distinct stages of every language change (emergence vs. spread of an innovation) be made, as Croft seems to assume? (“The phonetic and conceptual factors […] are responsible only for innovation, and social factors provide a selection mechanism for propagation” [Croft 2000: 39].) Or does a causal nexus of function and structure characterize the process by which linguistic innovations become established under the agency

203

204

II. Linguistic approaches to space of variation and selection (as in the theory of diachronic adaptation, cf. Haspelmath 1998; see also Andersen 1973 on abductive change)? The opportunities to clarify these fundamental questions on the basis of accurate empirical data have improved dramatically since the neogrammarian era. Whilst the individual “living” dialect, (apparently) independent of written language, could be viewed as a test bed for a theory of language (change) in 1876 and, much later, Moulton (1962: 25) could answer his own question of whether “there is any laboratory in which we can test the working hypotheses of structural linguistics” with the suggestion “that such a laboratory exists in the material of a linguistic atlas”, we have at our disposal today, for the first time in the history of linguistics, the opportunity to accurately trace the development of language in time and space using empirical data. This new, much improved “research laboratory” is introduced in this section.

2.1. The new research laboratory: Dynamic language atlases as broad-scale real-time studies The great dialect atlases of classical dialectology mapped (usually word-for-word) the areal distribution of samples of sectors of the linguistic competence (for example, the phonology and morphology of the “dialect” variety) of selected speakers from particular social groups (e.g., schoolchildren or sedentary farming folk) at particular times. Wherever we have such early surveys of linguistic competence at our disposal ⫺ whether very wide ranging (e.g., Wenker’s Sprachatlas des Deutschen Reichs [surveyed around 1880], Gillie´ron and Edmont’s Atlas Linguistique de la France [1902⫺1969], or Jaberg and Jud’s Sprach- und Sachatlas Italiens und der Südschweiz [1928⫺1960]) or confined to specific regions (e.g., Kurath’s [1939⫺1943] Linguistic Atlas of New England) ⫺ we have the opportunity to track the development of linguistic competence across space and time through a fresh survey of the same sectors of competence among the descendants of the original group of speakers. In some cases, such (in part multiple) reinvestigations have indeed been made on a grand scale. For instance, because of the obvious methodological weaknesses or gaps in the early supraregional atlases (Wenker: indirect data collection from elementary school teachers using the spelling rules of the time; Gillie´ron: just 992 survey locations for all of France; Jaberg and Jud: only 405 localities surveyed for Italy and southern Switzerland), regional atlases with a dense net of survey locations have been created in France, Germany and Italy from 1950 onwards. These partly incorporate the investigative programs of the old supraregional atlases but have, at least until now, only rarely been exploited for comparative mappings. Such linguistic atlases as do not just permit a comparison over time but systematically present directly comparable linguistic data collected at different times, we refer to as dynamic language atlases (sprachdynamische Atlanten). The earliest example of such an atlas is the Schlesische Sprachatlas (Bellmann 1967), which, for the vanished German dialects of Silesia, presents data from the Wenker survey of 1880 together with sound recordings of the Wenker sentences from 1962⫺1966 in a series of paired maps, the uppermost of which (the more recent data) is reproduced on tracing paper. The most comprehensive and sophisticated dynamic language atlas to date is the Digital Wenker Atlas, published across the internet (Schmidt and Herrgen 2001⫺2009; ). Its special value for precise linguistic dynamic analyses is closely bound up with the history of German dialectology. Because the 40 Wenker sentences are the basis of the only extant survey of the entire German-speaking area (Wenker’s Sprachatlas des Deutschen Reichs), because the latter with its c. 50,000 individual locations is also the world’s densest survey of the dialects of a language, and because of intrinsic methodological flaws, the Wenker lemmas have been sampled time and again in the history of German dialectology. They have, for example, been considered in most dialect monographs (since 1890) and included in extensive sound recordings from 1950 on; since 1970 they have formed an element of the German regional linguistic atlases, including the post-1980 polydimensional regional atlases that surveyed differing social groups; and they are often still an element in the extensive investigative armory of contemporary language variation studies. As a rule, the Digital Wenker Atlas presents the data collected at different times from a range of social groups in a manner which facilitates direct comparisons (cf. Kehrein, Lameli and Nickel 2005): the historical Wenker maps can be called up over the internet, the original survey forms can be viewed, more recent sound recordings containing the mapped lemmas can be played, and the corresponding maps from modern regional atlases can be overlaid semi-transparently. For all of the other data sets (traditional village monographs, dialect-geographical investigations, contemporary language variation studies) the internet publication offers georeferenced research aids. The enhanced research opportunities offered by this new geolinguistic laboratory are manifest: (a) real-time analysis of language (data segments) across space, (b) combination/cross-linkage of these data segments with thorough descriptions of the phonological and morphological subsystems of a variety at various points in time, and (c) combination/cross-linkage of the dialect variety data segments with a full description of the vertical dimension of a language across space (i. e., dialect⫺standard variation). Re (a): Dynamic language atlases’ most obvious advantage is the possibility of genuine real-time analysis they afford (see Labov 1994: 73⫺112 on the fundamental problems of real-time analyses, usually of a tightly constrained area). Traditional, unidimensional dialect atlases represent a sample of linguistic data at a particular “point in time” (or better, during a survey period). From a diachronic perspective, such a synchronic state represents a “frozen random sample” of the development of a language across space and time. Naturally, there have been (in part laborious) attempts to interpret the observable spatial distribution of the linguistic data as the result of antecedent processes of change. This can proceed in a relatively simple manner, as for instance when traditional dialect geography interprets wedge formations on maps as “vanguards of innovation” or marginal strips as “relicts” and “explains” them with the presumed prestige of historical language bearers and cultural currents (Kulturströmungen; cf. Schrambke in this volume). It can also proceed with virtuosity, as when the selection and adaptation (of the variants) of various innovative impulses is deduced from their functionality in comparison to the linguistic structure of antecedent systems (cf., for example, Seiler 2003 or Haas in this volume). The methodological problem remains the same: the assumptions regarding previous language states remain uncertain. As a rule, they can only be supported by isolated findings from older written sources or assumptions about historical frames of reference or be based on inferences from the synchronic data of an atlas. At best, a relative chronology can be postulated. In light of this problem, more recent geolinguistics has attempted to harness the sociolinguistic apparent-time approach (cf.

205

206

II. Linguistic approaches to space Labov 1994: 43⫺72) to the investigation of language change across space. For instance, for those regions of South America characterized by high mobility and intensive language contact, the pluridimensional language atlas has been developed (cf. Atlas Lingüı´stico Diato´pico y Diastra´tico del Uruguay [Thun and Elizaincı´n 2000⫺]). In contrast to unidimensional dialect atlases, the catalog of questions and the number of locations are reined in radically and phenomena believed to be currently subject to change are especially targeted. This makes it possible to extend the survey of linguistic competence to cover various social attributes (age, gender, education, mobility, etc.) and to collect performance data from these speakers in a range of contexts (interview, reading aloud, spontaneous material). The goal is to be able to analyze the sociocontextual and spatial distribution of data as a series of simultaneous windows onto a complex process unfolding in time (cf. Thun 2000: 83). Of course, the fundamental methodological problem for any apparent-time approach remains unresolved: synchronic variability is interpreted as change across time despite having no systematic control over the temporal dimension. With a dynamic language atlas, language change processes can be tracked precisely through space and time for various speaker types and contexts. Since these processes are observable through empirical spoken language data from a particular time, correlations with contemporaneous external factors can be studied exactly. Even more significant is the fact that the linguistic structure of antecedent systems, within the framework of which functional adaptations occur through language change processes, does not need to be posited, but is known from empirical observations. Re (b): Since historical language atlases always represent small samples of the linguistic competence of particular groups of speakers of a variety, knowledge of the full set, that is, the complete prosodic/phonological and morphological structure of the historical dialect, is of decisive significance in any linguistic dynamic analysis. This is exactly what is described in the phonetically precise village monographs and early dialect geographical studies, which also include a virtually complete catalog of the correspondences between phonemes and lexemes, described in terms of historical reference systems. In the German language area with its extremely dense net of such monographs, they also furnish a vital tool for validating Wenker’s survey method using (nearly) contemporaneous data, particularly his imprecise transcriptions. Re (c): Only when the vertical dimension of linguistic variation can be included is a full analysis of fundamental language change processes possible. It is not widely recognized that, to a degree, this can even be done for past language states. In Germany, socially differentiated sound recordings of dialect speakers began being made in 1922, capturing the “individual languages of the same location, of people of varying age, gender and status” (Wagner 1924⫺1925: 230; my translation). In the fifties, Zwirner made 5000 recordings of the “everyday language” of various social groups (cf. Zwirner, Maack and Bethge 1956). In the absence of appropriate analytic tools, these recordings have, however, barely been evaluated, but the opportunity for retrospective analyses with upto-date methods now exists. Far more imperative, however, is the opportunity to combine a systematic investigation of a language’s variation spectrum with an analysis of the material furnished by a dynamic language atlas. Any modern investigation of the overall variation structure of language areas needs to be designed to also function as the endpoint of a true panel (cf. Labov 1994: 76⫺77), i. e., to represent a fresh survey of the same data sample effectively tracked through time by a dynamic language atlas (cf. Elmentaler 2006 and, on the “REDE” project, and Kehrein 2008).

12. Language and space: The linguistic dynamics approach

2.2. Preliminary indings rom the new laboratory as a challenge or language (change) theory With the help of the Digital Wenker Atlas, the development of the German dialects over almost exactly a century can thus be closely followed, albeit for a restricted sample of linguistic data (viz., the lemmas of the 40 Wenker sentences) but a very large language area. For large parts of the Middle and Upper German language areas, regional atlases, whose data date from between 1970 and 1990, i. e., about a century after Wenker (1880), are available. Intermediary stages are documented in a relatively dense net of monographic studies (from between 1890 and 1960) and sound recordings (post-1950 in high quality). It will be some time until the final results of this first large-scale attempt at a linguistic dynamic real-time analysis are available. But the potential that these analyses harbor and the theoretical challenge they pose are already apparent in the preliminary findings which have emerged so far: despite the dramatic social upheavals of the twentieth century in Germany (the transformation from an agrarian to an industrial to a relatively mobile service-based society), there is no reason to believe in a rapid decline in the currency of dialects between 1880 and 1980. Overall, the situation is one of relative stability (cf. Herrgen in this volume). Far more fascinating is, however, the fact that while the dialects of an area have remained completely unchanged for over a century with regard to some phenomena, the repeatedly postulated convergence on the standard language can in fact be observed for other phenomena, and ⫺ most surprising of all ⫺ that yet other dialectal phenomena are diverging from both the standard and the dialect! Naturally, the theoretical challenge lies in explaining why the individual cases have developed so differently. The spatial distribution of the various developments also offers new potential insights. This is illustrated in the following in a simplified and in part schematized form. It should be made clear in advance that the analyses to date show that there is an absolutely dominant external factor driving dialect change in Germany in the twentieth century. It is not, as is widely expected, the familiarity of all speakers with the written and spoken standard. Rather, it is ⫺ as a result of broader social transformations ⫺ the supplantation of the local (locality-based) network of communicative relationships still dominant in the nineteenth century by the thoroughgoing regionalization of such interaction. (On this “supralocalization” or “de-localization” see the articles by Britain and Mæhlum in this volume, as well as Schmidt and Herrgen, to appear.)

Type 1: Absolute stability over a century Absolute constancy over a century without any variation in either language area, A or B, is primarily observed when the isogloss separating the two variants coincides with the historical boundary between two dialect formations (Dialektverbände). An example is southern was versus northern wat (‘what’) in West Middle German. This seems easy to explain with existing language change theories. The theoretical challenge first becomes clear when closer examination of an actual case reveals (1) that all speakers in areas A and B have active mastery of both variants (a is also the Standard German variant, b the dialectal variant), (2) that no barriers to intercourse (Verkehrsgrenzen) or other external

207

208

II. Linguistic approaches to space

Fig. 12.1: Absolute stability over a century (Type 1)

distinctions currently separate language areas A and B, and (3) that a and b are linguistically marginal (single-word) opposites, completely detached from the phonological structure of the dialects in A and B (cf. Schmidt 2005c: 24⫺30).

Type 2: Development away rom both the standard language and the old dialect Type 2 is illustrated by a verbal inflection example (cf. Schmidt and Herrgen, to appear: 4.2.1). In Standard German, bringen ‘to bring’ is neither a strong (class b, e.g., singen⫺ sang⫺gesungen ‘SING’) nor a weak verb (class c, e.g., sagen⫺sagte⫺gesagt ‘SAY’), but rather a rare case of mixed inflexion (bringen⫺brachte⫺gebracht; class a). Ongoing psycholinguistic studies (cf. Niedeggen-Bartke 2002, described in Schmidt and Herrgen, to appear: 4.2.1) show that German children acquire the past participle (gebracht) very late and that before the age of four, the verb is usually inflected following the rules for class

Fig. 12.2: Development away from both the standard language and the old dialect (Type 2)

12. Language and space: The linguistic dynamics approach c (*gebringt) or class b (*gebrungen). Such innovation thus arises spontaneously in the normal course of child language acquisition in line with the principle of analogy, well explored since the neogrammarian era. In dialect area B in 1880, the a and b forms stood juxtaposed, a as the old, standardconform dialect form, b a highly salient dialect form. In four sets of data from three different periods (1880, 1939, older speakers in 1980, younger speakers in 1980), the spread of the innovation can be traced very clearly. It can be empirically demonstrated that form b “spread out” in stages and ultimately took over the entire dialect area B (Rhine Franconian left of the Rhine). In that this innovation repeatedly recurs, it can be seen as a “natural” language change, upon closer investigation surprising only for the fact that the clear standard language norm and the negative evaluation of the variant have not hindered its spread. It is much more of an issue to explain why the innovation does not spread to the neighboring areas of A and C. These differ from B primarily in their phonological structure (they are “structural regions”, see section 3.3); the morphological structure in the innovation zone in B generally corresponds to that of A and C. Hence, the linguistic preconditions for a functional adaptation were identical. Nor can the differential effects of any external factors be discerned ⫺ here too, both variants are well known to all the speakers in all of the areas under observation.

Type 3: Broad-scale transormation o the phonological structure o dialect areas The examples for Types 1 and 2 depict relatively isolated phenomena, for which the ongoing influence of the historical boundaries between dialect formations or “structural boundaries” within them present explanatory challenges for linguistic theory. Type 3 involves broad-scale phonological change and shifts in the phonological boundaries between former dialect formations. One example is the change in the old dialectal diph-

Fig. 12.3: Broad-scale transformation of the phonological structure of dialect areas (Type 3)

209

210

II. Linguistic approaches to space thongal phoneme corresponding to MHG /o:/ (e.g., /ou/, /au/ or /oa/ in words like groß ‘big’, Rose ‘rose’, Loos ‘sow (female pig)’, rot ‘red’, tot ‘dead’, Brot ‘bread’, hoch ‘high’) in six different Middle and Upper German dialect areas (cf. Schmidt and Herrgen, to appear: 4.2.3). In 1880, all of the individual isoglosses for such words coincide in some of the dialect areas (e.g., area A in the schematic map shown as Figure 12.3). In these cases (e.g., in the southern Palatinate), there is no language change to be seen, even 100 years later: the dialect areas are stable, their phonological structure has not changed, the isoglosses for the individual lemmas are still superposed. The situation is completely different in dialect areas B and C, where in 1880 the isoglosses for the various words did not coincide. Here, in every case we have examined, the individual lemma boundaries have shifted over the course of the past century. In most cases, the erstwhile diphthong phoneme has been replaced by whichever former dialectal phoneme is phonetically most similar to the phonemes of the surrounding dialect areas (e.g., /oa/ in Bavarian Swabia replaced by /ou/, /ou / or /o:/). The old diphthong then either becomes a single lemma relict (area B, e.g., western Upper Bavarian) or vanishes completely (area C, e.g., the Bavarian Swabian transition zone). The clarity of this finding is striking. (1) For phonologically relevant phenomena (i. e., phonemic splits and mergers), the old aphorism “every word has a history of its own” implies that a very long term (in this instance phonological) change is in progress. A long-standing language change process has not yet been completed. (2) The phonological transformation proceeds via a lexeme-for-lexeme phonological redistribution. It is equally clear that, where the individual lemma boundaries of a phoneme in two different periods coincide, a (more recent) language change process has not yet set in.

Type 4: The role o unctionality in structural change in dialectal subsystems A further important early finding out of the new linguistic laboratory can be described without recourse to a map. In Rabanus (2008), changes in the inflectional morphology of person, number and case was investigated in 43 small-scale areas. German pronominal and verbal inflection has long been characterized by increasing syncretism (the collapse of inflective categories). The one-hundred-year comparison shows how this long-term change has unfolded in specific cases. For instance, there is total case syncretism of second-person plural pronouns in certain North Bavarian dialects ⫺ in 1880, nom IHR, acc/dat ENK; in 2000: nom/acc/dat ENKS. In other dialects the verbal inflexion has syncretized, e.g., beißen ‘to bite’ ⫺ in 1880: 3sg BAES-T, 2pl BAES-TS; in 2000, 3sg/ 2pl BAES-T. The astoundingly clear upshot of this very extensive analysis is that there must be an internal factor, for which the boundaries of the linguistic sub-areas are irrelevant, that acts to limit or stop the morphological change. It emerges that there is a functional lower limit to the spread of syncretism: the retention of a distinction when decoding semantic roles in minimal sentence pairs. Rabanus employs the methodological tool of the “minimal sentence”, in which the syntactic slots for predicate, subject and object(s) are filled with just the inflected verb plus pronouns. Through the interplay of the verbal and pronominal inflectional categories he can then show how preserving the functional lower limit controls language change (Rabanus 2008: 260⫺271). For the cho-

12. Language and space: The linguistic dynamics approach sen example, this means that complete case syncretism of second-person pronouns (i. e., ENKS in all cases) has not developed in a single village dialect in which verbal inflection syncretism also occurs ⫺ that would have led to ambiguous minimal sentence pairs of the type ENKS BAES-T DEI, in which case cannot be unequivocally assigned (Ihr beißt sie [2pl.nom bite.2pl 3sg.fem.acc] ‘You bite her’ or Euch beißt sie [2pl.acc bite.3sg 3sg.fem.nom] ‘She bites you’).

3. Theoretical implications The implications of these findings for a theory of language (change) have been explored by the Deutscher Sprachatlas team in Marburg in numerous publications under the rubric of the linguistic dynamics approach (LDA; cf. section 1). Linguistic dynamics (Sprachdynamik) is defined as the study of the influences acting on the constantly shifting complex of language and the resultant modifying and stabilizing processes. The starting point is the assumption that linguistic changes arise because (groups of) speakers in interaction with other (groups of) speakers, who have other lingual system and register competencies, apply cognitive, usually unconscious, optimization strategies in keeping with their communicative goals. The key theoretical gambits are fourfold: (1) supplanting the Saussurean dichotomy between synchrony and diachrony with the concept of synchronization; (2) resorting to the idea of individual competence as the basis for theoretical explication in linguistics; (3) propounding a simple, purely linguistic concept of variety that can discriminate between full and sectoral varieties; and (4) setting out a structural analysis of the entire dynamic language system which posits several clearly distinguishable vertical variety formations (regional languages) below the level of the standard language and its norms.

3.1. The concept o synchronization The concepts of system and of structure were introduced into linguistic theory in 1867 (cf. Whitney 1867: 50) and taken up by the neogrammarians (cf. Sievers 1876: 3⫺4). Not many people are aware that the famous charge of atomism was in the nineteenth century directed exclusively against the pre-neogrammarian theory of language, especially against August Schleicher (cf. Paul 1877: 322 and Gabelentz 1888 [1966: 258]). It was clear in the neogrammarian theory of language that the analysis of linguistic change had to involve the analysis of (synchronic) language states (see Reis 1978: 180⫺183 on Paul’s comprehensive concept of Sprachgeschichte, which included both synchrony and diachrony). It was Saussure who stridently rejected such an integrative analysis. Based on the justifiable assumption that systemic relationships can only pertain between elements that coexist (1916: 227), as is true of the individual competence vis-a`-vis a single specific act of language production, but not of langue, he believed he needed to demand a strict separation of synchrony and diachrony “at any price” (cf. the “panchronie” polemic, Saussure 1916: 226, 2003: 105⫺109, 154⫺155). In hindsight it is fascinating to reconstruct how prodigiously productive the “Copernican revolution in linguistics” (Verburg 1950: 224) associated with Saussure’s name has been in the history of the discipline,

211

212

II. Linguistic approaches to space even though it was very soon and very prominently made clear (cf. Schuchardt 1922 [1917]: 265⫺267), that the strict separation of synchrony and diachrony fell short of its intrinsically heterogeneous and constantly changing object, language ⫺ “actual synchrony is dynamic” (Jakobson 1980 [1962: 53]). Language is not a semidynamic system, as suggested by Saussure’s famous example of the chess game. In a semidynamic game of chess there are indeed static states which can be analyzed completely independently of their emergence (i. e., they are temporally isolated, atemporal). In contrast, truly dynamic systems like language, the solar system or the stock market are constituted by the time-dependent relations between their elements and can therefore not be adequately analyzed without involving the time dimension. The linguistic dynamics approach builds on Paul’s concept of dynamic individual competence (1880: 27⫺39 [1886: 21⫺34]; cf. Schmidt 1993: 182⫺185). Speakers do not progress from synchronic state to synchronic state, they actively and interactively “synchronize” their complex and distinct systems of linguistic knowledge (their individual competencies). By synchronization we mean “the calibration of competence differences in the performance act”, which results in a “stabilization and/or modification of the active and passive competencies involved” (Schmidt and Herrgen, to appear: 2.1.3). What does this mean? Partners in an interaction synchronize their language so as to conform to the cooperation principle (Grice 1975), that is, the desire to be understood, or at least not misunderstood. Interactions consist of acts of language production and of language comprehension. In every act of production a speaker applies the linguistic knowledge he/she has acquired over the course of a lifetime (individual competence) to match it to his/her partner’s communicative expectations and ability to understand. He/ she anticipates their capacity to comprehend (are they a child, a tertiary educated expert, …?) and their expectations regarding linguistic appropriateness (spouse, casual acquaintance, examiner, …) and activates appropriate segments of (conscious or unconscious) linguistic knowledge that may well have been acquired long before (e.g., knowledge of regional variants learned from grandparents). The dynamic of the individual interactions is essentially a result of feedback from the conversation partner (on interactionist approaches, see Auer 1990: 26⫺28; Milroy 2002; Meyerhoff 2002; Keller 2003: 132⫺143). Has the partner signaled a complete lack of comprehension (e.g., with a questioning glance), partial comprehension (You think?), the non-fulfillment of a behavioral expectation (e.g., through implicit or explicit correction), or complete comprehension and fulfillment of my behavioral expectations (e.g., through a confirmatory signal like the particle hm)? Each type of feedback effects a modification or stabilization of the applied language production strategy. Temporality and dynamism are thus constitutive characteristics of every interaction, no matter how elementary. We are dealing with the apriority of time in linguistic interaction. Whether such modifications and stabilizations remain temporary or whether they effect a profound cognitive reflex (usually a restructuring of individual competence) depends upon how the interaction is evaluated, the interaction partner and the context in which it takes place. Correction by a caregiver or role model during the long language acquisition phase or even an embarrassing use of language can invoke an instant and ongoing restructuring, whereas a failed interaction with a little valued counterpart often tends to remain inconsequential (cf. Schmidt and Herrgen, to appear: 2.1.2). The elementary synchronization act just described (the calibration of the individual competencies within a single interaction) is referred to as microsynchronization. It underlies all of the other types.

12. Language and space: The linguistic dynamics approach To understand the dynamic of a language and its varieties we need to observe microsynchronizations and their outcomes over a period of time. The time-dependent relationships which constitute truly dynamic systems are based not only on the immanent temporality of the individual interactions but also on periods in which subjects interact with varying frequency, thereby synchronizing their linguistic knowledge and situationdependent linguistic conventions. What do we mean by this? In the course of his/her life, any individual interacts with only a (small) fraction of the total language community; even with these he/she only interacts in certain periods (phases of life); the interactions in these phases only cover a portion of the situations relevant for the individual (e.g., occupational or domestic communication); and even the communicative contexts divided among particular phases of life vary in their subjective significance. It is of central importance for a realistic picture of the dynamic of an entire language to remember that individuals participate in the totality of synchronization acts in any period, whereby they find themselves in clearly different phases with regard to their linguistic knowledge, i. e., in periods of accelerated or retarded change. The periods in which speakers learn the fundamental rules of the semiotic system of language, the periods in which they actively and deliberately conform to the (language) behavioral expectations of others, the periods in which mutual unconscious accommodation takes place, and the periods of retardation in which changes are rare or language loss can even occur, all vary constantly from one individual to the next. The system of language is thus doubly dynamic ⫺ on the basis of individual factors and of socio-interactive factors. If we want to understand the dynamic system as a whole, we have to clarify which factors lead to divergence and which to convergence. Since the cooperation principle applies to every interaction, if individuals interact over an extended period in a context that is important for all participants (cf. Eckert and McConnell-Ginet 1992; Wenger 1998), e.g., a primary school class, a youth peer group, or other “communities of practice”), this leads to a series of parallel synchronizations. The cognitive repercussions of such a series of microsynchronizations depend on the length of the period over which the communication partners mutually interact (e.g., in a marriage), the frequency of the interactions (Kommunikationsdichte), and the biographical relevance the communication partners see in the enduring collective context. Where there is sustained duration, high communicative density and high individual significance, the participants develop similar optimization strategies. The result is a partial congruence in the linguistic knowledge of the participants, more precisely: an extensive congruence in that segment of their linguistic knowledge which enables the communication partners to successfully negotiate the situation in which they are participating. Such a series of parallel acts of synchronization, performed by individuals in personal contact situations, which lead to the establishment of common context-dependent linguistic knowledge, we term mesosynchronization (Schmidt and Herrgen, to appear: 2.1.3.2). Mesosynchronization is thus always an integrative force within a circumscribed segment of the complex dynamic of the language in toto. Mesosynchronizations are responsible for the establishment of group and context⫺dependent linguistic conventions and thus, in the final instance, for the formation of varieties. In terms of the language as a whole, they can have either a divergent or an integrative influence. They are divergent to the extent that the groups which enact them develop special knowledge and integrative to the extent that the mesosynchronizations are interlinked, i. e., where similar or even parallel optimization strategies are developed in interconnected situations. It should be

213

214

II. Linguistic approaches to space clear that the concept of synchronization is the key to a unified account of linguistic divergence and convergence. Macrosynchronizations play a decisive role in integrating the language as a whole. We use macrosynchronization to refer to synchronization acts via which the members of a linguistic community orient themselves to a common norm. Practically all members of a linguistic community or a larger social grouping in which personal contact between members is not a given undertake macrosynchronizations. Viewed over the long term, the boundaries of collective macrosynchronizations define the boundaries of the dynamic system of a particular language. For its speakers, they are (identical with) the boundaries of a particular language. We must differentiate between macrosynchronizations which align with the written norm and those which align with the oral norm (the pronunciation standard). In German, a common language only arose when writers developed a mutual written norm and then aligned their macrosynchronizations with it. Conversely, a linguistic community disintegrates to the degree that its major constituent groupings begin to orient their macrosynchronizations to differing target norms. The all-embracing effects of macrosynchronization processes can be clearly observed in the development of German and Dutch. Historically, there was a continuum between the two at the base dialect level (cf. Kremer 1979: 133⫺134). The long-term orientation toward two separate standard norms (German and Dutch) led the complete variety systems to diverge, so that today there is a linguistic divide (a “sprachliche Bruchstelle”, cf. Kremer 1993: 26, 36; Smits 2007: 312⫺313) between even the least standard varieties. The beginnings of a much more recent divergence can currently be observed between German and Luxembourgish (cf. Gilles 1999: 261; see Gilles and Moulin 2003 on the standardization of Luxembourgish). Within a dynamic language as a whole, the written standard exhibits the slowest rate of change, in that here all of the “conservative forces” come to bear. The primary function of writing is to transcend the immanent volatility of oral communication (shortterm memory) and ⫺ by preserving them ⫺ to extend the period over which linguistic acts can exercise communicative effect (as in the delivery time of a letter or the reception of a novel over decades or even centuries). The codification of a written language (e.g., in dictionaries) and the institutionally backed prescriptive norms (e.g., the maintenance of orthographic and grammatical rules over decades) lead millions of individuals to make, consciously but independently of one another, parallel retroactive microsynchronizations (looking things up, learning norms), which act together to reduce the speed of change. The oral norms on which the members of major groupings base their macrosynchronizations have been distributed by the mass media (radio and television) for around 70 years. Although participation in these macrosynchronizations is primarily receptive, it can still be demonstrated to have an impact on the development of the complete variety system of a language. This is true for both the near-standard varieties (significant reduction of regionalisms in the regiolect of speakers socialized in the radio era, cf. Lameli 2004: 108⫺111) and the spoken standard itself: since the oral norms have mainly been spread by nationally organized mass media, national pronunciation norms have been able to prevail, e.g., in American vs. British English or Austrian, Swiss and German standard pronunciations, despite the integrative influence of largely congruent written norms.

12. Language and space: The linguistic dynamics approach

3.2. The linguistic dynamic concept o variety The linguistic dynamic concept of variety was introduced in Schmidt (2005b). The key point is that varieties cannot be satisfactorily delimited from “outside”, i. e., through statements about the frequency of variants even when correlated with social and contextual factors. This temporal, social and contextual continuum stands in contrast to a clear linguistic/cognitive division. While lexical and pragmatic differences can be easily overcome in acts of synchronization, the same is not true of phonological/prosodic and morphosyntactic differences, because these are the fundamental sign-generating and linking rules which form the core of an individual’s linguistic competence. Actively overcoming a competence gap here requires restructuring one’s central linguistic knowledge: the mapping of phonemes to lexemes has to be relearned sign by sign, new morphological categories must be acquired, and the assignment of syntactic functions to both these and existing categories is reorganized. Most speakers do not fully succeed in learning new prosodemes, and they are clearly aware of this linguistic/cognitive barrier. In reception mode, they react to failed attempts (such as hypercorrections, cf. Herrgen 1986) with sanctions (linguistic ridicule, shibboleths); in production mode with avoidance strategies. For an individual to overcome gaps in fundamental competence thus requires extended phases of learning. The linguistic dynamic concept of variety therefore distinguishes between full varieties, which exhibit such fundamental dissimilarities, and sectoral varieties, which do not. Full varieties can be specified as sectors of linguistic knowledge defined by independent prosodic/phonological and morphosyntactic structures on the basis of which individuals or groups of speakers interact in particular situations. The full varieties of a language are semi-discrete and interdependent. The minimal and necessary criterion is the presence of at least one “idiovarietal” element or structural feature in the prosodic/ phonological or morphosyntactic subsystems. Full varieties must be distinguished from (1) sectoral varieties such as technical jargons (Fachsprachen), which are based on a full variety (whether standard or dialect) and feature restricted, sectoral, usually lexical distinctions and substitutions and in which an individual’s competence is subject to continual, lifelong change and from (2) speech levels (Sprechlagen), which arise from conventional allophonic and allomorphic variation within a full variety in correlation with social, contextual and spatial factors. Speech levels’ lack of inherent distinguishing features means that, in practice, the relationship between linguistic and socio-contextual constellations (i. e., types) has to be specified case-by-case using statistical tests (cf. Lenz 2003: 218⫺222).

3.3. Duplex (horizontal and vertical) variety ormations: Modern regional languages We are now in a position to explore a linguistic constellation that suggests an explanation for the unexpected results from the new “research lab” of the dynamic language atlas discussed in section 2.2. This is a two-dimensional (horizontal and vertical), duplex variety formation, for which we suggest the label of modern regional language. What is meant by this can be best explained by taking as an example the language most thoroughly investigated in this regard, German. Certain horizontal variety formations have long been recognized and extensively studied, viz., the old dialect formations.

215

216

II. Linguistic approaches to space If one takes into account all of the horizontal differences between the erstwhile local dialects, including the lexical distinctions and one-off phonological contrasts, the spatial distribution pattern which emerges is a continuum interwoven by a confusion of isoglosses. The achievement of structuralist dialectology was the creation of a dialect classification scheme that did not foreground isolated phenomena, but was instead based solely on linguistic/structural criteria. The classification principle was whether there were phonological/prosodic and morphosyntactic similarities between the dialects of an area A that distinguished them from those of area B. The dialect formations so defined in German, or the structural regions within them (cf. Wiesinger 1983) thus delimit full varieties, the speakers of which possess fundamental similarities in their core linguistic competence. Vertical variety formations have been subject to little study and our linguistic understanding of them is inadequate. They arose in German between 1720 and 1800 (cf. Schmidt 2005a), as larger groups of speakers, initially in the cities, began to base their speech on the written language that had arisen out of a purely written accommodation process. These speakers could only draw on their own fundamental levels of competence when pronouncing such (combinations of) graphemes, i. e., on the dialectal phonological/prosodic structures available to them. As a consequence of congruent mesosynchronizations all oriented to the written forms, oralization conventions for the new spoken variety emerged. Under the selective principle of maximal communicative range, consensus could only develop as far as the underlying dialectal uniformity of phonological structure permitted, i. e., within the old dialect formations. With the introduction of compulsory schooling from 1800 on, these oralization conventions became norms that were propagated in literacy lessons (speaking in chorus as a synchronization act). The new oral variety is known as regional (or landschaftliches ‘landscape-based’) High German. Spoken German thus became a two-variety language, shaped by the coexistence of a large number of duplex ⫺ horizontal and vertical ⫺ variety formations: the old horizontal dialect formations each with their own, structurally anchored regional Hochdeutsch. The duplex variety formation of the nineteenth century has been retained only in German-speaking Switzerland, although of course both varieties have changed here too, thanks to the regionalization of communication (cf. Christen 1998: 292⫺294 on dialect) and the general spread of the Swiss German pronunciation norm via the broadcast media. In contrast, a completely different linguistic constellation has emerged in Germany. In 1898, on the basis of the regional High German of northern Germany (cf. Vietor 1890: 12), Siebs established a pronunciation norm for the stage (Bühnenaussprache) that has been disseminated via radio since 1930 and (in modified form) via television since the second half of the twentieth century. This means that all speakers have been subject to constant macrosynchronization with the federal German spoken standard, which has at a very minimum altered their passive competence. Outside of northern Germany this led to a southwards diffusing devaluation of the former regional High Germans, now no longer seen as equal to Standard German, but rather as colloquial, linguistically nonstandard (more precisely substandard) forms, for which we suggest the label of regiolect. Thus, in Germany too, the old duplex variety formations have generally endured. But their linguistic status and dynamics have been fundamentally altered by the prevalence of the federal German spoken standard. They are now perceived to be regionally confined: they have become regional languages (Regionalsprachen; cf. Schmidt 2009).

12. Language and space: The linguistic dynamics approach A modern regional language can be defined in the light of the historical origins of these variety formations and their current linguistic status; it is an aggregate of varieties and speech levels unified by mesosynchronizations and bounded horizontally by the structural boundaries of the dialect formations or regions and vertically by its differences to the national oralization norms of the standard variety. The term thus encompasses all full varieties and speech levels that exist beneath the standard variety. A regional language is therefore also a language and not a regional variety. The (variation) linguistic structure of the various regional languages is primarily dependent upon whether additional vertical varieties exist alongside the regiolect, defined as a supraregional nonstandard full variety, namely dialects, understood as the least standard and most local (regionally restricted) full varieties.

3.4. Structure and dynamics o the modern regional languages Contemporary spoken German is in this view determined by the horizontal coexistence of regional languages, which in turn represent vertical variety formations within which the horizontal structural boundaries of the least standard varieties (the old dialect formations) coincide with the boundaries of the regiolects. This can be schematized as in Figure 12.4. With one notable exception (cf. Kehrein 2008), no studies directly comparing the linguistic structure of the various modern regional languages (cf. Schmidt 1998) exist. The most comprehensive study to date of the variation structure of a single regional language is Lenz (2003). Service sector employment now dominates the formerly agricultural hinterland of a small Moselle-Franconian town, the region in which Lenz’s 50 speakers were raised. None of them has an active command of the spoken standard, although the younger, better educated speakers in particular take pains to achieve it. They fail at the phonological boundary between the regiolect and the Standard German oralization norm, leading to hypercorrections, but all have command of the regiolect. A comparison of regiolect performance in various constellations of speakers and contexts reveals, with the help of multivariate and implication analyses, three distinct speech levels: the regional accent, the upper nonstandard and the lower nonstandard. Linguistically, the speech levels can be differentiated by virtue of the fact that, in addition to the common phonological features that define the entire Moselle-Franconian regional language and hence also the regiolect of this subregion, lone-word regionalisms occur in

Fig. 12.4: The linguistic structure of modern regional languages

217

218

II. Linguistic approaches to space the upper nonstandard, and with greater frequency in the lower. These regionalisms are of two kinds: either they are common to all the varieties of the greater region or they are subregional allophonic deviations from the standard. Viewed sociolinguistically, it is above all the younger speakers that live in town who also make active use of the nearstandard speech levels of the regiolect, while the older speakers from the surrounding villages are proficient only in the less standard speech levels. The key linguistic criterion for mastery of a speech level is primary language socialization. Are we dealing with bivarietal speakers who were raised in dialect and only subsequently acquired the regiolect, or with monovarietal speakers, whose primary language socialization was conducted exclusively in the regiolect? The key contextual criteria for the choice of speech level are the conversation partners and their usage: in interaction with speakers of one’s own regional language, the least standard speech levels of the regiolect are employed, but when talking with standard-speaking strangers, the regional accent is adopted. The ongoing reevaluation process is well illustrated by speakers’ subjective evaluations of the various speech levels. Whilst older, more rural, bivarietal speakers still consider the lower nonstandard to be Hochdeutsch (Standard German), the younger, town-dwelling, monovarietal speakers classify even the regional accent as bad or incorrect German (“schlechtes oder nicht richtiges Hochdeutsch”). Using the same method, Lenz was able to identify two speech levels within the dialect variety, the base dialect, which still includes truly local features (older speakers), and the regional dialect, characterized by (sub)regional features (younger speakers). It surely needs no spelling out that the dynamic of the regiolect described by Lenz can be interpreted in light of the apparent-time hypothesis as an ongoing convergence on the spoken standard. The two real-time studies to date have produced contradictory results. Whereas Spiekermann (2006) observed (among other things) an increase in regional allegro speech features in the Upper German area that he interpreted as increasing regionalization of the standard, Lameli (2004: 215⫺217), working with the same type of speaker in the same situation (city council meetings) in the Middle German area, was able to establish a convergence on the standard over forty years for one group of speakers, for another group even a shift from regiolect to the colloquial standard (Kolloquialstandard ), a speech level of Standard German that, although linguistically and objectively still characterized by regionalisms, is considered standard by listeners from the most diverse regions of Germany. In a North German city in contrast, Lameli recorded absolute stability in the speech levels over a forty-year period. While there are still too few studies to allow an exploration of the dynamic of the regiolect, an investigation of dialect dynamics is already possible on the basis of the very clear findings from the linguistic research laboratory introduced in section 2, the dynamic language atlas. How such an investigation might proceed with the help of the linguistic dynamics approach is illustrated in the following section.

4. A rudimentary explanation o twentieth-century dialect dynamics By the end of the twentieth century, all German dialect speakers had acquired active bivarietal competence (in dialect and regiolect) and at least passive competence in the standard spoken language. The comprehensive regionalization of communication had

12. Language and space: The linguistic dynamics approach decisively reshaped all interaction. Of the surprising findings reported in section 2.2, those which do not concern purely internal factors (i. e., all except Type 4) cannot be explained through an isolated analysis of the dialect variety and its dynamics. An explanation only becomes possible when the structure of the modern regional languages and their constitutive mesosynchronizations are included in the analysis. Let us begin with Type 1, the juxtaposition of individual lemmas on the borders of two dialect formations that remains stable over more than a century, although barriers to communication between the regions no longer exist and all speakers have an active command of the neighboring region’s variants, or can at least read them aloud. These oppositions, which reveal no connection whatsoever to the phonological structures of the dialect areas, involve those features common to both the dialect and the less standard speech levels of the regiolect. They are thus employed in all interactions within each of the regional languages, leading to a constant stabilization via mesosynchronizations conducted in the regional language (cf. Schmidt 2005c: 25⫺28). A very similar process is involved in Type 2, the single lemma change in inflection class that, within a century, has prevailed over both Standard German and the old dialect throughout an entire area and has only been halted at the former boundary of the neighboring structural region. Since this is a “natural” language change and the innovation zone (area B) has essentially the same morphological structure as its neighboring regions, here too an isolated analysis of the dialect variety cannot explain the change. Thanks to the thoroughgoing regionalization of interaction patterns, all of the speakers of areas A, B and C are familiar with the inflectional variants which deviate from both the dialect and the standard (passive competence). The decisive difference lies in their evaluation. While the speakers in the innovation zone consider the variants to be dialectal alternatives drawn from their own regional language, this is not the case in the adjoining areas. Hence, when children acquiring dialect overgeneralize in the direction provoked by the irregular German linguistic structure, they are not corrected by adults living in B, since the forms are treated as appropriate regional language variants (i. e., stabilizing mesosynchronizations occur). The natural language change can proceed untrammeled. The situation in the neighboring areas is completely different. Here the variants, which arise in the normal course of language acquisition, whilst recognized, are evaluated as extraneous to the resident’s own regional language and hence corrected (i. e., modifying mesosynchronizations occur). The corrective behavior of those effecting linguistic socialization, typical of the entire German language area apart from B, does not change, and a language change that might be thought natural is inhibited. The explanatory potential of the linguistic dynamics approach is seen most clearly in Type 3, where broad-scale phonological change progresses via word-by-word phonological redistribution (Umphonologisierung). The concurrence of the isoglosses of a phoneme at some point in time indicates that a process of language change has not yet started; “staggered” isoglosses mean that the change is in progress at the time of the survey. But this finding is outweighed by the new opportunity to use the dynamic language atlas to very precisely determine which factors are responsible for why in certain subregions (A) no language change takes place, while in others (B and C) it very much does. The external factors of “regionalization of communication” and “familiarity with standard language forms” apply equally across all subregions. Analysis shows that it is the specific nature of the phonological differences between each of the subregions A, B and C and their adjacent dialects (cf. the “white space” on the map of Figure 12.3) which triggers

219

220

II. Linguistic approaches to space language change in one situation and not in another. Because of the specific nature of the phonological differences (the “internal” factor), using the former dialect’s phoneme system in Areas B and C in regional communication with speakers of the surrounding dialects (the external factor) leads to faulty decodings (misunderstandings or incomprehension). For example, differences in the assignment of phonemes to lexemes automatically lead to faulty decodings in the Bavarian-Swabian transition zone, where the erstwhile dialectal phoneme /oa/ (in, e.g., Brot, Rose, tot, Seife, Weizen, or Geiß) is misunderstood in the neighboring Bavarian dialects, e.g., Broat is heard as breit ‘broad’, not Brot ‘bread’ and Roas as Reise ‘journey’ not Rose ‘rose’. This instigates a constant stream of modifying synchronizations that in the longer term sets a process of language change in motion. The situation in Area A is different: the phonological differences can be surmounted during the performance act, without giving rise to errant decodings (thus producing stabilizing synchronization acts). How the lexeme-for-lexeme phonological redistribution proceeds is also extremely revealing. The end target of the change is not the equivalent phoneme from the standard language, but rather that phoneme from the original local dialectal register that shares the greatest similarity with its equivalent in the neighboring regions (cf. Schmidt and Herrgen, to appear: 4.2.3.7). The role of the standard language, in which all bivarietal speakers have passive competence, is restricted to the mental connection (via lexemes) between two old dialectal phonemes. It functions as a cognitive “bridge” and determines the “target phoneme” of the phonological shift when an erstwhile dialectal phoneme causes interactive complications in variety contact situations. The degree to and order in which lexemes assigned to an old dialectal phoneme are involved in the phonological redistribution process depend upon their social and contextual relevance. Lexemes which can potentially be implicated in any regional language interaction (e.g., Brot ‘bread’) are subject to early and complete phonological reordering, whereas lexemes whose use is confined to narrow communicative contexts (e.g., Loos ‘sow’) are the last to be rephonologized. It is fascinating, finally, to reflect on the long-term communicative advantages of successive lexeme-based phonological rearrangement. Thanks to the fact that, initially, only a few of the lexemes relevant in intragenerational regional communication are implicated in the phonological change, the old dialectal phonemic system can generally be maintained over decades, minimizing interactive complications in communication with earlier generations. Clearly, we are dealing with a perfect, albeit unconscious, optimization strategy.

5. Reerences Andersen, Henning 1973 Abductive and deductive change. Language: Journal of the Linguistic Society of America 49(4): 765⫺793. Auer, Peter 1990 Phonologie der Alltagssprache: Eine Untersuchung zur Standard/Dialekt-Variation am Beispiel der Konstanzer Stadtsprache. Berlin/New York: de Gruyter. Auer, Peter 2005 Sound change / Lautwandel. In: Ulrich Ammon, Norbert Dittmar, Klaus J. Mattheier and Peter Trudgill (eds.), Sociolinguistics: An International Handbook of the Science of Language and Society, vol. 2, 1717⫺1727. 2nd ed. (Handbücher zur Sprach- und Kommunikationswissenschaft 3.2.) Berlin/New York: de Gruyter.

12. Language and space: The linguistic dynamics approach Bellmann, Günter 1967 Schlesischer Sprachatlas, vol. 1: Laut- und Formenatlas. Marburg: Elwert. Christen, Helen 1998 Dialekt im Alltag. Eine empirische Untersuchung zur lokalen Komponente heutiger schweizerdeutscher Varietäten. (Reihe Germanistische Linguistik 201.) Tübingen: Niemeyer. Croft, William 2000 Explaining Language Change: An Evolutionary Approach. (Longman Linguistic Library.) Harlow: Pearson Education. Eckert, Penelope and Sally McConnell-Ginet 1992 Think practically and look locally: Language and gender as community-based practice. Annual Review of Anthropology 21: 461⫺490. Elmentaler, Michael 2006 Sprachlagenspektren im arealen Vergleich. Vorüberlegungen zu einem Atlas der deutschen Alltagssprache. Zeitschrift für Dialektologie und Linguistik 73: 1⫺30. Frings, Theodor 1924 Rheinische Sprachgeschichte. Essen: Baedeker. Gabelentz, Georg von der 1888 [1966] Pott. Allgemeine deutsche Bibliographie 26, 478⫺485. [Reprinted in: Thomas A. Sebeok, Portraits of Linguists: A Biographical Source Book for the History of Western Linguists 1746⫺1963, vol. 1, From Sir William Jones to Karl Brugmann. (Indiana University Studies in the History and Theory of Linguistics.) Bloomington: Indiana University Press, 251⫺261.] Gilles, Peter 1999 Dialektausgleich im Le¨tzebuergeschen. Zur phonetisch-phonologischen Fokussierung einer Nationalsprache. (Phonai 44.) Tübingen: Niemeyer. Gilles, Peter and Claudine Moulin 2003 Language Standardization in Luxembourgish. In: Ana Deumert and Wim Vandenbusche (eds.), Germanic Standardizations. Past and Present, 303⫺329. Amsterdam: Benjamins. Gillie´ron, Jules and Edmond Edmont 1902⫺1969 Atlas Linguistique de la France (ALF). 13 vols. Paris: Champion. Grice, Herbert Paul 1975 Logic and conversation. In: Peter Cole and Jerry L. Morgan (eds.), Syntax and Semantics, vol. 3, Speech Acts, 41⫺58. New York: Academic Press. Haspelmath, Martin 1998 Does grammaticalization need reanalysis? Studies in Language 22: 315⫺351. Herrgen, Joachim 1986 Koronalisierung und Hyperkorrektion. Das palatale Allophon des /ch/-Phonems und seine Variation im Westmitteldeutschen. (Mainzer Studien zur Sprach- und Volksforschung 9.) Stuttgart: Steiner. Herrgen, Joachim 2006 Die Dynamik der modernen Regionalsprachen. In: Joachim Gessinger and Anja Voeste (eds.), Dialekt im Wandel. Perspektiven einer neuen Dialektologie, 119⫺142. (Osnabrücker Beiträge zur Sprachtheorie 71.) Bremen: Redaktion OBST. Jaberg, Karl and Jakob Jud (eds.) 1928⫺1960 Sprach- und Sachatlas Italiens und der Südschweiz (AIS). 8 vols. Zofingen: Ringler. Jakobson, Roman 1962 [1980] Zeichen und System der Sprache. In: Fritz Hintze, Georg Friedrich Meier, Eugen Seidel and Wolfgang Steinitz (eds.), Zeichen und System der Sprache, vol. 2. Veröffentlichung des 1. Internationalen Symposions “Zeichen und System der Sprache” vom 28. 9. bis 2. 10. 1959 in Erfurt, 50⫺56. (Schriften zur Phonetik, Sprachwissenschaft und Kommunikationsforschung 4.) Berlin: Akademie. [English translation: Sign and system of language: A reassessment of Saussure’s doctrine. Poetics Today 2(1): 33⫺38.]

221

222

II. Linguistic approaches to space Kehrein, Roland 2006 Regional accent in the German language area ⫺ How dialectally do German police answer emergency calls? In: Frans Hinskens (ed.), Language Variation: European Perspectives, 83⫺96. Amsterdam/Philadelphia: Benjamins Kehrein, Roland 2008 Regionalakzent und linguistische Variationsspektren im Deutschen. In: Frans Patocka and Peter Ernst (eds.), Dialektgeographie der Zukunft. Akten des 2. Kongresses der Internationalen Gesellschaft für Dialektologie des Deutschen (IGDD) am Institut für Germanistik der Universität Wien. 20. bis 23. September 2006, 131⫺156. (Zeitschrift für Dialektologie und Linguistik. Beihefte 135.) Stuttgart: Steiner. Kehrein, Roland, Alfred Lameli and Jost Nickel 2005 Möglichkeiten der computergestützten Regionalsprachenforschung am Beispiel des Digitalen Wenker-Atlas (DiWA). In: Georg Braungart, Peter Gendolla and Fotis Jannidis (eds.), Jahrbuch für Computerphilologie 7, 149⫺170. Paderborn: Mentis. Keller, Rudi 1990 [2003] Sprachwandel. Von der unsichtbaren Hand in der Sprache. (Uni-Taschenbücher 1567.) Tübingen: Francke. [3rd ed.] Kremer, Ludger 1979 Grenzmundarten und Mundartgrenzen. Untersuchungen zur wortgeographischen Funktion der Staatsgrenze im ostniederländisch-westfälischen Grenzgebiet. 2 vols. (Niederdeutsche Studien 28.) Köln/Wien: Böhlau. Kremer, Ludger (ed.) 1993 Diglossiestudien. Dialekt und Standardsprache im niederländisch-deutschen Grenzland. (Westmünsterland 1.) Vreden: Landeskundliches Institut Westmünsterland. Kurath, Hans 1939⫺1943 Linguistic Atlas of New England. 3 vols. (Linguistic Atlas of the United States and Canada.) Providence, RI: Brown University. Labov, William 1994 Principles of Linguistic Change, vol. 1, Internal Factors. (Language in Society 20.) Oxford: Blackwell. Lameli, Alfred 2004 Standard und Substandard. Regionalismen im diachronen Längsschnitt. (Zeitschrift für Dialektologie und Linguistik. Beihefte 128.) Stuttgart: Steiner. Lenz, Alexandra N. 2003 Struktur und Dynamik des Substandards. Eine Studie zum Westmitteldeutschen (Wittlich/ Eifel). (Zeitschrift für Dialektologie und Linguistik. Beihefte 125.) Stuttgart: Steiner. Malkiel, Yakov 1967 Every word has a history of its own. Glossa: An International Journal of Linguistics 1: 137⫺149. Meyerhoff, Miriam 2002 Communities of practice. In: J. K. Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.), The Handbook of Language Variation and Change, 526⫺548. (Blackwell Handbooks in Linguistics.) Malden, MA/Oxford: Blackwell. Milroy, Lesley 2002 Social networks. In: J. K Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.), The Handbook of Language Variation and Change, 549⫺572. (Blackwell Handbooks in Linguistics.) Malden, MA: Blackwell Moulton, William G. 1962 Dialect geography and the concept of phonological space. Word 18: 23⫺32. Niedeggen-Bartke, Susanne 2002 Williams syndrome and German participles. Poster presented at the International Symposium on Assessing the Dynamics of Human Brain Functions, August 2002 in Marburg.

12. Language and space: The linguistic dynamics approach Paul, Hermann 1877 Die Vocale der Flexions- und Ableitungssilben in den ältesten germanischen Dialekten. Beiträge zur Geschichte der deutschen Sprache und Literatur 4: 315⫺475. Paul, Hermann 1886 Principien der Sprachgeschichte. 2nd ed. Halle: Niemeyer. Pfalz, Anton 1918 Reihenschritte im Vokalismus. In: Beiträge zur Kunde der bayerisch-österreichischen Mundarten, 22⫺42. (Kaiserliche Akademie der Wissenschaften in Wien. Philosophischhistorische Klasse. Sitzungsberichte 190, vol. 2, Abhandlungen.) Wien: Rohrer. Rabanus, Stefan 2008 Morphologisches Minimum. Distinktionen und Synkretismen im Minimalsatz hochdeutscher Dialekte. (Zeitschrift für Dialektologie und Linguistik. Beihefte 134.) Stuttgart: Steiner. Reis, Marga 1978 Hermann Paul. Beiträge zur Geschichte der deutschen Sprache und Literatur 100: 159⫺ 204. Saussure, Ferdinand de 1916 [1968/1974] Cours de linguistique ge´ne´rale. [E´dition critique, vol. 1. Rudolf Engler (ed.). Wiesbaden: Harrassowitz.] Saussure, Ferdinand de 2003 [2002] Wissenschaft der Sprache. Neue Texte aus dem Nachlass. Ludwig Jäger (ed.), Elisabeth Birk and Mareike Buss (transl.). Frankfurt am Main: Suhrkamp. [Originally published as E´crits de linguistique ge´ne´rale. Simon Bouquet and Rudolf Engler (eds.) in collaboration with Antoinette Weil. Paris: Gallimard.] Schmidt, Jürgen Erich 1993 Die deutsche Substantivgruppe und die Attribuierungskomplikation. (Reihe Germanistische Linguistik 138.) Tübingen: Niemeyer. Schmidt, Jürgen Erich 1998 Moderne Dialektologie und regionale Sprachgeschichte. Zeitschrift für deutsche Philologie 117: 163⫺179. [Special issue: Werner Besch and Hans Joachim Solms (eds.), Regionale Sprachgeschichte.] Schmidt, Jürgen Erich 2005a Die deutsche Standardsprache: Eine Varietät ⫺ drei Oralisierungsnormen. In: Ludwig M. Eichinger and Werner Kallmeyer (eds.), Standardvariation. Wieviel Variation verträgt die deutsche Sprache?, 278⫺305 (Jahrbuch des Instituts für deutsche Sprache 2004.) Berlin/New York: de Gruyter. Schmidt, Jürgen Erich 2005b Versuch zum Varietätenbegriff. In: Alexandra N. Lenz and Klaus J. Mattheier (eds.), Varietäten ⫺ Theorie und Empirie, 61⫺74. (VarioLingua 23.) Frankfurt: Lang. Schmidt, Jürgen Erich 2005c Sprachdynamik. In: Eckhard Eggers, Jürgen Erich Schmidt and Dieter Stellmacher (eds.), Moderne Dialekte ⫺ Neue Dialektologie. Akten des 1. Kongresses der Internationalen Gesellschaft für Dialektologie des Deutschen (IGDD) am Forschungsinstitut für deutsche Sprache “Deutscher Sprachatlas” der Philipps-Universität Marburg vom 5.⫺8. März 2003, 15⫺44. (Zeitschrift für Dialektologie und Linguistik. Beihefte 130.) Stuttgart: Steiner. Schmidt, Jürgen Erich 2009 Die modernen Regionalsprachen als Varietätenverbund. In: Peter Gilles, Joachim Scharloth and Evelyn Ziegler (eds.), variatio delectat: Empirische Evidenzen und theoretische Passungen sprachlicher Variation. Klaus J. Mattheier zum 60. Geburtstag, 125⫺144. (Variolingua. Nonstandard⫺Standard⫺Substandard.) Frankfurt: Lang. Schmidt, Jürgen Erich and Joachim Herrgen to appear Sprachdynamik. Eine Einführung in die moderne Regionalsprachenforschung. (Grundlagen der Germanistik.) Berlin: Erich Schmidt.

223

224

II. Linguistic approaches to space Schmidt, Jürgen Erich and Joachim Herrgen (eds.) 2001⫺2009 Digitaler Wenker-Atlas (DiWA). Georg Wenkers Sprachatlas des Deutschen Reichs. Compiled by Alfred Lameli, Alexandra N. Lenz, Jost Nickel and Roland Kehrein, KarlHeinz Müller, Stefan Rabanus. Marburg: Forschungzentrum Deutscher Sprachatlas. Schuchardt, Hugo 1922 [1917] Sprachgeschichte und Sprachbeschreibung. In: Leo Spitzer (ed.), Hugo SchuchardtBrevier. Ein Vademekum der allgemeinen Sprachwissenschaft. Als Festgabe zum 80. Geburtstag des Meisters zusammengestellt und eingeleitet von Leo Spitzer, 264⫺269. Halle: Niemeyer. Seiler, Guido 2003 Präpositionale Dativmarkierung im Oberdeutschen. (Zeitschrift für Dialektologie und Linguistik. Beihefte 124.) Stuttgart: Steiner. Siebs, Theodor 1898 [1969] Deutsche Bühnenaussprache. Ergebnisse der Beratungen zur ausgleichenden Regelung der deutschen Bühnenaussprache, die vom 14. bis 16. April 1898 im Apollosaale des Königlichen Schauspielhauses zu Berlin stattgefunden haben. Berlin: Ahn. [New edition: Helmut de Boor, Hugo Moser and Christian Winkler (eds.), Deutsche Aussprache. Reine und gemäßigte Hochlautung mit Aussprachewörterbuch. Berlin: de Gruyter.] Sievers, Eduard 1876 Grundzüge der Lautphysiologie. Zur Einführung in das Studium der Lautlehre der indogermanischen Sprachen. Leipzig: Breitkopf and Härtel. [Later editions titled Grundzüge der Phonetik.] Smits, Tom F. H. 2007 Strukturwandel in Grenzdialekten. Die Konsolidierung der niederländisch-deutschen Staatsgrenze als Dialektgrenze. Unpublished PhD dissertation, Department of Linguistics, Faculty of Linguistics and Philosophy, University of Antwerp. [To appear in Zeitschrift für Dialektologie und Linguistik. Beihefte.] Spiekermann, Helmut 2006 Standardsprache als regionale Varietät ⫺ Regionale Standardvarietäten. In: Joachim Gessinger and Anja Voeste (eds.), Dialekt im Wandel. Perspektiven einer neuen Dialektologie, 81⫺101. (Osnabrücker Beiträge zur Sprachtheorie 71.) Bremen: Redaktion OBST. Thun, Harald 2000 Altes und Neues in der Sprachgeographie. In: Wolf Dietrich and Ulrich Hoinkes (eds.), Romanistica se movet. Festschrift für Horst Geckeler, 69⫺89. Münster: Nodus. Thun, Harald and Adolfo Elizaincı´n (eds.) 2000⫺ Atlas lingüı´stico diato´pico y diastra´tico del Uruguay (ADDU). Kiel: Westensee. Verburg, Pieter A. 1950 [1966] The background to the linguistic conceptions of Franz Bopp. Lingua 2, 438⫺468. [Reprinted in: Thomas A. Sebeok, Portraits of Linguists: A Biographical Source Book for the History of Western Linguists 1746⫺1963, vol. 1: From Sir William Jones to Karl Brugmann, 221⫺250. (Indiana University Studies in the History and Theory of Linguistics.) Bloomington: Indiana University Press.] Vietor, Wilhelm 1890 Die Aussprache des Schriftdeutschen. 2nd ed. Lepizig: Reisland. Wagner, Kurt 1924⫺1925 Grammophonische Aufnahmen deutscher Mundarten. Teuthonista 1: 229⫺231. Wenger, Etienne 1998 Communities of Practice: Learning, Meaning, and Identity. Cambridge: Cambridge University Press. Wenker, Georg 1878 Sprach-Atlas der Rheinprovinz nördlich der Mosel sowie des Kreises Siegen. Hand-drafted atlas held in the archives of the Forschungszentrum Deutscher Sprachatlas in Marburg.

12. Language and space: The linguistic dynamics approach Wenker, Georg 1887⫺1923 Sprachatlas des Deutschen Reichs. Laut- und Formenatlas. 1647 hand-drafted fullcolor maps held in the archives of the Forschungszentrum Deutscher Sprachatlas in Marburg and in the collection of the Staatsbibliothek Stiftung Preußischer Kulturbesitz in Berlin (2 copies extant). Wenker, Georg 1889 Sprachatlas des Deutschen Reiches. Einleitung. Handwritten manuscript held in the collection of the Staatsbibliothek Stiftung Preußischer Kulturbesitz in Berlin. Whitney, William Dwight 1867 Language and the Study of Language. Twelve Lectures on the Principles of Linguistic Science. London: Trübner. Wiesinger, Peter 1983 Die Einteilung der deutschen Dialekte. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke and Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, 807⫺900. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.2.) Berlin/New York: de Gruyter. Winteler, Jost 1876 Die Kerenzer Mundart des Kantons Glarus in ihren Grundzügen dargestellt. Leipzig/Heidelberg: C. F. Winter. Wrede, Ferdinand 1903 Der Sprachatlas des Deutschen Reichs und die elsässische Dialektforschung. Archiv für das Studium der neueren Sprachen und Literaturen 57(111): 29⫺48. Zwirner, Eberhard, Adalbert Maack and Wolfgang Bethge 1956 Vergleichende Untersuchungen über konstitutive Faktoren deutscher Mundarten. Zeitschrift für Phonetik und Allgemeine Sprachwissenschaft 9: 14⫺30.

Jürgen Erich Schmidt, Marburg (Germany)

225

III. Structure and dynamics o a language space 13. Identiying dimensions o linguistic variation in a language space 1. 2. 3. 4. 5. 6. 7.

Dimensions of variation and the architecture of a language Linguistic variables and linguistic varieties Diatopic varieties Diastratic variation and diastratic varieties Diaphasic variation and diaphasic varieties Variation in a language space as a multidimensional continuum References

1. Dimensions o variation and the architecture o a language A language space is made up of language varieties and at the same time a language is conceivable, broadly and metaphorically speaking, as a language space, i. e., a scene occupied by linguistic entities. This space, which constitutes a language, is by no means a homogeneous space; on the contrary, it is a realm of linguistic heterogeneity and differences of various kinds, which form on the whole the inner variation of a language. Variation permeates all languages, has a socio-cultural as well as a biological foundation and shows itself in multifarious manners with an adaptive significance (Chambers 1995: 206⫺253). For the purpose of the present article, it is in particular noteworthy that “the underlying cause of sociolinguistic differences, largely beneath consciousness, is the human instinct to establish and maintain social identity” (Chambers 1995: 250). The aim of this article is exactly to try establishing a certain order in this heterogeneity, which means identifying dimensions of variation and aiming at modeling them in a congruent way. In spite of the sometimes diverging terminology, there is a wide consent among linguists on most subjects and issues dealt with in this founding article: therefore, the following presentation will be mostly apodictic. As a first approach in order to capture the major generalizations pertaining the inner variability of a language or a language space it is useful to sketch out some main dimensions of variation, picking out classes of correlations (co-occurrences) between values of linguistic variables and relevant extra-linguistic (environmental, social, pragmatic) factors in a linguistic community. The main factors in the societal structure of a given linguistic community that can co-occur with (inner) linguistic differences fall into four types. First, time and space; then, social stratification; and last, social situations. Correspondingly, we can formulate four axioms: (i) a language varies with the passing of time; (ii) a language varies with the geographical distribution of its speakers; (iii) a language varies with the social class/ group of its speakers; (iv) a language varies with the communicative situations in which it is employed. Consequently, there are four main dimensions of variation: the temporal,

13. Identifying dimensions of variation historical dimension; the spatial, geographical dimension; the social dimension; and the situational dimension. These main dimensions representing the impact of extralinguistic macrofactors on linguistic structures have been singled out since the first theoretical approaches of linguistics to the study of the life of language in society, see, e.g., Gabelentz (1891: 267⫺ 283) in Europe or Whitney (1875: 153⫺160) and Bloomfield (1933: ch. 3.4⫺3.8) in America; and their phenomenology as it appears in linguistic facts has been widely analyzed in sociolinguistic theory and research from Labov (1966) on. In the continental European tradition the distinction has been often treated unitarily in terms of the “four dia-dimensions”, in accordance with the terminology of classical Greek flavor adopted by Eugenio Coseriu in the tradition of Leiv Flydal (see Coseriu 1980; Albrecht 1986: 2003): (i) diachronic dimension (diachronia): variation across time; (ii) diatopic dimension (diatopia): variation across space; (iii) diastratic dimension (diastratia): variation across socio-economic classes and social groups; (iv) diaphasic dimension (diaphasia): variation across situations. At any given time, diachronia is out of action, so that attention must be paid only to the three synchronic dimensions: diatopia, diastratia and diaphasia. However, studies by William Labov and other authors (cf., e.g., Labov 1994, 2001) have demonstrated how the synchronic dynamics of the inner variation often mirrors language change in progress, so that synchrony and diachrony should not be separated from each other. The synchronic dimensions work together inside a linguistic space that can be metaphorically depicted as a cube or a parallelepiped, in which the horizontal base axis (length) represents diatopia, the vertical axis (height) represents diastratia and the horizontal orthogonal axis (width) represents diaphasia. The whole is in a dynamic relationship and moves along the independent temporal axis. In synchrony, any point inside the cube, therefore, represents a certain combination of language variation factors concerning a given location in the geographical space, a given position on the social scale and a given type of socio-communicative situation. Of these three main dimensions, European dialectology has concentrated its attention at length upon geographical variation, whereas American variationists following the example of Labov (1966) have mostly focused on the social and stylistic (diaphasic) variation. The three dimensions together with their reciprocal relationships constitute what in the Coserian tradition is called Architektur der Sprache (‘a language’s architecture’). Every language or every language space has its own architecture, depending both on the extent of geographical variation or on the weight of social variation etc., as well as on the reciprocal correlation and the hierarchical relationship between the varieties belonging to different dimensions. In the Italian language space, for example, until recently diatopia appears to be a major differentiating factor (Berruto 1989a), whereas in the British and American language space diastratia seems to have more importance than the other dimensions, and in the French language space diaphasia (Gadet 2003) is prevailing. However, by and large, the diatopic dimension has a great relevance, at least because it is the first to draw the observer’s attention. It is common sense to say that “every place in the world has its own language, in every place in the world one finds a different way of speaking”. “Place, in one form or another ⫺ nation, region, county, city, or neighborhood ⫺ is one of the most frequently adduced correlates of linguistic variation” (Johnstone 2004: 65). The same author discusses also the double nature of place, “place as a location” and “place as meaning”, and points out that place is to be conceived not only as a physical entity, but also as a socially constructed context, shaped and constituted by people “through shared experiences and shared orientations” (2004: 69).

227

228

III. Structure and dynamics of a language space At a first and very wide glance, geographical diversity already works at the general level of the human language faculty (the Saussurean faculte´ de langage): the six thousand and more languages existing today on earth (approximately 6800 are listed by the web site Ethnologue) can be considered to be varied forms of the human verbal language. More significantly, all languages show widespread differentiation depending on the geographical zones where they are spoken. A standard language spoken as national language in different countries can easily develop partially different standard forms in each country, in particular as far as the lexicon is concerned, so that it ends up having different norms in different nations (“pluricentric language”, Clyne 1992). In a given country, a language or a language space (these two terms are employed in this article as nearsynonyms) displays regional variation; in a given region, a language displays subregional variation; and so forth, up to the minimal “language point”, even a very small community settled in a given geographical point, maybe a single individual. For instance, German is a typical pluricentric language with slightly different but clearly recognizable national standards in Germany, Austria and Switzerland. Thus, in the language (space) German (Ammon 2003: 5; also see Ammon 1995) one finds a geographically “unmarked form (general German), nationally marked forms (e.g., Austrian German), regionally marked forms (e.g., Northern Germany), and forms marked sub-regionally by the area of a single state (bayrisch referring to Bavaria) or the region of a dialect (schwäbisch ‘Swabian’)”. Swabian itself takes different colors in the various Swabian towns, villages and countryside. The diastratic dimension gives every geographically identified variety an internal social “depth”. It is commonplace to note that people reveal in their ways of speaking not only their geographical origin but also their social position and education, so that social class markers are recognizable in all languages. Thus this car, it needs washing said by a British subject, with pronunciation of postvocalic /r/, identifies the speaker as presumably Scottish (by a phonetic trait indicating geographical localization), while this car, it needs washed, with postvocalic /r/ and washed (a morphosyntactic feature implying social characterization) rather than washing (Trudgill 1995: 6), suggests the speaker is Scottish and of lower social class. A third very important range of variation within a language or a language space and even within a given socio-geographical variety is added by the intervention of the diaphasic dimension. Every type of social and communicative situation is characterized by a certain language use: every speaker in a speech community has the ability to change his/ her way of speaking in relation to the manifold factors present or activated in a situation, choosing the appropriate variety for that type of situation. Halliday (1978: 34⫺34, 225; also see Gregory 1967) has emphasized that a major distinction between diatopic and diastratic variations and varieties and diaphasic variation and varieties (both authors call the latter diatypic varieties) is that the former are chosen “according to the user” (because they are “the linguistic reflection of reasonably permanent characteristics of the user in language situations”) and the latter “according to the use” (because they are “the linguistic reflection of recurrent characteristics of the user’s use of language in situations”). This means that a given speaker speaks spontaneously one diatopic variety, which refers to his/her geographical provenance, and one diastratic variety, which correlates to his/her position in the social stratification within a society; simultaneously he/she is able to speak several diaphasic varieties according to the various communicative situations.

13. Identifying dimensions of variation

2. Linguistic variables and linguistic varieties Variation implies linguistic (or better, but less common, sociolinguistic) variables. The sociolinguistic variable is a very crucial concept for analyzing language internal variation. The rise of sociolinguistic, variationist studies in linguistics started in the 1960s with the seminal work by William Labov on linguistic variables (Labov 1966). A sociolinguistic variable is, in a nutshell, one point of the system of a language (a phonetic/ phonological unit, a morphological item, a syntactic structure, a construction, a semantic unit and so on) that admits and shows different realizations, with the same referential meaning, in correlation with extralinguistic (geographical, social, situational) factors and properties (see Tagliamonte 2006: 70⫺98). Thus, a sociolinguistic variable, as a linguistic form carrying social meaning, represents the minimal sociolinguistic unit in which language and society (in the broadest sense) closely correlate; it is the stitch that sews together language and society. Each realization of a given variable is one of the values of this variable. Therefore, each value is a variant of a variable; a sociolinguistic variable is a set of linguistic variants. This technical sense of “(socio)linguistic variable” has nothing to do with the broad meaning of the expression sociolinguistic variable that is sometimes found even in sociolinguistics; for instance, what is meant by a title such as Age as a Sociolinguistic Variable (Eckert 1998 is that age is a social variable possibly correlating with linguistic phenomena or showing linguistic significance, but by no means, of course, that age itself is a “sociolinguistic variable”. The pronunciation of the -ing suffix in British English represents a good, simple example of a variable; its final sound can have two different realizations: alveolar nasal [n] and velar nasal [n]; each realization bears a social significance, since [n] is the standard pronunciation and occurs, therefore, in educated speech, while the non standard [n] occurs in uneducated speech. Usually, notation marks for variables are put in round brackets: thus, in this case we have the (socio)linguistic variable (ng), with two variants, [n] and [n]. A sociolinguistic variable can occur at every level of the linguistic system and embody units of any extension; Cornips (1998) shows how the variable can concern even the setting of a parameter, in the sense of generative grammar, in syntactic variation (for a wider overview on this topic, see Cornips and Corrigan 2005). The distribution of variants in a language space is not random. On the contrary they are arranged in such a way that they tend to occur together with some given extralinguistic, social features. The tendential co-occurrence of variants gives rise to linguistic varieties. Therefore, a linguistic variety is conceivable as a set of co-occurring variants; it is identified simultaneously by both such a co-occurrence of variants, from the linguistic viewpoint, and the co-occurrence of these variants with extralinguistic, social features, from the external, societal viewpoint. Thus, a linguistic variety bears a double characterization, as a linguistic as well as a social entity, it is made of linguistic variants together with their social value. In simple words, a variety of language is “a set of linguistic items with similar social distribution” (Hudson 1996: 22). Linguistic varieties, i. e., the varieties of a given language, can be classified in relation to the main dimensions of the architecture of a language: we have indeed geographical or diatopic varieties, social or diastratic varieties, situational or diaphasic varieties. Some authors have raised doubts about the suitability of a variety-based model of language variation, since varieties are often difficult to identify and defining their presence in any given text can be problematic. Hudson (1996: 22⫺22, 48⫺49), for instance, argues rather

229

230

III. Structure and dynamics of a language space for a model based on the notion of “item”. Even if the definition of varieties raises doubts on many issues (but these are partially removed if varieties are conceived as particular points in a continuum, cf. section 6), the notion of variety brings about so many advantages in terms of generalization and abstraction when one deals with variation phenomena, that this notion seems difficult to replace or eliminate.

3. Diatopic varieties The most typical diatopic (geographical) variety is a “dialect”. Dialect seems to be a simple category, but it turns out to be a very controversial one. First of all, for American and British linguists dialect is often understood as a synonym for (language) variety, designating any particular language form with at least some differences in structure and grammar with respect to any other language form. Dialect is simply “a neutral label to refer to any variety of a language which is shared by a group of speakers” (Wolfram and Schilling-Estes 1998: 2). In other words, every socio-geographically recognizable language variety is a dialect. Similarly, Gregory (1967) distinguishes between “dialectal variation” and “diatypic variation” as the two basic aspects of language variation, with respect to society in a broad sense and to situations (see above). In this way, e.g., Standard English is to be conceived as a dialect of English: “Standard English is a dialect […]. Subvarieties of languages are usually referred to as dialects, and languages are often described as consisting of dialects” (Trudgill 2002: 165); a conception which undermines the fundamental opposition between standard language and dialect as two mutually self determining concepts. For linguists from the European continent, dialect (French dialecte, German Dialekt/Mundart, Spanish dialecto, Italian dialetto, Russian диал´кт, etc.) is better understood as any language variety spoken in a given place or region in concomitance with a more prestigious superimposed variety (the latter being a standard language), so that its diatopic characterization comes crucially to the fore. For a continental linguist “Standard English is a dialect” might therefore be a rather odd, confusing statement. Moreover dialect is a category sensitive to the different sociolinguistic situations and to the particular characteristics of linguistic repertoires, and can mean somewhat different things in different situations (Britain 2004). While, for example, in the USA dialects are simply spoken varieties of English with some differences in pronunciation and lexicon (cf. Chambers and Trudgill [1980] for a general Anglo-Saxon perspective, and Wolfram and Schilling-Estes [1998] for the USA), in Germany as well as in Italy dialects are mostly spoken regional linguistic systems with a noticeable structural distance from (Standard) German and (Standard) Italian, and with an autonomous history and development. As a consequence, it can be observed that in Italian and in German speaking countries not only two (dialects and standard) but three kinds of diatopic variety exist. Between dialect and standard there is an intermediate variety (cf. section 6), representing the way in which a standard language is regionally/locally spoken under the influence of a local dialect: the Umgangssprache or Substandard German in Germany or in Austria (see, e.g., Barbour and Stevenson 1990, Dittmar 1997: 193⫺201), the italiano regionale in Italy (Berruto 1989a). In a number of situations, in particular where we find a constellation of “diaglossia” (Auer 2005a) or “dilalia” (in which both the “low” and the “high”

13. Identifying dimensions of variation variety are employed in ordinary conversation; Berruto 1989b; Dittmar 1997: 150⫺151), intermediate varieties such as regional dialects and regional standards emerge from intensive contact between local dialects and national standard. The former are relatively standardized varieties of the dialects under interference from national language and having a superlocal range; the latter are varieties of the standard language with local coloring due to the interference of the dialect and representing the regional socially unmarked form of the national standard. It is useful at this point to introduce a distinction made by Coseriu (1980) between primary dialects and secondary dialects. Primary dialects are the varieties existing in a given language space before the formation of a standard, that is to say the “sister” varieties of the (future) standard, sharing the same origin with the variety which will later become the standard one. Secondary (and tertiary: but the present article will not deal with this further distinction proposed by Coseriu) dialects are regional varieties evolving after the establishment of a standard, as a result of its local differentiation. Another useful notion in this context is that of a “roof language/roofing language” (Dachsprache): with this term the German sociologist Heinz Kloss designates the (standard) language which, as a metaphorical roof, covers the dialects which are clearly related to it (existing in the same country and having the same origin), the language in question being taught at school to the native speakers of these dialects (see Ammon 1989: 38⫺43). Clearly, the distinction between dialect and (standard) language is by no means a linguistic one, but rather a by-product of social, historical and cultural factors. “Dialect” proves to be in a dialectical way a concept opposed to standard, in need of sociolinguistic definition. From a merely linguistic point of view there is no difference between standard and dialect; it is impossible to define an existing language form as a dialect or as a language on the grounds of purely linguistic features. The difference rests on their position in society: dialect is a geographically restricted variety, mainly spoken and lacking of (overt) prestige, occupying the “low” level in a linguistic repertoire; standard, instead, is a prestigious variety with a wider geographical range, which occupies the “high” level in a linguistic repertoire and is employed in written and more formal usages. In short, one could say with Hudson (1996: 36): “there is no real distinction to be drawn between ‘language’ and ‘dialect’ (except with reference to prestige)”. Normally a standard language is superimposed on many dialects (or, better, it “covers” them as a Dachsprache). What these dialects precisely are depends on the nature of the repertoire to which they belong and on the history of any given linguistic community. Finally, it is worth emphasizing that a dialect, as any language or language space, can easily display inner variation across the three synchronic dimensions: diatopia, diastratia and diaphasia. The extreme limit of sociogeographical variation is the single individual, and consequently the minimum sociogeographical variety is the so called idiolect, the language form of a single individual. The extreme limit of variation generally speaking, i. e., including the diaphasic dimension, is the variety spoken by a given single individual in a given actualized situation (for which no specific technical term is currently in use). The sociolinguistic counterpart of a dialect is a standard variety. A standard is a language form which, gaining prestige due to literary tradition, cultural acceptance, political or religious reasons, socio-economic factors and so on, has attained to a full elaboration (Ausbau) and has become through a codification process the reference norm (the

231

232

III. Structure and dynamics of a language space “good language”) in a linguistic community and the correct model taught at school. Thus a standard is ⫺ in contraposition to a dialect ⫺ supraregional, homogeneous, explicitly codified (i. e., with an overt norm and an acknowledged corpus of reference texts, grammars, dictionaries), typically written and normally linked to the upper middle class (Ammon 1986: 2004). However, besides dialect many other terms are in use to refer to diatopic varieties: vernacular, patois, basilect, koine´, regiolect and accent are amongst the most recurrent of them. Both vernacular and patois can be viewed as near-synonyms of dialect, bearing different nuances. By vernacular linguists usually mean a dialect in so far as it is the mother tongue of a definite group or community of speakers (see Macaulay 1988); patois designates, mostly in the French tradition or with reference to French or Galloroman speaking areas or also to Creole languages (and sometimes with a slight depreciatory nuance), a local dialect lacking any prestige. Basilect is instead a term adopted in creolistics (see, e.g., Bickerton 1973), where it has been employed to mean the most local, marked and least prestigious varieties within a variety continuum (cf. section 6), showing at the one end these basilects, at the other end the most prestigious standardized variety (the acrolect), with intermediate varieties (the mesolects) in between. The terms refer to the idea of the continuum as a metaphorical triangle, at whose base there are many basilects and at the top (Greek a¬κρον ‘extremity, peak’) one acrolect. While vernacular, patois and basilect are all varieties at the bottom of the sociogeographical scale, koine´ and regiolect (lect, from dia-lect and similar expressions, is sometimes employed as a neutral term for “language variety” and serves as formative for the names of other classes of varieties) are located at the top. A koine´, from Greek κοινη´ ‘common (language)’, is “the stabilized result of mixing of linguistic subsystems such as regional or literary dialects” (Siegel 1985: 363), often characterized in negativo by the reduction of too locally marked features: a typical koine´ is a regional koine´, that “usually serves as a lingua franca among speakers of the different contributing varieties” (Siegel 1985: 363), i. e., of the various partially different local dialects spoken in the region. Often, it is an “amended” variety of the dialect of the principal town that works as a koine´ in a given region. Regiolect is a term sometimes employed, mainly by German and Dutch speaking linguists, to refer to the dialect of a whole region, with a superlocal range. Finally, accent means a diatopic variety marked in pronunciation only, characterized exclusively by phonetic (segmental and/or suprasegmental) features.

4. Diastratic variation and diastratic varieties As American and British variationists have shown since Labov (1966) in New York City and Trudgill (1974) in Norwich, social variation in its strict sense often takes on the form of a “sociolinguistic pattern”, in which for each considered variable there is continuous variation (“fine stratification”) based on differences in frequency and showing no sharp breaks, rather than a plain opposition of the “all-or-nothing” kind: “A major finding of urban sociolinguistic work is that differences among social dialects are quantitative and not qualitative” (Romaine 1994: 70). The same source reports a simple example concerning two opposite sociolinguistic patterns: the variable “postvocalic r” as analyzed by Labov and Trudgill (Table 13.1).

13. Identifying dimensions of variation

233

Tab. 13.1: Percentage of post-vocalic /r/ pronounced in New York and Reading (Romaine 1994) New York City

Reading

Social class

32 20 12 0

0 28 44 49

upper middle class lower middle class upper working class lower working class

The inverse distribution of variants is socially significant insofar as the amount of the standard, more prestigious form in both speech communities (i. e., the realization of /r/, as prevalent in General American, in New York, and its deletion, according to the Received Pronunciation, in Reading) progressively increases along the social scale. In Reading, no tokens of nonstandard pronunciation were recorded among the informants of the upper middle class, while the greatest proportion of occurrences was found in the lower working class. This pattern is overtly symmetrical, with inverse values (no realization of /r/ by members of the lower working class, the greatest amount by the upper middle class), to the one found in New York City by Labov. There is no stabilized term for diastratic varieties, analogue to dialect for diatopic varieties. A diastratic variety is normally referred to as a social dialect, or more specifically as a social-class dialect. However, the term sociolect is also employed, in particular by German speaking sociolinguists, either as a synonym of social dialect or with more specific meanings (Durrell 2004). On the diastratic dimension, a language can co-vary with many different social factors. Besides social class, the main social factors which intervene to determine diastratic variation are age, sex or better gender (the sex of a person as reflected in social position, status and role and their attributes), ethnicity and social network. In many societies membership in social or professional groups or religious faith can also be relevant factors of language differentiation. Over the last decades a type of variety which for some authors can be traced back in extenso, and in certain regards, to the diastratic dimension has increasingly gained in importance: the interlanguages (learner varieties) of foreign immigrants. In spite of its importance as an obvious first reference point in diastratic variation, social class is by no means a clear-cut and undisputable category: it is often difficult to establish in what precisely a social class consists, being undoubtedly in itself a plurifactorial concept, that includes in various mixtures ingredients such as education, occupation, income, attitudes, life styles and so on (some of them hardly quantifiable; Georg 2004). Social class membership moreover depends on the different social patterns and the different shapes the social stratification takes in different societies. Establishing how many social classes should be taken into account in correlating language and society, as well as defining exact boundaries between social classes within a society, are also questionable matters. In early sociolinguistics, four classes were preferably taken into account (see Table 13.1); a higher number of classes, five or six or even nine (upper class, middle class and lower class, each subdivided in upper, middle and lower), was also sometimes taken into consideration. Trudgill (1974) in his survey in Norwich assumed five socio-economic classes, calculated on the basis of an index resulting from six components: occupation, fathers’ occupation, income, education, locality, housing. In Romance sociolinguistics a more descriptive, albeit methodologically less articulated, representation of social class

234

III. Structure and dynamics of a language space has been often adopted: Sanga (1981) for instance deals with Italian sociolinguistic stratification in the 1970s by means of six classes such as bourgeoisie (⬵ middle class), petite bourgeoisie du tertiaire (⬵ lower middle class, traders), classe ouvrie`re (working class), artisans et couches moyennes pre´-industrielles (⬵ craftsmen and pre-industrial lower middle class), paysans (peasants) and marginaux (outcasts). Lately, though, it has become increasingly difficult to deal with the social stratification in globalized societies through categories of this kind. Since the work of Milroy (1980) sociolinguists often prefer to work with social networks rather than social classes as the main social factor correlating with language variation. Social network, a structured set of social relations connecting a person and the people with whom this person interacts, seems to correlate very well with the distribution of socially significant linguistic features and has the advantage of being less abstract and more flexible than the concept of class. It involves not only the social position of an individual in a group but also the actual interactions in which he/she takes part. Closeknit networks, i. e., dense and multiplex networks (“density” and “multiplexity” are two important criteria in defining types of network) appear to correlate with conservative variants, while loose-knit networks (networks with little density and low multiplexity) appear to correlate with innovative variants in the dynamics of language change (Milroy 1992). Furthermore, a social network perspective is perfectly compatible with a social class perspective, for, on the one hand, certain network shapes turn out to be (more) typical of certain classes and, on the other hand, socialization patterns as well as communicative habits are important in characterizing both social class and social network.

5. Diaphasic variation and diaphasic varieties The first sociolinguistic surveys already reported a recurrent behavior of many variables. A variable such as (ng) in Norwich (Trudgill 1974; cf. section 2) varies at the same time with social stratification and, for each socio-economic class, with what has been called style: the more nonstandard realizations [n] are recorded, the less controlled language use is, i. e., the less formal the situation is perceived. This means that (ng) presents a stylistic variation pattern. Each language variety that depends on the relative formality of a communicative situation is a “style” (sometimes specified as “contextual style”; cf. Eckert and Rickford 2001 on the whole issue of style). Relative formality is an important parameter correlating with the diaphasic dimension. There are obviously other parameters which play an important part in diaphasic variation. The most remarkable among them are the activity carried out in interaction and the subject matter of discourse, both needing typical syntactic and textual patterns and a particular lexicon. Any language variety depending on the activity and topic dealt with in a communicative situation is a “register”. If all linguists agree in distinguishing register and style as the two main genres of diaphasic varieties, the definition of these notions is slightly blurry, and the terminology is in need of further elucidation. According to many sociolinguists, register is rather an overarching term, generally designating any “variety according to use” and, therefore, on the same rank as the notion of dialect, the “variety according to the user”. In a functional perspective, Halliday (1978; see also Gregory 1967) identifies within the situa-

13. Identifying dimensions of variation tional or contextual variation three interacting dimensions, called field, tenor and mode. Field refers to the nature of the social action carried out in the situation and includes the subject matter of discourse as a special aspect; tenor concerns the role structure governing the relationships between participants; mode has to do with the function the discourse is assigned to in the given situation and also includes channel or medium, i. e., the distinction between speaking/speech and writing. Any variety selected on these three interplaying dimensions is a “register”; within register, style concerns the aspect corresponding to tenor. Any language includes numerous registers, chosen by the manifold situational configurations in different domains (Biber and Finegan 1994). A special type of register is the so-called “foreigner talk”, namely the simplified language native speakers use in certain circumstances with foreigners little competent in the native speakers’ mother tongue (sometimes also called xenolect). The distinction between written and spoken language is such a pervasive differentiating factor in all languages that some authors have suggested a fourth main synchronic dimension of variation, namely the “diamesic” dimension. Even if the medium of communication is in principle selected by the situation, a reasonable ground for postulating an independent “diamesia” lies in the wide range of differences one normally finds between written and spoken language, including pragmatics and textuality, lexicon, morphosyntax, etc. (Halliday 1985; Koch and Österreicher 1990). In order to avoid confusion or overlapping with the meaning of “style” as a literary and rhetorical term, other sociolinguists call registers what have been called styles above, i. e., different ways of speaking according to the degree of formality of a given situation and the relationship with the addressee. According to this terminology (Berruto 1995: 148⫺150), the varieties depending primarily upon the type of activity and upon the subject matter of discourse (and hence governed by the “field”) are variously called sectorial languages, subcodes, or special languages. Subtypes of these with a link to the technical or professional areas in which they are employed and which require a specific lexicon are called Fachsprachen, Sondersprachen, langues de specialite´, languages for special/specific purposes, technolects, jargons and so forth. In this approach, no general term for the varieties on the diaphasic dimension is employed. A particular notion of “sociolinguistic style” has been developed by German sociolinguists from an interpretative perspective. Dittmar (1995: 156) considers sociolinguistic style as a complex and ordered system of preferences of language use, selecting forms of expression from the individual linguistic variety space and combining them by means of co-occurrence restrictions (see Auer 1997), in order to realize/achieve speakers’ goals and purposes through accommodating to the interactional partners’ speech.

6. Variation in a language space as a multidimensional continuum The arrangement of varieties in the language space constituting a language takes the form of continua. The concept of continuum implies an ordered set of elements arranged in such a way that between two adjacent entities of the set (in this case, language varieties) there are no sharp boundaries, but rather a gradual, fuzzy differentiation, each variety sharing some sociolinguistically marked features with adjacent varieties. The very notion of a continuum in variational linguistics arose in geolinguistics, where the dialec-

235

236

III. Structure and dynamics of a language space tal landscape is often viewed as a dialect continuum: between two neighboring village dialects in a linguistic area one finds little difference, while differences increase cumulatively as one considers progressively more distant villages, the greatest difference being recorded at the extremities of the continuum. In other words, linguistic distance between local dialects seems to be a function of their geographical distance, no clear-cut boundaries being perceived between contiguous dialects. To this view two considerations must be added. First of all, dialectometric analysis of phonological distance between Dutch dialects (Heeringa and Nerbonne 2001) has shown that differences along the continuum are not simply cumulative, but aggregate in relation to recognizable dialect areas, so that the continuum presents a certain discontinuity at the points coinciding with isoglosses defining traditional dialect areas. These appear to be separated by unsharp borders. Jörgensen and Kristensen (1995) likewise showed by means of quantitative methods that within a dialect continuum there is at least one clear boundary, i. e., the one between one of the poles (extremities) of the continuum and the rest. On the other hand, the very notion of border deserves some reflections, since borders turn out to be not only plainly objective boundaries in space, based on external criteria, but also depend on subjective attitudes and perceptions defining an identity and resulting in a socially constructed, cognitive space (cf. section 1). Applying such a cognitive approach to space, Auer (2005b) shows how German dialect continua across state borders develop (converging or diverging both in linguistic features and in the structure of the linguistic repertoires) in relation to ideological constructed borders delimiting “imagined communities”. Gerritsen (1999), moreover, has examined how very similar dialects around the Belgian⫺Dutch⫺German border diverge because of their exposure to the influence of different standard languages. The geographical dimension is only one of the dimensions in the architecture of a language. In order to capture the complex nature of the latter, it is useful to consider polarized continua such as those suggested by creolistics (Rickford 1987): in fact, each dimension of variation is conceivable as a continuum, so that the resulting general picture of variation in a language space takes the form of a sum of intercrossing continua, one nonpolarized (the diatopic variation) and two polarized (the diastratic and diaphasic variation), as sketched in Figure 13.1. As Downes (1984: 28) points out, “the linguistic side of any variety […] is a clustering tendency within a continuum”. Thus, the result is a Kontinuum mit Verdichtungen (Berruto 1987, “continuum con addensamenti”); varieties in this continuum represent concentration areas, where a variety, though not clearly-cut separated from other varieties, is identified by a particular frequency of certain variants, by the co-occurrence of several features and possibly by some diagnostic traits, which appear in that variety only. A variety appears where such a concentration, or condensation, takes place. A variant, an item or a linguistic feature can spread along a certain sector of a dimension or even over several dimensions. This is for instance evident in the case of diastratia and diaphasia, because an item (i. e., a certain pronunciation or a certain morphemic opposition or construction or syntactic structure or lexeme etc.) can appear in the lower part of the diastratic dimension as much as in the lower part of the diaphasic dimension. There is no room here for discussing the idea of Bell (1984), that stylistic variation is a subset of social variation. When such a particular item occurs with high frequency and in cooccurrence with other features, thus forming a belt of concentration, the whole may constitute a variety.

13. Identifying dimensions of variation

Fig. 13.1: Architecture of a language as a multidimensional continuum

The intersecting continua are additionally structured by the fact that the diastratic and diaphasic dimensions stretch from a high pole to a counterposed low pole. The high pole corresponds to a prestigious, socially preferred position, being occupied by formal, written and elaborated varieties on the situational dimension and by the variety of educated, upper class people on the social dimension; the low pole corresponds to low prestige, socially dispreferred positions, occupied by informal, spoken varieties (casual speech, slang, etc.) and by the variety of uneducated, lower class people respectively. Thus, every linguistic item is simultaneously characterized by a position over the three interplaying continua. For instance, a particular pronunciation of a phoneme or a particular morphological realization could be, roughly sketching, “Scottish, lower class, informal” or “Swabian, middle class, formal” and so on. This is the status of all sociolinguistically marked elements. There are of course also a good number of linguistic traits that are neutral with respect to the social variables and are, therefore, sociolinguistically unmarked. A speaker adopting a certain variant or employing a certain variety places him/herself not only at a given point of the linguistic space he/she lives in, but also in a particular position within the speech community, because each variety has a symbolic value such that its use amounts to an “act of identity” (Le Page and Tabouret-Keller 1985). The standard variety occupies the high poles of the diaphasic and diastratic dimensions. What does not belong to the standard is labeled nonstandard by Anglo-Saxon linguists, while in the European continent the term substandard is often preferred. In German dialectology, there is a tendency to see substandard varieties as a series of more or less marked, intermediate varieties between standard and dialect, from the near-stan-

237

238

III. Structure and dynamics of a language space dard to the most localized and typical dialect (Bellmann 1998), but there is wide agreement on the fact that the concept of substandard is not exhausted by the dichotomy standard/(geographical) dialect, since everything, in the architecture of a language or a language space, which is simply below the standard, can be referred to as substandard, whether it is dialectal or not (Albrecht 1986; Mattheier 1990).

7. Reerences Albrecht, Jörn 1986 “Substandard” und “Subnorm”. Die nicht-exemplarischen Ausprägungen der “Historischen Sprache” aus varietätenlinguistischer Sicht. In: Holtus and Radtke (eds.), 65⫺88. Albrecht, Jörn 2003 Die Standardsprache innerhalb der Architektur europaı¨scher Einzelsprachen. Sociolinguistica 17: 11⫺30. Ammon, Ulrich 1986 Explikation der Begriffe “Standardvarietät” und “Standardsprache” auf normtheoretischer Grundlage. In: Holtus and Radtke (eds.), 1⫺63. Ammon, Ulrich 1989 Towards a descriptive framework for the status/function (social position) of a language within a country. In: Ulrich Ammon (ed.), Status and Function of Languages and Language Varieties, 21⫺106. Berlin: de Gruyter. Ammon, Ulrich 1995 Die deutsche Sprache in Deutschland, Österreich und der Schweiz. Das Problem der nationalen Varietäten. Berlin: de Gruyter. Ammon, Ulrich 2003 On the social forces that determine what is a standard in a language and on conditions of successful implementation. Sociolinguistica 17: 1⫺10. Ammon, Ulrich 2004 Standard variety. In: Ammon et al. (eds.), 273⫺283. Ammon, Ulrich, Norbert Dittmar, Karl J. Mattheier and Peter Trudgill (eds.) 2004 Sociolinguistics/Soziolinguistik: An International Handbook of the Science of Language and Society, vol. 1. (⫽ Handbücher zur Sprach- und Kommunikationswissenschaft 3.1.) Berlin: de Gruyter. Auer, Peter 1997 Co-occurrence restrictions between linguistic variables. In: Frans Hinskens, Roeland van Hout and Leo Wetzels (eds.), Variation, Change and Phonological Theory, 69⫺99. Amsterdam: Benjamins. Auer, Peter 2005a Europe’s sociolinguistic unity, or: A typology of European dialect/standard constellations. In: Nicole Delbecque, Johan van der Auwera and Dirk Geeraerts (eds.), Perspectives on Variation, 7⫺42. Berlin: Mouton de Gruyter. Auer, Peter 2005b The construction of linguistic borders and the linguistic construction of borders. In: Markku Filppula, Juhani Klemola, Marjatta Palander and Esa Penttilä (eds.), Dialects Across Borders, 3⫺30. Amsterdam: Benjamins. Barbour, Stephen and Patrick Stevenson 1990 Variation in German: A Critical Approach to German Sociolinguistics. Cambridge: Cambridge University Press. Bell, Allan 1984 Language style as audience design. Language in Society 13: 145⫺204.

13. Identifying dimensions of variation Bellmann, Günter 1998 Between base dialect and standard language. Folia Linguistica 32: 23⫺34. Berruto, Gaetano 1987 Sociolinguistica dell’italiano contemporaneo. Rome: Carocci. Berruto, Gaetano 1989a Main topics and findings in Italian sociolinguistics. International Journal of the Sociology of Language 76: 7⫺30. Berruto, Gaetano 1989b On the typology of linguistic repertoires, In: Ulrich Ammon (ed.), Status and Function of Languages and Language Varieties, 552⫺569. Berlin: de Gruyter. Berruto, Gaetano 1995 Fondamenti di sociolinguistica. Rome: Laterza. Biber, Douglas and Edward Finegan 1994 Sociolinguistic Perspectives on Register. Oxford: Oxford University Press. Bickerton, Derek 1973 On the nature of a Creole continuum. Language 49: 640⫺669. Bloomfield, Leonard 1933 Language. New York: Holt. Britain, David 2004 Dialect and accent. In: Ammon et al. (eds.), 267⫺273. Chambers, Jack K. 1995 Sociolinguistic Theory. Linguistic Variation and its Social Significance. Oxford: Blackwell. Chambers, J. K. and Peter Trudgill 1980 Dialectology. Cambridge: Cambridge University Press. Clyne, Michael (ed.) 1992 Pluricentric Languages: Differing Norms in Different Nations. Berlin: Mouton de Gruyter. Cornips, Leonie 1998 Syntactic variation, parameters, and social distribution. Language Variation and Change 10: 1⫺21. Cornips, Leonie and Karen P. Corrigan (eds.) 2005 Syntax and Variation. Reconciling the Biological and the Social. Amsterdam. Benjamins. Coseriu, Eugenio 1980 “Historische Sprache” und “Dialekt”. In: Joachim Göschel, Pavle Ivic´ and Kurt Kehr (eds.), Dialekt und Dialektologie, 106⫺122. Wiesbaden: Steiner. Dittmar, Norbert 1995 Theories of sociolinguistic variation in the German context. In: Stevenson, Patrick (ed.), The German Language and the Real World: Sociolinguistic, Cultural, and Pragmatic Perspectives on Contemporary German, 135⫺167. Oxford: Oxford University Press. Dittmar, Norbert 1997 Grundlagen der Soziolinguistik ⫺ Ein Arbeitsbuch mit Aufgaben. Tübingen: Niemeyer. Downes, William 1984 Language and Society. Cambridge: Cambridge University Press. Durrell, Martin 2004 Sociolect. In: Ammon et al. (eds.), 200⫺205. Eckert, Penelope 1998 Age as a sociolinguistic variable. In: Florian Coulmas (ed.), The Handbook of Sociolinguistics, 151⫺167. Oxford: Blackwell. Eckert, Penelope and John R. Rickford (eds.) 2001 Style and Sociolinguistic Variation. Cambridge: Cambridge University Press. Gabelentz, Georg von der 1891 Die Sprachwissenschaft, ihre Aufgaben, Methoden und bisherigen Ergebnisse. Leipzig: Weigel.

239

240

III. Structure and dynamics of a language space Gadet, Franc¸oise 2003 La variation sociale en franc¸ais. Gap: Ophrys. Georg, Werner 2004 Schicht. In: Ammon et al. (eds.), 378⫺383. Gerritsen, Marinel 1999 Divergence of dialects in a linguistic laboratory near the Belgian⫺Dutch⫺German border: Similar dialects under the influence of different standard languages. Language Variation and Change 11: 43⫺65. Gregory, Michael 1967 Aspects of varieties differentiation. Journal of Linguistics 3: 177⫺198. Halliday, M. A. K. 1978 Language as Social Semiotic. The Social Interpretation of Language and Meaning. London: Arnold. Halliday, M. A. K. 1985 Spoken and Written Language. Geelong: Deakin University Press. Heeringa, Wilbert and John Nerbonne 2001 Dialect areas and dialect continua. Language Variation and Change 13: 375⫺400. Holtus, Günter and Edgar Radtke (eds.) 1986 Sprachlicher Substandard, vol. 1. Tübingen: Niemeyer. Hudson, Richard A. 1996 Sociolinguistics. 2 ed. Cambridge: Cambridge University Press. Johnstone, Barbara 2004 Place, globalization, and linguistic variation. In: Carmen Fought (ed.), Sociolinguistic Variation: Critical Reflections, 65⫺83. Oxford: Oxford University Press. Jørgensen, J. Normann and Kjeld Kristensen 1995 On boundaries in linguistic continua. Language Variation and Change 7: 153⫺168. Koch, Peter and Wulf Österreicher 1990 Gesprochene Sprache in der Romania: Französisch, Italienisch, Spanisch. Tübingen: Niemeyer. Labov, William 1966 The Social Stratification of English in New York City. Washington, DC: Center for Applied Linguistics. Labov, William 1994 Principles of Linguistic Change, vol. 1, Internal Factors. Oxford: Blackwell. Labov, William 2001 Principles of Linguistic Change, vol. 2, Social Factors. Oxford: Blackwell. Le Page, Robert B. and Andre´e Tabouret-Keller 1985 Acts of Identity. Cambridge: Cambridge University Press. Macaulay, Ronald K. S. 1988 The rise and fall of the vernacular. In: Caroline Duncan-Rose and Theo Vennemann (eds.), On Language. Rhetorica, Phonologica, Syntactica: A Festschrift for R. P. Stockwell, 106⫺115. London/New York: Routledge. Mattheier, Karl J. 1990 Überlegungen zum Substandard im Zwischenbereich von Dialekt und Standardsprache. In: Günter Holtus and Edgar Radtke (eds.), Sprachlicher Substandard, vol. 3, 1⫺16. Tübingen: Niemeyer. Milroy, James 1992 Linguistic Variation and Change: On the Historical Sociolinguistics of English. Oxford: Blackwell. Milroy, Lesley 1980 Language and Social Networks. Oxford: Blackwell. Rickford, John R. 1987 Dimensions of a Creole Continuum. Stanford, CA: Stanford University Press.

14. Horizontal convergence

241

Romaine, Suzanne 1994 Language in Society. An Introduction to Sociolinguistics. Oxford: Blackwell. Sanga, Glauco 1981 Le dynamiques linguistiques de la socie´te´ italienne (1861⫺1980): de la naissance de l’italien populaire a` la diffusion des ethnicismes linguistiques. Langages 61: 93⫺115. Siegel, Jeff 1985 Koines and koineization. Language in Society 14: 357⫺378. Tagliamonte, Sali A. 2006 Analysing Sociolinguistic Variation. Cambridge: Cambridge University Press. Trudgill, Peter 1974 The Social Differentiation of English in Norwich. Cambridge: Cambridge University Press. Trudgill, Peter 1995 Sociolinguistics: An Introduction to Language and Society. Harmondsworth: Penguin. Trudgill, Peter 2002 Sociolinguistic Variation and Change. Edinburgh: Edinburgh University Press. Whitney, William Dwight 1875 The Life and Growth of Language: An Outline of Linguistic Science. New York: Appleton & Co. Wolfram, Walt and Natalie Schilling-Estes 1998 American English: Dialects and Variation. Malden, MA/Oxford: Blackwell.

Gaetano Berruto, Torino (Italy)

14. Horizontal convergence o linguistic varieties in a language space 1. 2. 3. 4. 5. 6.

Introduction Some terms and definitions Diachronic and synchronic evidence Factors leading to convergence Conclusions References

1. Introduction In traditional diachronic map-based dialectology, priority is given to divergence in an originally uniform language space (cf. Schrambke, this volume; Harnisch, this volume). Dialect divergence is explained in terms of natural or man-made borders which limit the spread of a change in that they impede communication and interaction (Bach 1969: 80⫺ 81; Murray, this volume; Paul 1920: § 22⫺25; Trudgill 1986). Auer (2004) suggests a cognitive interpretation of these borders as mental boundaries that crystallize out of cultural and political borders.

242

III. Structure and dynamics of a language space In contrast to this, the focus of this article is on the convergence of varieties in a language space. Convergence is a process by which similarities between varieties increase. It implies a historical dimension and therefore language change. If one reverses the above argument, convergence and the leveling of varieties should be the result of communication between speakers, or of the speakers’ sense of a common belonging. Leveling and mixing lead to less diversity and greater uniformity, which can imply convergence towards a common standard variety along a dialect ⫺ standard language dimension (cf. Røyneland, this volume). On the other hand, there are processes of mixing or leveling between dialects that lie beyond the dialect ⫺ standard language dimension. This article on horizontal convergence deals with such processes. However, in areas with an established standard language it is often difficult to clearly separate horizontal convergence from vertical convergence (cf. Hinskens, Auer and Kerswill 2005; Radtke 2006).

2. Some terms and deinitions Convergence in a broad sense refers to the reduction of differences between varieties. Divergence, in contrast, refers to an increase in such differences. But convergence and divergence are often just two sides of the same coin: when the differences between one variety and another are reduced, the differences between the first variety and yet another are usually increased, as already argued by Martinet (1962: 105). For instance, the convergence of the varieties of Le¨tzebuergesch towards central Luxemburgish involves the abandonment of east Luxemburg features and, therefore, a divergence away from the Moselle-Franconian dialects of German (Gilles 1998: 1999). Similar developments are shown in Siebenhaar (2008), where particular aspects of regional convergence in Bern in Switzerland imply a divergence from features of neighboring varieties, even though these are closer to standard German. Developments of regional convergence which imply a divergence from the standard variety have also been documented in the West Middle German area (Bellmann 1998).

2.1. Maniestations o convergence In historical linguistics, in the aftermath of the neogrammarian approach (cf. Murray, this volume), a principal focus was laid on explaining language change in terms of a concept of divergence reducible to sound laws that allow of no exception. This concept is visualized and idealized in the family tree diagrams of historical linguistics. However, the sound laws do show exceptions. Besides analogy, the main explanation for these exceptions was sought in lexical borrowings. Borrowings are induced by interaction between speakers and were therefore conceived as externally motivated language change. The consequences of borrowings are dialect mixing and leveling between dialects and thus a reduction of differences, i. e., convergence. Convergence on the horizontal level can manifest itself in three forms. Firstly, there may be a displacement of isoglosses. In these cases, the variety borrowing a feature from a neighboring variety converges on this variety and at the same time diverges away from other varieties that maintain this feature. The displacement of isoglosses has often been

14. Horizontal convergence described in martial metaphors as driving a wedge into a linguistic area, repelling or driving back a variant (cf. the models of areal distribution of linguistic forms in Goossens 1977). A classic example is the traditional interpretation, in the work of Theodor Frings for instance, of the Rhenish fan, which represents the High German consonant shift across space. As a second manifestation, there can be a dialect leveling, i. e., a convergence of varieties in an area. Each of these varieties loses those features that distinguish it from the surrounding majority of varieties. Leveling can therefore be seen as the minority variety converging upon the majority variety. In this sense leveling is unilateral and corresponds to Mattheier’s (1996: 34) concept of advergence. For example, Hinskens (1996) discusses the loss of the /x ~ c¸ / allophony rule in the Limburg dialect region of Holland, where /c¸ / wins out and occupies the place of /x/ in standard Dutch. Christen (1998: 179⫺185) documents the loss of local phonetic and morphological variants in favor of regional forms in Swiss-German dialects. For instance, the local diphthongs [e=4 /ow4] for MHG eˆ / oˆ, aˆ in parts of the canton of Schwyz (as in Leirer ‘teacher’, Oubed ‘evening’) are given up in favor of Leerer, Aabed, which are general Zürich variants. Bothorel-Witz and Huck (2000) document among other phenomena the loss of the Alsatian palatalization of MHG u: Hünd becomes Hund ‘dog’, Lümpe becomes Lumpe ‘floorcloth’. All these “winning” forms have a broader acceptance in the dialectological space and partially correspond to standard German. In these cases it is therefore difficult to separate horizontal and vertical convergence. A third form of convergence is koineization, which is the emergence of a de-localized variety. Unlike leveling, where one dialect loses features not present in the other, koineization involves a mixing of features from different dialects leading in a stabilized compromise dialect as a result of contact between speakers of mutually intelligible varieties of a language. Koineization therefore cannot be described at the level of single features. It necessarily involves a set of different features. Kerswill (2002) describes the emergence of a de-localized variety in the new town of Milton Keynes. Trudgill (2008) argues that dialect mixture in the colonial varieties of European languages, such as Brazilian Portuguese, Canadian French and Australian English, is an inevitable result of dialect contact. A koine´, however, does not necessarily imply that the basic dialects are abandoned. In ancient Greek, the koine´ did not replace the dialects, it was used as a lingua franca between speakers of different dialects. In New Zealand, on the other hand, where the speakers of the traditional British English dialects were delocalized, the English koine´ emerged as a new vernacular, as the old dialects had lost their speech communities (cf. Lenz, this volume).

2.2. Linguistic variables: From the lexicon to phonetics and pragmatics In principle, convergence in space can be found at every level of the linguistic system. However, research has concentrated on the traditional dialectological fields of phonetics, morphology, and lexicon, and it is harder to find studies on phonology and pragmatics. Convergence in syntax is often the focus of typological studies. Since the neogrammarians, lexical borrowing has been seen as a primary trigger of horizontal convergence. Borrowings are easily integrated into the open system of the receiving language’s lexicon. Such borrowings are more often observed in vernacular

243

244

III. Structure and dynamics of a language space varieties near the language boundary than in the interior of a language area (cf. the documentations of French borrowings in Swiss German by Steiner 1921, in Alsatian by Bothorel-Witz and Huck 2000, in the Saarland dialects by Braun and Treib 2008). Lexical borrowings, however, are more likely in closely related varieties than across languages. The Swiss-German dialects are all Alemannic dialects, and mutually intelligible. In the linguistic atlas of German-speaking Switzerland (SDS: IV, 83) the verb ‘to nudge’ is represented by 15 lexical types with some phonetic variants. In the midlands there is an eastern type stupfe/stüpfe and a western type mupfe/müpfe. The areal distribution with its wedge-shaped isogloss suggests that the eastern type stupfe/stupfe has been borrowed into the western area, thereby driving back the traditional mupfe/müpfe type. The lexical difference between the two dialects is decreasing, and the dialects are leveling out, i. e., they are converging. Borrowings on the lexical level normally mark the first step of convergence, referred to as Mischung ‘mixing’ by the Neogrammarians, while morphemes and sounds are only transferred later, on the basis of a large number of borrowed words (Paul 1920: § 274⫺ 285). An example of such a morphological convergence emerging from language contact is the import of French derivational suffixes into medieval German and English. A multitude of French loan words from the courtly milieu ending in -erie caused the transfer of this French suffix into MHG (-erıˆe > -erei) and ME -erie > -ery. The suffixes are still productive and can be used with Germanic stems as in Bäckerei/bakery, Schweinerei/piggery. Again, morphological borrowing from one dialect to another within a language is more common than borrowing between languages because the morphemes can be transferred directly and do not need to be derived from a class of borrowed words. For instance, Renn (1994: 106⫺107) reports borrowing of the Bavarian verbal flexion -s for the second-person plural into eastern Swabian dialect areas around Augsburg: Swabian ihr geh⫹t ‘you (pl.) go’, which corresponds to standard German ihr geh⫹t, is pushed back in space by the Bavarian form ihr geh⫹ts. It is one of many variables that show a convergence of this traditionally Alemannic dialect region onto the dialect of the political centre within the state of Bavaria (cf. SBS 1997⫺2008). Another example comes from the Swiss midlands, where there are two paradigms for the verbal plural: the western dialects still show the two forms -e, -ed, -e as in standard German, while in the eastern dialects we find a newer uniform suffix (-ed ) for all three persons. However, there are two distinct isoglosses, one for the suffixes of the full verbs and one for the monosyllabic short verbs (such as sii ‘to be’, haa ‘to have’, tue ‘to do’, gaa ‘to go’, staa ‘to stand’, choo ‘to come’). While the isogloss for the full verbs follows the river Reuss, the isogloss for the short verbs lies forty kilometers to the west. Between these two isoglosses there is a region with two different paradigms, one for full verbs and one for short verbs. Bangerter (1951: 6 and 110) interprets this situation as one in which the western forms are advancing while Hotzenköcherle (1961 [1986: 90]) believes that the eastern forms are spreading. His interpretation is supported by recent empirical evidence (Siebenhaar 2000: 141⫺146). In the central part of this region a new compromise form has even emerged for the full verbs: the combination of the eastern uniform plural and the western morphemes of the first and third persons plural has resulted in an -e-plural for all three persons, accommodating the paradigm of the full verbs to that of the short verbs. Leveling in this region is not only borrowing, it results in a fusion of the two paradigms. Therefore, paradoxically, convergence results in a new paradigm that leads to divergence from both neighboring areas in this area.

14. Horizontal convergence Convergence on the phonetic and phonologic level can follow the same patterns. In neogrammarian terminology, this is referred to as sound substitution (Lautersatz) in opposition to sound change (Lautwandel), which is discussed below. Once again, convergence between dialects is more common but often less evident than convergence between languages, because in most cases the phoneme system is not affected; it is only the substitution of a sound in specific words (Lautersatz) or the substitution of words (Wortverdrängung) that increases the relative frequency of a sound. However, these substitutions may lead to phonological reorganization. For example, the region around Zürich has a highly differentiated system of closing diphthongs with traditionally six closing diphthongs (/ow4/ /æw4/ /ø=4/ /œ=4/ /e=4/ /æ=4/) (cf. Fleischer and Schmid 2006). But western dialects of Switzerland have only three closing diphthongs with a close-mid first segment (/ow4/ /ø=4 ~ œ=4/ /e=4/), while we find three corresponding variants with an openmid or open first segment in the eastern Swiss-German dialects (/æw4/ /œ=4/ /æ=4/). The Zürich system is now in a stage of reorganization as the close-mid diphthongs (/ow4/ /ø=4/ /e=4/), which have a low type frequency, are replaced by the more open variants (/æw4/ /œ=4/ /æ=4/) that have a high type frequency. Fleischer and Schmid (2006: 248) already report the merger of /ø=4/ and /œ=4/. Recent recordings in the eastern (Siebenhaar, unpublished) and western parts (Siebenhaar 2000) also unveil insecurities in the use of traditional /ow4/ and /e=4/. So frei ‘free’ can now be found with the traditional /e=4/ as well as with a new /æ=4/, Sou ‘sow’ is realized with traditional /ow4/ and with new /æw4/. This substitution of sounds will result in a merger of /e=4/ and /æ=4/ and /ow4/ and /æw4/, respectively. The mergers level out a difference between the Zürich dialect region and the eastern dialects, where these sounds merged long ago; we are therefore dealing with an instance of convergence in space. However, it is also possible to link this change to the standard German system, which for the closing diphthongs is the same as that of the eastern Swiss-German dialects; in this case this change would have to be interpreted as a case of vertical convergence (Røyneland, this volume) that would have nothing to do with space. The Atlas of North American English (Labov, Ash and Boberg 2005) provides modern examples of the spreading across space of sound changes that seem not to originate from a word-by-word adoption, but from a system-internal principle of maximal dispersion corresponding to the neogrammarian concept of Lautwandel. These maps document the phonetic dynamics of sound changes in progress and their phonological implications. Labov, Ash and Boberg base their maps not only on auditory transcriptions but also on acoustic measurements of formants, which allows them to discover sound changes in progress and to reconstruct the progression of a sound change in geographical space. Implicational relationships between phonologically coherent developments shed light not only on the notion of divergence from a uniform American English, but also on regional convergence in the new vernaculars. As historical pragmatics has only moved into focus in recent decades, aspects of pragmatic convergence have not yet been studied in detail. For instance, address systems have changed in most European languages (as documented in the volume edited by Taavitsainen and Jucker 2003), but this prominent pragmatic change is not discussed within the concept of convergence, although interrelations between the different systems are often mentioned. An example may be found in the German dialects of Bern and Freiburg, which are adjacent to the French language area. These dialects have conserved the honorific second-person plural pronoun and verbal form, while in most German-

245

246

III. Structure and dynamics of a language space speaking areas the third-person plural forms are used to express politeness. Convergence towards the new German system has therefore not yet had an influence in this area. However, the penetration of the third-person plural forms has been deplored by dialect purists over the years. Likewise, changes in communicative genres can be analyzed as part of a cultural history of communication (Linke 2007). It is possible to interpret these changes as pragmatic convergence. To date, they have rarely been linked to the spatial dimension. Yet, this connection could revitalize the interdisciplinary cooperation between cultural anthropology and dialectology that was characteristic of the first half of the twentieth century in the German-speaking area. The following example is illustrative of this phenomenon. On the level of dialect contact, Cheshire, Kerswill and Williams (2005) discuss the use of the discourse particles like and innit in three British towns: Hull, Milton Keynes and Reading. Like as a focus marker and a marker of reported speech has been observed not only in Britain but in urban centers throughout the English-speaking world. Its origin is supposed to lie in southern California. The rapid spread of like has been associated with youth culture with an international dimension. Space can only be referred to here in the metaphoric sense of a global village, where distance no longer plays a crucial role. However, in young people’s speech there is a social class difference in Hull, with the middle class using like more often than the working class, while in the other towns there are no such differences. The discourse marker innit has been reported to be in use among young people in London for twenty years, and it is now also used by middle class Londoners. London working-class youth use it non-paradigmatically, and it is exclusive to workingclass speakers in Hull, Milton Keynes and Reading. Cheshire, Kerswill and Williams (2005: 156) suggest that, due to their international dissemination by US films and TV shows, the convergence in the use of globally innovative features such as like is not led by any single social group. The spread of like demonstrates that a convergence need not be bounded by space or social class; the communicative space transcends the traditional geographical understanding of space. Innit, on the other hand, spreads in a regionally and socially defined way. Convergence here follows the classic path from one social group to another and from one region to another. However, Cheshire, Kerswill and Williams (2005: 158) problematize their results, asking if convergence in this case can be tied to the use of a new, invariant tag, or whether it represents a convergence in interactional style, which becomes more involved and addressee-oriented and includes new politeness strategies. Up to now, convergence has been discussed mainly with a focus on individual phenomena. But convergence understood as a loss of structural differences normally affects more than individual features. For instance, von Polenz’ (1954) description of the relict area of Altenburg in Thüringen shows a bundle of concentric circles around the town. Swiss-German dialectology has described the Brünig-Napf-Reuss boundary (Haas 2000; Weiss 1947), where many isoglosses separating eastern and western linguistic forms coincide. Wiesinger’s (1983) map of the German dialect areas shows core dialect areas with only a few isoglosses and transition zones that are marked by isogloss bundles. Such distributions, where many isoglosses coincide, suggest that the spread of new forms will also stop at the isogloss bundles. Additional isoglosses therefore increase divergence at these borders and increase convergence within the more coherent areas. Actual data on this aspect is rare, albeit some evidence can be found in the MRhSA (1994⫺2002). Here, ongoing changes in the realization of MHG aˆ, MHG a or MHG üe seem to bundle at

14. Horizontal convergence the isogloss dividing the voiced and unvoiced variants of /z/. Similar results for the Main-Franconian area are reported by Schunk (1999). In this area, the isogloss separating palatal and velar realizations of a is reinforced by the actual development of regional dialects, themselves the result of dialect convergence.

3. Diachronic and synchronic evidence Convergence has an intrinsically diachronic dimension. Research on language in space collects data in different places so as to represent spatial diversity. Adding a diachronic perspective at least doubles the amount of data needed. Moreover, over the years research interests change, so that directly comparable data recorded with the same method at different times and places are rarely available. In most cases the old questionnaires and methods are adapted to the new questions (e.g., Bigler 1979; Cornips and Corrigan 2005; Schifferle 1995). Yet Bailey (2002) stresses that even small differences in sampling procedures may have significant effects on the results. Furthermore, demographic changes influence the results and must not be confused with actual linguistic change. Panel studies, conducted with the same subjects after a certain interval, pose practical problems. It has to be kept in mind when interpreting the results that individual speakers can change their linguistic behavior in different ways, as documented in Bausch (2000) or Siebenhaar (2002), to name two of the few panel studies on German dialects. Although a direct analysis of convergence or, generally, language change in space is therefore often not possible, synchronic evidence can be substituted for the real-time diachronic dimension. Since Labov’s (1963, 1966) famous studies on Martha’s Vineyard and in New York City, the apparent-time construct, which interprets synchronic variation as change over time, has become an established substitute for real-time observations of language change. One of the atlases that use the apparent-time method is the Linguistic Atlas of the Middle Rhine (MRhSA 1994⫺2002, cf. the methodological discussion in Bellmann 1986). Many phenomena are presented in two maps, one displaying the traditional dialect of the NORMs, while the other represents the realizations of the more mobile middle-aged generation. Differences between corresponding locations can be interpreted as language change. When these changes occur in coherent areas or when they bundle to one side of the isoglosses, they can be interpreted as changes in space. Map 14.1 gives an example of biserial mapping, where the data from the younger and the older subjects are combined. The red symbols indicate places where the younger, more mobile subjects use another variant than the older informants. The map shows that along the lower Moselle the younger mobile subjects replace the voiceless labiodental fricative ⫺ dominant across the greater part of the investigation area and, moreover, consonant with the standard German form ⫺ (here, in the intervocalic position within the word pfeifen ‘to whistle’) with a new, voiced bilabial or labiodental fricative. The apparent-time construct allows this difference to be interpreted as language change, in this case convergence towards a regional dialect. Another, more traditional substitute for the temporal dimension is of course areal distribution. Based on the assumption that language change spreads through space, linguistic maps can be interpreted historically. An example is given in Map 14.2 which shows the distribution of the personal pronoun uns ‘us (dat and acc)’ in the SwissGerman dialects. The Old High German accusative form unsih is first transformed to

247

248

III. Structure and dynamics of a language space

Map 14.1: West Germanic p: pfeifen ‘to whistle / to pipe’ (MRhSA 1994⫺2002: vol. 4, 320); the red symbols mark differences between the first (older speakers) and the second (younger speakers) data series

üns (following German sound laws: umlaut [u > ü before i] plus syllable reduction). This form of the pronoun is still found to the southwest of Bern. Emanating from this form we find two developments. The result of the first development is insch/ünsch in eastern Wallis and Graubünden. Here, s is palatalized to sch [s] and ü [y] is in part unrounded to i. The second development again starts from the üns form: the area with the form üs [y6s] shows elision of n and a lengthening of the vowel, a development known as Staub’s law. In the area with öis, then, the long vowel is diphthongized. Finally, in the northeast (Basel), the standard German form is adopted, which is the reflex of the dative form OHG uns. Following this line of argument, the areal distribution reflects diachronic change. Any of these steps is a divergent step away from the previous variant. Map 14.3 shows the classic example of the Rhenish fan, the gradual progression of the High German consonant shift in space which results in a graded landscape (Staffellandschaft). In the southern dialects all voiceless plosives are affected by the consonant shift. The northern German dialects do not take part in this sound change, whereas in the central German dialects, the sounds are only partially affected, and furthermore only in certain contexts. In the entire upper and central German area, intervocalic voiceless plosives became geminated fricatives (maken/machen in Map 14.3), and postvocalic ones,

14. Horizontal convergence

Map 14.2: uns ‘us (dat/acc)’ (Hotzenköcherle 1961: 221)

Map 14.3: Rhenish fan (Wolf 1983: 1118)

249

250

III. Structure and dynamics of a language space in final position, single fricatives. However, the isoglosses for the highly frequent articles (dat/das in Map 14.3) run further to the south and those for the pronouns (sik/sich, ik/ ich) further north. In initial position, in the geminate and after a consonant, the voiceless plosives /t, p, k/ became affricates /ts, pf, kx/; /pf/ was later simplified to /f/ after liquids. The shift /t/ > /ts/ affected all central and upper German dialects, while /k/ > /kx/ only affected the south Alemannic and south Bavarian dialects. The shift /p/ > /pf/ shows a gradual distribution in the central German dialects: a shift after a liquid consonant south of the Eifel (dorp/dorf in Map 14.3), an additional shift after nasals and in the geminate to the south of Speyer (appel/apfel in Map 14.3), and finally some kilometers further south, a shift in all positions (pund > pfund). Frings (1957, based on earlier studies) interprets this areal distribution as a development that started in the southern dialects and affected more and more northern dialects; it therefore reflects a convergence of the northern dialects towards the southern ones. However, this monogenetic approach is not uncontroversial today (cf. the discussion in Wolf 1983).

4. Factors leading to convergence Convergence is generally understood to be contact induced and therefore externally motivated. Yet internal factors such as homogenization, simplification and the reduction of system-internal quantitative variation can also play a part in convergence. Internal and external factors are often hard to distinguish. Moreover, what begins as an internal sound change in one variety can be transferred to another by borrowing and lexical diffusion. The contraction of the diphthong system in Zürich discussed above is an example of this pattern of an ongoing contact-induced phonemic merger. On the basis of the opening and closing of vowels, Haas (1978) has shown how internal changes within one dialect can spread to other dialects through contact. In the different dialects which have adopted the features, the new sounds can lead to different system-internal homogenizations and thus to different systems. And these new systems can in turn be diffused again. The following sections try to differentiate internal and external motivations for convergence in space.

4.1. Sociolinguistic actors Traditionally, differences between dialects are explained in terms of natural or manmade borders which impede or restrict communication (Bach 1969: 80⫺81; Murray, this volume; Paul 1920: § 22⫺25; Trudgill 1986). As a consequence, linguistic features are kept from spreading, and the coherence of the systems on both sides of the border is strengthened, as Schifferle (1995) shows for the border separating Germany and Switzerland, and Harnisch (this volume) for the former border between East and West Germany. Schönfeld, Reiner and Grünert (2001) document the linguistic differences in the formerly divided city of Berlin, where the dialect converged on FRG or GDR norms on the western and eastern side of the border, respectively. Another example is the Northern Cities Shift in American English. The shift does not cross the political border between

14. Horizontal convergence the United States and Canada (Boberg 2000; Labov, Ash and Boberg 2005). In reversing the argument, Aubin, Frings and Müller (1926) use linguistic borders to reconstruct and explain cultural history (cf. Knobloch, this volume). Since the political boundaries in Europe have become less relevant in recent decades, and since modern means of transportation easily surmount natural barriers, one would expect the linguistic effects of such borders to disappear; nevertheless, the linguistic borders coinciding with the political borders remain and are even reinforced, as Auer (2004) shows for the political borders of Germany. He argues that political borders do not directly influence communicative and linguistic borders, but that divergence at these borders is instead a result of cognitive structures; the idea of a linguistic space, which is bound to (former) political borders, also creates linguistic borders. The consequence is that speakers who believe they speak the same dialect abandon linguistic features that separate them from the group they want to belong to. Convergence is therefore a consequence of speakers’ belief that they speak the same variety. Mobility and migration favor convergence as the contact between formerly distinct varieties increases. This aspect is important not only for language contact, but also for convergence between dialects, which are affected by commuters’ movements and by small-scale internal migration. Convergence occurs as migrants from nearby areas do not or only partially adapt to local forms that then disappear as a result. Wolfensberger (1967) has illustrated this process in an early apparent-time study in the small town of Stäfa that came within the linguistic sphere of influence of the Zürich metropolis. The opposite holds for commuters (Muhr 1981; Siebenhaar 2000), who disfavor local variants and adopt variants with greater regional acceptance, again leading to convergence. Convergence is a notion operating on an interactional, an individual and a grammatical level and it is one of the key concepts of communication accommodation theory (Giles and Powesland 1975: 154⫺170). Within this interactionist theory, the focus lies on the performance of the individual. In an interaction, an individual accommodates to his interlocutors in either a convergent or a divergent way. In converging, he or she avoids the distinctive features of his/her own variety or even adopts those of his/her coparticipant. Converging accommodation ⫺ linguistically and on other behavioral dimensions ⫺ is seen as a psychologically motivated effort to gain the interlocutor’s respect or to show solidarity. Based on Bell’s audience design (Bell 1984: 2006) and LePage and Tabouret-Keller’s acts of identity model (1985), Auer, Barden and Großkopf (1998) propose an identity projection model. In this model speakers converge not only towards their interlocutors, but also towards linguistic stereotypes of positively connotated groups. On the other hand, accommodation can also be divergent: in order to indicate dissent, a speaker can increase the linguistic and behavioral distance from his or her interlocutor. This short-term accommodation within an interaction is not permanent. However, if speakers regularly accommodate in a similar way over a longer period of time, this can lead to a long-term accommodation (Trudgill 1986: 11⫺38). If it stabilizes, it leads to a convergence at the level of the linguistic system. In recent years, many publications have identified prestige and attitudes as driving forces in language change. Along the vertical dimension prestige seems to be one of the prominent factors in convergence. Up until now, however, there is no clear support for this claim on the horizontal dimension. Three recent studies on language change and positive attitudes towards a place (Ortsloyalität) carried out in Switzerland (Hofer 1997; Leuenberger 1999; Siebenhaar 2000) are quite critical with respect to the causal link

251

252

III. Structure and dynamics of a language space between attitudinal factors and language change. On the other hand, Barden and Großkopf (1998) can establish attitudes as an explanatory factor for long-term accommodation among former GDR citizens in the west of Germany. For Schirmunski (1930), the salience of primary dialect features is a major stimulus for convergence. Without reference to Schirmunski, Trudgill (1986) identifies salience as one of the factors favoring accommodation as well, thereby moving the focus from language change to individual adaptation. However, salience is still a difficult concept to define, as subjective factors such as attitude, prestige and stigmatizing effects, as well as objective factors like frequency, system-internal transparency and linguistic contrast are involved independently of one another. Furthermore, Auer, Barden and Großkopf (1998) have shown that objective and subjective salience factors play different roles in the accommodation of lexicalized and non-lexicalized features. Salience thus seems to be important for convergence of dialects, but it is nevertheless only a necessary, not a sufficient condition for a linguistic feature to be affected by accommodation.

4.2. Structural actors Although external factors are seen as the main reason for convergence, there are some structural aspects that have to be taken into account. Horizontal convergence in space may be supported by the standard variety, if the standard has the same form as the converged-upon dialect. The opposite may hold if dialects converge in order to increase the distance from the standard. The expansion of l-vocalization from the western to the eastern Swiss-German dialects (Christen 2001), the emergence of regional substandard varieties in northern Germany in the sixteenth century (Elmentaler 2005) and the spread of the coronalization of the voiceless velar fricative [c¸] > [C] in western Germany (Herrgen 1986) are examples. In general, structures that are already similar favor convergence while distinct structures favor divergence. For the structuralists, intrasystemic coherence is the main factor in internally motivated language change. As early as 1918, Pfalz formulated the theory that vowels change in “rows” (Reihenschritte), meaning that front and back vowels of the same height and tension undergo the same sound changes. Wiesinger (1970) applied this concept to all of the German dialects. Likewise, Labov’s principles of vowel shifting (1994: 116) for English can be seen in the context of intrasystemically coherent sound change. As internal and external factors co-occur (cf. Torgersen and Kerswill 2004), it can be observed that a change is often not of a solitary nature; instead, structurally related aspects change along with it in a similar way. This applies not only to the sound system but also to morphology. Siebenhaar (2000) has shown that convergence affects the parts of a subsystem in a similar way; convergence in one subsystem, however, need not affect other subsystems. For the dialect of the small town of Aarau in Switzerland, factor analysis revealed that the vowel system tends to converge towards the western dialects and standard German, while morphology tends to converge on the eastern dialects. A growing number of studies within the framework of Optimality Theory discuss phonological and morphological change in regional substandard varieties (cf. Bresnan, Deo and Sharma 2007; Herrgen 2005; Wegener 1999). Convergence within this model can be seen as a transfer of constraints or rankings of constraints from one variety to another, which may ⫺ but need not ⫺ result in convergence in the surface systems.

14. Horizontal convergence

4.3. Frequency Another relevant factor is the frequency of the respective form, as already pointed out by Jakobson (1931). Highly used positions with a high functional load in a phoneme system are strengthened, while weakly used positions with a low functional load tend to be lost (see Martinet 1955: 91 for another formulation of this position). Frequency in this sense means the frequency of linguistic structures. However, the frequency of social contacts can also be important. Bloomfield (1933: 476) postulated a principle of density which says that people automatically influence each other’s language whenever they speak to each other. This principle is applied again by Labov (2001) who believes that the diffusion of a linguistic change can be reduced to a simple calculation and that the principle of density implicitly asserts that we do not have to search for a motivating force behind the diffusion of linguistic change. The effect is a mechanical and inevitable one; the implicit assumption is that social evaluation and attitudes play a minor role. (Labov 2001: 20)

However, the concept of frequency is not as simple as it seems: type frequency and token frequency have to be separated, different levels of the linguistic system may interfere with different frequencies, and measuring frequency in spoken interaction is still a challenge. Frequency effects have thus rarely been analyzed in actual research on convergence.

5. Conclusions Convergence is a process by which varieties become more similar due to frequent communication between speakers or due to the speakers’ belief that they speak a common variety (and belong to one social group). It is observed both on the vertical dialect ⫺ standard language dimension and on the horizontal dialect ⫺ dialect dimension. A clear distinction is often not possible. Convergence can be observed at all levels of the linguistic system, from phonetics to pragmatics. Nevertheless, lexical borrowings are often seen as the starting point of convergence. Out of lexical borrowings, morphological or phonological elements can be generalized, which may lead to grammatical convergence. However, if the varieties are closely related, it is also possible that morphemes, phonemes and phonetic features are directly borrowed from one system into another. On the horizontal level, convergence becomes manifest in (a) a displacement of isoglosses, (b) unilateral convergence or advergence or (c) koineization, which is the emergence of a delocalized variety. Most often convergence can be found in closely related varieties whose speakers are in intensive contact. Methodologically, convergence can be documented through real-time comparison. The apparent-time construct interprets synchronic variation as a consequence of changes over time and is used as a substitute for real-time observation. The areal distribution of linguistic features may also be interpreted as a function of time. Convergence is most often explained as contact induced and therefore externally caused. Explanations for convergence in space can be seen in such structural factors as a roofing language that influences horizontal convergence or frequency. However, external aspects are generally believed to be the main factors: socio-geographic or

253

254

III. Structure and dynamics of a language space mental structures, migration, overt and covert prestige, attitudes and salience have been considered to have an impact on convergence. Most of these explanations can be directly or indirectly related to an accommodation model of speech.

6. Reerences Aubin, Hermann, Theodor Frings and Josef Müller 1926 Kulturströmungen und Kulturprovinzen in den Rheinlanden. Geschichte, Sprache, Volkskunde. Bonn: Röhrscheid. Auer, Peter 2004 Sprache, Grenze, Raum. Zeitschrift für Sprachwissenschaft 23(2): 149⫺179. Auer, Peter, Birgit Barden and Beate Großkopf 1998 Subjective and objective parameters determining “salience” in long-term dialect accommodation. Journal of Sociolinguistics 2: 163⫺187. Auer, Peter, Frans Hinskens and Paul Kerswill (eds.) 2005 Dialect Change: Convergence and Divergence in European Languages. Cambridge: Cambridge University Press. Bach, Adolf 1969 Deutsche Mundartforschung. Ihre Wege, Ergebnisse und Aufgaben. 3rd ed. Heidelberg: Winter. Bailey, Guy 2002 Real and apparent time. In: J. K. Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.), The Handbook of Language Variation and Change, 312⫺332. (Blackwell Handbooks in Linguistics.) Malden, MA/Oxford: Blackwell. Bangerter, Arnold 1951 Die Grenze der verbalen Pluralendungen im Schweizerdeutschen. (Beiträge zur schweizerdeutschen Mundartforschung 4.) Frauenfeld: Huber. Barden, Birgit and Beate Großkopf 1998 Sprachliche Akkommodation und soziale Integration. Sächsische Übersiedler und Übersiedlerinnen im rhein-/moselfränkischen und alemannischen Sprachraum. (Phonai 43.) Tübingen: Niemeyer. Bausch, Karl-Heinz 2000 Dialektologie und interpretative Soziolinguistik am Beispiel des Sprachwandels im Rhein-Neckar-Raum. In: Dieter Stellmacher (ed.), Dialektologie zwischen Tradition und Neuansätzen. Beiträge der Internationalen Dialektologentagung, Göttingen, 19.⫺21. Oktober 1998, 78⫺98. (ZDL Beihefte 109.) Stuttgart: Steiner. Bell, Allan 1984 Language style as audience design. Language in Society 13: 145⫺204. Bell, Allan 2006 Speech accommodation theory and audience design. In: Keith Brown (ed.), Encyclopedia of Language and Linguistics, 648⫺651. 2nd ed. Oxford: Elsevier. Bellmann, Günter 1986 Zweidimensionale Dialektologie. In: Günter Bellmann (ed.), Beiträge zur Dialektologie am Mittelrhein, 1⫺55. (Mainzer Studien zur Sprach- und Volksforschung 10.) Stuttgart: Steiner. Bellmann, Günter 1998 Between base dialect and standard language. Folia Linguistica 32: 23⫺34. Bigler, Niklaus Jakob 1979 Mundartwandel im mittleren Aargau. Eine Untersuchung zu den heutigen Sprachverhältnissen im Spannungsfeld zwischen Ost- und Westschweizerdeutsch. (Europäische Hochschulschriften I, 264.). Frankfurt am Main: Lang.

14. Horizontal convergence Bloomfield, Leonard 1933 Language. New York: Henry Holt. Boberg, Charles 2000 Geolinguistic diffusion and the U.S.-Canada border. Language Variation and Change 12: 1⫺24. Bothorel-Witz, Arlette and Dominique Huck 2000 Die Dialekte im Elsaß zwischen Tradition und Modernität. In: Dieter Stellmacher (ed.), Dialektologie zwischen Tradition und Neuansätzen. Beiträge der Internationalen Dialektologentagung, Göttingen, 19.⫺21. Oktober 1998, 143⫺155. (ZDL Beihefte 109.) Stuttgart: Steiner. Braun, Edith and Evelyn Treib 2008 Keine Fisimatenten: Französische Wörter in saarländischen Mundarten. Merzig: Gollenstein. Bresnan, Joan, Ashwini Deo and Devyani Sharma 2007 Typology in variation: A probabilistic approach to be and n’t in the Survey of English Dialects. English Language and Linguistics 11(2): 301⫺346. Cheshire, Jenny, Paul Kerswill and Ann Williams 2005 Phonology, grammar, and discourse in dialect convergence. In: Auer, Hinskens, and Kerswill (eds.), 135⫺167. Christen, Helen 1998 Dialekt im Alltag. Eine empirische Untersuchung zur lokalen Komponente heutiger schweizerdeutscher Varietäten. (RGL 201.) Tübingen: Niemeyer. Christen, Helen 2001 Ein Dialektmarker auf Erfolgskurs: Die /l/-Vokalisierung in der deutschsprachigen Schweiz. Zeitschrift für Dialektologie und Linguistik 68: 16⫺26. Cornips, Leonie and Karen P. Corrigan 2005 Convergence and divergence in grammar. In: Auer, Hinskens, and Kerswill (eds.), 96⫺ 134. Elmentaler, Michael 2005 Die Rolle des überregionalen Sprachkontakts bei der Genese regionaler Umgangssprachen. Zeitschrift für deutsche Philologie 124: 395⫺415. Fleischer, Jürg and Stephan Schmid 2006 Zurich German. Journal of the International Phonetic Association 36(2): 243⫺253. Frings, Theodor 1957 Grundlegung einer Geschichte der deutschen Sprache. 3rd ed. Halle an der Saale: Niemeyer. Giles, Howard and Peter F. Powesland 1975 Speech Style and Social Evaluation. London: Academic Press. Gilles, Peter 1998 Virtual convergence and dialect levelling in Luxemburgish. Folia Linguistica 32(1/2): 69⫺82. Gilles, Peter 1999 Dialektausgleich im Le¨tzebuergeschen. Zur phonetisch-phonologischen Fokussierung einer Nationalsprache. (Phonai 44.) Tübingen: Niemeyer. Goossens, Jan 1977 Deutsche Dialektologie. (Sammlung Göschen 2205.) Berlin/New York: de Gruyter. Haas, Walter 1978 Sprachwandel und Sprachgeographie. Untersuchung zur Struktur der Dialektverschiedenheit am Beispiel der schweizerdeutschen Vokalsysteme. (ZDL Beihefte 30.) Wiesbaden: Steiner. Haas, Walter 2000 Die deutschsprachige Schweiz. In: Hans Bickel and Robert Schläpfer (eds.), Die viersprachige Schweiz, 57⫺138. (Reihe Sprachlandschaft 25.) Aarau: Sauerländer.

255

256

III. Structure and dynamics of a language space Herrgen, Joachim 1986 Koronalisierung und Hyperkorrektion. Das palatale Allophon des /CH/-Phonems und seine Variation im Westmitteldeutschen. (Mainzer Studien zur Sprach- und Volksforschung 9.) Stuttgart: Steiner. Herrgen, Joachim 2005 Sprachgeographie und Optimalitätstheorie. Am Beispiel der t-Tilgung in Auslaut-Clustern des Deutschen. Zeitschrift für Dialektologie und Linguistik 72: 278⫺317. Hinskens, Frans 1996 Dialect Levelling in Limburg: Structural and Sociolinguistic Aspects. (Linguistische Arbeiten 356.) Tübingen: Niemeyer. Hinskens, Frans, Peter Auer and Paul Kerswill 2005 The study of dialect convergence and divergence: conceptual and methodological considerations. In: Auer, Hinskens, and Kerswill (eds.), 1⫺50. Hofer, Lorenz 1997 Sprachwandel im städtischen Dialektrepertoire. Eine variationslinguistische Untersuchung am Beispiel des Baseldeutschen. (Stadtsprache ⫺ Sprachen in der Stadt am Beispiel Basels 1 / Basler Studien zur deutschen Sprache und Literatur 72.) Tübingen/Basel: Francke. Hotzenköcherle, Rudolf 1961 Zur Raumstruktur des Schweizerdeutschen. Zeitschrift für Mundartforschung 28: 207⫺ 227. (Reprinted in: R. Hotzenköcherle, Dialektstruktur im Wandel. Ed. by R. Schläpfer and R. Trüb. Aarau: Sauerländer.) Jakobson, Roman 1931 Prinzipien der historischen Phonologie. Travaux du Cercle Linguistique de Prague 4: 247⫺267. Kerswill, Paul 2002 Models of linguistic change and diffusion: New evidence from dialect levelling in British English. Reading Working Papers in Linguistics 6: 187⫺216. Labov, William 1963 The social motivation of sound change. Word 19: 273⫺309. Labov, William 1966 The Social Stratification of English in New York City. Washington: Center for Applied Linguistics. Labov, William 1994/2001 Principles of Linguistic Change, vol. 1: Internal Factors, vol. 2: Social Factors. Oxford: Blackwell. Labov, William, Sharon Ash and Charles Boberg 2005 The Atlas of North American English: Phonetics, Phonology, and Sound Change. Berlin/ New York: Mouton de Gruyter. LePage, Robert B. and Andre´e Tabouret-Keller 1985 Acts of Identity: Creole-Based Approaches to Ethnicity and Language. Cambridge: Cambridge University Press. Leuenberger, Petra 1999 Ortsloyalität als verhaltens- und sprachsteuernder Faktor. Eine empirische Untersuchung. (Basler Studien zur deutschen Sprache und Literatur 74.) Tübingen/Basel: Francke. Linke, Angelika 2007 Communicative genres as categories of a cultural history of communication. In: Stephan Elspaß, Nils Langer, Joachim Scharloth and Wim Vandenbussche (eds.), Germanic Language Histories “from Below” (1700⫺2000), 473⫺494. (Studia Linguistica Germanica 86.) Berlin: de Gruyter. Martinet, Andre´ 1955 E´conomie des changements phone´tiques. Bern: Francke. Martinet, Andre´ 1962 A Functional View of Language. Oxford: Clarendon Press.

14. Horizontal convergence Mattheier, Klaus J. 1996 Varietätenkonvergenz. Überlegungen zu einem Baustein einer Theorie der Sprachvariation. Sociolinguistica 10: 31⫺52. MRhSA ⫽ Bellmann, Günter, Joachim Herrgen and Jürgen Erich Schmidt 1994⫺2002 Mittelrheinischer Sprachatlas. 5 vols. Tübingen: Niemeyer. Muhr, Rudolf 1981 Sprachwandel als soziales Phänomen. Eine empirische Studie zu soziolinguistischen und soziopsychologischen Faktoren des Sprachwandels im südlichen Burgenland. (Schriften zur deutschen Sprache in Österreich 7.) Wien: Braunmüller. Paul, Hermann 1920 [1880] Prinzipien der Sprachgeschichte. 5th ed. Tübingen: Niemeyer. Pfalz, Anton 1918 Reihenschritte im Vokalismus. Beiträge zur Kunde der bayerisch-österreichischen Mundart I, 22⫺42. (Kaiserliche Akademie der Wissenschaften in Wien ⫺ Philosophisch-historische Klasse ⫺ Sitzungsberichte, 190/2.) Wien: Rohrer. Polenz, Peter von 1954 Die altenburgische Sprachlandschaft. Untersuchungen zur ostthüringischen Sprach- und siedlungsgeschichte. (Mitteldeutsche Forschungen 1.) Tübingen: Niemeyer. Radtke, Edgar 2006 Konvergenz und Divergenz regionaler Varietäten. In: Ulrich Ammon, Norbert Dittmar, Klaus J. Mattheier and Peter Trudgill (eds.), Sociolinguistics / Soziolinguistik, 2nd ed., 2189⫺2197. (Handbücher zur Sprach- und Kommunikationswissenschaft 3.3.) Berlin/ New York: de Gruyter. Renn, Manfred 1994 Die Mundart im Raum Augsburg. Untersuchungen zum Dialekt und zum Dialektwandel im Spannungsfeld großstädtisch-ländlicher und alemannisch-bairischer Gegensätze. (Sprache ⫺ Literatur und Geschichte 9.) Heidelberg: Winter. SBS ⫽ Werner König (ed.) 1997⫺2008 Sprachatlas von Bayerisch-Schwaben. Heidelberg: Winter. Schifferle, Hans-Peter 1995 Dialektstrukturen in Grenzlandschaften. Untersuchungen zum Mundartwandel im nordöstlichen Aargau und im benachbarten südbadischen Raum Waldshut. (Europäische Hochschulschriften I, 1538.) Frankfurt: Lang. Schirmunski, Viktor M. 1930 Sprachgeschichte und Siedlungsmundarten, I⫺II. Germanisch-Romanische Monatsschrift 18, 113⫺113: 171⫺188. Schönfeld, Helmut, Ruth Reiner and Sabine Grünert 2001 Berlinisch heute: Kompetenz ⫺ Verwendung ⫺ Bewertung. Frankfurt: Lang. Schunk, Gunther 1999 Regionalisierung von Dialekten. Ein lautlicher Stadt-Land-Vergleich in Mainfranken. (Schriften zum Bayerischen Sprachatlas 1.) Heidelberg: Winter. SDS ⫽ Rudolf Hotzenköcherle [in collaboration with Konrad Lobeck, Robert Schläpfer, Rudolf Trüb and with the assistance of Paul Zinsli] (ed.) 1962⫺1997 Sprachatlas der deutschen Schweiz (SDS). Bern (vols. 7 & 8 Basel): Francke. Siebenhaar, Beat 2000 Sprachvariation, Sprachwandel und Einstellung. Der Dialekt der Stadt Aarau in der Labilitätszone zwischen Zürcher und Berner Mundartraum. (ZDL Beihefte 108.) Stuttgart: Steiner. Siebenhaar, Beat 2002 Sprachwandel von Sprachgemeinschaften und Individuen. In: Annelies Häcki Buhofer (ed.), Spracherwerb und Lebensalter, 313⫺325. (Basler Studien zur deutschen Sprache und Literatur 83.) Tübingen/Basel: Francke.

257

258

III. Structure and dynamics of a language space Siebenhaar, Beat 2008 Sprachwandel und Sprachgeographie ⫺ der Einfluss der Stadt Bern auf die Region. In: Krefeld, Thomas (ed.), Sprachen und Sprechen im städtischen Raum, 173⫺195. (Spazi comunicativi ⫺ kommunikative Räume 2.) Frankfurt am Main: Lang. Steiner, Emil 1921 Die französischen Lehnwörter in den alemannischen Mundarten der Schweiz; kulturhistorisch-linguistsche Untersuchung mit etymologischem Wörterbuch. Wien/Basel: A. Holzhausen/Wepf-Schwabe. Taavitsainen, Irma and Andreas H. Jucker (eds.) 2003 Diachronic Perspectives on Address Term Systems. (Pragmatics & Beyond, new series 107.) Amsterdam: Benjamins. Torgersen, Eivind and Paul Kerswill 2004 Internal and external motivation in phonetic change: Dialect levelling outcomes for an English vowel shift. Journal of Sociolinguistics 8(1): 23⫺53. Trudgill, Peter 1986 Dialects in Contact. Oxford: Blackwell. Trudgill, Peter 2008 Colonial dialect contact in the history of European languages: On the irrelevance of identity to new-dialect formation. Language in Society 37(2): 241⫺280. Wegener, Heide 1999 Die Pluralbildung im Deutschen ⫺ ein Versuch im Rahmen der Optimalitätstheorie. Linguistik online 4. Available at ; last accessed 8 October 2008. Weiss, Richard 1947 Die Brünig-Napf-Reuß-Linie als Kulturgrenze zwischen Ost- und Westschweiz auf volkskundlichen Karten. Geographica Helvetica 2: 153⫺175. Wiesinger, Peter 1970 Phonetisch-phonologische Untersuchungen zur Vokalentwicklung in den deutschen Dialekten. (Studia Linguistica Germanica 2/1, 2/2.) Berlin: de Gruyter. Wiesinger, Peter 1983 Ergebnisse dialektologischer Beschreibungen: Areale Bereiche deutscher Dialekte im Überblick. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke and Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, 807⫺900. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.2.) Berlin/New York: de Gruyter. Wolf, Norbert Richard 1983 Durchführung und Verbreitung der zweiten Lautverschiebung in den deutschen Dialekten. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke, and Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, 1116⫺1121. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.2.) Berlin/ New York: de Gruyter. Wolfensberger, Heinz 1967 Mundartwandel im 20. Jahrhundert. Dargestellt an Ausschnitten aus dem Sprachleben der Gemeinde Stäfa. (Beiträge zur schweizerdeutschen Mundartforschung 14.) Frauenfeld: Huber.

Beat Siebenhaar, Leipzig (Germany)

15. Vertical convergence

15. Vertical convergence o linguistic varieties in a language space 1. 2. 3. 4. 5.

Introduction Standardization and national spoken standard varieties in Europe Examples of vertical convergence and its results Conclusion References

1. Introduction This article deals with the notion of vertical convergence, how it comes about, the sociocultural mechanisms behind it, and how it is manifested both linguistically and sociolinguistically. After a general discussion of what vertical convergence is and how it is related to horizontal convergence, a series of illustrative examples of vertical convergence and its results will be presented. There will also be a discussion concerning the hegemonic status of some linguistic varieties, how they have achieved this status (historically), and how this status may be changing. A language space is understood here as a multidimensional space which includes both geographical and social-cultural axes.

1.1. Vertical convergence The term vertical convergence refers to the process whereby a certain range of linguistic features of a variety is substituted by features that enjoy higher standing than the original ones. Hence, vertical convergence arises from direct or indirect contact between varieties where the converged-to variety holds a higher status in social space than the converging variety. In many cases the converged-to variety is an overarching spoken standard variety, but it does not have to be. The main point is the asymmetrical relationship between the converging and the converged-to variety, be it a national standard, an urban high-status variety or another hegemonic variety. The most prominent example of such an asymmetrical relationship between varieties is, however, that between traditional dialects and a national spoken standard variety. Several of the examples in this article will therefore present dialect-standard relationships in different parts of Europe, particularly Scandinavia (cf. section 3). In many cases, however, one may question what the target of vertical (or semi-vertical) convergence is: the standard proper or a substandard (but still socially prestigious) variety. Vertical convergence is intimately connected to its counterpart, horizontal convergence, and the two types of linguistic convergence are in many cases difficult to distinguish. Due, for instance, to isomorphism it may be impossible to determine whether a certain contact innovation should be mapped as vertical or horizontal convergence. It is therefore often an empirical challenge to find unambiguous examples of either type. Moreover, given that most cases of convergence may be traced back to asymmetrical relationships between varieties ⫺ that is, to differences in status or prestige ⫺ one could

259

260

III. Structure and dynamics of a language space argue that all cases of convergence might in that sense be labeled vertical. For the sake of conceptual clarity, however, it may be useful to reserve the term vertical to convergence between dialect and a standard or standard-like variety and horizontal to convergence between basilectal varieties where the relative difference in prestige is smaller. Even so, as these considerations indicate, it is not always easy or trivial to draw a clearcut line between vertical and horizontal convergence. Vertical convergence of linguistic varieties will normally reduce the degree of intersystemic variation. Intra-systemically, however, the degree of variation may increase since speakers can choose linguistic variants from a larger repertoire. Decreasing contrast between varieties does not necessarily imply a loss of traditional forms. Old and new forms may co-exist for some time, at least in the beginning of the process. However, in the long run, features from the converging variety tend to be abandoned. The longterm effect of the process is therefore most probably also a reduction of variety-internal variation. In discussing the phenomenon of vertical convergence ⫺ and the relationship between inter- and intra-systemic variation ⫺ it is thus important to distinguish between the process and its outcome (cf. Auer 2000). Obviously the outcome of vertical convergence will be a reduced distance between asymmetrically placed varieties, but it almost always involves a reduced distance between horizontally placed varieties as well (cf. Auer and Hinskens 1996). This means that vertical convergence has the potential to change a language space, both geographically and socially speaking, and maybe even to create new language spaces. By contrast, reduced horizontal distance ⫺ that is, between dialects ⫺ does not necessarily imply less distance between dialect and standard. Horizontal dialect-dialect convergence may in fact increase the structural distance to the standard, as shown by Hinskens (1996) in his study of dialect leveling in Limburg, the Netherlands (cf. Siebenhaar in this handbook).

1.2. A multidimensional model o linguistic convergence Auer and Hinskens (1996) propose three different scenarios of the process of vertical and horizontal convergence and its products. Figure 15.1 shows different degrees of focusing and reduction of the linguistic repertoire, which can be seen as three stages of development (from left to right). If we consider the repertoire of a speech community with extensive vertical and horizontal convergence, variants from different varieties will compete on the linguistic market. Hence, the normal situation will be that the amount of variation within the repertoire increases in the beginning of the process (as displayed in Figure 15.1, [i] and [ii]). At the termination of the process, however, some variants will normally win out and others will disappear, meaning that the total scope of variation will be reduced (as shown in Figure 15.1 [iii]). Generally, features with a restricted geographical and social distribution tend to disappear first, whereas features with a wider extension in language space have a greater potential of surviving. The expected result of the process is thus the emergence of new lects (see Figure 15.1 [ii], where the bars in the rectangle illustrate new, more or less focused lects), but more diffuse situations may also emerge with a range of non-discrete structures within the dialect-standard continuum (as illustrated in Figure 15.1 [i]).

15. Vertical convergence

Fig. 15.1: Multidimensional model of linguistic convergence (adapted from Auer and Hinskens 1996)

1.3. Dialect leveling An important consequence of vertical convergence between dialect and standard (or semi-standard) is that of dialect leveling. Dialect leveling is probably the most distinct tendency in modern European dialect change, and vertical convergence perhaps the most powerful contributing force. Dialect leveling is understood here as a dynamic and multidimensional dialect contact phenomenon that leads to the reduction of inter-systemic variation by a gradual abandonment of local dialect features in favor of more regional or standard ones. This definition of dialect leveling is inspired by Trudgill (1986) and Hinskens (1996), but it does not include the dimension of “markedness”. Trudgill (1986: 98) defines dialect leveling as “reduction or attrition of marked variants”. As pointed out by Auer (2000: 13), there are several examples of convergence where the structurally more marked feature is the converged-to feature: for instance in the case of phoneme splits, as they occur in dialect convergence towards a standard variety. By leaving out the question of markedness, leveling and convergence may be equated. The term dialect leveling is used here in the sense of “bottom-up” ⫺ that is, as the formation of regiolects, and not “top-down”, in the course of demotization or de-standardization (see section 3.1). An important question is whether vertical convergence necessarily implies a direct adoption of the target forms, and hence dialect loss, or if vertical convergence may result in the emergence of new forms (compromise forms, innovations) that are neither present in the dialect nor in the target variety. In the following we will take a closer look at some examples of dialect leveling in different parts of Europe with particular focus on vertical convergence between dialects and a hegemonic variety. As noted above, the existence of a hegemonic spoken variety is a prerequisite for vertical convergence. The most typical example of such a hegemonic variety is a spoken standard norm (i. e. orthoepy). For this reason, before turning to specific cases, we will briefly discuss the concept of an oral standard and present a brief overview of the status of some spoken standards in present-day Europe (for a detailed account of dialect/standard constellations in Europe see Auer 2005).

261

262

III. Structure and dynamics of a language space

2. Standardization and national spoken standard varieties in Europe Broadly put, a standard language may be defined as that variety of a language which has gained literary and cultural supremacy over the other varieties and which is accepted also by the speakers of the other varieties as a more proper form of that language than other varieties (cf. Bull 1992). As Haugen (1972: 246) puts it: “It is a significant and probably crucial requirement for a standard language that it be written” [my italics]. Nevertheless, a standard language may also be spoken. In Europe the emergence of oral standards is intimately linked to and dependent on the formation and codification of written standards. Thus, we may take the existence of a written standard as a prerequisite for an oral one. When discussing the ontological status of a spoken standard within a language community, not necessarily as a codified norm, but as an operative norm, it is important to have a look at the history and status of the written standard. By definition, an oral standard covers a wide range of functions. It is typically involved in code switching and is clearly the preferred variety in formal situations; it has the position of the H-variety in diglossic speech communities; and it is often the converged-to variety in dialect leveling situations. It has an overarching function with respect to other varieties within the same dia-system and sometimes even over varieties of other dia-systems (e.g., exoglossic standards such as Latin and Arabic). A spoken standard may be regionally unmarked in the sense that it may be impossible to place a standard speaker geographically. However, as we will see below, a spoken standard norm is often based on the speech of the upper classes in the capital or in the most important cities of a country and as such it may be highly geographically localizable. In most cases it is also possible to place the speakers of an oral standard in social space. The spoken standard is often regarded as a symbol of good manners and a signal of a particular sociocultural, symbolic capital that gives the speaker access to certain social circles (see Mæhlum in this volume for a more detailed discussion of the indexicality of spoken standards as compared to dialects). The model presented in section 1.2 presupposes that a national spoken standard exists, that there is at least one dialect, and that both are part of the same dia-system. As pointed out by Auer and Hinskens (1996: 6), the two last presuppositions do not necessarily hold for all language communities in Europe today. One may, however, question even the assumption that spoken standards exist in all parts of Europe. As we will see below, this issue has been a hotly debated point within the Norwegian language community. As stated by Deumert and Vandenbussche (2003: 2) in their introduction to a volume on the different histories of standardization, the Germanic languages “provide a wide range of highly diverse standardization scenarios”. Some Germanic languages developed written standards in medieval times (like Danish and Icelandic), others during the nineteenth century (like Norwegian (Nynorsk and Bokma˚l) and Afrikaans), and some languages are still in the process of developing a standard language (like Luxembourgish and Scottish).

2.1. Contrasting history and status o spoken standards Schematically, one may represent the communities in Europe today on a continuum, running from those with indisputable and generally acknowledged spoken standards to

15. Vertical convergence communities where the standard is weak and contestable with respect to both content and status. In this article this contrast will be illustrated through examples taken from the Scandinavian language area. Here we find linguistic communities that have been intimately linked to each other, but that have had very different developments regarding the emergence and status of oral standards, as well as with regard to the status and sociolinguistic situation of the traditional dialects and the amount of vertical and horizontal convergence in the communities’ linguistic repertoires. The linguistic communities I will focus on are Denmark and Norway ⫺ countries with a shared national, political history until 1814, when Norway gained its independence from Denmark. While the existence of an overarching oral standard is treated as an accepted and indisputable “truth” in present-day Denmark, it is a disputed issue in Norway. With respect to this contrast, the Danish language situation may be taken as the “normal case”. As far as the status of the spoken standard and the dialect-standard relationship are concerned, most European countries are undoubtedly closer to the Danish end of the spectrum than the Norwegian. In Denmark, upper class Copenhagen speech is clearly associated with the standard and is generally acknowledged as the “best” and most “normal” way of speaking Danish. In Norway, on the other hand, whether the upper class variety spoken in the capital, Oslo, may be conceived of as standard speech is a contentious issue. While dialects in Norway enjoy a relatively high status and are accepted as legitimate means of communication within most social domains, negative attitudes towards dialects and pronounced dialect leveling are typical of Denmark. It is not surprising, therefore, that vertical convergence is a prominent force in modern Danish dialect change, whereas its role is disputable in Norway. A look at the countries’ respective standardization processes reinforces this point. In Denmark the first developments towards a spoken standard were undertaken in the sixteenth century ⫺ that is, more than three centuries after the emergence of the written standard, and social prestige was the most important element in its formation (see Kristiansen 2003a: 73; Skautrup 1947). Over the centuries, written Danish has undergone relatively small changes whereas spoken Danish, in all its varieties, has moved further away from the common Nordic origin than the other Scandinavian languages have done (cf. Kristiansen 2003a: 78; Vikør 1995; Haugen 1976). Since the phonetic spelling approach has had only minimal impact on the codification of Danish orthography, there is a substantial gap between written and spoken Danish, and Brink and Lund (1975) show that the majority of the changes in spoken standard Danish during the last two centuries have increased rather than decreased the distance between the written and the spoken standard (for examples see Kristensen 2003). During the second half of the eighteenth century, a social stratification of Copenhagen speech took place, and two separate varieties ⫺ so-called high and low Copenhagen speech ⫺ gradually emerged. High and low refer to the position of the varieties in social space. The national spoken standard variety, which took shape during the first half of the nineteenth century, is often referred to as rigsma˚l and was primarily based on high Copenhagen speech. Today, however, rigsma˚l as spoken by the younger generations is also influenced by what used to be considered low Copenhagen features ⫺ that is, downward convergence towards the low-Copenhagen variety has taken place (cf. Kristensen 2003: 33; see section 3 below). The case provides an illustration of the dynamic relation between convergence and standard formation. Synchronically, two ver-

263

264

III. Structure and dynamics of a language space sions of the spoken standard exist; the traditional one based on high Copenhagen speech and a more recent version which also includes re-allocated low Copenhagen features. The former version is primarily used by the older generation whereas the latter is more common among younger speakers ⫺ hence, the different versions of rigsma˚l are agespecific. This has led to a situation where features from older rigsma˚l have been substituted by new features among young people in Copenhagen, while still being used by young people outside of Copenhagen: young people in peripheral areas may use old Copenhagen features, which were once part of rigsma˚l, but which have now fallen out of use even by the oldest rigsma˚l speakers in Copenhagen (cf. Kristensen 2003: 33). This may potentially create new distinctions ⫺ social, cultural, generational and geographic. However, as we will see below, in present-day dialect convergence in Denmark, the more influential of the two varieties is the younger rigsma˚l. In Norway the question of a spoken standard norm is more problematic. Norway was under Danish rule from 1450 until 1814. During this time no separate written standard of Norwegian existed; that is, the written language in Norway was Danish, and the dialects spoken in what constitutes Norway today were regarded as dialects of the common Dano-Norwegian language. After the separation from Denmark, a long-lasting language struggle started, which resulted in the recognition of two written standards in 1885 (now Bokma˚l and Nynorsk). Bokma˚l was based on the Danish written standard, but was gradually Norwegianized through a series of official language reforms. The target of this reform process was the Dano-Norwegian dialect of the upper middle classes in the capital Christiania (today Oslo). Nynorsk, on the other hand, was based on a wide range of rural, Norwegian dialects. During the first half of the twentieth century, the goal of the official language planning policy was to unify the two standards into a single written standard by using linguistic material from rural and urban dialects. This policy, however, has now been officially abandoned (see Jahr 2003 for details). Today there are two official written standards of Norwegian ⫺ Bokma˚l and Nynorsk. Both standards are quite heterogeneous, to the extent that it is possible to identify two major varieties within each of them. Unsurprisingly, agreeing on a common spoken standard is problematic. Like the situation in many other European countries, there is no officially sanctioned spoken norm in Norway: there are no authoritative dictionaries of how “proper”, “correct” Norwegian should be pronounced. In fact, the Norwegian Parliament decided in 1878 that no particular spoken standard should be taught in elementary and secondary schools. This principle is still valid today, and in Norwegian schools there is no tradition of correcting pupils’ dialects. This historical background has no doubt been essential for the continued use of local dialects in Norway even in polylectal communication and it explains the relatively weak position of the spoken standards. Attempts at establishing a spoken Nynorsk standard have never met with great success. A spoken standard based on Bokma˚l, on the other hand, has been much more successful. Spoken Bokma˚l may be pronounced with a range of different accents, but south-eastern Norwegian phonology, particularly as it appears in and around the capital, Oslo, seems to have achieved a hegemonic status and is commonly regarded to be the most “normal”, “neutral” and even “correct” way of speaking Norwegian (cf. Røyneland 2009 for further discussion). It is often referred to as the east Norwegian standard, it is the mother tongue of many people, and it is used by many speakers in the region in order to conceal exactly where they come from (cf. Sandøy 1998). This variety is often

15. Vertical convergence reported to be the converged-to variety (cf. Røyneland 2005). It has rather fuzzy borders and, as is the case in Denmark, one may distinguish between two versions of the spoken standard: a “conservative” standard associated with upper middle class Oslo speech and a “modern” standard that reflects younger Oslo speech. The status of these spoken standards is somewhat unclear ⫺ not all members of the Norwegian speech community would acknowledge them as such. Nevertheless, while they may not fully meet the criteria of a standard proper, they certainly share, as demonstrated, some of the key characteristics of oral standards. In the following we will take a closer look at the role of vertical convergence in communities where the status of both the standard and the dialects differs substantially. We will consider particular results of vertical convergence such as dialect leveling, shift or death, the formation of new (regional) varieties and the emergence of intermediate and hybrid forms.

3. Examples o vertical convergence and its results Vertical convergence between a converging and a converged-to variety may have a range of different results. As will be demonstrated by the examples below, vertical convergence ⫺ or more precisely, advergence ⫺ towards another variety may (but does not necessarily have to) imply a direct adoption of the target variants. In many cases, convergence is only partial and may even lead to the emergence of completely new, hybrid forms. Vertical convergence is a powerful force in language change ⫺ not only at linguistic markets where there is a commonly accepted, strong and unquestioned standard spoken norm roofing the dia-system, but also in communities where the standard norm is fuzzier and its status more unclear. In some cases the target of vertical (or semivertical) convergence may not even be a standard proper, but a socially prestigious variety which has achieved some sort of hegemonic standard-like status. Vertical convergence may contradict principles of language change, and may produce a re-categorization of constraints within non-standard dialects, even where this leads to more marked or complex forms. It is, for instance, a common belief that innovations can create mergers, but not reverse them. If two words have become identical through a phonetic change, they can never be differentiated by phonetic means (cf. Gardes’ principle, Labov 1994: 311). However, mergers may be undone in vertical convergence, as Labov (1994: 342) argues: “Given the right social conditions, it is reasonable to think that a distinction can be reintroduced into a speech community in a consistent way”. An example of this may be found in the south of Spain where overt prestige pressure on speakers in Andalucı´a has enhanced vertical convergence towards the standard distinction of coronal fricatives. In the traditional dialects of the area, the alveolar fricative /s/ as in tasa (tax) and the dental fricative /u/ as in taza (cup) are merged into /s/; that is, both words would traditionally be pronounced as /tasa/. At present this merger is not viewed as particularly trendy, and young, urban Andalusians tend to avoid it. This trend of vertical convergence is relatively recent, is favored by inter-dialect contact and seems to have affected the youngest and most educated speakers (Villena-Ponsoda and Vida Castro 2005: 5⫺6).

265

266

III. Structure and dynamics of a language space In the following, examples from the Danish and Norwegian speech communities will be brought forward as illustrative examples of the varied results that vertical convergence may have. The differences in results reflect the contrasting poles of the spectrum of dialect-standard relationships in present-day Europe.

3.1. Dialect leveling, dialect death and the ormation o regional (standard) varieties The Danish case provides an example of a language space where vertical convergence has lead to massive dialect leveling and even to dialect death and where one may see the emergence of regional (standard) varieties (cf. scenario [iii] in Figure 15.1). In their introduction to the sociolinguistics of Danish, Kristiansen and Jørgensen (2003: 1) state that “almost all living mother-tongue speakers of Danish speak the Copenhagen standard or a variety very close to it”. Until 1900 Denmark was a fairly heterogeneous linguistic area, compared to the geographical space it occupies. By the end of the twentieth century, the situation has dramatically changed (cf. Kristensen 2003: 29). Language variation has decreased substantially along the geographical dimension in the direction of a de-dialectalization of the area, i. e., most dialect features are on the brink of extinction. Only minor segmental differences are found today, though some supra-segmental differences may still be encountered. However, language variation has increased within varieties. Thus, individual speakers are offered the possibility of a wider linguistic repertoire. The radical de-dialectalization and standardization of Denmark is “the result of pervasive influence from the spoken language of Copenhagen” (Kristensen 2003: 30), combined with “a strong standard ideology” (Pedersen 2003: 10), or, in other words, a result of extensive vertical convergence. From the late nineteenth century until the 1960s, much of the Danish dialect vocabulary was lost and specific local variants were substituted by variants with a wider geographical range. These variants were often, but not always, closer to the standard language. The dialects converged to each other and to the standard, that is, both horizontally and vertically, resulting in leveled or modernized dialects (cf. Pedersen 2003: 21). At this time, dialects were still transmitted from one generation to the next. During the 1960s, however, the number of dialect speakers declined dramatically. New generations simply acquired a (regional) standard language instead of their parents’ local dialect. Pedersen explains this dramatic shift as a consequence of large-scale societal changes such as urbanization, modernization and a restructuring of the country’s labor force. The primary industries underwent a precipitous decline; the number of people employed in agriculture fell rapidly and their sociocultural standing was eroded. Since dialect use in Denmark was connected largely with rural, agricultural ways of life, an effect of this restructuring was that dialects quickly lost ground as a means of communication beyond the private sphere. This kind of dialect loss, however, cannot be described as a result of convergence. Rather, it is a case where abrupt societal changes disrupt the traditional ways of life sustaining dialect use. The predicted development in Denmark is that the traditional dialects will disappear completely, the linguistic centralization and standardization will proceed and the few inter-systemic differences between different regional varieties that still exist will diminish

15. Vertical convergence (Kristiansen 2003b; Kristensen 2003; Pedersen 2003: 2008). One may, in other words, expect a development that corresponds to the third and last triangle (iii) in Figure 15.1 (the multidimensional model of linguistic convergence by Auer and Hinskens 1996) where the base dialects are lost and new (regional) standard varieties with a few regional traces, particularly supra-segmental features such as regional intonational patterns, have emerged. These regiolectal forms may now be considered the most basilectal way of speaking. Danes seem to have become highly sensitive to subtle phonological and phonetic differences. Thus, even if the linguistic differences are small and as such may seem insignificant, they can still function as important signals of social and geographical distinction. An important question is whether regional standard varieties have come into existence as a means of signaling regional affiliation and identity. At the level of language use, non-arbitrary but fuzzy boundaries within the continuum between dialect and standard may be detected, as shown by Kristensen and Thelander (1984). According to Kristensen (2003), the regional standard varieties of Danish are defined by particularly resistant dialect features that are shared by a larger region and combined with features from rigsma˚l and low Copenhagen speech. The users of these regional standard varieties are found in larger geographical regions of Denmark such as Zealand, Funen, northern, western, eastern and southern Jutland, and are neither (base) dialect nor standard rigsma˚l speakers. However, there are no sharp boundaries between the different regions, the varieties are rather heterogeneous both inter- and intra-individually and one may question whether there are clear norms and social functions associated with them. Kristiansen (2003b) raises doubts as to whether such regional standard varieties have in fact emerged in Denmark ⫺ at least if we consider their social reality for the speakers and their function as symbolic markers of regional identity. According to Kristiansen (2003b), there is no regional linguistic consciousness among Danes, and regional varieties of Danish lack any positive group-marking function of the sort that may be used as a means of distinction or to display regional allegiance in opposition to a centralized culture of the capital. As pointed out also by Kristensen (2003: 41), the regional standard varieties tend to be perceived as a variety of the standard, or even as the standard proper, by the language users themselves; it is simply perceived as the “correct” language. At the socio-psychological level, then, the regional varieties are not varieties in their own right. Nevertheless, at the level of language use, regional standard varieties can be distinguished. It is primarily vertical convergence that has reshaped the Danish linguistic landscape, although horizontal convergence also came into play particularly at an early stage. However, as stated above, the present-day development is not only characterized by vertical convergence towards the standard, but also by a restructuring of the standard itself through what one may denote a process of demotization (“Demotisierung”; Mattheier 1997). This is a process of downward convergence where features from so-called lowCopenhagen speech are included in young people’s perception of standard Danish. The standard does not necessarily become less important or lose any of its functions (i. e., it is not de-standardized), but it becomes more popular and less connected to a high socioeconomic stratum of the population. It is, as pointed out both by Pedersen (2003) and Kristiansen (2003b), still an open question whether Danes outside the capital actually strive towards total convergence with high or low Copenhagen speech, or if they are trying to maintain certain subtle

267

268

III. Structure and dynamics of a language space phonetic features as a means of displaying regional affiliation. If the former is the case, Denmark would, as pointed out by Pedersen (2003: 26), “be a good candidate to be the first country to accomplish total linguistic standardization.” Norway provides a contrasting case. Generally speaking, dialects in Norway are wellpreserved and, hence, dialect diversity is still considerable as compared, for instance, to modern Sweden and Denmark (e.g., Pedersen 2003; Kristiansen 2003b; Edlund 2003; Vikør 1995; Kristensen/Thelander 1984; Thelander 1980). Whereas dialect use in Denmark is seen as appropriate only within the private sphere, dialects in Norway are generally used within all social domains (cf. Røyneland 2009). However, the differences may be more quantitative than qualitative. In both countries, dialect use is associated with a traditional, rural lifestyle and with informal, everyday speech events, whereas standard speech is associated with a modern, urban lifestyle and formal, official speech events (cf. Mæhlum in this handbook). The conceptions of the prototypical dialect speaker versus the standard speaker have, thus, some of the same characteristics in the two countries, but dialect use is still generally perceived more positively and has a higher value on the linguistic marketplace in Norway than in Denmark. However, not all dialects in Norway have the same social standing. Dialects close to the larger cities and particularly those close to the capital, Oslo, are more exposed to negative stereotyping than, for instance, dialects from the west coast or from mid-Norway (for a more detailed discussion see for instance Mæhlum 2007; Røyneland 2009). In spite of the supposed egalitarian linguistic ideal and the uncertain status of a spoken standard norm in Norway, vertical convergence is also a powerful force in modern Norwegian dialect change ⫺ not necessarily towards an acknowledged spoken standard, but towards a hegemonic operative norm. As noted, a Bokma˚l-based variety with south-eastern Norwegian intonation shares important characteristics with a spoken standard. The current linguistic situation, then, is no doubt marked by increasing dialect leveling and even in some areas by regular dialect shift. In western regions horizontal convergence seems to dominate (cf. Akselberg 2005; Sandøy 2004; and for a diverging account see Solheim 2006), whereas vertical convergence is clearly the dominating force in other parts of the country (cf. Mæhlum 2007; Røyneland 2005: 2009). Changes in the vowel system provide a clear instance of vertical convergence. Throughout the country, dialects with a phonemic inventory containing more vowel phonemes than the east Norwegian urban standard (with its nine monophthongs and four diphthongs) show a clear tendency towards reducing the existing vowel system, bringing it closer to the standard system. In the consonant system we also find examples of mergers which reduce the number of phonemes or allophones. As we will return to below, however, convergence towards the standard is not always complete (see section 3.2). As in the case of the Danish standard, there has been a restructuring of the Norwegian standard through demotization. Today one may distinguish between two versions of the spoken standard: a “conservative” standard associated with upper middle class Oslo speech and a “modern” standard that reflects younger Oslo speech. Studies of dialect leveling in Norway indicate that the “modern” standard is clearly the most expansive, whereas the “conservative” standard is losing territory (see Vikør 1999; Skjekkeland 2000 for further references). As noted above, this also seems to be a tendency in other parts of Europe (e.g., Pedersen 2003; Kerswill 2000). In summary, although vertical convergence is a powerful force in modern Norwegian, it is not nearly as dominating as it is in Denmark. The standard itself and the standard

15. Vertical convergence ideology do not have a particularly strong position in Norway. Dialects in Norway enjoy a relatively high status and are relatively well preserved. Still, the Norwegian linguistic landscape is undergoing a restructuring which is marked by dialect leveling, but also by a demotization of the standard. It is, as yet, unclear whether regiolects or regional standard varieties are in the process of establishing themselves ⫺ both at the level of language use and at the level of attitudes. The first triangle, (i) in Figure 15.1, where there are still no focused new lects, but rather a range of non-discrete structures within the dialectstandard continuum, is probably the one that describes the Norwegian language space most accurately.

3.2. Incomplete convergence and the emergence o hybrid orms As discussed extensively by, for instance, Trudgill (1986) and Auer (2000) the adoption of a target variant does not have to be complete ⫺ either because speakers only converge partially to the target variant or because they simply miss the target and in turn create a new variant. One example of incomplete adoption mentioned by Trudgill (1986: 22, 58) is the variable use of [d] rather than [d4 ] as a realization of intervocalic /d/ by middle class Americans living in England. Another example mentioned by Trudgill (1986: 59) is incomplete adoption linked to lexical diffusion. Northern English accents have a fivevowel system (/=, ε, a, w, c/) whereas southern accents have a six-vowel system (/=, ε, æ, w, v, A/). Hence, in southern accents could and cud, put and putt are distinguished /w, v/, whereas northern accents have /w/ throughout. The southern six-vowel system is gradually spreading northwards, but in a transition zone speakers replace /w/ by /v/ only in certain words. These speakers do not modify their phonological system as such to resemble the target variety. Rather, they change the pronunciation of individual lexical items to make them sound more like those of the target variety; hence, speakers’ motivation is, according to Trudgill (1986: 58), phonetic rather than phonological. These are both examples of partial convergence where the original and the new variant co-exist. This co-existence may last over time and create what Trudgill (1986) calls mixed dialects. Increased intra-systemic variability is also a typical result of vertical convergence ⫺ at least in the beginning of the process. However, features from the converging variety tend to be abandoned over time (see discussion in section 1.1). Adoption of a target variant may also be phonetically, and not only lexically, partial and hence give rise to what Trudgill (1986) calls fudged lects ⫺ that is, phonetically intermediate lects. Thus, the linguistic form which is transmitted from a converged-to variety does not have to be identical to the form eventually adopted in a converging variety. This may occur when converging speakers miss the target or when speakers more or less intentionally choose a compromise between the converging and the converged-to variant. The result may be the emergence of compromise forms ⫺ either completely new forms that previously did not exist in either of the contact varieties, or forms that could already be found in the converging variety but only in other contexts. An example of phonetically intermediate variants mentioned by Trudgill (1986: 60) is the emergence of the intermediate vowel quality [¥] in the /w/ ⫺ /v/ transition zone between northern and southern England. In cases of vertical convergence, speakers may overshoot the target and generalize the new variant to contexts where it is not used in the converged-to variety (i. e., hypercorrection). An example mentioned by Auer (2000: 12) is the avoidance of

269

270

III. Structure and dynamics of a language space /c¸ /-coronalization; in many German dialects convergence towards standard phonology has lead to hypercorrect /c¸/ in words such as Menschen ‘people’ (standard: /mεnsn/). Another example of incomplete vertical convergence towards the standard may be found in dialects in northern and central Norway where the dialectal palatals [M], [Y], [c], [n] are replaced by the retroflexes [U], [l], [t], [K] instead of the corresponding standard alveolar variants [n], [l], [t], [d]. Palatalization of long alveolars and alveolar clusters used to be a very widespread rural feature in the north, west and central parts of Norway, but it is currently disappearing in most of these areas (e.g., Hanssen 1990; Mæhlum 1992; Røyneland 2005). The feature deviates from the standard, and from most urban varieties, even those of the urban centers in the palatalizing area. Generally, there is a high social awareness of this feature, and it is often regarded as “broad”, “rural” and “peasant-like”. The retroflex variants represent a regional development and can, perceptually speaking, be understood as compromises both in linguistic and social terms. Instead of converging completely to the standard alveolar variants, or sticking to the dialectal palatal forms, an intermediate form (which already existed in the dialects, but in other contexts) is preferred. This demonstrates that vertical convergence may result in the emergence of regiolectal forms, and maybe even regional varieties as in the Danish case. It is possible, however, that the emergence of intermediate forms is an unintended result. As mentioned above, it may well be that the speakers are actually aiming at the standard, but are simply missing the target. Hence, it is an open question whether or not the emergence of regiolectal variants or varieties is intentional and whether they are used as a means of displaying regional affiliation and identity. In some cases, what may be understood as deviance from the standard at one linguistic level may be analyzed as vertical convergence towards the standard at another. An example of this may be found in historical reflexes of the Old Norwegian syllable balance rule. The syllable balance rule has had different historical effects in different parts of Norway, and has traditionally been the most important variable that dialectologists have used to divide Norway into different dialect regions. The east Norwegian urban standard does not have any reflexes of the syllable balance rule. In central Norway, the traditional system is an apocope in most infinitives and nouns (those which had a long root syllable in Old Norwegian) and maintenance of the final vowel in the minority (those which had a short root syllable in Old Norwegian), ex. [hc6p] ‘to jump’ versus [vær6a] ‘to be’. However, apocopated forms are spreading in the region so that infinitives and nouns that traditionally should have retained the final vowel are apocopated, thus resulting in a more homogenous class of items: [vcr6c/vær6a] > [væ6r], standard [væ6re] ‘to be’, [scvc] > [sc6v], standard [sc6ve] ‘to sleep’, [v⁄k6⁄/væk6a] > [⁄6k], standard [⁄6ke] ‘week’ (cf. Røyneland 2005: 2009). This development diverges from the standard as far as the number of syllables in the affected words is concerned, but at the level of phonology there seems to be a tendency to prefer root vowels whose quality and quantity correspond to that of the standard. In this sense, one could argue that there are both vertical and horizontal convergences at work; vertical convergence towards standard phonology and horizontal convergence towards regional dialect morpho-phonology (cf. [hc6p] ‘to jump’, [kast] ‘to throw’, [vi6s] ‘a song’). This combination of vertical and horizontal convergence results in the emergence of a new hybrid variant that did not exist in either of the contact varieties. In some cases, it may be difficult or even impossible to determine whether vertical convergence is the driving force behind a given change due to isomorphism between the

15. Vertical convergence involved varieties. If the standard, the urban dialects in the area, as well as the surrounding dialects all have the same variant, it is of course impossible to decide what kind of convergence we are dealing with. In those cases, it is reasonable to assume that several processes are at work at the same time. Many studies demonstrate that local and regional varieties deviate substantially from the standard at the level of the lexicon (Sandøy 2008). It is probably at this linguistic level that we find the most pronounced vertical convergence, in the sense of direct adoption of standard words. At this linguistic level, too, it is easiest to determine whether the standard is unambiguously the ideal norm, and hence the target of convergence.

4. Conclusion As we have seen, vertical convergence is a powerful force in modern European dialect change. The impact of such convergence, however, is not the same in all linguistic communities ⫺ as the cases of Denmark and Norway clearly demonstrate. Denmark provides an example of a language space that has changed dramatically over the last 40⫺ 50 years ⫺ both geographically and socioculturally ⫺ particularly due to vertical convergence. Linguistic diversity has decreased substantially along the geographical dimension, and regional standard varieties covering larger geographical areas have emerged at least at the level of language use, (but maybe not as a means of signaling a regional identity). Linguistic diversity has also decreased along the social dimension where LowCopenhagen features have extended their reach at the expense of High-Copenhagen features. In sum, this leaves a much more homogenous language space. Nevertheless, new distinctions may emerge where old ones disappear and more subtle linguistic differences may function as symbolic markers of social, cultural, ethnic or geographical distinctions. One recent development which contributes to a more heterogeneous linguistic picture is the emergence of multiethnolectal speech styles among adolescents in multicultural environments in urban Copenhagen (cf. Quist 2005). In Oslo as well, multiethnolectal speech styles are emerging among young people in culturally diverse areas (cf. Svendsen and Røyneland 2008). At the same time, traditionally low-status Oslo features are reallocated and may now arguably be seen as part of the “modern” standard. The Norwegian language space has also undergone restructuring due to vertical convergence, but the result in Norway has not been the emergence of regional varieties and/or dialect death, but rather a more moderate dialect leveling. The developments are not the same throughout the country, but tendencies towards vertical convergence may be detected in all areas. The result, however, often differs from the standard itself and may contribute to the establishment of new intermediate variants and to the creation of new distinctions. The Danish and the Norwegian cases demonstrate that vertical convergence may appear both within language spaces that are characterized by a well-defined spoken standard and a strong standard ideology, and in language spaces where the spoken standard and the standard ideology have a relatively weak status. Vertical convergence toward a hegemonic variety may be complete (or almost complete), or it may be realized to a significantly smaller degree. In the latter case, the result may be the formation of intermediate forms and eventually intermediate varieties. Vertical convergence, thus, results in a restructuring of the language space. As the Danish and Norwegian cases clearly

271

272

III. Structure and dynamics of a language space show, the result of vertical convergence may be the emergence of new lects, dialect death or simply dialect leveling. In many cases, developments such as the ones we have seen in Denmark and Norway may amount to linguistic simplifications. However, as the case of the split of coronal fricatives in Andalucı´a clearly illustrates, simplification need not be the result. Here, in fact, vertical convergence has produced a system of greater complexity. The impact of vertical convergence on language spaces can vary greatly, and will be mediated by such sociocultural factors as the relative social standing and indexicality of the converging and converged-to varieties.

5. Reerences Akselberg, Gunnstein 2005 Dialects and regional varieties in the 20th century II: Norway. In: Oskar Bandle, Kurt Braunmüller, Ernst Ha˚kon Jahr, Allan Karker, Hans-Peter Naumann, Ulf Telemann (eds.), The Nordic Languages. An International Handbook of the History of the North Germanic Languages, vol. 2, 1707⫺1721. (⫽ Handbücher zur Sprach- und Kommunikationswissenschaft 22.2.) Berlin/New York: de Gruyter. Auer, Peter 2000 Processes of horizontal and vertical convergence in present day Germany. Ma˚lbryting 4: 9⫺26. Auer, Peter 2005 Europe’s sociolinguistic unity, or: A typology of European dialect/standard constellations. In: Nicole Delbecque, Johan van des Auwera and Dirk Geeraerts (eds.), Perspectives on Variation, 6⫺42. Berlin/New York: Mouton de Gruyter. Auer, Peter and Frans Hinskens 1996 The convergence and divergence of dialects in Europe. New and not so new developments in an old area. Sociolinguistica 10: 1⫺25. Brink, Lars and Jørn Lund 1975 Dansk Rigsma˚l: Lydudviklingen Siden 1840 med Særligt Henblik pa˚ Sociolekterne i København. [Danish Rigsma˚l: The Development of Sounds Since 1840 with Particular Regard to the Sociolects of Copenhagen.] Copenhagen: Gyldendal. Bull, Tove 1992 Dialect and standard in a language contact area in northern Norway. In: Jan A. van Leuvensteijn and Jan Berns (eds.), Dialect and Standard Language, 365⫺378. Amsterdam: Royal Netherlands Academy of Arts and Sciences. Deumert, Ana and Wim Vandenbussche (eds.) 2003 Germanic Standardizations. Past to Present. Amsterdam/Philadephia: Benjamins. Edlund, Lars-Erik 2003 Det svenska spra˚klandskapet. De regionala spra˚ken och deras ställning i dag ⫺ och i morgon. [The Swedish linguistic landscape. The regional languages and their position today ⫺ and tomorrow.] In: Gunnstein Akselberg, Anne Marit Bødal and Helge Sandøy (eds.), Nordisk Dialektologi, 11⫺51. Oslo: Novus. Hanssen, Eskil 1990 Nordland. In: Ernst Ha˚kon Jahr (ed.), Den store Dialektboka [The Big Book of Dialects], 141⫺155. Oslo: Novus. Haugen, Einar 1972 [1968] The Scandinavian Languages as cultural artefacts. In: The Ecology of Language: Essays by Einar Haugen, 265⫺286. Stanford: Stanford University Press.

15. Vertical convergence Hinskens, Frans 1996 Dialect Levelling in Limburg. Structural and Sociolinguistic Aspects. Tübingen: Niemeyer. Jahr, Ernst Ha˚kon 1997 On the use of dialects in Norway. In: Heinrich Ramisch and Kenneth Wynne (eds.), Language in Time and Space: Studies in Honour of Wolfgang Viereck on the Occasion of his 60th Birthday, 363⫺369. Stuttgart: Steiner. Jahr, Ernst Ha˚kon 2003 Norwegian. In: Ana Deumert and Wim Vandenbussche (eds.), Germanic Standardizations: Past to Present, 331⫺355. Amsterdam/Philadelphia: Benjamins. Kerswill, Paul 2000 Mobility, meritocracy and dialect levelling: The fading (and phasing) out of Received Pronunciation. Paper given at “British Studies in the New Millennium: the Challenge of the Grassroots”, University of Tartu, 21 August 2000. Kristiansen, Tore 2003a Danish. In: Ana Deumert and Wim Vandenbussche (eds.), Germanic Standardizations. Past to Present, 69⫺93. Amsterdam/Philadephia: Benjamins. Kristiansen, Tore 2003b Sproglig regionalisering i Danmark? [Linguistic regionalization in Denmark?]. In: Gunnstein Akselberg, Anne Marit Bødal and Helge Sandøy (eds.), Nordisk Dialektologi [Nordic dialectology], 115⫺149. Oslo: Novus. Kristiansen, Tore and J. Normann Jørgensen 2003 Introduction: The sociolinguistics of Danish. International Journal of the Sociology of Language 159: 1⫺9. Kristensen, Kjeld 2003 Standard Danish, Copenhagen sociolects, and regional varieties in the 1900s. International Journal of the Sociology of Language 159: 29⫺43. Kristensen, Kjeld and Mats Thelander 1984 On dialect levelling in Denmark and Sweden. Folia Linguistica 28(1⫺2): 223⫺246. Labov, William 1994 Principles of Linguistic Change. Internal Factors. Oxford: Blackwell. Mæhlum, Brit 1992 Dialektal Sosialisering: En Studie av Barn og Ungdoms Spra˚klige Strategier i Longyearbyen pa˚ Svalbard. [Dialectal Socialization: A Study of Children and Adolescents’ Linguistic Strategies in Longyearbyen at Svalbard]. Oslo: Novus. Mæhlum, Brit 2007 Konfrontasjoner. Na˚r spra˚k møtes. [Confrontations. When Languages Meet]. Oslo: Novus. Mattheier, Klaus J. 1997 Über Destandardisierung, Umstandardisierung und Standardisierung in modernen europäischen Standardsprachen. In: Klaus J. Mattheier and Edgar Radtke (eds.), Standardisierung und Destandardisierung europäischer Nationalsprachen, 1⫺11. Frankfurt am Main: Lang. Pedersen, Inge Lise 2003 Traditional dialects of Danish and the de-dialectization 1900⫺2000. International Journal of the Sociology of Language 159: 9⫺29. Pedersen, Inge Lise 2008 Two processes of de-dialectalisation ⫺ and a historical explanation for both. Paper presented at “Sociolinguistic Symposium 17”, Amsterdam, 3⫺5 April. Quist, Pia 2005 Stilistiske praksisser i storbyens heterogene skole ⫺ en etnografisk og sociolingvistisk undersøgelse af sproglig variation. [Stylistic practices in the urban heterogeneous school ⫺ an ethnographic and sociolinguistic research of linguistic variation]. Unpub-

273

274

III. Structure and dynamics of a language space lished PhD dissertation, Nordisk Forskningsinstitut, Avd. for dialektforskning, Københavns Universitet. Røyneland, Unn 2005 Dialektnivellering, ungdom og identitet. Ein komparativ studie av spra˚kleg variasjon og endring i to tilgrensande dialektomra˚de, Røros og Tynset. [Dialect leveling, adolescents and identity. A comparative study of linguistic variation and change in two neighboring communities, Røros and Tynset]. Unpublished PhD dissertation, Faculty of Humanities, University of Oslo. Røyneland, Unn 2009 Dialects in Norway ⫺ catching up with the rest of Europe? International Journal of the Sociology of Language 196/197: 7⫺30. Sandøy, Helge 1998 The diffusion of a new morphology in Norwegian dialects. Folia Linguistica 32: 83⫺100. Sandøy, Helge 2004 Types of society and language change in the Nordic countries. In: Britt-Louise Gunnarsson et al. (eds.), Language Variation in Europe: Papers from the Second International Conference on Language Variation in Europe, ICLaVE 2, Uppsala University, Sweden, June 12⫺12, 2003, 53⫺76. Uppsala: Universitetstryckeriet. Sandøy, Helge 2008 Standardtalema˚l [Spoken standard?]. Paper presented at the “Standard eller ikke standard? Har vi et standardtalema˚l i Norge?” [Standard or not? Do we have a spoken standard in Norway?] conference, Røros, February 8⫺10. Skautrup, Peter 1947 Det Danske Sprogs Historie. [The History of the Danish Language], vol. II. Copenhagen: Gyldendal. Skjekkeland, Martin 2000 Dialektutviklinga i Noreg dei Siste 15 A˚ra ⫺ Drøfting og Analyse [The Dialect Development in Norway of the Last 15 Years ⫺ Discussion and Analysis]. (Skriftserien 67.) Kristiansand: Høgskolen i Agder. Solheim, Randi 2007 Spra˚kbruk og spra˚khaldningar i Høyanger etter andre verdskrigen. Eit sosiolingvistisk særtilfelle eller eit lokalt døme pa˚ ein allmenn tendens? [Language use and language attitudes in Høyanger after the Second World War. A sociolinguistic peculiarity or a local example of a general tendency?] Ma˚lbryting 8: 7⫺29. Svendsen, Bente Ailin and Unn Røyneland 2008 Multiethnolectal facts and functions in Oslo, Norway. International Journal of Bilingualism 12(182): 63⫺85. Thelander, Mats 1980 De-dialectisation in Sweden. (FUMS Rapport 86.) Uppsala: Uppsala Universitet. Trudgill, Peter 1986 Dialects in Contact. (Language in Society 10.) Oxford: Blackwell. Vikør, Lars S. 1995 The Nordic Languages: Their Status and Interrelations. (Nordic Language Secretariat 14.) Oslo: Novus. Vikør, Lars S. 1999 Austlandsma˚l i endring. [Eastern dialects in change.] In: Turid Kleiva, Ingeborg Donali and Trygve Nesset (eds.), Austlandsma˚l i endring, 13⫺48. Oslo: Samlaget. Villena-Ponsoda, Juan Andre´s and Matilde Vida Castro 2005 The effect of social prestige on reversing phonological changes: Universal constraints on speech variation in southern European Spanish. Unpublished manuscript.

Unn Røyneland, Oslo (Norway)

16. Divergence

16. Divergence o linguistic varieties in a language space 1. 2. 3. 4. 5.

Linguistic divergence: Term and concept Research overview and examples Types of linguistic divergence Parameters of linguistic divergence References

1. Linguistic divergence: Term and concept Divergence is a kind of change by which relatively uniform linguistic varieties become more dissimilar. Divergence is therefore a diachronic process per se and, as such, is dynamic over time. The divergence of linguistic varieties becomes apparent within a geographical space and/or within a social stratification, thus it shows diatopic and/or diastratic dynamics as well. We speak of “horizontal” divergence as far as diatopic dissimilation is concerned, and “vertical” divergence concerning diastratic dissimilation. As the term “relative uniformity” shows, the starting point of dissimilation is not identity. Absolute uniformity may occasionally occur when a variety is split into several subvarieties, for instance as a consequence of the migration of its speakers (see Riehl in this volume). “Relative uniformity” means that in the unmarked case the varieties before dissimilation form a continuum. In this sense, divergence amounts to a break within a continuum. Varieties may diverge in particular structural features or in bundles of such features to different degrees. Only in theory will a linguistic space split in all structural features which then define and delimit the emerging new spaces exhaustively. For this reason, the denominations of linguistic spaces according to the dialects which “occupy” them ⫺ e.g., “Bavarian” or “Eastern Franconian” ⫺ give an unrealistic, highly idealized picture.

2. Research overview and examples Research on the divergence of dialects has been disregarded for a long time. The reason for this, according to Kremer (1990: 86), is that during the past decades dialectology has concentrated on the so-called base dialects which seemed to fulfill the ideal of purity (Henzen 1954: 41). But it is also true that traditional dialectology had an elementary interest in dialectal dissimilarity in so far as it attempted to explore the boundaries (isoglosses) of “old” language spaces along which the variants of certain linguistic features or bundles of features differed. After establishing these language spaces and their boundaries, traditional dialectology sought to explain the conditions and factors of their emergence (e.g., political or economic territories, areas of communication, lines of transportation). On the whole, spatial difference was considered very static. There was more emphasis on the results of divergence, i. e., the existing dialect spaces, than in dialectal divergence itself (cf. Radtke 2006: 2189).

275

276

III. Structure and dynamics of a language space After the sociolinguistic turn in dialectology, phenomena of recent change came to the fore and with them the questions relating to their social and pragmatic conditions. Social dialectology turned to “dialect leveling” (Dialektabbau) by researching the vertical convergences of dialects and standard languages, mostly in the form of unilateral advergences of the dialects to the standard languages (Auer 2004: 170). Later, phenomena such as the “restructuring” of former local dialects (Dialektumbau) and their “regionalization” (Schunk 1999) were considered as well, i. e., cases of horizontal convergence of neighboring dialects. Only recently, however, have the processes of dialect change that led to these divergences been analyzed. The impetus came from a school of research that worked on the problem of how recent political borders can turn into language boundaries. The first to deal with this question were Kremer and Niebaum (1990) with a volume on central and western central European language situations where former dialect continua are cut through by political borders. Kremer (1990: 86) points out that the European tradition of organizing linguistics according to national language philologies has for a long time been “fatal” for research of this kind. Taking the border zone between Germany and the Netherlands for an example, which has become a model for recent research into linguistic divergence, he states that German and Dutch dialectology confined themselves to their national territories and for too long were unable to look beyond them. This was hardly reasonable and unwarranted, at least up until the middle of the twentieth century, in that Dutch and (Low) German dialects at the Dutch/German border could be called “Dutch” and “German” respectively only on the basis of sociolinguistic criteria ⫺ that is, because they were roofed by the Dutch standard language on the one side and the German standard language on the other. Apart from that, the two dialect areas were part of a continuous Continental West-Germanic language space in which, as a rule, the state frontiers were not identical with significant dialect boundaries. Auer, Hinskens and Mattheier (1996), Auer (1998) and Kallen, Hinskens and Taeldeman (2000) contain studies on the borders between Ireland and Northern Ireland, Scotland and England, Finland and Russia, etc. In another anthology dedicated to this subject (Auer, Hinskens and Kerswill 2005), Woolhiser (2005) deals with divergence in the eastern and southeastern European language space, such as in the Czech Republic, former Yugoslavia and its now independent republics (particularly Serbia and Croatia), in the contact areas of Macedonia and Bulgaria, etc. In this article I confine myself to the Continental West Germanic language space and its neighboring areas. The essential parameters and their constellations can be shown in this restricted language space as well. Divergence of varieties in a language space will be described on the basis of recent developments in the following sections, because in these instances the data base is much more solid than in historical situations. Accordingly, the factors that will be shown to cause and influence such processes, e.g., state borders, political territories (in contrast to the medieval forms of personally organized governance), are “modern” ones. Nevertheless, identical or at least comparable factors hold for processes of divergence in historical times as well, that is, e.g., for the emergence of the traditional dialect areas which were the objects of traditional dialectology. Political borders, albeit not yet “political” in the modern sense of the word, but governmental with reference to pre-modern fragmented areas of power, have always played a role in the formation of distinct, likewise fragmented, language areas. Some factors, e.g., religions denomination, have lost their force

16. Divergence over the course of time. Others, including factors of economic exchange, interpersonal communication and, within certain limits, factors of “natural” geomorphology, are still now as influential on the emergence of separate (divergent) linguistic spaces as they were in the past.

3. Types o linguistic divergence The types of linguistic divergence will be exemplified by cases in the Continental WestGermanic language space. The criterion will be the kind of language roof (cf. Kloss 1978 for the concept of linguistic roofing): ⫺ varieties beneath one roofing language (e.g., German at the German⫺Swiss border along the Upper Rhine) ⫺ varieties beneath two cognate but different roofing languages (e.g., German and Dutch at the border between Germany and the Netherlands along the Lower Rhine) ⫺ varieties beneath non-cognate roofing languages (e.g., German and French at the Baden⫺Alsatian border along the Upper Rhine). The focus of this article is horizontal divergence in space. Vertical divergence will not be treated in the same depth, but will be taken into account in relevant cases. Before we enter the discussion of present-day examples, two historical instances of vertical divergence will be mentioned: the emergence of French and German standard languages. Endeavors to establish common (standard) languages above dialects which continue in existence always lead to vertical divergence. When the dialect of the Iˆle de France (the Paris basin) was installed on top of the French dialects (patois) and extended to the whole area of the langue d’oı¨l, a diglossic situation emerged, in which two vertically divergent varieties, a higher and a lower one, co-exist. Vertical divergence arises, too, when a koine´ on the top of the dialects is established, as within the German language space from the fifteenth century onwards. In principle, the French, centralistic way of raising the dialect of a politically and economically powerful region to a nationwide standard language above the patois had the same effect as the German way of creating a compromise language (Ausgleichssprache) out of and above the local and regional dialects. Linguistic stratification and vertical divergence were the result of both processes. Once established, however, any standard variety causes linguistic homogenization. Standard language and the dialects below it vertically converge, mostly by advergence of the low varieties to the high one. Horizontal convergence of diatopic low varieties is a secondary result in the long run.

3.1. Divergence beneath the same rooing language 3.1.1. Divergence in the German language space Divergence beneath the roof of the German standard language will be described with reference to the situation at the former inner-German state border between Thuringia

277

278

III. Structure and dynamics of a language space (German Democratic Republic, GDR) and Bavaria (Federal Republic of Germany, FRG). Following World War II, this border completely separated a formerly homogeneous dialect space for four decades (Petzold 1993; Harnisch, Reinhold and Schnabel 2008). The resulting divergence shows, for example, that when one compares the former syllable-initial r-articulation with the current one (Reinhold 1997; Schnabel 2006: 41, following Harnisch 2004a in methods). The traditional spatial distribution of this feature is as follows. In East Franconian apical [r] was spoken, in Thuringian uvular [r]. Four decades of state and language separation had the effect of aligning the borderline between the two r-articulations with the political border: regions in the GDR in which apical [r] had been spoken adapted themselves to the Middle German hinterland with its uvular [r] articulation, and regions in the FRG in which uvular [r] had been spoken adapted themselves to the Upper German hinterland with its apical [r] realization. Thus divergence resulted from forces of adaptation on both sides of the border, which were opposite in effect but similar in manner (cf. Maps 16.1a and 16.1b). The horizontal divergence of the former homogeneous base dialect at the border is conditioned by processes of horizontal convergence between the diverging dialect zones

Map 16.1a: r-boundary in 1935

Map 16.1b: r-boundary in 1994 (young generation)

16. Divergence and their hinterlands. On the FRG side, speakers took over the [r]-sound of their EastFranconian hinterland; on the GDR side, the [r]-sound of the Thuringian hinterland. It cannot be ruled out that the process was supported by the r-variants of regional varieties: in East Germany by the uvular [r] spoken in the East Middle German vernacular and favored by the German orthoepic standard norm, in West Germany by the apical [r] used in the East-Franconian language space as well as in other Upper German vernaculars. Even in the South German standard variety, including the Austrian and Swiss German language space, it is at least socially unmarked (cf. Part IV of Ammon et al. 2004, in particular Harnisch 2004b). The German standard pronunciation norm, however, favors the northern and western uvular, more prestigious [r]-variant. In this regard, the new [r]-articulation that was introduced into the dialects on GDR soil close to the Franconian border after World War II at the expense of the older [r]-articulation vertically converges with the Standard German target norm, too. The new [r]-articulation, however, that entered FRG soil after World War II and replaced the old [r]-articulation, does not correspond with the Standard German norm, but implies vertical divergence from the standard. This horizontal divergence of linguistic features at a political border is usually accompanied by a change in linguistic attitudes. For the inhabitants of the GDR, the articulation of the apical [r] was generally associated with “Bavarian” (Bayerisch), here understood as the name of the political territory. (The dialect space in the FRG next to the border, which is part of the state of Bavaria, belongs to the East-Franconian area.) In other words, when “Bavarian” was used as an ethnolectal term, its meaning was “a variety spoken by people from the Federal State of Bavaria that differs from our own variety”. After the alignment of the dialectological border between the different r-articulations with the political border between East and West Germany (or nowadays between Thuringia and Bavaria), the expression Bavarian in the sense of the political term Bayerisch based on the feature of the apical [r] is even more justified. The example demonstrates that speakers can be aware of the form of their own language and that of the language of others and can evaluate them in an ethnolinguistic manner: from over here vs. from over there (von hüben vs. von drüben), in the political sense of from the East vs. from the West; in the geographical sense as northern vs. southern; or as closer to vs. further away from the written standard language in the sociolinguistic sense. Thus, the change from the [r] of the elder generations to the [r] of the younger generations on the Bavarian side can be interpreted as a turning away from a shibboleth sound associated with coming from over there (the GDR) and as the simultaneous turning to a South German vernacular, which is more prestigious in the eyes of the younger generations. It therefore seems to be rather improbable that this re-orientation took place unconsciously, i. e., as a consequence of turning to what now is heard more often and turning away from what now is heard less frequently. At any rate, the [r]-variant preferred in the standard language was not taken as a model. The same political border was the scene of a process of horizontal divergence of a different kind. It was driven not by horizontal convergences with differing hinterlands at a base dialectal level, but by vertical convergences of the base dialect towards higher varieties whose strength differed on each side of the border. Before the division of Germany, speakers from that area used words like [veis] ‘meadow’ and [geisn] ‘to water’, [neibe] ‘over there’ and [kei] ‘cows’, [stoup] ‘room’ and [fous] ‘foot’, i. e., a mid-rising diphthong. In the course of 40 years of political separation, however, a horizontal diver-

279

280

III. Structure and dynamics of a language space gence emerged as the consequence of the fact that the diphthong remained stable on the FRG side whereas the younger generation on the GDR side gave up this feature and replaced the diphthong with a long monophthong, as found in the East Middle German vernacular and Standard German (Schnabel 2006): [vi:s, gi:sn, ni:be], etc. Traditional (socio-) dialectology would have accredited this process to differing amounts of dialect leveling. On the one hand, it can be assumed that the rising diphthong is a “primary dialect feature” in the sense of Schirmunski (1930), which became stigmatized and was consciously abandoned for this reason (upward vertical convergence). On the other hand, one could claim that the decline of the dialects on the GDR side was due to the influence of the Thuringian hinterland whose dialects also have long monophthongs and that therefore the loss of the diphthongs is supported or even caused by horizontal convergence, as in the case of [r]. However, in the FRG no convergence took place, neither vertically towards “higher” varieties nor horizontally towards neighboring East-Franconian dialects, which also lack the rising diphthongs and have long monophthongs, as in the standard language, instead. Another process of divergence had the opposite effect: in Sparnberg in the GDR, the old centralization of all vowels and the [A]-articulation of unstressed final -er in words like Eier ‘eggs’, Fuder ‘cartload’ etc. were preserved, whereas Rudolphstein in the FRG adapted to the linguistic hinterland, which has standard, non-centralized vowels and [B] for final -er. In the pre-unification FRG, both Middle German features carried the stigma of “sounding GDR”. Schnabel (2006: 41) analyzed the data of the five oldest and the five youngest interviewees and found the distribution shown in Table 16.1 (every “⫹” or “⫺” represents one person). The language change on the FRG side is therefore complete.

Tab. 16.1: Articulation of unstressed final -er The five oldest interviewees

The five youngest interviewees

Former GDR

Sparnberg

⫹⫹⫹⫹⫹

⫹⫹⫹⫹⫹

Former FRG

Rudolphstein

⫹⫹⫹⫹⫹

⫺⫺⫺⫺⫺

Divergences also took place in the urban variety systems of the politically divided city of Berlin (Schlobinski 1997). These are linked to the social prestige of the Berlin vernacular. During the post-war decades the interactional style associated with this Berlin vernacular, the Berliner Schnauze (Berlin bluntness), became stigmatized in the middle and upper classes of West Berlin, whereas it persisted as a variety of broad social acceptance in East Berlin. In the western part of the city this devaluation of the vernacular led to the emergence of a prestigious urban standard language and therefore to vertical divergence. With regard to their present-day systems of urban varieties, West and East Berlin diverged horizontally as well (cf. Figure 16.1). Horizontal divergence on the level of vernaculars, sometimes also on the level of standard languages (in the sense of regional or national standard varieties), can be observed on both sides of the German⫺Austrian border (Scheuringer 1990: 372). In this area, on the Lower Bavarian and the Upper Austrian side of the river Inn, different vernaculars have emerged, the one being based on the model of the Vienna urban variety,

16. Divergence

281 West Berlin L East Berlin

H-variety L-variety

higher urban variety C Berlin jargon

L horizontal divergence

Berlin jargon C vertical divergence

Fig. 16.1: Divergences in the Berlin vernacular

the other on the model of the Munich urban variety. Influenced by the Viennese urban variety, the vernacular speakers on the Austrian side tend to articulate au and ei as monophthongs ([hc:s] ‘house’, [ve:s] ‘white’) whereas the speakers on the Bavarian side preserve the articulation of these segments as diphthongs ([haws], [va=s]). A particularly salient feature of this horizontal divergence at the state border is the reading pronunciation of the a-vowel as [a] on the Austrian side and as [c] on the Bavarian side. Another kind of horizontal divergence between the Austrian and the Bavarian language space is caused by different ranges which linguistic features can occupy on the standard/dialect scale. For instance, in Austria, Kren ‘horseradish’ occupies all levels on this scale, including the national standard variety, whereas in Bavaria (in the Bavarian and the East-Franconian dialect space) this lexeme is replaced by Meerrettich in standard-near ways of speaking (Meerrettich is the standard word in the Federal German language space). Considering the long-standing political inclusion of the now Austrian Inn district (Innviertel) in Bavaria (until 1779), it is very likely that the common Bavarian variety used as a lingua franca before that time was closer to the regiolect spoken on the Bavarian side today than to the one now spoken on the Austrian side. With the new political inclusion, the distance between the Inn district base dialect and the Austrian common variety based on the Viennese urban variety was much wider than that between the base dialect and common Bavarian based on the Munich urban variety. To this day it remains true that when speakers of the Austrian Inn district switch from their base dialect to common Austrian, more ⫺ and more salient ⫺ linguistic features are involved than when speakers of the Lower Bavarian Inn area switch from their base dialect to Munich-based common Bavarian. It is thus not the “natural border” of the river Inn that caused a linguistic divergence, but rather the fact that this river was declared a political border. Divergence is also taking place at the border between Austria and Italy, between North and South Tyrol. In the old undivided Tyrol, dialectal borders ran between East and West, but none separated North and South (Mall and Plagg 1990). After the annexation of South Tyrol by Italy after World War I, however, a particular sociolinguistic space emerged due to the fact that urban vernaculars could not develop in South Tyrol. One reason for this was that the German-speaking population was rather immobile. Another reason was that the South Tyrolean dialects were so similar that they were mutually comprehensible and did not cause the speakers to switch to a super-local variety. Furthermore the South Tyroleans quickly shifted to Italian in public domains. With regard to the language system itself, the South Tyrolean varieties inevitably diverge from the North Tyrolean ones by integrating large amounts of lexical material into the standard as well as into the dialect. This is due to the fact that in germanophone South Tyrol

282

III. Structure and dynamics of a language space the heterogenetic Italian language constitutes a second roof on top of the homogenetic German language. A vertical divergence between base dialect and standard variety can also increase when speakers in a “pluricentrically” organized language space (cf. Ammon 1995 with regard to germanophone states, Ammon 2005 and Clyne 1991, 2004 in general) switch to another standard variety further away from the dialects than the old one. An example is provided by the German-speaking South Tyroleans who are said to adapt their standard to that of Germany (instead of the Austrian standard traditionally used) because of the many German tourists coming into the region. Another example of horizontal divergence at a political border is the Swiss-German state border in the Upper Rhine area (Schifferle 1990: 318). The difference between the German and the Swiss language space is already hinted at in the terms for the respective dialectal varieties. On the Swiss side the dialects are called Schwyzerdütsch ‘Swiss German’, a political reference; on the German side they are called Alemannic, without any reference to a political space. On the Swiss side there is strict medial diglossia characterized by the use of dialect(s) in any oral, even public and official, domain and the use of the standard language in written domains only. On the German side, the use of standard and dialect is controlled by situational and social factors. The dialect is not spoken in public situations, but remains restricted to more private spheres. The dialects on both sides of the border are structurally similar and mutually comprehensible without problems. But they are not used by the Germans (from Baden) in conversation with their Swiss neighbors, nor vice versa, because the diglossic Swiss situation is not compatible with the diaglossic German situation (cf. Bellmann 1998 for the concept of diaglossia). In addition to this divergence in variety use, in recent times divergences concerning the dialectal features themselves have come into being. On the German side the spirantization or affrication of Germanic k is retreating word by word in a conscious process. Apical [r] is retreating, too, in favor of the uvular [r] that is preferred by the standard language norm in Germany. This feature is very salient and has far-reaching consequences for the perception of the language spaces. Linguistic innovations advancing from north to south in Germany stop at the political border nowadays. For instance, the new preterite form of ‘to be’ (war ‘was’) is replacing the old forms with the perfect

Tab. 16.2: Diglossia and diaglossia in the contact zone of Switzerland, Liechtenstein, and Austria written

spoken

The Swiss and Liechtenstein Rhine valley: H

standard German

M L

dialect —

Vorarlberg (Austria): H

standard German

M L

vernacular(s) —

dialect

16. Divergence participle gsii (< gesıˆn ‘been’) in Germany but does not cross the border into Switzerland. The contrast between medial diglossia on the one hand and diaglossia with vernacular intermediate levels on the other also holds for a linguistic region in which three countries meet: Switzerland (Rhine valley of the St. Gallen canton), Liechtenstein and Austria (Vorarlberg). In Liechtenstein there is medial diglossia as in Switzerland, whereas Vorarlberg goes along with the Austrian diaglossia (Banzer 1990). In a cross-classification of diastratic levels (H ⫽ high, M ⫽ middle, L ⫽ low) and medial representations (written versus spoken), two models oppose each other at the border (see Table 16.2). As far as the repertoire and the social regulation of variety use are concerned, the Vorarlberg speakers of Austria behave like the Baden speakers of Germany, i. e., they differ from the Swiss model.

3.1.2. Divergence in the Dutch language space Structural divergence on the level of regional standard varieties (cf. section 3.5.) as well as on the dialectal level can also be observed along the border between the Netherlands and Belgium (Taeldeman 1990). For instance, ⫺ the Dutch dialects in the Netherlands gave up syntactic patterns that they originally had in common with the Flemish dialects at their western outskirts (durven ‘to dare’ ⫹ infinitive) in favor of standard Dutch patterns (durven ⫹ te ‘to’ ⫹ infinitive); ⫺ word geographic areas on both sides expanded towards the political border (e.g., Netherlands Dutch spijker ‘nail’ vs. Flemish nagel); ⫺ the Flemish variety, in contrast to the Dutch variety in the Netherlands, has integrated French loan words on a huge scale and, when word stress was variable (moto´or vs. mo´tor), it has generalized the French word-final stress (Flemish moto´or/mote´ur vs. Netherlands Dutch mo´tor); ⫺ conversely on the Flemish side, anti-French purism replaced loan words from French with Germanic words (general Dutch punaise ‘thumb tack’ > Flemish duimspijker); ⫺ speakers of Netherlands Dutch prefer to construct new terms out of Dutch elements (koppeling ‘clutch’) and resort to loan words from English (rubber), whereas on the Belgian side the French terms are used (embrayage, caoutchouc). With regard to the socio-dialectal situation, there is divergence at the political border, too. The dialect is used more on the Belgian side than in the Netherlands, and the standard on the Belgian side shows interference from the dialect, above all in the Brabant, whereas the standard language in the Netherlands does not. Finally, different attitudes on both sides of the border bring about divergence. The Dutch are indifferent towards their Flemish fellow speakers at best or they look down on them and their language. The Flemish view the Dutch and their “smart” articulation with reserve or even disapproval. Dutch is also telling when we consider what influence the age of a border has. For instance, Map 16.2 on the vowels in the Dutch word drie ‘three’ shows that the phonological and political boundaries run parallel along the older border between the Nether-

283

284

III. Structure and dynamics of a language space

Map 16.2: Boundaries of vocalism in West Germanic ‘three’ (based on Taeldeman 1990: 282)

lands and Belgium (Old Flanders and Old Brabant, see left and middle third of the map), whereas phonological boundaries disregard the political ones where the border is more than 250 years younger (Limburg, see rightmost third of the map).

3.2. Divergence beneath homogenetic but dierent rooing languages 3.2.1. The contact zone o Germany and the Netherlands Divergence beneath the roof of different but homogenetic languages can be observed at the Dutch⫺German border inter alia. Again structural features as well as socio-pragmatic factors are involved. Since World War II with its negative consequences on (language) contact across the border, the former dialect continuum has broken down. The example of the Dollart and Vechte region shows how a former diglossic situation on both sides of the border is transformed into a quasi-monoglossic situation in which mostly (regional) standard German or Dutch is spoken (Niebaum 1990). The German and the Dutch standard languages are of course very different (though homogenetic), and the political border has therefore turned into a linguistic boundary ⫺ not only between two standard languages but also between the substandard varieties that have assimilated more (Germany) or less (Netherlands, cf. Giesbers 2008) to the respective standard languages. In this respect, a further divergence can be observed. The dialects on the Dutch side are much more tolerant of transferences from the standard because dialect and standard are similar in vocabulary and grammatical structure. This does not apply to the German side because the Low German variety and the High German standard language differ greatly in these respects. For instance, in the Netherlands, the old cross-border serialization of perfect participle ⫹ finite verb in the verb phrase of subordinated clauses, as in dat [de Eerappels] aanbrannt binn’n lit.: ‘that [the potatoes] burnt

16. Divergence are’ is transformed into the standard language serialization of finite verb ⫹ perfect participle (dat [de Eerappels] binn’n aanbrannt ‘that [the potatoes] are burnt’). Investigating the section of the German-Dutch border that follows in the south, Kremer (1990) adds further observations concerning the question of how a language continuum can break up at a state border. On the basis of examples of vocabulary, phonology and morphology, he proves that the spread of innovations caused by transferences from the standard language into the dialects has stopped at the border in recent times. Processes of interchange and balance that are essential for continua cease to apply. For instance, the distributionally conditioned reflexes of West-Germanic sk adapt to the respective standard languages on both sides of the state border. The standard Dutch diminutive ending -tje advances from the west towards the state border and replaces the diminutive ending -ken which was the former cross-border variant. The same holds for the standard Dutch future tense with gaan-auxiliary, which replaces the old base dialect future tense with the auxiliary zullen or the use of the present tense to denote events in the future. On the eastern side, the serialization type dat heff he nich betalen wollen lit.: ‘that has he not (to) pay wanted’, i. e., ‘he did not want to pay that’ in verb phrases formed by negation ⫹ infinitive ⫹ infinitive-like participle which is influenced by Standard German replaces the old pattern negation ⫹ infinitive-like participle ⫹ infinitive (i. e., dat heff he nich wollen betalen), which the Low German dialects had shared with dialects on the Dutch side. The state border is perceived as a strong dialect border in the subjective view of the speakers on both sides. When asked in which villages the (approximately) same dialect was spoken, informants positioned almost all villages on their side of the border in their answers.

3.2.2. The contact zone o Germany and Denmark At the German⫺Danish border, divergence can be observed with regard to several varieties (Dyhr 1990). North of the state border, the South Jutland variety assimilates to the standard Danish language; south of the border it shows interferences from German varieties. The Standard Danish language in its South Slesvig variant (in Germany) deviates from Denmark Standard Danish on all structural levels. Interferences from Standard German and North German substandard varieties are responsible for this.

3.3. Divergence beneath heterogenetic rooing languages 3.3.1. The contact zone o France and Germany The language space along the Upper Rhine, with the Alsatian area roofed by French and the Baden area roofed by German, may serve as an example of divergences of linguistic features and of language use beneath heterogenetic roofing languages (Klausmann 1990: 2000). On the Baden, but not on the Alsatian side, for instance, linguistic features from higher varieties leak into lower ones and influence the dialects, while on the Alsatian, but not on the Baden side, the dialect borrows French words. These lexical innovations from French stop at the river Rhine today, whereas in the past AlsaceGerman (elsasserdütsche) lexemes could sometimes cross the Rhine because of the lin-

285

286

III. Structure and dynamics of a language space guistic influence of the big Alsatian cities Strasbourg and Colmar. Due to the inclusion of germanophone Alsace in France, Alsatian became a “roofless outer dialect” (dachlose Außenmundart in Kloss’s [1978] terms), and no German regiolect (and therefore no diaglossia) could emerge on the French side of the Rhine as it did on the German side (see Table 16.3). Tab. 16.3: Types of variety stratification in the contact zone of Germany and France Alsace

Baden

Standard French

Standard German

Level(s) of vernaculars

⫺

regiolectal German

Level of base dialects

Alsatian (Alemannic)

dialectal German (Alemannic)

Level of roofing languages

Horizontal divergence is caused by the fact that there is a continuum of forms between the standard variety and dialect on the Baden side, whereas on the Alsatian side the speakers who want to switch from the dialect into a higher level are compelled to use French (see Harnisch [1996: 419⫺422] on the two-language type of diglossia). The origin of the Alsatian situation after the Thirty Years’ War is another example of how a so-called natural border was turned into a political one and finally developed into a linguistic boundary along which structural, sociolinguistic and pragmatic features increasingly diverge.

3.3.2. The contact zone o France and Belgium The divergence of originally similar dialects at the border of two states with different roofing languages can be a preliminary stage of divergence of a language as a whole at this border. With regard to Flemish, this can be observed at the Belgian⫺French border. Innovations in phonology, grammar or vocabulary start in the eastern (Belgian West Flanders) as well as western (French Flanders) part of the language space, but these usually stop at the state border nowadays. The eastern innovations (e.g., vis, mens replacing visch, mensch ‘fish’, ‘human being’; brandde/ebrand replacing bron/ebronnen ‘burnt. pret/p.p.’) turn the French part of the Flanders area into a relic zone, whereas the western innovations that also come to a standstill at the border (e.g., the weak preterite ending in -ste; increasing number of borrowings from the French which are not used in Belgian West Flanders) cause the same area to appear to be a divergent progress zone (cf. Ryckeboer 1990: 2000).

3.4. Divergence in contact zones with dierent types o linguistic roos 3.4.1. The contact zone o Germany, Luxembourg, and France In the so-called SaarLorLux region, three areas (belonging to different states) that share the Moselle-Franconian base dialect border on one another: East Lorraine (originally German speaking and roofed by the French language), Saarland (roofed by German)

16. Divergence

287

and Luxembourg (with its threefold roof of German, French and Luxembourgesch [Letzebue¨rgesh]) (cf. Hoffmann 1990). Due to these quite different types of roofing, the dialects are increasingly drifting apart in the respective areas. Heterogenetic French as the only roof in East Lorraine exerts the strongest effects and causes the most far-reaching retreat of dialect use. In Saarland, the dialect does not retreat in the same way, as German is the roofing language, but the dialect is less stable than in Luxembourg where it became an ausbau language and thereby a third roof (Kloss 1967; as far as Luxembourg is concerned, cf. Kloss 1978, ch. 2.1.2). The different types of local roofing between Saarland and Eastern Lorraine have become so strong that the state border will eventually become the Germanic⫺Romanic language boundary (see Table 16.4). Tab. 16.4: Emerging differences of dialect use in the Saar-Lor-Lux region East Lorraine French as heterogenetic roofing language

Saarland

Luxembourg

German as homogenetic roofing language

Roofing languages: French (heterogenetic), German, Letzebue¨rgish (homogenetic)

Sociopragmatic divergence of dialect use strong retreat of dialect

moderate retreat of dialect

relative stability of dialect

Structural divergence by transference from … … French

… Standard German

… French

Result: state border as language boundary

In Luxembourg one can see how recent efforts of language planning have produced vertical divergence. The Letzebue¨rgesch standard language was created out of and installed above the West Moselle-Franconian dialects by koine´ization, following the concept of “willful and deliberate trilingualism” (Kramer 1986; Fröhlich 1996). Now this Letzebue¨rgesch standard language diverges vertically from the dialects that were its base (see Figure 16.2). This vertical, Luxembourg-internal, divergence results in horizontal German⫺Luxembourgish divergence. The regional varieties above the local dialects differ in status German as supranational and national language Luxembourginternal processes of vertical divergence

Letzebue¨rgesch elaborated to become a national roofing language A

German regiolect(s)

Luxembourg-wide koine´ Alzette dialect

Sauer dialect

Moselle dialect

Oesling dialect

West Moselle-Franconian dialects

…

…

… …

Fig. 16.2: Vertical Luxembourg-internal and horizontal German⫺Luxembourgish divergence

288

III. Structure and dynamics of a language space since Letzebue¨rgesch resulted from language-planning processes, whereas the German regiolects emerged without such planning and have no official status. In addition, a diglossic relationship between the Letzebue¨rgish ausbau language and the Luxembourgian base dialects has developed, whereas on the German side the West Moselle-Franconian regiolect (Umgangssprache) covers the intermediate space between standard and local base dialects on a diaglossic scale.

3.4.2. The contact zone o Germany, Luxembourg, the Netherlands and Belgium Another case in point is the contact zone between Germany, Luxembourg, Belgium and the Netherlands. Using lexical criteria, Cajot (1990) statistically proved a certain degree of divergence at the respective borders, both between states like Germany and Belgium and between administrative regions within a state, like the originally germanophone socalled “Old” and “New Belgium” (cf. Hinderdael and Nelde 1996). The continua on the dialect level are increasingly interrupted by horizontal convergences of dialects on all sides of the political border, above all, however, by the dialects converging vertically with the standard (or roofing) languages of the respective states. The Limburg dialect area is mainly stable on both sides of the Belgian⫺Dutch state border, because Belgian Flanders and the Netherlands are roofed by more or less the same language (Gerritsen 1999). In contrast, there are tendencies to dialectal divergence at the Dutch⫺German border. They result from a unilateral loss of dialects (Dialektabbau) on the German side.

3.5. Diverging processes within one rooing language On the level of standard languages, horizontal divergence can often be observed when states with a formerly homogeneous standard language fall apart. A linguistic disintegration of the standard which orients itself towards the recently emerged state borders is the result. In the beginning, this disintegration proves to be attitudinal only. Salient linguistic differences are still absent. Examples for this early phase are the attempts within early GDR language policy to establish two German state or even national varieties (cf. Ammon [1995, ch. 8] on the difference between the two) and to claim diverging sociolectal diastrata in the GDR and in the FRG (Lerchner 1974: 264). However, beyond the institutional, political and ideological nomenclature of the East German state (Staatsrat ‘council of state’, Arbeiterklasse ‘working class[es]’, Kollektiv ‘team’), not even an everyday vocabulary typical of the GDR developed (cf. Fleischer 1987), to say nothing of a GDR-typical grammar. At most, words such as Broiler ‘roast chicken’ vs. (Brat-)Hähnchen (Eichhoff 1977⫺2000, vol. 4, map 36), die Plaste, less frequent der Plast ‘plastic’ vs. Plastik (Eichhoff 1977⫺2000, vol. 2, map 77) and Zielstellung ‘target, objective’ vs. Zielsetzung (Ammon 1995: 390) could be given as examples. Only a few old regional expressions such as Eierkuchen ‘pancake’ vs. Pfannkuchen formed a wordgeographical area that became congruent with the political territory of the GDR. Words from the language of administration such as Fahrerlaubnis ‘driver’s licence’ (GDR) vs. Führerschein (FRG, Austria) vs. Fahr-/Führerausweis (German-speaking Switzerland) vs. Patent (South Tyrol) clearly demonstrate that areas of governmental and administrative

16. Divergence terminology are congruent with the respective governmental and administrative territories and thus likely to linguistically diverge from each other (Eichhoff 1977⫺2000, vol. 3, map 48). Another example of divergence on the standard language level is provided by efforts in the Slovak language policy to stress the relatively minor differences between Slovak and Czech. They already existed at the time of the Czechoslovakian confederation but were reinforced after Slovakia gained autonomy in 1993. Czech influences were eliminated in order to purify the language as a symbol of national sovereignty and identity (Lubas´ 2006: 1824). After the disintegration of the Yugoslavian state, efforts were made to split up the Serbo-Croatian language ⫺ itself only the product of convergence over a relatively long period ⫺ into separate Serbian and Croatian standard languages again and possibly to expand these endeavors to the creation of additional standard languages such as Bosnian or Montenegrin (Greenberg 2000; Woolhiser 2005: 244): first by renaming the separate languages, but later by language planning, too. In post-war Yugoslavia, Serbo-Croatian was not an official state language but a lingua franca consisting of variants of equal status. In the sovereign states that came into being after the disintegration of the Yugoslavian confederation the status of Serbo-Croatian changed. It was divided into two (and more) official state languages, and this separation was connected to the search for divergent linguistic features themselves. In order to increase the differences, features of the other variety were sorted out and the variety’s own archaic and dialectal words were revitalized (Kunzmann-Müller 2000; Neweklowsky 2006: 1827⫺1827, 1834). This amounts to a de facto dissimilation from below. Differences between the national standard varieties of German in Germany and Austria (as well as German-speaking Switzerland), are not limited to the terminology of political and other state institutions but reach the level of ordinary vocabulary. For example, the area in which the word Tischler is used for ‘carpenter’ is congruent with the state territory of Austria (including South Tyrol). It is sharply limited against South German Schreiner. In Alemannic-speaking Western Austria the variants Tischler and Schreiner co-exist, making for a transition towards Swiss German Schriener ⫽ Schreiner (Eichhoff 1977⫺2000, vol. 1, map 20). Food vocabulary (e.g., Austrian Karfiol vs. Federal German Blumenkohl ‘cauliflower’) also shows many divergences at the level of national varieties and has a high symbolic value for Austrian identity (Muhr, Schrodt and Wiesinger 1995; Muhr and Schrodt 1997). Attempts of language patriots to establish grammatical features as symbols of a national variety have failed, however (e.g., Austrian ich pendel-Ø ‘I commute’ instead of standard German ich pendl-e) (Ammon 1995: 387).

4. Parameters o linguistic divergence Based on the empirical analysis of concrete instances of divergence, generalizations can be made about divergence. The relevant parameters, binary or ternary for the most part, are italicized in the following. Horizontal divergence takes place on all diastratic levels: on the base dialectal level on the one hand, on the standard language level on the other. It also occurs between these extremes, in the regiolectal sector.

289

290

III. Structure and dynamics of a language space The processes of horizontal (i. e., diatopic, geographic) and of vertical (i. e., diastratic, social) divergence of linguistic varieties are mostly intertwined with processes of horizontal and vertical convergence. Thus, divergence and convergence cannot be strictly separated from one another, and the same holds for the horizontal and vertical orientation of these processes. This intertwining is clearly flagged in the title of the paper “Converging divergence and diverging convergence” by Trumper and Maddalon (1988). Horizontal divergence may emerge because a formerly uniform variety splits into parts that either converge horizontally with differing neighboring varieties or converge vertically with a higher variety to differing degrees. Horizontal convergence that causes divergence in a part of a formerly uniform dialect area can be, and often is, accompanied by vertical convergence, i. e., by advergence to the standard language. But sometimes horizontal convergences that cause divergence in a part of a formerly uniform dialect area imply vertical divergence from the standard language, because substandard features of a certain prestige are adapted (see below). Features of divergence can be structural in nature. Here, grammar in the broadest sense (phonology, morphology, syntax) and/or the lexicon (vocabulary, word formation) are concerned. They can be pragmatic in nature as well. This concerns the composition and the structure of the repertoire of varieties and the domains of their use. Finally, divergence can be attitudinal in nature. This includes the perceptual evaluation of the varieties by linguistic laypeople. These evaluations appear as mental linguistic maps when they apply to the spatial distribution of the varieties, and they appear as sociolinguistic stereotypes when they apply to the social value of the varieties. These attitudes influence the use of the varieties and can have effects on their structural shape. For instance, so called “primary” dialect features of phonology, morphology and syntax (Schirmunski 1930; Hinskens 1986) may be more or less consciously avoided and replaced by the speakers, due to the overt or covert prestige of these items. In contrast, there is also an unconscious change of phonological, morphological and syntactic items which is not influenced by attitudinal factors. The more consciously speakers change their language, the more they orient themselves to social or socio-geographic linguistic models such as super-regional or standard varieties or a neighboring dialect of high prestige. Sometimes the speakers also orient themselves to urban varieties, often over great geographic distances. But a change that causes divergence, especially an unconscious change, can also be autochthonous, i. e., without influences from prestigious neighboring or more distant varieties. In this case, the reasons are intra-linguistic (structural), not extra-linguistic (social). In some language spaces, the speakers’ repertoire has only two poles: standard and (base) dialect (strict diglossia). In other language spaces, the repertoires include intermediate levels (regional dialects, regional standards). Structurally, these intermediate levels are less rigidly organized than base dialects or standard varieties. As a result, they are suited to form a continuum between the polar ends of the repertoire (diaglossia). By definition, varieties in a diglossic model diverge to a higher degree than those in a diaglossic model. They can even become more divergent, for example when processes of standardization or (re-)dialectalization occur. On the other hand, very often the varieties in a repertoire converge. These processes of vertical convergence can cause horizontal divergence when they differ in strength in the different areas of a language space. Such differing degrees of vertical convergence, which make a formerly uniform language space diverge horizontally, cannot develop as easily under strict diglossia as under diaglossia.

16. Divergence Diglossia (in the post-Ferguson conception) exists within one and the same language (a repertoire of homogenetic varieties) as well as in a sociolinguistic situation where more than one language is involved (a repertoire of heterogenetic varieties). Diaglossia, however, can only exist within one and the same language. The languages “roofing” the substandard varieties (dialects, vernaculars) play a decisive role in the determination of types of linguistic divergence. Either the roofing language itself can be affected by divergence, for instance, when a uniform language space splits up, first politically and then linguistically or the roofed substandard varieties can be affected by divergence. In the latter case, divergence can either occur beneath the same (homogenetic) standard language or beneath a different (heterogenetic) standard language. Different but cognate roofing (standard) languages, i. e., languages belonging to one and the same language family such as Dutch and German, take an intermediate position. Intra-linguistic factors of structural, pragmatic and attitudinal divergence correlate to a greater or lesser degree with extra-linguistic factors. Extra-linguistic factors of divergence are so-called natural spaces and boundaries as well as political spaces and boundaries in the widest sense. In some instances, such political spaces and boundaries are declared to be natural ones in order to pursue a policy of (linguistic/political) expansion. Depending on their strength, extra-linguistic boundaries can more or less support or obstruct trade, contact and communication. The speakers’ knowledge of extra-linguistic spaces and boundaries, particularly political boundaries, plays an important role in speakers’ ideas about linguistic spaces and boundaries and in their linguistic behavior, which may even result in structural divergence.

5. Reerences Ammon, Ulrich 1995 Die deutsche Sprache in Deutschland, Österreich und der Schweiz. Das Problem der nationalen Varietäten. Berlin/New York: de Gruyter. Ammon, Ulrich 2005 Pluricentric and divided languages. In: Ammon et al. (eds.), 1536⫺1543. Ammon, Ulrich, Norbert Dittmar, Klaus J. Mattheier and Peter Trudgill (eds.) 2004 Sociolinguistics: An International Handbook of the Science of Language and Society, vol. 1. 2nd ed. (⫽ Handbücher zur Sprach- und Kommunikationswissenschaft 3.1) Berlin/New York: de Gruyter. Ammon, Ulrich, Norbert Dittmar, Klaus J. Mattheier and Peter Trudgill (eds.) 2005 Sociolinguistics: An International Handbook of the Science of Language and Society, vol. 2. 2nd ed. (⫽ Handbücher zur Sprach- und Kommunikationswissenschaft 3.2) Berlin/New York: de Gruyter. Ammon, Ulrich, Norbert Dittmar, Klaus J. Mattheier and Peter Trudgill (eds.) 2006 Sociolinguistics: An International Handbook of the Science of Language and Society, vol. 3. 2nd ed. (⫽ Handbücher zur Sprach- und Kommunikationswissenschaft 3.3) Berlin/New York: de Gruyter. Auer, Peter (ed.) 1998 Dialect levelling and the standard varieties in Europe. Folia Linguistica 32 (special issue). Auer, Peter 2004 Sprache; Grenze, Raum. Zeitschrift für Sprachwissenschaft 23: 149⫺179. Auer, Peter, Frans Hinskens and Paul Kerswill (eds.) 2005 Dialect Change: Convergence and Divergence in European Languages. Cambridge: Cambridge University Press.

291

292

III. Structure and dynamics of a language space Auer, Peter, Frans Hinskens and Klaus J. Mattheier (eds.) 1996 Convergence and divergence of dialects in Europe. Sociolinguistica 10. Banzer, Roman 1990 Pragmatik und Interferenzen der Mundarten des Fürstentums Liechtenstein, des St. Galler Rheintals und Vorarlbergs. In: Kremer and Niebaum (eds.), 341⫺359. Bellmann, Günter 1998 Between base dialect and standard language. In: Auer (ed.): 23⫺34. Cajot, Jose´ 1990 Neue Sprachgrenzbildung an der deutschen Staatsgrenze zu Niederländisch-Ostlimburg, Ostbelgien und Luxemburg. In: Kremer and Niebaum (eds.), 125⫺152. Clyne, Michael (ed.) 1991 Pluricentric Languages: Differing Norms in Different Nations. Berlin/New York: Mouton de Gruyter. Clyne, Michael 2004 Pluricentric language. In: Ammon et al. (eds.), 296⫺300. Dyhr, Mogens 1990 Hybridisiertes Südjütisch. In: Kremer and Niebaum (eds.), 25⫺47. Eichhoff, Jürgen 1977⫺2000 Wortatlas der deutschen Umgangssprachen. 4 vols. Bern/München: Francke/Saur (from vol. 3). Fleischer, Wolfgang 1987 Wortschatz der deutschen Sprache in der DDR. Fragen seines Aufbaus und seiner Verwendungsweise. Leipzig: Bibliographisches Institut. Fröhlich, Harald 1996 Luxemburg. In: Hinderling and Eichinger (eds.), 459⫺478. Gerritsen, Marinel 1999 Divergence of dialects in a linguistic laboratory near the Belgian⫺Dutch⫺German border: Similar dialects under the influence of different standard languages. Language Variation and Change 11: 43⫺66. Giesbers, Charlotte 2008 Dialecten op de grens van twee talen. Een dialectologisch en sociolinguistisch onderzoek in het Kleverlands dialectgebied. Unpublished PhD dissertation, University of Nijmegen. Greenberg, Robert D. 2000 Language politics in the Federal Republic of Yugoslavia: The crisis over the future of Serbian. Slavic Review 59: 625⫺640. Harnisch, Rüdiger 1996 Das Elsass. In: Hinderling and Eichinger (eds.), 413⫺457. Harnisch, Rüdiger 2004a Junger Dialektwandel in der Bavaria Thuringica. In: Maik Lehmberg (ed.), Sprache, Sprechen, Sprichwörter. Festschrift für Dieter Stellmacher zum 65. Geburtstag, 453⫺463. Stuttgart: Steiner. Harnisch, Rüdiger 2004b [Soziale Implikationen sprachlicher Konstruktionsebenen:] Morphologie. In: Ammon et al. (eds.), 522⫺530. Harnisch, Rüdiger, Frank Reinhold and Michael Schnabel 2008 Neue Dialektgrenzen an der ehemaligen deutsch-deutschen Grenze? In: Peter Ernst and Franz Patocka (eds.), Dialektgeographie der Zukunft. Akten des 2. Kongresses der Internationalen Gesellschaft für Dialektologie des Deutschen (IGDD), 203⫺218. Stuttgart: Steiner. Henzen, Walter 1954 Schriftsprache und Mundarten. Ein Überblick über ihr Verhältnis und ihre Zwischenstufen im Deutschen. 2nd ed. Bern: Francke.

16. Divergence Hinderdael, Michael and Peter Nelde 1996 Deutschbelgien. In: Hinderling and Eichinger (eds.), 479⫺495. Hinderling, Robert and Ludwig M. Eichinger (eds.) 1996 Handbuch der mitteleuropäischen Sprachminderheiten. Tübingen: Narr. Hinskens, Frans 1986 Primaire en secundaire dialectkenmerken. Een onderzoek naar de bruikbaarheed van een vergeten(?) onderscheid. In: Jos Creten, Guido Geerts and Koen Jaspaert (eds.), Werkin-uitvoering. Momentopname van de sociolinguı¨stiek in Belgie¨ en Nederland, 135⫺158. Leuven/Amersfoort: Acco. Hoffmann, Jean-Paul 1990 Sind Staatsgrenzen auch Mundartgrenzen? Zum Verhältnis von Standardsprache und Dialekt im saarländisch-lothringisch-luxemburgischen Grenzgebiet. In: Kremer and Niebaum (eds.), 153⫺191. Kallen, Jeffrey L., Frans Hinskens and Johan Taeldeman (eds.) 2000 Dialect convergence and divergence across European borders. International Journal of the Sociology of Language 145 (special issue). Klausmann, Hubert 1990 Staatsgrenze als Sprachgrenze? Zur Entstehung einer neuen Wort- und Sprachgebrauchsgrenze am Oberrhein. In: Kremer and Niebaum (eds.), 193⫺215. Klausmann, Hubert 2000 Changes of dialect, code-switching, and new kinds of usage: The divergence of dialects along the border between Germany and France in and around the region of the Oberrhein. In: Kallen, Hinskens and Taeldeman (eds.), 109⫺130. Kloss, Heinz 1967 Abstand languages and ausbau languages. Anthropological Linguistics 9(7): 29⫺41. Kloss, Heinz 1978 Die Entwicklung neuer germanischer Kultursprachen seit 1800. 2nd ed. Düsseldorf: Schwann. Kramer, Johannes 1986 Gewollte Dreisprachigkeit ⫺ Französisch, Deutsch und Le¨tzebuergesch im Großherzogtum Luxemburg. In: Robert Hinderling (ed.), Europäische Sprachminderheiten im Vergleich. Deutsch und andere Sprachen, 229⫺249. Stuttgart: Steiner. Kremer, Ludger 1990 Kontinuum oder Bruchstelle? Zur Entwicklung der Grenzdialekte zwischen Niederrhein und Vechtegebiet. In: Kremer and Niebaum (eds.), 85⫺123. Kremer, Ludger and Hermann Niebaum (eds.) 1990 Grenzdialekte. Studien zur Entwicklung kontinentalwestgermanischer Dialektkontinua. Germanistische Linguistik 101⫺103. Kunzmann-Müller, Barbara 2000 Sprachliche Wende und Sprachwandel im Kroatischen/Serbischen. In: Barbara Kunzmann-Müller (ed.), Die Sprachen Südosteuropas heute. Umbrüche und Aufbruch, 42⫺65. Frankfurt: Lang. Lerchner, Gotthard 1974 Zur Spezifik der Gebrauchsweise der deutschen Sprache in der DDR und ihrer gesellschaftlichen Determination. Deutsch als Fremdsprache 11: 259⫺265. Lubas´, Wladyslaw 2006 Die westslawischen Sprachen. In: Ammon et al. (eds.), 1837⫺1851. Mall, Josef and Waltraud Plagg 1990 Versteht der Nordtiroler die Südtirolerin noch? Anmerkungen zum Sprachwandel in der deutschen Alltagssprache Südtirols durch den Einfluß des Italienischen. In: Kremer and Niebaum (eds.), 217⫺239.

293

294

III. Structure and dynamics of a language space Mattheier, Klaus J. (ed.) 2000 Dialect and Migration in a Changing Europe. Frankfurt: Lang. Muhr, Rudolf and Reinhard Schrodt (eds.) 1997 Österreichisches Deutsch und andere nationale Varietäten plurizentrischer Sprachen in Europa. Wien: Höder⫺Pichler⫺Tempsky. Muhr, Rudolf, Reinhard Schrodt and Peter Wiesinger (eds.) 1995 Österreichisches Deutsch. Linguistische, sozialpsychologische und sprachpolitische Aspekte einer nationalen Variante des Deutschen. Wien: Hölder-Pichler-Tempsky. Neweklowsky, Gerhard 2006 Die südslawische Region. In: Ammon et al. (eds.), 1824⫺1836. Niebaum, Hermann 1990 Staatsgrenze als Bruchstelle? Die Grenzdialekte zwischen Dollart und Vechtegebiet. In: Kremer and Niebaum (eds.), 49⫺83. Petzold, Rainer 1993 Gespaltener Dialekt? Untersuchungen zur Sprachsituation im Grenzsaum beiderseits der Grenze zwischen Thüringen und Bayern (früher zwischen BRD und DDR). In: Wolfgang Lösch (ed.), Beiträge zur Dialektforschung in Thüringen 1993, 45⫺52. Jena: Bussert und Stadeler. Radtke, Edgar 2006 Konvergenz und Divergenz regionaler Varietäten. In: Ammon et al. (eds.), 2189⫺2196. Reinhold, Frank 1997 Schwindendes Zungenspitzen-R [sic!] als Merkmal der Sprachsituation an der ehemaligen deutsch-deutschen Grenze in Blankenberg/Saale. In: Wolfgang Lösch (ed.), Beiträge zur Dialektforschung in Thüringen 1997, 139⫺143. Jena: Bussert und Stadeler. Ryckeboer, Hugo 1990 Jenseits der belgisch-französischen Grenze: Der Überrest des westlichsten Kontinentalgermanischen. In: Kremer and Niebaum (eds.), 241⫺271. Ryckeboer, Hugo 2000 The role of political borders in the millennial retreat of Dutch (Flemish) in the north of France. In: Kallen, Hinskens and Taeldeman (eds.), 79⫺108. Scheuringer, Hermann 1990 Bayerisches Bairisch und österreichisches Bairisch. Die deutsch-österreichische Staatsgrenze als “Sprachgrenze”? In: Kremer and Niebaum (eds.), 361⫺381. Schifferle, Hans-Peter 1990 Badisches und schweizerisches Alemannisch am Hochrhein. In: Kremer and Niebaum (eds.), 315⫺340. Schirmunski, Viktor M. 1930 Sprachgeschichte und Siedelungsmundarten. Germanisch-Romanische Monatsschrift 18: 113⫺122 and 171⫺188. Schlobinski, Peter 1997 The Sociolinguistics of Berlin Urban Vernacular. Berlin: Mouton de Gruyter. Schnabel, Michael 2006 Dialektspaltung im thüringisch-bayerischen Grenzgebiet am Beispiel des Ortspaars Sparnberg/Rudolphstein. Wie eine politische Grenze zur Sprachgrenze wurde. Zeitschrift für Dialektologie und Linguistik 73: 30⫺54. Schunk, Gunther 1999 Regionalisierung von Dialekten. Ein lautlicher Stadt-Land-Vergleich in Mainfranken. (Schriften zum Bayerischen Sprachatlas 1.) Heidelberg: Winter. [PhD dissertation, Würzburg 1998.] Taeldeman, Johan 1990 Ist die belgisch-niederländische Staatsgrenze auch eine Dialektgrenze? In: Kremer and Niebaum (eds.), 275⫺314.

17. Emergence of varieties

295

Trumper, John and Marta Maddalon 1988 Converging divergence and diverging convergence: The dialect⫺language conflict and contrasting evolutionary trends in modern Italy. In: Peter Auer and Aldo di Luzio (eds.), Variation and Convergence. Studies in Social Dialectology, 217⫺259. Berlin/New York: de Gruyter. Woolhiser, Curt 2005 Political borders and dialect divergence/convergence in Europe. In: Auer, Hinskens and Kerswill (eds.), 236⫺262.

Rüdiger Harnisch, Passau (Germany)

17. Emergence o varieties through restructuring and reevaluation 1. 2. 3. 4. 5. 6. 7.

Introduction The emergence of varieties through reevaluation The emergence of standard varieties through restructuring und reevaluation The emergence of dialects through restructuring und reevaluation The emergence of regiolects through restructing and reevaluation The reevaluation and restructuring cycle References

1. Introduction The history of the cross-linguistically varied dialect/standard constellations we encounter in contemporary European languages is dynamic on (at least) two different but closely connected levels. Simply put, these levels can be identified as “objective” versus “subjective”. Whereas linguistic variation and change can be considered objective processes, attitudinal dynamics play out on a subjective level. Objective and subjective dynamics can not only give rise to structural and attitudinal changes in existent varieties, they can also lead to the emergence ⫺ the formation and entrenchment ⫺ of new varieties or types of varieties. In line with the focus of this handbook, this contribution concentrates on the emergence of the most important varieties within dialect-standard constellations, namely on (new) standard varieties, (new) regiolects and (new) dialects. The analysis of the emergence of varieties presupposes the formulation of a distinction between the emergence of a new variety and processes of linguistic change within an existing one. As the boundaries between these types of processes are fluid (cf. Kerswill and Trudgill 2005), they are not easily distinguishable. From the abundance of hotly debated concepts of variety within variational linguistics (cf. Berruto 2004 and in this volume; Lenz and Mattheier 2005) it seems best to select a narrow concept in the quest for clarification. Here, variety is defined as a subsystem of a language characterized by

296

III. Structure and dynamics of a language space internal linguistic cohesion, clear system boundaries (separating it from other varieties), well-defined pragmatic functions and an emic status (cf. Lenz 2003, 2004b, 2005a). Building on this definition of variety, dialect can in turn be defined as the “lowest” (most linguistically divergent from the standard) and the most spatially confined variety in the dialect⫺standard constellation. Like a variety in general, a dialect is not a homogeneous construct, but a variable subsystem with a varying degree of internal substructure, which can exist alongside other dialects or “beneath” other diatopic varieties with a greater areal distribution. Divergent attitudinal, functional and structural characteristics of dialects on the one hand, and coexistent “middle” varieties on the other, motivate a terminological distinction between these two types of varieties, frequently designated dialect and regiolect in the research literature (cf. also Schmidt, this volume). Among the European languages, the most disparate dialect⫺standard constellations can be found, especially with regard to the number of varieties and their linguistic and sociolinguistic structure. The extreme poles of standard and dialect are the constitutive elements of these dialect⫺standard variety constellations, whereby the standard-language pole is historically younger. According to Auer (2005: 8), a standard variety is a “variety of a language […], which is characterized by the following three features: (a) it is orientated to by speakers of more than one vernacular variety (which does not necessarily imply that it is mastered by everybody), (b) is looked upon as an H-variety and used for writing, and (c) it is subject to at least some codification and elaboration”. In the prevalent triangle or pyramid model (cf. Chambers and Trudgill 1980: 10⫺11), the status of the (“highest”) high variety, the standard, is symbolized by its placement at the apex, canopying the other varieties (Überdachung). The horizontal level of this model stands for the areal/diatopic dimension of language; the vertical level represents the diastratic dimension. The analysis of the overall linguistic and attitudinal dynamics of a complex dialect⫺ standard constellation presupposes a fundamental knowledge of the dynamic structure of the individual varieties and their complex relations. Research desiderata here especially concern the vertical axis between dialect and standard and the “intermediate varieties” (regiolects) which have to be positioned on this axis. Language variation research to date can be seen to have focused mainly on variation and change at the extremes, whereas the “mittlere Bereich” (‘middle zone’; Bellmann 1983) has increasingly attracted interest in the past few decades. Extensive research literature can be found concerning the emergence of standard varieties and the consequences for their dialects. The fact that less research has been conducted on the emergence of “intermediate/middle varieties” can surely be attributed to the fact that regiolects often emerge out of a dialect/standard “diglossia” (Ferguson 1959) and are thus “younger” than the poles. In addition, the analysis of emerging varieties has generally concentrated on objective/linguistic dynamics, with only secondary attention devoted to attitudinal aspects (including processes of reevaluation).

2. The emergence o varieties through reevaluation A look into the research literature reveals that in the past the emergence of varieties has mainly been described on the objective level. The hypothesis that attitudinal processes also make a central contribution to the emergence of varieties has been hitherto more or

17. Emergence of varieties less neglected. Attitudes are commonly conceptualized as beliefs and valuations that ⫺ according to the mentalistic concept of Allport (1954) ⫺ are compounds of cognitive, evaluative and conative elements (cf. Rosenberg and Hovland 1960; Hewstone, Manstead and Stroebe 1997). They are not innate, predefined and invariable constants, but dynamic and processual features which emerge, further develop and vary in socially interactive processes from sedimentations of individual and transferred experience (cf. Deprez and Persoons 1987). Variety reevaluation processes, and the restructuring which they imply, are subject to the general dynamics of attitudes. This includes all forms of attitudinal dynamics, especially the ebb and flow of prestige. Linguistic and attitudinal dynamics form a complex relational network from which new varieties can emerge, which in turn can initiate and support new linguistic and attitudinal processes. Within the history of European dialect⫺standard constellations, three primary patterns for the emergence of varieties through reevaluation can be reconstructed: Type A: Reevaluation of a vernacular variety as a high variety. Type B: Reevaluation of a vernacular variety as a low variety. Type C: Reevaluation of a high variety as a low variety. The first type of reevaluation process (Type A) is found in the case of a monocentric selection and implementation of a standard variety which emerges from the growing prestige of an existing variety, for example. This variety increasingly becomes an orientation norm for other diatopic varieties among which it originally existed as more or less an equal. The reevaluation of a vernacular variety as a high variety is sketched in Figure 17.1. This depiction, like Figures 17.2, 17.3 and 17.7, draws on Auer 2005, but here the aspect of (re)evaluation is emphasized. A more detailed discussion of this reevaluation type follows in section 3.

Fig. 17.1: Reevaluation of a vernacular variety as a high variety (Type A: Emergence of a standard variety via monocentric selection)

The concomitant reevaluation of the other vernacular varieties as dialects represents an example of the second type of reevaluation process (Type B, see Figure 17.2). It is only through the existence of their standard counterpart that dialects become a relational and constitutive element of the dialect⫺standard constellation (see section 4). Similarly, regiolects ⫺ which can be located on the vertical axis as “middle” varieties, “above” the dialects but “below” an overarching standard variety ⫺ presuppose the

297

298

III. Structure and dynamics of a language space

Fig. 17.2: Reevaluation of vernacular varieties as low varieties (Type B: Emergence of dialects, in the sense of relational diasystems)

existence of both relational poles. In the research literature, the emergence of regiolects is mostly explained as a complex interplay of vertical and horizontal convergence processes (cf. Bellmann 1983; Siebenhaar, this volume; Røyneland, this volume). Alongside convergence processes which effect structural changes on the dialect⫺standard axis, reevaluation processes have to be taken into consideration in explaining the emergence of regiolects. As can be illustrated by the test case of German (see section 5.3), intermediate varieties may be identified as “old” high varieties, which are devalued as a result of the “superimposition” of a new high variety and take on a modified position and function within the variety spectrum (Type C, see Figure 17.3).

Fig. 17.3: Reevaluation of an erstwhile high variety as a low variety (Type C: Emergence of regiolects by “superimposition”)

In general, the evaluation and reevaluation of varieties presents a complex and highly challenging object of research (cf. Auer, Hinskens and Kerswill 2005: 38⫺39). Evidence for a variety’s changing prestige can be provided by so-called “hyperforms”, which are (primarily) the result of a speaker’s inadequate approximation of an intended target variety and which can in certain cases even attain system-internal status (see Herrgen 1986; Pargman 1998). In linguistic terms, hyperforms are “motivated” by a partial contrast between varieties, i. e., where no one-to-one correspondence between the varieties in contact can be formulated. The generation of hyperforms can be explained as false

17. Emergence of varieties analogy, namely the overgeneralization of rules of correspondence (“Korrespondenzregeln”, Auer 1993: 8⫺9) between these varieties. According to Auer (1990: 274), rules of correspondence have the cognitive function of connecting the phonemes and morphemes of one variety with those of another. In the case of hyperdialectalisms, the low variety is the intended but only partially mastered variety, whereas in hypercorrections, a high variety is the target variety (cf. Lenz 2004b, 2005a). Since hyperforms are sociolinguistically motivated by the intention to reach/approximate a particular variety, they can tell us something about the prestige of varieties, about change in prestige and not least about cognitive processes. For example, not only do hyperforms provide evidence for an increasing prestige of high varieties (especially standard varieties), they can also be interpreted as evidence of a speaker’s minimal, emergent competence in the intended prestige variety (cf. Besch 2000: 188; see also section 5.3).

3. The emergence o standard varieties through restructuring and reevaluation Standardization is primarily a sociopolitical phenomenon. Its starting point is comparable across the European languages (for the Germanic languages see Deumert and Vandenbussche 2003; for standardization in general see Haugen 1966: 1987). At the beginning of a standardization process, different autochthonous local and highly regional varieties, none of which has the status of a norm, coexist. Horizontal processes of leveling between these vernacular varieties occur only marginally. During the first phase of standardization, the selection phase, one variety increasingly assumes the role of a supraregional norm. In the case of polycentric selection, this is a new, previously nonexistent and more-or-less constructed lead variety (Leitvarietät). Examples for polycentric selection, which “seems to be rather more common in language history” (Deumert and Vandenbussche 2003: 4), include the standard varieties of German and Norwegian (see Mattheier 2003 and Jahr 2003, respectively). Whereas the selection of the New High German standard variety had already begun in the fifteenth century, the philologically supported, planned construction of the Norwegian standard variety only has a relatively short history which does not predate Norwegian independence from Denmark in 1814. From the outset, two different paths were chosen in the search for a Norwegian standard, resulting in two official standard varieties which still coexist (cf. Røyneland 2009; Jahr 2003) In contrast to polycentric selection, in which a previously nonexistent variety is also “materially” created, the monocentric selection of a standard variety is primarily an act of reevaluation, in which a variety is extricated from its old horizontal variety constellations and imposed “above” the other varieties (Type A reevaluation, see Figure 17.1). An example of monocentric selection is provided by the standardization history of Danish (see Pedersen 2005; Kristiansen 2003), characterized by the increasing supraregional prestige and orientation norm status of the dialects of Zealand/Copenhagen. The onset of the monocentric selection process is marked by a reevaluation in which the selected prestige variety’s [⫹diatopic] dimension increasingly becomes a background feature while its [⫹diastratic] and [⫹diaphasic] features gain emphasis (for the dimensions see Coseriu 1988: 24). The validity of the new lead variety initially expands within the

299

300

III. Structure and dynamics of a language space written domain where it gradually supersedes the former allochthonic high varieties (exoglossic standards). Through the emergence of an endogenous standard variety the previously monoglossic language community becomes increasingly diglossic (cf. Auer 2005: 12). In the standardization process, the selection phase is followed by a codification phase in which the new prestige variety is structurally elaborated via the codification of linguistic and communicative norms. Official institutions support this codification. The gradual spread and implementation of the codified norm across the diatopic, diastratic, diaphasic and medial (from oral to written) domains takes place in the implementation phase, which Deumert and Vandenbussche (2003: 7) call the “Achilles heel” of the standardization process. For instance, it took until the second half of the eighteenth century for the East Middle German written variety of Meissen to diatopically supersede the competing common German (Gemeine Deutsch) with Upper German flavor (cf. Mattheier 1997, 2003). In Norway, two competing standard varieties are still confronting each other and the medial spread to the oral domain is only in its initial phase (cf. Røyneland 2009). Standard varieties are subject to attitudinal dynamics, which produce fluctuations in prestige. Their “value” ⫺ together with the evaluation of varieties in general ⫺ is not separable from the evaluations of the speakers prototypically associated with them. For example, from the point of view of Middle and Upper German speakers, spoken “proper” Standard German was and is connected with the German of the North Germans whilst their own, regionally colored Hochdeutsch is contrasted with the Northern German prototype (cf. Lenz 2003: 313⫺327; König 1997: 250). The belief that oral standard competence is limited is still widespread in the German language area (cf. Huesmann 1998: 105). Counter-developments to standardization have been observed in a whole range of standard languages, especially in the latter half of the twentieth century. These developments involve highly disparate linguistic and attitudinal processes often lumped together under the heading of destandardization (see Auer 1997; Mattheier and Radtke (eds.) 1997; Deumert and Vandenbussche 2003; Danesˇ 2006; for Dutch see Willemyns 1997, 2003; Stroop 1998; for Italian see Scholz 1997; for German see Mattheier 1997, 2003; Spiekermann 2005; for Danish see Kristiansen 2003; Pedersen 2005; for English see Nevalainen 2003; for Swedish see Teleman 2003; Sandøy 2002; for Norwegian see Røyneland, in press; for Polish see Mazur 1997). For example, Danesˇ (1968/1982) observes a diastratic narrowing of the Czech standard variety which involves decreasing competence in and prestige of the standard among the younger generation. Stroop (1992: 162) suggests that destandardization tendencies are causally connected with dialect loss in the Dutch situation (see also the discussion about “Poldernederlands” and “Schoon Vlaams”, Stroop 1997; Goossens 2000). Extralingual factors such as better education, greater social mobility and growing norm skepticism are also thought to contribute to destandardization.

4. The emergence o dialects through restructuring and reevaluation As mentioned, areal linguistic research has focused primarily on processes at the objective/linguistic level. “New dialect formation”, narrowly defined, “refers to the emergence of distinctive, new language varieties following the migration of people speaking mutu-

17. Emergence of varieties ally intelligible dialects to what, to all intents and purposes, is linguistically territory” (Kerswill and Trudgill 2005: 196). In this narrow conception, the emergence of a new dialect is interpreted as the result of an extreme, often very rapid process of convergence. The foremost scenarios for the emergence of a dialect are new territorial settlements, the founding of towns or language and dialect islands (see Kerswill and Williams 2000; Trudgill 2006). The emergence of a new dialect can be divided into partially overlapping subprocesses (cf. Kerswill and Trudgill 2005; cf. also Trudgill 1986, 1998; Trudgill et al. 2000). At the beginning one finds a coexistence of variants (“mixing”), which can still be attributed to dialect varieties in contact. Certain variants are then selected and incorporated in the emerging dialect variety (on leveling and koineization see also Hinskens 1996; Gilles 1999). The factors which support the selection of a particular variant (or set of variants) can vary. Sometimes subjective/attitudinal considerations are involved, of which the factor “salience” has received special attention within the research literature in recent decades (cf. Auer, Barden and Großkopf 1996; Auer 2001; Kerswill and Williams 2000; Lenz 2003: 21⫺21, 194⫺203; Auer, Hinskens and Trudgill 2005: 43⫺45). It is not uncommon for “old” variants to survive in the new dialect variety after receiving an attitudinal boost (see reallocation, Trudgill 1986: 110). Aside from attitudinal aspects, linguistic factors must of course also be considered as forces in the leveling process. Among these, Kerswill and Trudgill emphasize the factor of “simplification” which “means a decrease in irregularity in morphology and an increase in invariable word forms, as well as the loss of categories such as gender, the loss of case marking, simplified morphophonemics (paradigmatic leveling), and a decrease in the number of phonemes” (Kerswill and Trudgill 2005: 198). Alongside “old” variants which can be attributed to the original source dialects, a “new” dialect (in the sense defined above) also features new variants not found in any of the sources. The final subprocess of “new dialect formation” is that of “focusing”, which contributes to the sociolinguistic stabilization of the newly emerged dialect (cf. Le Page 1980). This stabilization is manifest in the furnishing of the new variety with (at least subsistent) norms, additional emotional value (e.g., “covert prestige”, Trudgill 1983: 177) and explicit communicative and sociopragmatic functions. A definition of dialect as the “deepest” pole of the dialect⫺standard spectrum presupposes the existence of such a spectrum. In terms of the emergence of dialects (understood as relational diasystems), this means that this type of diasystem can only occur when a dialect⫺standard axis exists. The emergence and generalization of standard varieties (see section 3) has attitudinal, not just linguistic consequences for the entire constellation which they crown. These consequences are not confined to current destandardization tendencies, but are also evident at the outset of the standardization process, when the selection phase begins. The emergence of standard varieties is accompanied by massive reevaluation and reorganization processes in the previously nonsubordinate diatopic varieties. Herrgen (2001: 1515) sketches the effects which the emergence of standard varieties have on the reevaluation and the initial vertical grading of dialects as diasystems (see Figure 17.2) for the German language area: With the step-by-step establishment of a supraregional compromise variety (New High German Standard), but especially through the formation of this variety’s own pronunciation norm, the dialects first gained the counterpole which allowed them to command awareness as systematically distinct and areally restricted. The New High German standardization proc-

301

302

III. Structure and dynamics of a language space ess was not simply the historical precondition for the perception of the dialects as German areal varieties, it was simultaneously the trigger for the displacement of the spectrum of the dialectal domain.

5.

The emergence o regiolects through structuring and reevaluation

5.1. Terminological aspects Not only did the emergence of standard varieties have far-reaching consequences for the erstwhile vernacular varieties (lantsprachen), it was (and is) the precondition for the formation of regiolects, the latter being standard-divergent varieties with broader regional distribution that can be located between dialects and standard varieties. Just as the perception of dialects as systemically different and regionally limited varieties presupposes the emergence of a standard counterpart, regiolects too are first perceptible as “intermediate” varieties in the presence of standard varieties. Whereas the terms dialect and standard have reached some degree of international validity, regiolects are subject to varied terminologies, even within the same philological tradition, e.g., (regionale) Umgangssprache (Dittmar 2004), regiolect/Regiolekt (Auer, Hinskens and Kerswill 2005; Lenz 2005a), Alltagssprache (Elspaß 2007), Neuer Substandard (Bellmann 1983), regional varieties (Hughes, Trudgill and Watt 2005), intermediate variety (Coetsem 1992), urban vernacular (Milroy 1982), italiano regionale (Berutto, in this volume). This terminological effusion can be interpreted as a reflection of the often discussed difficulty in separating intermediate varieties from dialects or the standard variety (on this discussion, see Gilles 1997: 164). The knotty question of where standard varieties or dialects end and where middle varieties start is often avoided through the adoption of a continuum model: “[m]ore usually, the space between base dialect and standard is characterized by nondiscrete structures (standard/dialect continuum)” (Auer 2005: 22, see also Figure 17.5). But, as a rule, the continuum posited by linguists stands in stark contrast to speakers’ very clear ideas about a structured variety spectrum (cf. Mattheier 1990a: 8). Current results from the German language area, based on linguistic and sociolinguistic data, provide evidence for the thesis that there is no necessary contradiction between a continuum and a variety model (cf. Lenz 2003). For example, in large parts of the West Middle

Fig. 17.4: Variety structure of the West Middle German nonstandard (cf. Lenz 2004b, 2006: 2008)

17. Emergence of varieties German area, we find evidence for a continuum from the local base dialects up to the Regionalakzent ‘regional standard’, which can however be divided into two varieties, dialect versus regiolect (see Figure 17.4). Neither variety is a homogeneous entity; both feature internal variability. With the aid of various linguistic variationist methods (e.g., cluster analysis, implicational scaling, hearer evaluations), Sprechlagen ‘speech levels’ (also termed Verdichtungsbereiche ‘concentration zones’) characterized by relative internal cohesion can be identified within both the dialect and the regiolect. In other words, the variation within a concentration zone is smaller than the variation between two concentration zones. Further, for speakers/hearers, the speech levels within a variety are not separated by categorical boundaries, whereas a clear and conscious boundary is held to divide dialect (known as Platt) from regiolect (Umgangssprache). Regiolects (as defined above) assume different positions and functions which can vary from case to case (cf. Scholz 1997 for Italian; Lenz 2003 and Figure 17.4 for German). They can function as the highest individual variety of a “bivarietal” speaker who also has an active dialect competence; secondly they may represent the lowest pole of an individual’s variety spectrum where the highest variety is the (interference-free) oral standard; thirdly, the regiolect may be the only variety within an individual speaker’s competence. In that case, we deal with a “monovarietal” speaker who “shifts” (cf. Auer 1986: 1990) between a “lower” and a “higher” concentration zone within the regiolect. It is possible that there are also “trivarietal” speakers, who have active mastery of the standard, the regiolect and a dialect, but as yet there is no evidence of them.

5.2. Regiolects as a result o horizontal/vertical convergence There is no consensus on exactly how a diaglossic spectrum with (more or less continuous) intermediate forms evolves from a diglossic situation, in which a clearly discrete standard variety and dialects structurally and functionally coexist (see Figure 17.5). But such a diaglossic repertoire is found in large parts of the European language area, in those areas of Norway in which Bokma˚l is the standard variety (particularly in the southeastern part), in some parts of the Netherlands excluding the Randstat area, in most of Flemish-speaking Belgium, in many southern parts of Germany (Alemannic and Eastern Franconian, as well as the Rheno-Franconian and Moselle-Franconian area), in the larger Vienna area in Austria, in Scotland, in Moravia […], in Poland with the exception of the previously German parts, in which the loss of the dialects has further progressed […], in the south of Spain […], in the Catalan language area, in Cyprus […], in Bulgaria […], in Sweden, in Finland, in most parts of England and presumably in areas of Italy, etc. (Auer 2005: 24⫺25)

As the arrows in Figure 17.5 suggest, regiolects are usually interpreted as the result of a complex dynamics both at the extremes of the dialect⫺standard spectrum and within the intermediate zone. Among these processes are horizontal convergence between dialects and vertical convergence between dialects and standard varieties (see Siebenhaar, this volume; Røyneland, this volume; Auer, Hinskens and Kerswill 2005). But it is impor-

303

304

III. Structure and dynamics of a language space

Fig. 17.5: From diglossia to diaglossia repertoires (Auer 2005: 23, Figure 5)

tant to recognize that it is sometimes very difficult to distinguish between vertical and horizontal dynamics (cf. Auer 1997: 131). On a linguistic level, Dialektabbau ‘structural dialect loss’ can often be observed as a trigger of or concomitant with the emergence of regiolectal varieties, beginning with variation and possibly leading to change in the dialectal system, i. e., to a reduction in local dialectal elements and their transformation into (regionally) more widely distributed varieties, usually nearer to the standard. The linguistic convergence processes are often accompanied by a sociolinguistic Dialektaufgabe ‘functional dialect loss’ which entails an increasingly curtailed use of dialect or special dialect features (cf. Mattheier 2004). The two types of process, linguistic and sociolinguistic, are mutually conditioning. If the use of a dialect variety is reduced, its linguistic structure is also affected (cf. Hinskens 1996: 11). The filling of the functional gap created by dialect loss is sometimes looked upon as a decisive factor driving the emergence and success of mid-range varieties (cf. Stroop 1992: 162). Alongside convergence within the dialectal range, processes active at the other end of the dialect⫺standard axis are also cited as invoking or supporting the emergence of regiolects. These processes involve attitudinal/functional and structural changes in the standard variety, mostly subsumed under the label of destandardization (see above). Together, horizontal interdialectal convergence and vertical convergence on the dialect⫺ standard axis lead to a reduction in the gap between the previously distinct and diglossically distributed extremes (cf. Scholz 1997 for Italian; Bellmann 1983 for German).

5.3. Regiolects as a result o superimposition processes  German as a test case German can serve as a test case for the thesis that, processes of horizontal/vertical convergence aside, diatopic exchange processes ⫺ more precisely, processes of the superimposition of external prestige varieties ⫺ also have to be taken into consideration as explanations for the emergence of regiolects. Mihm (2000) argues that the emergence of the modern German regiolects (Umgangssprachen) is due to (at least) two superimposition processes. The first superimposition phase (starting in the sixteenth century) was due to the increasing prestige of High German, especially East Middle German, varieties. In this phase, High German high varieties with varying Low German substrates spread

17. Emergence of varieties out across the Low German area, first in written, then in oral domains. At the same time, West Middle and Southern German varieties also drew on East Middle German elements, leading to a horizontal convergence of the varieties. However, these new regional standards were not yet overarched by a spoken national standard. This overarching structure developed within the framework of a second superimposition phase beginning in the early eighteenth century. During this phase, the upper bourgeoisie of Northern and Middle Germany aspired to a pronunciation of German that was closer to the written form, and which had emerged among the educated classes of the Eastphalian towns. This oral standard variety increasingly dominated the older regional high prestige varieties, which were reevaluated and converted to intermediate varieties: [t]he second-order superimposition of a (new) standard variety on an already existing standard-dialect repertoire redefined the former standard variety as an intermediate, non-standard variety in terms of its areal and social reach, and its prestige. Thus, the first-order standard variety remained in use, but was “lowered” to a regional standard or regiolect. The result of the second superimposition of a national standard variety is a continuous, nondiglossic repertoire structure (diaglossia […]). (Auer 2005: 23)

Current studies demonstrate that, within the German language area, the second reevaluation process is not yet complete. The present variety spectra represent the result of centuries of linguistic and attitudinal dynamics which have taken different courses in the various regions of the German language area and have reached different stages (see Ammon 2003; Besch 1983, 2003; Mattheier 1990a, 1990b, 2000 and 2003; Mihm 2000; Schmidt 1998; Wiesinger 2000). It is likely that the diatopic extension of the various regional pronunciation norms of present-day German continue to coincide with the dialect areas (Dialektverbände) as illustrated in Figure 17.6. (For further details, see Schmidt, this volume and 2005a.)

Fig. 17.6: Reevaluation of the older German variety clusterings as modern regional languages (Regionalsprachen; Schmidt 2008)

305

306

III. Structure and dynamics of a language space The Middle German area is a particularly suitable research site for the analysis of reevaluation and restructuring processes. Different types of speakers can currently be found here in synchronic coexistence. Their interindividually distinct variety spectra effectively represent different diachronic stages of vertical reevaluation in apparent time. The “second superimposition” (Mihm 2000: 2112) phase of the subordination of the former high varieties to the supraregional oral standard variety can be observed here in vivo, both “objectively”, via speech data, and “subjectively”, on the basis of the speakers’ language concepts and attitudes. People who have been linguistically socialized in a West Middle German region usually have at least an active regiolectal competence but no dialect competence. They usually shift within the regiolect in all communicative contexts, whereby their “deeper” regiolectal speech levels perform the sociocontextual functions the dialects formerly had. The level closest to the standard is usually the Regionalakzent ‘regional standard’, but a few speakers can, with effort, attain a supraregional standard variety (interferenzfreie Standardsprechsprache; see König 1989). While older West Middle German speakers do not normally further differentiate their Hochdeutsch ‘Standard German’ category, most younger informants do. The various types of subjective Hochdeutsch categorization are summarized in Figure 17.7 (cf. Lenz 2003: 263⫺387). Education, mobility, interregional contact and other factors all contribute to challenge the initial “normality” of regiolectal elements. Some elements attain the status of manifest, highly salient regional identity markers employed under certain situational/pragmatic conditions. Lexicalized regiolectal variants in particular ⫺ mostly also elements of the dialect ⫺ can be found among these “stereotypes” in the Labovian sense (1972: 237, 248). Among those who differentiate their Hochdeutsch further, we find both speakers who (can) only vaguely and tentatively describe their subcategories and others who draw clear distinctions between their subjective categories. On the diachronic axis, the latter can be identified as younger standard-oriented speakers who in Figure 17.7 would be at Steps 2 to 4.

Fig. 17.7: Steps in the reevaluation processes in the West Middle German regiolect (schematized)

17. Emergence of varieties For speakers at Step 2, their subjective best Standard German (“bestes Hochdeutsch”) can be identified linguistically as the regional accent (regional standard). The more definitely the subjective boundary is drawn and hence metacommunicated, the more likely the deeper speech levels of the regiolect are classified by the speaker as colloquial language (“Umgangssprache”) and therefore as non-standard (see Step 3). In a further step along the reevaluation process (see Step 4), the former category of Standard German (“Hochdeutsch”) is completely reevaluated as colloquial language (“Umgangssprache”), and only the supraregional (accent-free) standard variety is called Standard German within the evaluation schema of these most progressive speakers. The growing sensitivity towards regiolectal features is accompanied by an increasing awareness of the features which distinguish the regiolect from the supraregional (accent-free) standard variety. This awareness of variety boundaries sometimes finds expression in hypercorrections (cf. Lenz 2004b, 2005a). The various objective and subjective speech level spectra made evident in synchronic interindividual comparisons of West Middle German speakers schematize the different stages of the superimposition process posited for the superimposition of the Northern flavored New High German oral standard variety upon the antecedent high varieties of Central Germany. In Central Germany, this process is ongoing. Whereas speakers at Step 4 are still a minority in the rural regions of the West Middle German area (cf. Lenz 2003), they are increasingly prevalent in the larger towns of the area, as Lameli’s (2004) real-time study has shown for the city of Mainz.

6. The reevaluation and restructuring cycle Regiolects are not simply the products of reevaluation processes. Their emergence in turn can have massive reevaluation and restructuring effects on the variety spectrum to which they belong. Whilst the emergence of intermediate varieties presupposes the existence of both dialects and standard varieties, it is not rare for these to compete with the former at a later stage. For instance, structural and attitudinal changes at a dialectal level may play a key role in the emergence of or attitudinal support for regiolects. But later, these new intermediate varieties can replace the old dialects. The devaluation of dialect varieties can be seen as an indicator preceding a general dialect loss (cf. Willemyns 1997: 130). The devaluation of dialects is followed in a next step by a deletion of the whole dialectal system (see Figure 17.8; Auer 2005: 28). Dialects, which in antecedent genera-

Fig. 17.8: Loss of local low varieties (cf. Auer 2005: 28)

307

308

III. Structure and dynamics of a language space tions still had covert prestige, are functionally and attitudinally replaced by regiolects. For instance, in the more urban regions of the West Middle German area, the regiolect currently fulfills the sociocontextual functions and bears the covert prestige which in more rural regions is still predominantly the preserve of the dialect. While dialect here is called for in special domains and local networks, there the regiolect takes over the emotional/integrative and other functions formerly performed by the dialect (cf. Lenz 2003: 412). Comparable processes of dialect loss are further advanced in East Middle German (cf. Schönfeld 1992). It is not uncommon for selected dialectal variants to “survive” within the regiolect, when their [⫹diatopic] feature is reevaluated via “reallocation” (Trudgill 1986: 110) to the benefit of other sociopragmatic connotations. Covert prestige aside, other saliency factors which disfavor loss (Abbauresistenz; Lenz 2003: 187) of a (set of) dialectal variant(s) have to be considered (cf. above). Just as dialects and regiolects can support and undermine one another, the relationship between standard varieties and regiolects is complex and ambivalent. On the one hand, standard varieties can function as a motor of horizontal and vertical interdialectal convergence processes which lead to the emergence or stabilization of regiolects. On the other hand, processes of superimposition and destandardization contribute to the emergence and stabilization of regiolects. While the standard variety thus directly or indirectly supports the regiolects, the growing prestige of the standard varieties can also weaken or even completely sidetrack the regiolects. But they in turn can also weaken the standard variety. In the most extreme scenario, regiolects are broad-based prestigious varieties which compete with the standard variety (cf. Auer 1997: 137).

7. Reerences Allport, Gordon W. 1954 Attitudes in the history of social psychology. In: Neil Warren and Marie Jahoda (eds.), Attitudes: Selected Readings, 19⫺25. Harmondsworth: Penguin. Ammon, Ulrich 2003 Dialektschwund, Dialekt-Standard-Kontinuum, Diglossie: Drei Typen des Verhältnisses Dialekt-Standardvarietät im deutschen Sprachgebiet. In: Jannis K. Androutsopoulos and Evelyn Ziegler (eds.), “Standardfragen”: Soziolinguistische Perspektiven auf Sprachgeschichte, Sprachkontakt und Sprachvariation, 163⫺171. (VarioLingua 18.) Frankfurt am Main: Lang. Ammon, Ulrich, Norbert Dittmar and Klaus J. Mattheier (eds.) 1987 Soziolinguistik. Ein internationales Handbuch zur Wissenschaft von Sprache und Gesellschaft. vol. 1. (Handbücher zur Sprach- und Kommunikationswissenschaft 3.1.) Berlin: de Gruyter. Ammon, Ulrich, Norbert Dittmar, Klaus J. Mattheier and Peter Trudgill (eds.) 2004 Sociolinguistics. An International Handbook of the Science of Language and Society. 2nd ed. (Handbücher zur Sprach- und Kommunikationswissenschaft 3.1.) Berlin: de Gruyter. Auer, Peter 1986 Konversationelle Standard/Dialekt-Kontinua (Code-Shifting). Deutsche Sprache 14: 97⫺ 124. Auer, Peter 1990 Phonologie der Alltagssprache. Eine Untersuchung zur Standard/Dialekt-Variation am Beispiel der Konstanzer Stadtsprache. (Studia Linguistica Germanica 28.) Berlin/New York: de Gruyter.

17. Emergence of varieties Auer, Peter 1993 Zweidimensionale Modelle für die Analyse von Standard/Dialekt-Variation und ihre Vorläufer in der deutschen Dialektologie. Ein Beitrag zur historischen Kontinuität. In: Wolfgang Viereck (ed.), Verhandlungen des Internationalen Dialektologenkongresses. Bamberg, 29. 7.⫺4. 8. 1990, vol. 2, 3⫺22. (Zeitschrift für Dialektologie und Linguistik. Beihefte 75.) Stuttgart: Steiner. Auer, Peter 1997 Führt Dialektabbau zur Stärkung oder Schwächung der Standardvarietät? Zwei phonologische Fallstudien. In: Mattheier and Radtke (eds.), 129⫺162. Auer, Peter 2001 Schirmunskis Unterscheidung zwischen primären und sekundären Dialektmerkmalen und ihre Bedeutung für die heutige Sozio-Dialektologie. In: Swetlana M. Pankratowa (ed.): Nemetskaja Fililigija i Sankt-Peterburgskom gosudarstwennom universitete [German Philology at the National University of St Petersburg], 198⫺214. St. Petersburg: Universitätsverlag. Auer, Peter 2005 Europe’s sociolinguistic unity, or: A typology of European dialect/standard constellations. In: Nicole Delbecque, Johan van der Auwera and Dirk Geeraerts (eds.), Perspectives on Variation: Sociolinguistic, Historical, Comparative, 8⫺42. (Trends in Linguistics, Studies and Monographs 163.) Berlin/New York: Mouton de Gruyter. Auer, Peter, Birgit Barden and Beate Grosskopf 1996 Dialektanpassung bei sächsischen “Übersiedlern” ⫺ Ergebnisse einer Longitudinalstudie. In: Norbert Boretzky, Werner Enninger and Thomas Stolz (eds.), Areale, Kontakte, Dialekte. Sprache und ihre Dynamik in mehrsprachigen Situationen, 139⫺166. (Bochum-Essener Beiträge zur Sprachwandelforschung 24.) Bochum: Studienverlag Brockmeyer. Auer, Peter and Frans Hinskens 1996 The convergence and divergence of dialects in Europe. New and not so new developments in an old area. In: Auer, Hinskens and Mattheier (eds.), 1⫺30. Auer, Peter, Frans Hinskens and Paul Kerswill 2005 The study of dialect convergence and divergence: conceptual and methodological considerations. In: Auer, Hinskens and Kerswill (eds.), 1⫺48. Auer, Peter, Frans Hinskens and Paul Kerswill (eds.) 2005 Dialect Change: Convergence and Divergence in European Languages. Cambridge: Cambridge University Press. Auer, Peter, Frans Hinskens and Klaus J. Mattheier (eds.) 1996 Konvergenz und Divergenz von Dialekten in Europa. (Sociolinguistica 10.) Tübingen: Niemeyer. Bellmann, Günter 1983 Probleme des Substandards im Deutschen. In: Klaus J. Mattheier (ed.): Aspekte der Dialekttheorie, 105⫺130. (Reihe Germanistische Linguistik 46.) Tübingen: Niemeyer. Berruto, Gaetano 1987 Varietät. In: Ammon, Dittmar and Mattheier (eds.), 263⫺267. Berruto, Gaetano 2004 Linguistic Variety ⫺ Language (Whole Language, Historical Language). In: Ammon, Dittmar, Mattheier and Trudgill (eds.), 188⫺195. Besch, Werner 1983 Dialekt, Schreibdialekt, Schriftsprache, Standardsprache. Exemplarische Skizze ihrer historischen Ausprägung im Deutschen. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke and Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 2, 961⫺990. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.2.) Berlin/New York: de Gruyter.

309

310

III. Structure and dynamics of a language space Besch, Werner 2000 Variantentyp Hyperkorrektion. In: Annelies Häcki Buhofer (ed.), Vom Umgang mit sprachlicher Variation. Soziolinguistik, Dialektologie, Methoden und Wissenschaftsgeschichte. Festschrift für Heinrich Löffler zum 60. Geburtstag, 187⫺193. (Basler Studien zur deutschen Sprache und Literatur 80.) Tübingen: Francke. Besch, Werner 2003 Die Entstehung und Ausformung der neuhochdeutschen Schriftsprache/Standardsprache. In: Werner Besch, Anne Betten, Oskar Reichmann and Stefan Sonderegger (eds.), Sprachgeschichte. Ein Handbuch zur Geschichte der deutschen Sprache und ihrer Erforschung, vol. 3, 2252⫺2296. 2nd ed. (Handbücher zur Sprach- und Kommunikationswissenschaft 2.3.) Berlin/New York: de Gruyter. Besch, Werner, Anne Betten, Oskar Reichmann and Stefan Sonderegger (eds.) 2000 Sprachgeschichte. Ein Handbuch zur Geschichte der deutschen Sprache und ihrer Erforschung, vol. 2. 2nd ed. (Handbücher zur Sprach- und Kommunikationswissenschaft 2.2.) Berlin/New York: Walter de Gruyter. Chambers, Jack and Peter Trudgill 1980 Dialectology. Cambridge: Cambridge University Press. Coseriu, Eugenio 1988 Sprachkompetenz. Grundzüge der Theorie des Sprechens. Bearbeitet und herausgegeben von Heinrich Weber. Tübingen: Francke. Coetsem, Frans van 1992 The interaction between dialect and standard language, and the question of language internationalization. Viewed from the standpoint of the Germanic languages. In: Jan Arnoldus van Leuvensteijn and Jan Berns (eds.), Dialect and Standard Language in the English, Dutch, German and Norwegian Language Areas, 15⫺70. Amsterdam: North-Holland. Danesˇ, Frantisˇek 1968/1982 Dialektische Tendenzen in der Entwicklung der Literatursprache. In: Jürgen Scharnhorst and Erika Ising (eds.), Grundlagen der Sprachkultur. Beiträge der Prager Linguistik zur Sprachtheorie und Sprachpflege. Teil 1, 92⫺113. Berlin: Akademie-Verlag. Danesˇ, Frantisˇek 2006 Herausbildung und Reform von Standardsprachen und Destandardisierung ⫺ Development and Reform of Standard Languages and Destandardization. In: Ulrich Ammon, Norbert Dittmar, Klaus Mattheier and Peter Trudgill (eds.), Sociolinguistics. An International Handbook of the Science of Language and Society, vol. 3, 2197⫺2209. 2nd ed. (Handbücher zur Sprach- und Kommunikationswissenschaft 3.3.) Berlin/New York: de Gruyter. Deprez, Kas and Yves Persoons 1987 Attitude. In: Ammon, Dittmar and Mattheier (eds.), 125⫺132. Deumert, Ana and Wim Vandenbussche 2003 Standard language. Taxonomies and histories. In: Deumert and Vandenbussche (eds.), 1⫺14. Deumert, Ana and Wim Vandenbussche (eds.) 2003 Germanic Standardization. Past to Present. (Impact: Studies in Language and Society 18.) Amsterdam: Benjamins. Dittmar, Norbert 2004 Umgangssprache ⫺ Nonstandard / Vernacular ⫺ Nonstandard. In: Ammon, Dittmar, Mattheier and Trudgill (eds.), 251⫺262. Elspaß, Stephan 2007 Variation and change in colloquial (Standard) German ⫺ The Atlas zur deutschen Alltagssprache (AdA) project. In: Christian Fandrych and Reinier Salverda (eds.), Standard, Variation und Sprachwandel in germanischen Sprachen / Standard, Variation and Language

17. Emergence of varieties Change in the Germanic Languages, 201⫺216. (Studien zur deutschen Sprache 41.) Tübingen: Narr. Ferguson, Charles A. 1959 Diglossia. Word 15: 325⫺340. Gilles, Peter 1997 Bemerkungen zum Verhältnis zwischen “Dialektveränderung” und “Standardisierung”/ “Destandardisierung”. In: Mattheier and Radtke (eds.), 163⫺169. Gilles, Peter 1999 Dialektausgleich im Letzebuergeschen. Zur phonetisch-phonologischen Fokussierung einer Nationalsprache. (Phonai 44.) Tübingen: Niemeyer. Goossens, Jan 2000 De toekomst van het Nederlands in Vlaanderen [The future of Dutch in Flanders]. Ons Erfdeel 43: 2⫺13. Haugen, Einar 1966 Language Conflict and Language Planning. The Case of Modern Norwegian. Cambridge, MA: Harvard University Press. Haugen, Einar 1987 Language planning. In: Ammon, Dittmar and Mattheier (eds.), 626⫺637. Herrgen, Joachim 1986 Koronalisierung und Hyperkorrektion. Das palatale Allophon des /CH/-Phonems und seine Variation im Westmitteldeutschen. Stuttgart: Steiner. Herrgen, Joachim 2001 Die Dialektologie des Deutschen. In: Sylvain Auroux, E. F. K. Koerner, Hans-Josef Niederehe and Kees Versteegh (eds.), Geschichte der Sprachwissenschaften. Ein internationales Handbuch zur Entwicklung der Sprachforschung von den Anfängen bis zur Gegenwart, vol. 2, 1513⫺1535. 2nd ed. (Handbücher zur Sprach- und Kommunikationswissenschaft 18.2.) Berlin/New York: de Gruyter. Hewstone, Miles, Antony S. R. Manstead and Wolfgang Stroebe 1997 The Blackwell Reader in Social Psychology. Oxford: Blackwell. Hinskens, Frans 1996 Dialect Leveling in Limburg. Structural and Sociolinguistic Aspects. Tübingen: Niemeyer. Huesmann, Anette 1998 Zwischen Dialekt und Standard. Empirische Untersuchung zur Soziolinguistik des Varietätenspektrums im Deutschen. (Reihe Germanistische Linguistik 199.) Tübingen: Niemeyer. Hughes, Arthur, Peter Trudgill and Dominic Watt 2005 English Accents and Dialects: An Introduction to Social and Regional Varieties of English in the British Isles. 4th ed. London: Hodder Arnold. Jahr, Ernst Ha˚kon 1997 On the use of dialects in Norway. In: Heinrich Ramisch and Kenneth Wynne (eds.), Language in Time and Space: Studies in Honour of Wolfgang Viereck on the Occasion of his 60th Birthday, 363⫺369. (Zeitschrift für Dialektologie und Linguistik. Beihefte 97.) Stuttgart: Steiner. Jahr, Ernst Ha˚kon 2003 Norwegian. In: Deumert and Vandenbussche (eds.), 331⫺353. Kerswill, Paul and Peter Trudgill 2005 The birth of new dialects. In: Auer, Hinskens and Kerswill (eds.), 196⫺220. Kerswill, Paul and Ann Williams 2000 Creating a new town koine: Children and language change in Milton Keynes. Language and Society 29(1): 65⫺115. König, Werner 1989 Atlas zur Aussprache des Schriftdeutschen in der Bundesrepublik Deutschland. Ismaning: Hueber.

311

312

III. Structure and dynamics of a language space König, Werner 1997 Phonetisch-phonologische Regionalismen in der deutschen Standardsprache. Konsequenzen für den Unterricht “Deutsch als Fremdsprache”? In: Gerhard Stickel (ed.), Varietäten des Deutschen. Regional- und Umgangssprachen, 246⫺270. (IDS-Jahrbuch 1996.) Berlin/ New York: de Gruyter. Kristiansen, Tore 2003 Danish. In: Deumert and Vandenbussche (eds.), 69⫺91. Labov, William 1972 Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Lameli, Alfred 2004 Standard und Substandard. Regionalismen im diachronen Längsschnitt. (Zeitschrift für Dialektologie und Linguistik. Beiheft 128.) Stuttgart: Steiner. Lenz, Alexandra N. 2003 Struktur und Dynamik des Substandards. Eine Studie zum Westmitteldeutschen (Wittlich/ Eifel). (Zeitschrift für Dialektologie und Linguistik. Beihefte 125.) Stuttgart: Steiner. Lenz, Alexandra N. 2004a Verdichtungsbereiche und Varietätengrenzen im Methodenvergleich. In: Helen Christen (ed.), Dialekt, Regiolekt und Standardsprache im sozialen und zeitlichen Raum, 199⫺220. Vienna: Edition Präsens. Lenz, Alexandra N. 2004b Hyperforms and Variety Barriers. In: Britt-Louise Gunnarsson, Lena Bergström, Gerd Eklund, Staffan Fridell, Lise H. Hansen, Angela Karstadt, Bengt Nordberg, Evan Sundgren and Mats Thelander (eds.), Language Variation in Europe. Papers from the Second International Conference on Language Variation in Europe, ICLaVE 2, Uppsala University, Sweden, June 12⫺12, 2003, 281⫺294. Uppsala: Universitetstryckeriet. Lenz, Alexandra N. 2005a Hyperdialektalismen und Hyperkorrektionen. In: Lenz and Mattheier (eds.), 76⫺95. Lenz, Alexandra N. 2005b Zur Struktur des westmitteldeutschen Substandards ⫺ Dynamik von Varietäten. In: Eckhard Eggers, Jürgen Erich Schmidt and Dieter Stellmacher (eds.), Moderne Dialekte ⫺ Neue Dialektologie. Akten des 1. Kongresses der Internationalen Gesellschaft für Dialektologie des Deutschen (IGDD), 229⫺252. (Zeitschrift für Dialektologie und Linguistik. Beihefte 130.) Stuttgart: Steiner. Lenz, Alexandra N. 2006 Clustering linguistic behaviour on the basis of linguistic variation methods. In: Markku Filppula, Juhani Klemola, Marjatta Palander and Esa Penttilä (eds.), Topics in Dialectal Variation. Selection of Papers from the Eleventh International Conference on Methods in Dialectology, 69⫺98. (Studies in Languages 40.) Joensuu: Joensuu University Press. Lenz, Alexandra N. in press “Von Erp nach Wittlich und zurück” ⫺ Substandardsprachliche Strukturen des Mittelfränkischen. In: Peter Gilles, Joachim Scharloth and Evelyn Ziegler (eds.), Variatio delectat. (VarioLingua.) Frankfurt am Main: Lang. Lenz, Alexandra N. and Klaus J. Mattheier (eds.) 2005 Varietäten ⫺ Theorie und Empirie. (VarioLingua 23.) Frankfurt am Main: Lang. Le Page, Robert 1980 Projection, focussing, diffusion, or, steps towards a sociolinguistic theory of language, illustrated from the Sociolinguistic Survey of Multilingual Communities. Stage I: Cayo District, Belize (formerly British Honduras), and II: St Lucia. York Papers in Linguistics 9: 9⫺32. Leuvensteijn, Jan Arnoldus van and Jan Berns (eds.) 1992 Dialect and Standard Language in the English, Dutch, German and Norwegian Language Areas. Amsterdam: North-Holland.

17. Emergence of varieties Mattheier, Klaus J. 1990a Überlegungen zum Substandard im Zwischenbereich von Dialekt und Standardsprache. In: Günter Holtus and Edgar Radtke (eds.), Sprachlicher Substandard, vol. 3, Standard, Substandard und Varietätenlinguistik, 1⫺16. (Konzepte der Sprach- und Literaturwissenschaft 45.) Tübingen: Niemeyer. Mattheier, Klaus J. 1990b Dialekt und Standardsprache. Über das Varietätensystem des Deutschen in der Bundesrepublik. International Journal of the Sociology of Language 83: 59⫺81. Mattheier, Klaus J. 1997 Über Destandardisierung, Umstandardisierung und Standardisierung in modernen europäischen Standardsprachen. In: Mattheier and Radtke (eds.), 1⫺9. Mattheier, Klaus J. 2000 Die Durchsetzung der deutschen Hochsprache im 19. und beginnenden 20. Jahrhundert: sprachgeographisch, sprachsoziologisch. In: Besch, Betten, Reichmann and Sonderegger (eds.), 1951⫺1966. Mattheier, Klaus J. 2003 German. In: Deumert and Vandenbussche (eds.), 211⫺244. Mattheier, Klaus J. 2004 Tod der Dialektologie. In: Alexandra Lenz, Edgar Radtke and Simone Zwickl (eds.), Variation im Raum. Variation and Space, 233⫺244. (VarioLingua 20.) Frankfurt/Main: Lang. Mattheier, Klaus J. and Edgar Radtke (eds.) 1997 Standardisierung und Destandardisierung europäischer Nationalsprachen. (VarioLingua 1.) Frankfurt am Main: Lang. Mazur, Jan 1997 Destandardisierungserscheinungen im gegenwärtigen Polnischen. In: Mattheier and Radtke (eds.), 221⫺236. Mihm, Arend 2000 Die Rolle der Umgangssprachen seit der Mitte des 20. Jahrhunderts. In: Besch, Betten, Reichmann and Sonderegger (eds.), 2107⫺2137. Milroy, Lesley 1982 The effect of two interacting extra-linguistic variables on patterns of variation in urban vernacular speech. In: David Sankoff and Henrietta Cedergren (eds.), Variation Omnibus, 161⫺170. Edmonton, Alberta: Linguistic Research. Nevalainen, Terttu 2003 English. In: Deumert and Vandenbussche (eds.), 127⫺156. Pargman, Sheri 1998 On the regularity of hypercorrection in phonological change. Diachronia 15(2): 285⫺307. Pedersen, Inge Lise 2003 Traditional dialects of Danish and the de-dialectalization 1900⫺2000. International Journal of the Sociology of Language 159: 9⫺28. Pedersen, Inge Lise 2005 Processes of standardisation in Scandinavia. In: Auer, Hinskens and Kerswill (eds.), 171⫺195. Romaine, Suzanne 1998 Introduction. In: Suzanne Romaine (ed.), The Cambridge History of the English Language, vol. 4, 1776⫺1997, 1⫺56. Cambridge: Cambridge University Press. Rosenberg, Milton J. and Carl I. Hovland 1960 Cognitive, affective, and behavioral components of attitudes. In: Milton J. Rosenberg, Carl I. Hovland, William J. McGuire, Robert P. Abelson and Jack W. Brehm (eds.), Attitude Organization and Change. An Analysis of Consistency Among Attitude Components, 15⫺30. New Haven: Yale University Press.

313

314

III. Structure and dynamics of a language space Røyneland, Unn 2009 Dialects in Norway. Catching up with the rest of Europe? International Journal of the Sociology of Language 196⫺197: 7⫺30. Sandøy, Helge 2002 Nordic language history and current trends in dialectology. In: Oskar Bandle, Kurt Braunmüller, Ernst Ha˚kon Jahr, Allan Karker, Hans-Peter Naumann and Ulf Teleman (eds.), The Nordic Languages, 304⫺312. (Handbücher zur Sprach- und Kommunikationswissenschaften. 22.1.) Berlin/New York: de Gruyter. Schmidt, Jürgen Erich 1998 Moderne Dialektologie und regionale Sprachgeschichte. Zeitschrift für deutsche Philologie 117: 163⫺179 (special issue: Regionale Sprachgeschichte). Schmidt, Jürgen Erich 2005a Die deutsche Standardsprache: Eine Varietät ⫺ drei Oralisierungsnormen. In: Ludwig M. Eichinger and Werner Kallmeyer (eds.), Standardvariation ⫺ Wie viel Variation verträgt die deutsche Sprache?, 278⫺305. (Institut für deutsche Sprache. Jahrbuch 2004.) Berlin/New York: de Gruyter. Schmidt, Jürgen Erich 2005b Versuch zum Varietätenbegriff. In: Lenz and Mattheier (eds.), 61⫺74. Schmidt, Jürgen Erich in press Die modernen Regionalsprachen als Varietätenverbund. In: Peter Gilles, Joachim Scharloth and Evelyn Ziegler (eds.), Variatio delectat. (VarioLingua.) Frankfurt am Main: Peter Lang. Scholz, Arno 1997 Das Varietätenspektrum des Italienischen im Wandel. In: Mattheier and Radtke (eds.), 61⫺86. Schönfeld, Helmut 1992 Dialekt, Umgangssprache und Standardsprache auf dem Gebiet der ehemaligen DDR im 20. Jahrhundert. In: van Leuvensteijn and Berns (eds.), 256⫺270. Spiekermann, Helmut 2005 Regionale Standardisierung, nationale Destandardisierung. In: Ludwig M. Eichinger and Werner Kallmeyer (eds.), Standardvariation ⫺ Wie viel Variation verträgt die deutsche Sprache?, 100⫺125. (Institut für deutsche Sprache. Jahrbuch 2004.) Berlin/New York: de Gruyter. Stroop, Jan P. A. 1992 Towards the end of the standard language in the Netherlands. In: van Leuvensteijn and Berns (eds.), 162⫺175. Stroop, Jan 1997 Wordt het Poldernederlands model? Taal en Tongval 10: 10⫺29 (special issue: Standaardisering in Noord en Zuid). Teleman, Ulf 2003 Swedish. In: Deumert and Vandenbussche (eds.), 405⫺429. Trudgill, Peter 1983 On Dialect. Social and Geographical Perspectives. Oxford: Blackwell. Trudgill, Peter 1986 Dialects in Contact. (Language in Society 10.) Oxford: Blackwell. Trudgill, Peter 1998 The chaos before the order: New Zealand English and the second stage of new-dialect formation. In: Ernst Ha˚kon Jahr (ed.), Advances in Historical Sociolinguistics, 1⫺11. Berlin: Mouton de Gruyter. Trudgill, Peter 2006 New-Dialect Formation: The Inevitability of Colonial Englishes. Edinburgh: Edinburgh University Press.

18. Urban and rural language

315

Trudgill, Peter, Elizabeth Gordon, Gillian Lewis and Margaret A. Maclagan 2000 Determinisms in new-dialect formation and the genesis of New Zealand English. Journal of Linguistics 36: 299⫺318. Wiesinger, Peter 2000 Die Diagliederung des Neuhochdeutschen bis zur Mitte des 20. Jahrhunderts. In: Besch, Betten, Reichmann and Sonderegger (eds.), 1932⫺1951. Willemyns, Roland 1997 Dialektverlust im niederländischen Sprachraum. Zeitschrift für Dialektologie und Linguistik 64(2): 133⫺154. Willemyns, Roland 2003 Dutch. In: Deumert and Vandenbussche (eds.), 93⫺125.

Alexandra N. Lenz, Groningen (Netherlands)

18. Urban and rural language 1. 2. 3. 4. 5.

Urban and rural language: Changing perspectives The expansion of urban language Urban insularity Rural dynamics References

1. Urban and rural language: Changing perspectives Until a few decades ago dialectologists generally focused on geographical variation in rural areas. Although traditional dialectology produced a number of studies on city dialects as well (e.g., Sivertsen 1960; Heike 1964; Taeldeman 1985), the discipline manifested itself essentially as “rural dialect geography”: space was a central issue and research was carried out mainly in the countryside. One of its main aims was the recording of the authentic local dialect and rural speakers were assumed to be the best guardians of the old local varieties. Urban speakers were supposed to mix with other social groups and were therefore considered less representative of local speech. Moreover it was expected that language changes occurred “at a relatively slower rate in areas of lower population density relative to areas of high population density” (Hamilton 2001: 418⫺ 419). The preference for rural informants manifests itself from the beginning of the twentieth century in larger dialectological research projects such as Gillie´ron and Edmont’s Atlas Linguistique de la France. Furthermore informants usually were NORMs (non-mobile old rural males, see Chambers and Trudgill 1980). For instance, for the Survey of English Dialects (SED), carried out by the University of Leeds between 1950 and 1961, the fieldworkers were instructed to select elderly men and women. Since women were said to encourage the social upgrading of the speech of their families, there was a preference for male informants who were born locally (Orton, Sanderson and Widdowson 1978: 3). Rural life was and is often associated with lack of mobility and

316

III. Structure and dynamics of a language space (linguistic) stability. Apart from reflections on the ideal profile of the informants, traditional dialectology showed little interest in the social determinants of language variation. The innovative work of Labov in the US in the early 1960s and the subsequent rise of the discipline of sociolinguistics involved a shift from the study of geographic language variation towards the study of social language variation. This shift was paralleled by a shift from research on rural language towards research on language variation in urban speech communities. Symptomatic of the changing perspective is the fact that early sociolinguistic research by Labov and his colleagues has often been labeled “urban dialectology”. The name reveals both the roots of the discipline and the “new” focus on urban speech. Since research within this branch focused (and still focuses) on the correlation between social variation and linguistic variation, from now on the social profile of the informants (including parameters such as age and gender, but also social class and ethnicity) was diversified systematically (Foulkes and Dochterty 1999: 2). The focus of traditional dialectology on the recording and preservation of local speech and the related preference for rural and older informants may have reinforced an assumption which pervades a great many dialectological and sociolinguistic studies: the idea of rural language being static and urban language being innovative. The latter even more so since the use of the apparent time method in many (urban) sociolinguistic studies led to a distinct interest in the dynamics of speech and opened up new perspectives for the study of language change in progress. “Researching in the city was most probably seen as the way to gain access to the most fluid and heterogeneous communities, and therefore to tackle the issue of the social embedding of change ‘where it’s all happening’”. The rural, on the other hand, is “still portrayed as the insular, the isolated, the static, as an idyll of peace and tranquility rather than as composed of heterogeneous communities, of contact, of change and progress, and of conflict” (Britain 2002: 607⫺ 608). At the same time, recent research has shown a growing awareness of the unfoundedness of this dichotomy (e.g., Besch 1981, 1983; Hinskens 1996; Vandekerckhove 2000): “Language varies and changes in rural as well as urban communities” (Britain 2002: 608). The changing perception is also linked to the fact that, in many respects, the dichotomy between rural and urban areas and their respective communities proves difficult to hold nowadays. As Auer and Hinskens (1996: 4) point out, processes of industrialization, urbanization, increasing geographical mobility and ease of communication at regional and supraregional levels have brought about dialect contact to an extent hitherto unknown (see also Watt and Milroy 1999: 26). In Europe at least, living in the countryside does not imply immobility and isolation anymore. In the past decades, many closed rural village communities have been broken open by migration from rural into urban areas and vice versa. Many people from the country move to the city for their professional career while others commute there every day for professional and other purposes. On the other hand, city people seek living space in the surrounding countryside, a process which has been called “counterurbanization”. Moreover, in the worldwide process of urbanization, many small villages have been swallowed by neighboring towns and cities. All these processes lead to intensive dialect contact and result in the dissolution of closed networks and the local language norms which need such close-knit networks in order to remain intact (Watt and Milroy 1999: 42). The developments mentioned above do not imply that the distinction between urban and rural or non-urban language varieties has been completely blurred in Europe. Nor

18. Urban and rural language does it mean that city dialects and non-urban varieties do not develop their own dynamics anymore. But their dynamics is often the product of contact between several varieties. Even in former times there was linguistic exchange between city and countryside on the one hand and between several cities on the other hand (cf. Goossens 1992; Taeldeman 2005), but this contact is more intensive today than ever. This might explain why recent research shows a major interest in the interaction between urban and rural language and in the influence of urban varieties on their rural surroundings. These topics will be the main focus of this article.

2. The expansion o urban language 2.1. Contagious and hierarchical diusion Cities are commercial, economic and cultural centers and as such they are often attributed a certain prestige by the rural hinterland and/or by other cities. The inhabitants of a city generally derive prestige from being part of that city and sharing an urban life style. Among the many (possible) markers of urban life style, speech may play a prominent role. And since prestige is a well-known catalyst of imitation, the adoption of characteristics of urban speech often appears to be part of a (conscious or subconscious) strategy of people living in small towns or in the countryside aimed at acquiring a share in the prestige associated with urbanity and urban life style. The expansion of urban vernacular features has been attested in many (socio-)dialectological studies in the past decades (e.g., Milroy 1992; Bailey, Wikle and Tillery 1993; Taeldeman 2005). In the literature a distinction is made between two models, the contagious diffusion model, better known as the wave model, and the hierarchical diffusion model. The wave model is less well represented in the literature than the hierarchical diffusion model, but as Britain (2002: 623) puts it, it is a kind of “iconic representation of diffusion”, “with diagrams resembling the ripples created by raindrops falling in a puddle of water”. In the wave model, distance is the major determinant: innovations spread from an urban center to the hinterland, first affecting the regions neighboring the city and step by step also affecting more distant locations. The bigger the distance between the “trend-setting” city and a community in its hinterland, the longer it takes before the expansive features reach the place. Wave-like diffusions from city to hinterland have been attested in several socio- and geolinguistic studies (cf. infra), but most changes and innovations do not spread like a wave or an oil stain, i. e., gradually, with decreasing force. By adopting gravity models from social geography, Trudgill (1974) was able to demonstrate that the size of the communities involved interacts with the distance between them: linguistic innovations follow a pattern of urban hierarchy, from larger city, to smaller city, to larger town and smaller town till they finally reach the countryside via those smaller towns or villages. The formula which Trudgill used to predict the likely influence of place x on place y incorporated both the distance between both places and their relative sizes. In recent years the predictive power of the gravity model has been criticized because of its neglect of structural-linguistic factors on the one hand and socio-psychological factors, like the mutual appreciation of urban centers and hinterland, on the other. Both factors, especially the latter, are hard to account for in a formal model. And yet, as was

317

318

III. Structure and dynamics of a language space demonstrated by Taeldeman (2005), these factors may either stimulate the spread of a feature or, conversely, inhibit the spread of an innovation which, judging from the gravity model, would be extremely likely to diffuse. However, this has not undermined the value of the hierarchical diffusion model. On the contrary: [I]nteraction between urban centers in modern societies is likely to be greater, and therefore a more frequent and effective conduit for accommodation and transmission of innovations, than between urban and rural. Transportation networks tend to link urban with urban, the socioeconomic and consumer infrastructure tends to be based in and oriented towards urban centers, with the ensuing consequences for employment and commuting patterns, and these obviously feed the hierarchical nature of diffusion. (Britain 2002: 623⫺624)

Many recent studies reveal an intensive linguistic influencing of urban centers (cf. section 2.2). In modern societies, marked by great geographic mobility and, on the linguistic side, dialect loss and dialect leveling, the influence of some cities and their characteristic speech appears to be strong enough to create a new dynamics which can stand up or even run counter to the increasing impact of the standard language.

2.2. Both diusion models exempliied The expansion of Cockney features to other cities, some of them quite remote from London, offers a well-known example of innovations spreading according to the hierarchical diffusion model. Foulkes and Docherty (1999: 11) state that research conducted in the past decades produces plenty of evidence to support Wells’ speculation that London’s “working-class accent is today the most influential source of phonological innovation in England” (Wells 1982: 301). Glottalization of intervocalic and word-final /t/, th-fronting, labiodental realizations of /r/ and vocalization of /l/ are the most prominent features which seem to have spread from London, first affecting large urban centers before reaching smaller towns or villages. Some authors warn not to be too eager to interpret the presence of these features in other urban vernaculars as direct borrowings from Cockney. In some cases, they may be the result of independent developments in different locations (Britain, to appear). Moreover the source of these changes is often not identified as Cockney but as “Estuary English”, which is said to occupy an intermediate position between Cockney and RP and contains some features present in most of southeastern England. Foulkes and Docherty (1999: 11) claim that Estuary English features are spreading “not because Estuary English is a coherent and identifiable influence, but because the features represent neither the standard nor the extreme non-standard poles of the continuum”. So some caution may be needed when interpreting the diffusion of features which could have originated in Cockney. And yet, there can hardly be any doubt that the prestige of the capital in many cases formed the basis for the spread of London features. For instance, Mees and Collins (1999: 201) discuss glottalization in Cardiff English. They find that it is an attractive feature for speakers of Cardiff English because it is associated with “sophisticated and fashionable speech”: “Nowadays glottalization is associated with London Life, metropolitan fashions and trend-setting attitudes”. It has become a prestige feature, although in London and in the south of England it is a marker of male and working class speech (cf. Britain 2009: 140). This interpretation demonstrates the importance of so-called socio-psychological factors (cf. supra: Taelde-

18. Urban and rural language man 2005). London life style is regarded as worth imitating and so is London speech, even if the borrowed features are (or were) not prestigious at all in the place of origin. From the 1980s onwards, Trudgill has described the diffusion of London forms across urban East Anglia (see Britain 2009 and Trudgill 1983, 1986, 1988, 1999). In Trudgill 1986 for instance, the shift in the production of voiced and voiceless /th/ attested in the language of Norwich teenagers is directly linked to the Cockney merger between /v/ and /Î/ on the one hand and /f/ and /θ/ on the other. Other linguists have described the presence of this feature and other prototypical London or southeastern features like labiodental /r/, vocalized /l/ and t-glottalization in cities and towns all over England, more specifically in Colchester, Reading, Milton Keynes, The Fens, Derby, Birmingham, Hull, Liverpool, Sheffield, Middlesbrough and Newcastle (Britain 2009: Table 1). Once again, not all of the innovations in these city vernaculars can be attributed to (direct) London influence, but in many cases London features have been found to jump to other urban centers, often leaving the hinterland unaffected (at least for some time). In some cases the hinterland may lag behind for a very long time. Milroy (1992: 197) refers to the historical spread of /h/-dropping in British English, which also appears to have proceeded along an urban hierarchy and adds: “despite the long history of /h/-dropping, there are still rural (but not urban) dialects in southern England which do not /h/-drop”. The same holds for parts of East Anglia, where /h/-dropping is a more recent phenomenon which has spread from London: in urban centers such as Norwich it was already well established in the 1980s but it did not occur in the rural areas of the region (Taeldeman 2005: 270, see Trudgill 1983, 1986). A similar pattern can be found all over western Europe. Røyneland (2009) describes the rapid expansion of the regional East Norwegian “urban standard”, with the capital Oslo as its center, to many southeast Norwegian cities and towns. Strikingly similar to the Cockney case is the re-evaluation or re-allocation of low status urban features (cf. the popularity of London Cockney features in England): In Norway in general, as in the other Scandinavian countries, we observe a development where traditional low status urban features appear to be spreading at the expense of traditional high status ones, both in the cities and in the surrounding areas. (Røyneland 2009: 20)

These features no longer function as markers of social class, but have become markers of urbanity and urban life style (Røyneland 2009; see also Røyneland 2005). France has been subject to more dramatic dialect loss than any other western European countries due to the long lasting expansion of the Parisian vernacular all over the country. But today the homogenizing effects of Paris may meet “countervailing centrifugal pressures” which might result in the development to supra-local regional dialects. The gradual emergence of these supra-local varieties, leading to greater regional diversification in the linguistic landscape in France, is rooted in some larger urban centers and restricted to the northern and southern peripheries. Cities like Lille in the north of France and Marseille in the south are far enough away from Paris “to resist the pressure from the capital and sizeable enough to diffuse local forms outwards to their own hinterlands” (Hornsby 2007, 2009). So time and again, recent research confirms the trend-setting role of urban centers and the linguistic interaction between bigger and smaller cities or towns. Time and again

319

320

III. Structure and dynamics of a language space

Map 18.1: Northern Belgium (Dutch-speaking Belgium, Flanders) with its five provinces; the left one is the province of West Flanders

also, urban centers are the driving force for the expansion of urban vernacular features at the expense of (more) standard colloquial speech, even and especially in recent times. These features diffuse from major urban centers to smaller cities or towns, but that does not mean they never reach the smaller communities in between them. In many cases the diffusion is a combination of the hierarchical and contagious pattern. Taeldeman (2005) illustrates this complementarity, but at the same time the dominance of the hierarchical over the contagious pattern, on the basis of data from the Dutch-speaking part of Belgium. According to Taeldeman (2005: 265) uvular /R/ was still confined to urban centers in the 1970s in the Flemish and Brabantine dialect area. It had become general in Brussels and in the East-Flemish city of Ghent, and it was also well represented in some other large centers such as Antwerp, Bruges and Leuven too (see Map 18.1). In some smaller towns, however, it scored only ten percent at that time. Twenty years later, the uvular /R/ had spread from the city of Ghent to its hinterland. Rogier (1994) distinguished three concentric circles around the city and studied the presence of /R/ in the speech of the youngest generation in each of these circles. He found that 83.9 percent of the children and teenagers living in the circle closest to Ghent realized /R/ instead of /r/, in the middle circle 75.7 percent of the youngsters had /R/ and, finally, in the zone most remote from the city 47.6 percent produced /R/. This gradual expansion of /R/ at the expense of the /r/ offers a perfect example of contagious (or wave) diffusion of a linguistic innovation. But it is clear that in the first phase of the expansion of the feature in the Flemish and Brabantine dialects, only city dialects were involved. It took some time before the innovation spread to the hinterland of some cities (see also Tops 2005). Vandekerckhove (2000, 2005a, 2005b) also offers several examples of contagious diffusion. The research area was in this case the south-east of the province of West Flanders, and more particularly the small town of Deerlijk, in the eastern hinterland of the West Flemish city of Kortrijk (see Map 18.1). The study, which had both a real time and an apparent time dimension, showed that both Kortrijk city dialect features (like the word initial consonant cluster /sk/ for Dutch /sx/ in words like schoen ‘shoe’) and general West Flemish features (like the diphthong /oe / for Dutch /o/ in words like boot ‘boat’) had gradually spread east of the city since the beginning of the twentieth century. The

18. Urban and rural language city dialect of Kortrijk appeared to be the donor language for most of the changes observed. However, the role of this urban center could not be interpreted unidirectionally. It functioned both as a source language and as a recipient language (Van Coetsem 1988), in some cases passing on dialect features which it had borrowed itself from other West Flemish dialects (Vandekerckhove 2005a, 2005b). The city of Kortrijk offers an interesting case since it illustrates the complex behavior of city dialects. Kortrijk is a provincial city of a modest size, but it has a long history as an urban and economic center in the south of West Flanders. Bruges is its counterpart in the northern part of the province. Kortrijk may look up to the prestigious city of Bruges, but seems to have been self-confident enough to resist innovations spreading from Bruges. For instance, Goossens (1992) discusses the diffusion of compound pronouns of the type of winder instead of wider ‘we’, with an inserted /n/ (derived from the object and possessive pronoun hunder ‘them’ or ‘their’), an innovation which appears to have originated in Bruges and consequentially spread to some smaller West Flemish towns like Roeselare, but was not adopted by the city dialect of Kortrijk. And yet, in its recent history the Kortrijk dialect in some respect converged to the western and northern West Flemish dialects, in a complex interaction of attraction and distancing and of hierarchical and contagious diffusion.

2.3. The explanatory orce o social network theory Social network theory as designed and applied by J. Milroy and L. Milroy offers a good model for interpreting patterns of hierarchical diffusion, with linguistic innovations hopping from one urban center to another, and for explaining why hierarchical diffusion is dominant over contagious diffusion (see J. Milroy and L. Milroy 1985; L. Milroy 1987; J. Milroy 1992). The explanatory value of social network theory for linguistic innovation is to be found in the use of the relative strength of network ties as a predictor of language change. The model is based on the idea that “to the extent that ties are strong, linguistic change will be prevented or impeded, whereas to the extent that they are weak, they will be more open to external influences, and so linguistic change will be facilitated” (Milroy 1992: 176). In close-knit communities strong network ties often function as a barrier against innovations which originate outside the network. Strong networks ties are “norm-enforcing mechanisms”. Milroy (1992: 176) adds: “behind this lies an idealization which predicts that in a community bound by maximally dense and multiplex networks ties linguistic change would not take place at all”. Such communities do not exist, but some communities are marked by relatively weak networks and others by relatively strong ones and that may explain why linguistic changes are more likely to affect the first type of community than the latter. On the level of the individual, the network model predicts that “mobile individuals who have contracted many weak ties, but who as a consequence of their mobility occupy a position marginal to some cohesive group, are in particularly strong position to carry information across social boundaries and to diffuse innovations of all kinds” (Milroy 1992: 180⫺181). When confronting these principles with the geographic diffusion of many linguistic innovations, it is easy to see why innovations tend to spread according to an urban hierarchy. Although in modern times we should not overemphasize the gap between urban and non-urban ways of living (cf. section 1 of this article), geographic and social mobility and the weak network ties

321

322

III. Structure and dynamics of a language space which are the product of both are still more typical of urban communities than of rural ones. Strong social networks can certainly be found both within urban and within rural communities. And so can people with weak network ties, occupying a marginal position with respect to one or more strong social networks, but it is beyond doubt that there are many more of them in urban centers than in rural communities. An innovation cannot be said to have been integrated in the speech of place x or y, if it is only used by individuals that are operating at the fringe of the community. The innovation should be adopted by a considerable part of the population, which means that it should be used by central members of the community as well. But it may seem unlikely that a central member, being part of a strong social network, is willing to converge towards the speech of a peripheral individual. This apparent bottleneck leads J. Milroy to add two “conditions” that are crucial for the practicability of network theory as an explanatory model of language innovation and change: first of all, peripheral people will only be successful in transmitting innovations to central members of the group if they are quite numerous. “The existence of numerous weak ties is a necessary condition for innovations to spread”, because “adopting an innovation that is already widespread on the fringes of the group is much less risky” for central members of the group than adopting an innovation which is used by a (very) small number of people weakly linked to the community (Milroy 1992: 181⫺182). Since transportation networks tend to link cities to cities and since there is much economic and cultural interaction between urban centers, it is easy to understand that this results in numerous weak ties between them. The individuals embodying these weak ties transmit innovations from city x to city y. They may be successful in getting them accepted if they are well-represented numerically in city y. Once again, this certainly does not imply that the non-urban hinterland of a city does not establish weak links with the urban center. But the people establishing (weak) ties both with the donor city and with the recipient hinterland may in many cases form too small a minority within the rural or at least non-urban community. The second necessary condition for an innovation to be adopted is a psycho-social one: “this is that speakers from the receptor community want to identify for some reason with speakers from the donor community” (Milroy 1992: 182). In other words, they should benefit from adopting an innovation. Speakers from a smaller town or city may adopt features from a larger city because they look up to that city and its urban life style (cf. section 2.2). By adopting a fashionable urban feature, people may aim at sharing the prestige of a large city. If however, by adopting an innovation, people will be considered disloyal in their local community, violating norms of group solidarity, they will not be inclined to do so. The pressure of local solidarity norms is likely to be stronger in nonurban than in urban communities, but again we should be careful not to present differences of this type in absolute terms: non-urban communities may be susceptible to the prestige of urban centers as well and that prestige may prevail over local solidarity norms. The differences between the urban and the non-urban as dealt with in this section are always relative, which may explain the dominance of the hierarchical diffusion of linguistic innovation, without excluding contagious diffusion. Network theory also appears to be a most valuable explanatory tool for explaining dialect maintenance versus dialect loss in contexts of large-scale migration. BortoniRicardo (1985) illustrates the correlation between social network indices and “dialect diffusion”. His subjects were rural migrants in the Brazilian city of Brazlaˆndia. Like other Brazilian cities, Brazlaˆndia was exposed to massive rural migration in the second

18. Urban and rural language half of the twentieth century. In the 1920s only ten percent of the Brazilian population lived in a city, in the 1980s this figure had risen to 67.6 percent (Bortoni-Ricardo 1985: 20). Whereas the studies of J. and L. Milroy focus on network strength as a predictor of language maintenance and change, with a strong focus on multiplex and dense networks as norm enforcement mechanisms or ⫺ using the terminology of Bortoni-Ricardo ⫺ mechanisms that support dialect focusing, Bortoni-Ricardo develops network indices designed to deal with dialect diffuseness. In his study, dialect focusing implies the use of “a highly focused form of rural vernacular, which contains virtually the whole set of non-standard features that define a sharp distinction between rural and urban varieties” (Bortoni-Ricardo 1985: 20). Diffuseness is not viewed necessarily as an assimilation of the standard language but rather as a movement away from the stigmatized rural dialects, which manifests itself as a gradual shift from isolated rural vernacular towards urban speech (Bortoni-Ricardo 1985: 105). In order to predict dialect diffusion, Bortoni-Ricardo uses an “integration index” and an “urbanization index”. The former measures the number of persons with whom his subjects have a personal relationship, the latter represents an urbanization profile of the members of each migrant’s personal network. These indices function as indicators “of the migrants’ process of transition from an insulated network of kinsfolk and pre-migration acquaintances to a more heterogeneous integrated network” (Bortoni-Ricardo 1985: 167). As migrants switch to more heterogeneous networks they are more exposed to mainstream urban culture and urban language. As a consequence their dialect becomes more “diffuse”. Bortoni-Ricardo shows that “the migrants’ urbanization is a complex process during which the migrant [is] permanently faced with the ambivalence between out-group identification, motivated mainly by pragmatic reasons, and need for the in-group psychological and social support” (Bortoni-Ricardo 1985: 119). Consequently, although the loss of rural vernacular due to massive migration into large metropoles is inevitable, it is no straightforward process.

3. Urban insularity 3.1. Three types o urban insularity As urban centers are generally ahead of their hinterland in adopting linguistic innovations, they often constitute islands on language maps which reflect the spatial diffusion of linguistic features. In section 2.2, several examples of this phenomenon were discussed: in the past two centuries h-dropping diffused from London to cities in East Anglia (e.g., Norwich) but not (or to a minor extent) to their hinterland, uvular /R/ spread to several cities in Flanders long before it reached the rural surroundings of these cities, etc. In many cases this so-called urban insularity is a temporary pattern, since hierarchical diffusion is often followed by contagious diffusion. Taeldeman (2005) presents three types of urban insularity, which are distinguished on the basis of the factors causing the isolated position of urban dialects. Urban insularity as the inevitable side effect of hierarchical diffusion constitutes one of these types. As the latter phenomenon was dealt with extensively in the preceding paragraph, we focus on the two other types in the rest of this section.

323

324

III. Structure and dynamics of a language space The second type of urban insularity discussed by Taeldeman (2005) is “innovative urban insularity”. In this case, “urban dialects can themselves generate an innovation” (Taeldeman 2005: 268). Cities often play a pioneering role, not only in adopting innovations from other cities, but also in generating innovations themselves. For instance, the Flemish city dialect of Ghent has been a very innovative dialect in its relatively recent history. Most of these innovations have been adopted by the hinterland and have to a large extent determined the make-up of the present day East-Flemish dialects. Quite surprisingly, however, some of these innovations which clearly originated in Ghent, were rejected by the city once they had reached the non-urban hinterland, thus giving the city its island-like position (see Taeldeman 2005: 271⫺273). This remarkable urge to keep up a distinct profile for the Ghent city dialect will also be apparent from the third type of urban insularity discussed below. Taeldeman (2005: 269) also refers to a special type of innovative urban insularity caused by the emergence of new towns and the consequential emergence of new dialects through a process of koineization: “As a result of enduring contact between relocated speakers of mutually intelligible varieties of the same language a “new” koine comes into being through a process of mixing, leveling, and simplification”. Of course this process, which does not take place in the hinterland of the new town, gives the new urban center an island-like position, at least for some time. In a later phase, the leveled variant may spread to the hinterland. Kerswill (2002) deals with new dialect formation through koineization in the case of Milton Keynes, a large town about 45 miles northwest of London (see also Kerswill and Williams 2000). Milton Keynes, designed as a new town in the 1960s to relieve housing congestion in London, attracted people from several towns in northeast England, thus stimulating the emergence of a linguistic koine. As it is situated within the sphere of influence of several cities, the most prominent one being London, and consequentially adopted linguistic features from those cities, Milton Keynes seems to combine innovative urban insularity with urban insularity caused by hierarchical urban diffusion of linguistic innovations (cf. the presence of certain London Cockney features or Estuary English features in Milton Keynes, section 2.2). A remarkable type of urban insularity is the third and final one to be discussed here, i. e., “conservative urban insularity”. In some cases urban dialects do not fulfill their prototypical pioneering role, but, on the contrary, resist innovations, even when these innovations have been adopted by the dialects of the hinterland. A striking example of urban conservatism can be found in the recent linguistic history of the German city of Hamburg. Goossens (1988) describes the gradual replacement of Low German dialect features by their High German counterparts in northern Germany in the twentieth century. The process started in the cities, with the bigger cities being ahead of the smaller cities, but Hamburg, the biggest city in northern Germany, showed strong resistance to the loss of Low German. This finally gave the city dialect an isolated position in the northern German linguistic landscape (see also Taeldeman 2005: 270). Conservative urban insularity also applies to the city dialect of Ghent. The realization of old intervocalic /sk/ in the East-Flemish dialects illustrates once again the peculiar behavior of this dialect. In most Flemish dialects old intervocalic /sk/ has been replaced by /s/. Verstegen (1942) found that a large area north and south of Ghent, including Ghent itself, preserved an intermediate stage of the development, /sx/ (e.g., in words like vissen ‘to fish’: vi[sk]en > vi[sx]en > vi[s]en). More recent data show that the intermedi-

18. Urban and rural language ate realization has been replaced by /s/, except in Ghent, which has resisted the final phase in the development of the cluster and still has intervocalic /sx/ (Taeldeman 2005: 276, 1985: 231).

3.2. Urban language as an urban identity marker Whenever the dynamics of city dialects, whether innovative or conservative, leads to urban insularity, the effect is that the city dialects (or their speakers) acquire a distinct profile with respect to the hinterland and that may be exactly their intention. In the interpretation of this recurrent dynamics, the socio-psychological dimension undoubtedly holds a prominent place. Since the pioneering study of W. Labov on language change on Martha’s Vineyard (Labov 1963), it is beyond doubt that displaying a “local identity” may be the driving force behind language behavior and language change. The notion of identity is a complex one, and often “identity” is reformulated as the “awareness of identity”, or the identification and recognition of a number of characteristics as being constitutive for an individual or a group. Language may be one of these characteristics, since it is not only an instrument to talk about one’s identity but also a component of the process of identity construction itself (Frijhoff 2004). For town people it may be an important marker of their “urban identity”, expressing their “großstädisches Selbstbewußtsein” (‘urban self-awareness’; Goossens 1988). This may lead to divergent behavior: Divergence is deemed a tactic of intergroup distinctiveness of individuals in search of a positive social identity. By diverging one’s communicative style […], members of an ingroup can accentuate the differences between themselves and a relevant outgroup on, usually, a highly valued dimension of their group identity. (Giles 2001: 195)

In practice, feelings of superiority combined with a positive evaluation of belonging to the same urban community often appear to lead to a collective distancing of the urban population from the dialects and vernaculars of the non-urban hinterland. Country people and people from smaller towns for their part often look up to the city life and all characteristics associated with urban life style (including the vernacular of the city) and are eager to adopt these characteristics (see sections 2.2 and 2.3). This process of distancing and assimilation leads to the continuous dynamics of innovation and change in all the varieties involved (Goossens 1992: 30). The importance of linguistic identity markers for creating a distinct urban profile has been attested both in older dialectological studies and in present day sociolinguistic research. To give just one example for the latter: Eckert (2000, and in this volume) discusses the language behavior of children and teenagers in the Detroit area. “Urban” and “suburban” for these youngsters are identity markers in many respects and “urban and suburban kids” can be discriminated on the basis of linguistic features. A very interesting symptom of the linguistic tension between several communities (whether urban or non urban) are mocking rhymes (Weijnen 1961). However, they do not function as descriptive evidence of dialect differences, since the speakers aim at distancing themselves from the “owners” of the vernacular which they try to portray. Thus, in former days, citizens of the West Flemish city of Bruges expressed the superiority of their dialect to that of the city of Kortrijk by producing the following nonsensical

325

326

III. Structure and dynamics of a language space sentence: ’t is een skande van die skone skorte te skeuren [Dutch: Het is een schande om die mooie schort te scheuren ‘it’s a shame to tear up that beautiful apron’]. In the sentence the cluster /sk/ is clearly overrepresented. This consonant cluster was a typical Kortrijk city dialect feature, which has spread to the hinterland by now. In former days, however, it functioned as a shibboleth of the Kortrijk vernacular. By overemphasizing this feature, the Bruges citizens distanced themselves from Kortrijk and its dialect and reinforced their own urban identity and their allegiance to their own dialect. In the past decades, the Kortrijk cluster [sk] spread to the smaller city of Roeselare (see Map 18.1), but it never conquered the city of Bruges (Vandekerckhove 2003). Town people may display a remarkable linguistic solidarity, not only in distancing themselves from other (smaller) towns, but even more so in distancing themselves from the non-urban hinterland. Preserving a structural distance between the city dialect and the dialects of the hinterland contributes to preserving a distinct urban, or non-rural or even non-peasant identity. This may explain urban conservatism with respect to language innovations present in the hinterland. It also explains why some cities get rid of features which once were markers of the city dialect but which have spread to the hinterland and consequentially can no longer fulfill their function as urban identity markers (see the case of Ghent discussed in section 3.1).

4. Rural dynamics 4.1. Rural prevailing over urban: Contrahierarchical diusion In the preceding sections it has been shown that linguistic innovations prototypically spread from urban to non-urban or less urban(ized) centers, along an urban hierarchy. Exceptionally, however, innovations diffuse in the opposite direction. Bailey, Wikle and Tillery (1993: 371⫺373), who describe the diffusion of the quasi-modal fixing to from rural to urban Oklahoma, label this pattern “contrahierarchical diffusion”. Trudgill (1986) presents another example: smoothing processes found in rural north Norfolk diffuse southwards to urban centers in the country of Suffolk (see Britain 2002: 626). The West Flemish city of Bruges has been subject to contrahierarchical diffusion as well. In former days the Bruges dialect was a trendsetting dialect, the innovations of which were eagerly adopted by the hinterland (cf. section 2.2). Though they are not well documented, there are some indications that in recent times many of the typical features of the old Bruges city dialect tend to be replaced by more “general” West Flemish features, present in the Bruges hinterland. The darkening of /e/ in front of /l/ (gald instead of geld ‘money’), for example, was a typical feature of the Bruges city dialect which was resisted by the hinterland. Nowadays, probably due to pressure from the hinterland, the city dialect itself has nearly given up this feature (Devos and Vandekerckhove 2005: 55). The pronouns with inserted /n/ (section 2.2) offer another example: they originated in Bruges and they are still well represented in the Bruges dialect, but nowadays they compete with their West Flemish counterparts without /n/ (Vandekerckhove 2005b). The motivation behind this relatively unusual process of city dialects being (partially) swallowed by the dialects of the (rural) hinterland may well be found in the status of the particular city dialect features. The darkening of [e] in the Bruges dialect was and is typically associated with working class speech. Middle class people moving into the city have not been willing

18. Urban and rural language to adopt it and the autochthonous middle class has probably been eager to drop it. But explanations like this remain tentative and rather ad hoc. The examples discussed in 2.2 show that the working class nature of urban features need not be an impediment for their survival and diffusion to other areas. So, most probably, there are other factors which contribute to the negative evaluation of these city dialect features. The make up of the population of the city may be one of them, but we cannot enlarge on that aspect here.

4.2. From local to regional dialects Urban language can generally be said to be more susceptible to change than rural language. However, this does not imply that rural language is static. Stable dialects, resisting any kind of external influence, are an unrealistic idealization. In former days (for many regions in western Europe until the beginning, for others until the middle of the twentieth century), the inhabitants of rural settlements operated in an area with a radius of five to maximally seven kilometers on a more or less daily basis, a distance which corresponded to approximately an hour’s walk, provided that there were no obstacles of whatever kind (rivers, mountains …). The neighboring settlements were generally situated within this distance: patterns of settlements with an average distance of about five to six or seven kilometers from each other were found in many western European regions, e.g., in certain regions of the Netherlands and in southern Germany (Hoppenbrouwers 1990: 16; Pacione 1984: 28; Dieleman 1971). Regional markets in the nearby town or city forced country people to cover longer distances, but they certainly did not do so on a daily basis and the maximal distance people tended to bridge for these markets was probably approximately 20 kilometers, though 10 or 15 kilometers may have been more common (see for Flanders Taeldeman 1987). So country people were not immobile, but their geographic mobility was very limited. Except for some villages that were completely isolated due to geographical circumstances (e.g., some mountain villages), most villages had at least some contacts with other rural or urban communities. This may explain why even in “pre-modern” times there was a dynamics of convergence and divergence between country dialects and city dialects, leading to changes in both. But these changes certainly did not take place at the same rate and with the same intensity as they do nowadays, since contact with other communities was not only much more limited than it is at present, the make up of the villages also guaranteed the preservation of local features: old villages were small, tightly knit and marked by strong social networks, which are very efficient at reinforcing local norms and resisting external influences (cf. 2.3). They offered the ideal habitat for the local dialect. These circumstances often brought about enormous dialect diversity within one and the same region (“every village having its own dialect”) and they guaranteed that this diversity was not threatened at all for centuries. Urbanization, counterurbanization, industrialization, increased mobility and increased ease of communication led to the dissolution of many old rural villages and consequentially to dialect contact to an extent hitherto unknown (Auer and Hinskens 1996: 4, see also section 1). In some cases rural or non-urban dialects simply appear to have been swallowed up by the city dialect of an expansive city, the inhabitants of which often moved to the neighboring towns or villages in search of space (i. e., so-called

327

328

III. Structure and dynamics of a language space “counterurbanization”). A well-known case in point is the city dialect of Antwerp: the former dialects of many of the (also former) villages neighboring Antwerp have completely disappeared. Nowadays, the dialect spoken in these places, many of which are highly urbanized, is the Antwerp city dialect (De Schutter and Nuyts 2005: 26). In many rural or formerly rural areas, however, the socio-economic processes named above, have not led to such a drastic dialect loss. In the past decades, all over western Europe, dialect contact in rural areas led to dialect leveling, i. e., a gradual reduction of interdialectal differences. The sociolinguistic and sociodialectological literature offers numerous examples. Vandekerckhove (2005b) shows that in the West Flemish dialect area dialect features with a limited geographical diffusion have been replaced by dialect features with a wider diffusion. This tendency could be illustrated for several phonological and morphological features. One is the diminutive morpheme. Of the many variants of the diminutive morpheme present in the West Flemish dialects in former days, only those with a relatively wide diffusion have “survived”. As Hinskens (1996) put it: in dialect geography, leveling results in an increase in scale. When communicating with peers (but not with their children), young West Flemish adults nowadays still use dialect, but dialect forms which are marked in supraregional contacts because of their limited diffusion do not longer belong to their idiolect. The dialect of the younger generations is not a copy of the dialect of their parents and grandparents. It is gradually developing into a dialect with a wider geographical range. People still operate in strong networks, but these networks are often less “local” than before. As a consequence the medium of communication is less local too. When old networks are disrupted, people seek to develop new network ties and these new ties may result in the “(re)establishment and subsequent social enforcement of a more focused koineized linguistic system” (Britain 2002: 616). Ortsloyalität (or ‘loyalty to one’s hometown’, cf. Mattheier 1980) in its narrow sense is replaced by a more general regional affiliation (cf. Giddens 1991). Research on dialect leveling in Norway also leads Røyneland to the conclusion that identification with the local town or village is decreasing. She assumes that regionalization explains why “mixed ways of speaking” are more accepted among young people: “The region gradually replaces the local community as the important linguistic and mental orientation unit ⫺ it becomes gradually more important to display regional affiliation and identity than strictly local” (Røyneland 2009: 8). Whether rural dialects are subject to leveling, koineization or mixing, or, in many cases all three of these, the outcome is the same: the old authentic village dialect becomes a relic of the past. Time and again studies on dialect change and dialect loss lead to the conclusion that dialects or regional varieties (“regiolects”, cf. Hoppenbrouwers 1990) still survive or even revive in Europe (cf. Auer 2005, Britain and Vandekerckhove 2009), but the small scale local village dialect marked by many features with a limited diffusion is definitely dying.

5. Reerences Auer, Peter 2005 Europe’s sociolinguistic unity, or: A typology of European dialect/standard constellations. In: Nicole Delbecque, Johan van der Auwera and Dirk Geeraerts (eds.), Perspectives on Variation: Sociolinguistic, Historical, Comparative, 7⫺42. Berlin: Mouton de Gruyter.

18. Urban and rural language Auer, Peter and Frans Hinskens 1996 The convergence and divergence of dialects in Europe. New and not so new developments in an old area. Sociolinguistica 10: 1⫺30. Bailey, Guy, Tom Wikle and Jan Tillery 1993 Some patterns of linguistic diffusion. Language Variation and Change 5: 359⫺390. Besch, Werner (ed.) 1981 Sprachverhalten in ländlichen Gemeinden. Forschungsbericht Erp-Projekt, vol. 1, Ansätze zur Theorie und Methode. Berlin: Erich Schmidt. Besch, Werner (ed.) 1983 Sprachverhalten in ländlichen Gemeinden. Forschungsbericht Erp-Projekt, vol. 2, Dialekt und Standardsprache im Sprecherurteil. Berlin: Erich Schmidt. Bartoni-Ricardo, Stella Maris. 1985 The Urbanization of Rural Dialect Speakers. Cambridge: CUP. Britain, David 2002 Space and spatial diffusion. In: Jack K. Chambers, Peter Trudgill and Nathalie SchillingEstes (eds.), The Handbook of Language Variation and Change, 603⫺637. London: Blackwell. Britain, David 2009 One foot in the grave?: Dialect death, dialect contact and dialect birth in England. International Journal of the Sociology of Language 196/197: 121⫺155. Britain, David and Reinhild Vandekerckhove 2009 Dialects in western Europe: A balanced picture of language death, innovation, and change. International Journal of the Sociology of Language 196/197: 1⫺6. Chambers, Jack K. and Peter Trudgill 1980 Dialectology. Cambridge: Cambridge University Press. De Schutter, Georges and Jan Nuyts 2005 Stadsantwerps [The Antwerp City Dialect]. Tielt: Lannoo. Devos, Magda and Reinhild Vandekerckhove 2005 West-Vlaams [West Flemish]. Tielt: Lannoo. Dieleman, Frans M. 1971 De centrale plaatsen theorie van Christaller [Christaller’s theory of central places]. Geografisch Tijdschrift 5, 543⫺563. Eckert, Penelope 2000 Linguistic Variation as Social Practice: The Linguistic Construction of Identity in Belton High. Malden/Oxford: Blackwell. Eckert, Penelope 2005 Variation, convention, and social meaning. Paper presented at the annual meeting of the Linguistic Society of America, Oakland CA January 7. Available from . Foulkes, Paul and Gerard Docherty (eds.) 1999 Urban Voices: Accent Studies in the British Isles. London: Arnold. Frijhoff, Willem 2004 Hoe talig is groepsidentiteit? Reflecties vanuit de geschiedenis [How linguistic is group identity? Reflections from history]. Taal en Tongval 17 [Taalvariatie en groepsidenteit]: 9⫺29. Giddens, Anthony 1991 Modernity and Self-Identity. Self and Society in the Late Modern Age. Cambridge: Polity Press. Giles, Howard 2001 Speech accommodation. In: Rajend Mesthrie (ed.), Concise Encyclopedia of Sociolinguistics, 193⫺197. Amsterdam/New York: Elsevier.

329

330

III. Structure and dynamics of a language space Goossens, Jan 1988 Zur Lage des Niederdeutschen und ihrer Erforschung. Michigan Germanic Studies 12: 1⫺20. Goossens, Jan 1992 Dialecten in het centrale Zuidnederlandse stedennetwerk [Dialects in the central southern Dutch city network]. Taal en Tongval 5 [Stadsdialecten]: 29⫺47. Hamilton, Anne Marie 2001 Urban and rural forms of language. In: Rajend Mesthrie (ed.), Concise Encyclopedia of Sociolinguistics, 418⫺419. Amsterdam/New York: Elsevier. Heike, Georg 1964 Zur Phonologie der Stadtkölner Mundart. Marburg: Elwert. Hinskens, Frans 1996 Dialect Levelling in Limburg. Structural and Sociolinguistic Aspects. Tübingen: Niemeyer. Hoppenbrouwers, Cor 1990 Het Regiolect. Van Dialect tot Algemeen Nederlands [The Regiolect. From Dialect to Standard Dutch]. Muiderberg: Coutinho. Hornsby, David 2007 Regional dialect levelling in urban France and Britain. Nottingham French Studies 46(2): 64⫺81. Hornsby, David 2009 Dedialectalization in France: Convergence and divergence. International Journal of the Sociology of Language 196/197: 157⫺180. Kerswill, Paul 2002 Koineization and accomodation. In: Jack K. Chambers, Peter Trudgill and Nathalie Schilling-Estes (eds.), The Handbook of Language Variation and Change, 669⫺702. London: Blackwell. Kerswill, Paul and Anne Williams 2000 Creating a new koine: Children and language change in Milton Keynes. Language and Society 29(1): 65⫺115. Labov, William 1963 The social motivation of a sound change. Word 19, 273⫺309 Mattheier, Klaus J. 1980 Pragmatik und Soziologie der Dialekte: Einführung in die kommunikative Dialektologie des Deutschen [Pragmatics and Sociology of Dialects: An Introduction to Communicative Dialectology of German]. Heidelberg: Quelle und Meyer. Mees, Inger M. and Beverley Collins 1999 Cardiff: A real time study of glottalization. In: Foulkes and Dochterty (eds.), 185⫺202. Milroy, James 1992 Linguistic Variation and Change. On the Historical Sociolinguistics of English. Oxford: Blackwell. Milroy, James and Lesley Milroy 1985 Linguistic change, social network and speaker innovation. Journal of Linguistics 21: 339⫺384. Milroy, Lesley 1987 Language and Social Networks. Oxford: Blackwell. Orton, Harold, Stewart Sanderson and John Widdowson 1978 The Linguistic Atlas of England. London: Routledge. Pacione, Michael 1984 Rural Geography. London: Harper and Row. Rogier, Doreen 1994 De verspreiding van een sociaal hooggewaardeerd taalkenmerk: de huig-R rond Gent. [The diffusion of a language feature with social prestige: Uvular R in the Ghent hinterland]. Taal en tongval 7 [R⫺zes visies op een kameleon]: 45⫺53.

18. Urban and rural language Røyneland, Unn 2005 Dialektnivellering, ungdom og identitet. Ein komparativ studie av spra˚kleg variasjon og endring i to tilgrensande dialektomra˚de, Røros og Tynset. [Dialect leveling, youth and identity. A comparative study in language variation and change in two neighboring dialect areas, Røros and Tynse]. Unpublished PhD dissertation, Faculty of Humanities, University of Oslo. Røyneland, Unn 2009 Dialects in Norway ⫺ Catching up with the rest of Europe? International Journal of the Sociology of Language 196/197: 7⫺30. Sivertsen, Eva 1960 Cockney Phonology. Oslo: Oslo University Press. Taeldeman, Johan 1985 De klankstruktuur van het Gentse dialect. Een synchrone beschrijving en een historische en geografische situering [The phonological structure of the dialect of Ghent. A synchronic description against a historical and geographic background]. Ghent: Rijksuniversiteit. Taeldeman, Johan 1987 Extern-linguı¨stische achtergronden van het Vlaamse dialektlandschap. [The external linguistic basis of the Flemish dialect landscape]. Leuvense Bijdragen 76: 315⫺334. Taeldeman, Johan 2005 The influence of urban centres on the spatial diffusion of dialect phenomena. In: Peter Auer, Frans Hinskens and Paul Kerswill (eds.), Dialect Change: Convergence and Divergence in European Languages. Cambrigde: Cambridge University Press. Tops, Evie 2005 Variatie en Verandering van de /r/ in Vlaanderen [Variation and Change of /r/ in Flanders]. Unpublished PhD dissertation, Faculty of Humanities, Vrije Universiteit Brussel. Trudgill, Peter 1974 Linguistic change and diffusion: Description and explanation in sociolinguistic geography. Language in Society 1: 179⫺195. Trudgill, Peter 1983 On Dialect: Social and Geographical Perspectives. Oxford: Blackwell Trudgill, Peter 1986 Dialects in Contact. Oxford: Blackwell. Trudgill, Peter 1988 On the role of dialect contact and interdialect in linguistic change. In: Jacek Fisiak (ed.), Historical Dialectology. Regional and Social, 547⫺563. Amsterdam: Mouton de Gruyter. Trudgill, Peter 1999 Norwich: Endogenous and exogenous linguistic change. In: Foulkes and Docherty (eds.), 124⫺140. Van Coetsem, Frans 1988 Loan Phonology and the Two Transfer Types in Language Contact. Dordrecht: Foris. Vandekerckhove, Reinhild 2000 Structurele en Sociale Aspecten van Dialectverandering. De Vitaliteit van het Deerlijkse Dialect [Structural and Social Aspects of Dialect Change. The Vitality of the Dialect of Deerlijk]. Ghent: Koninklijke Academie voor Nederlandse Taal- en Letterkunde. Vandekerckhove, Reinhild 2003 West-Vlaams dialectbewustzijn van Noordzee tot Leie [West Flemish awareness of dialect differences from North Sea to Leie]. In: Veronique De Tier and Reinhild Vandekerckhove (eds.), Aan Taal Herkend. (Het dialectenboek 7.) Groesbeek: Stichting Nederlandse Dialecten. Vandekerckhove, Reinhild 2005a Interdialectal convergence between West-Flemish Urban Dialects. In: Nicole Delbecque, Johan Van der Auwera and Dirk Geeraerts (eds.), Perspectives on Variation, 111⫺127. (Trends in Linguistics. Studies and Monographs 163.) Berlin: Mouton de Gruyter.

331

332

III. Structure and dynamics of a language space Vandekerckhove, Reinhild 2005b Patterns of variation and convergence in the West-Flemish dialects. In: Eckhard Eggers, Jürgen Erich Schmidt and Dieter Stellmacher (eds.), Moderne Dialekte ⫺ Neue Dialektologie, 535⫺552. Stuttgart: Steiner. Verstegen, V. 1942 De Westgermaanse sk in de Zuidnederlandse dialecten [Westgermanic sk in the southern Dutch dialects]. Handelingen Koninklijke Commissie voor Toponymie en Dialectologie 16: 31⫺42. Watt, Dominic and Lesley Milroy 1999 Patterns of variation and change in three Newcastle vowels: Is this dialect levelling? In: Foulkes and Dochterty (eds.), 45⫺46. Weijnen, Antonius A. 1961 Het Bewustzijn van Dialectverschil [The Awareness of Dialect Differences]. (Voordrachten gehouden voor de Gelderse leergangen te Arnhem, 5.) Groningen: Wolters. Wells, John C. 1982 Accents of English. 3 vols. Cambridge: Cambridge University Press.

Reinhild Vandekerckhove, Antwerp (Belgium)

19. Discontinuous language spaces (Sprachinseln) 1. 2. 3. 4. 5. 6. 7. 8.

Definition The history of language island research (Sprachinselforschung) The history of language islands The current state of language islands Factors determining language maintenance in discontinuous language spaces Other types of discontinuous language spaces Conclusion References

1. Deinition The concept of Sprachinsel is difficult to define. As Mattheier (1996: 812) points out, the numerous attempts which have been made to devise a definition are closely connected with certain research traditions on the one hand and with the history of language islands on the other (see section 2). A widely accepted definition among German linguists was suggested by Wiesinger (1983: 901), who asserts that language islands are either discrete or areal and are relatively small, enclosed linguistic settlements situated within a relatively large area where another language is used (for a similar definition, see Hutterer 1982). Mattheier (1996), however, questions some of the criteria specified in this definition. For example, Wiesinger’s restriction to either rural communities or miner settlements tends to overlook the importance of urban settlements, which already existed in the

19. Discontinuous language spaces past and have become more common in the modern era. Moreover, the construct of Geschlossenheit (‘unity’) includes the notion of a linguistically and ethnically homogenous area. Many settlements, however, are characterized by multiethnic constellations and are primarily defined in terms of tightly-knit socio-communicative networks and an awareness of ethnic difference (e.g., in Hungary). Other definitions stress the aspects of “isolation” and “roofing” by a majority language, e.g., Protze (1995: 55). In this definition, language islands are mainly characterized by minimal contact with not only the motherland but also the surrounding speech community. The minority language is roofed by the standard variety of the majority language and the communities are linguistically and culturally isolated. The roofing argument, however, is not historically valid. Prior to the Second World War, a stable diglossia existed in many German-speaking language islands throughout Eastern Europe. This diglossia included dialectal German and Standard German; even speakers of the majority group had a solid command of the German language (Riehl to appear; Mattheier 1996: 814; for further definitions Földes 2006). The difficulties in defining a Sprachinsel are primarily due to the fact that the term cannot be readily detached from the historical-political constellations and social changes connected with the emergence of language islands (i. e., the colonization of Eastern Europe and the New World, cf. Mattheier 1996: 815). With the rapid change of these constellations and conditions which took place after World War II, some of the defining criteria became inadequate (see section 4.1). If, however, the notion Sprachinsel is not to be considered obsolete (a suggestion posed by Földes 2006), the focus should be placed on other, mainly sociolinguistic aspects. As Mattheier (1994: 334) suggests: A language island is a linguistic community which develops as a result of disrupted or delayed linguistic cultural assimilation. Surrounded or enclosed by a linguistic and/or ethnic dominant culture, such an island is made up of a linguistic minority that has become separated off from its original roots; and, as a function of its unique socio-psychological disposition (i. e., its “island mentality”) is held separate and apart from the majority culture with which it maintains tangential contact. [Translation by Iman Laversuch]

Mattheier (1994: 335) argues that a language island cannot only be defined in terms of linguistic and dialectological criteria. The main characteristic is a “delayed assimilation” which is based on many different factors. Chief among them is the speakers’ attitude toward the linguistic majority. Therefore, language islands exist “rather in the minds of speech-islanders than on the landscape” (Mattheier 1994: 106; see also Eichinger 2003). Mattheier even goes on to assert that members of language islands are able to mark their ethnic identity without a language of their own, but through its reflexes in vocabulary, intonation as well as personal and place names. As attitudinal factors are also relevant for other types of minorities, it is important to consider the ways in which language islands differ from them. Edwards (1990) distinguishes between three different types: (i) minorities who are unique within one nation (Bretons); (ii) non-unique minorities dispersed across several states (Catalans, Basques); (iii) non-local minorities who are minorities in one setting, but a part of the majority elsewhere. Further parameters are adjoining vs. non-adjoining and cohesive vs. noncohesive. According to this typology, language islands can be defined as “non-local, nonadjoining minorities”. They have to be distinguished mainly from two other types of minorities (cf. Riehl 2004: 55⫺57):

333

334

III. Structure and dynamics of a language space 1. ethnic minorities without motherland (e.g., Sorbs, Ladinians or Bretons), who are defined as unique and non-adjoining. They are old settlements that were excluded from the nation-building processes and therefore do not have a linguistic motherland that provides a standard variety and literary tradition. Consequently, they have to codify their own variety as a written language (cf. Pusch 2005). By comparison, language islands can use an external standard and may even converge toward it. They are influenced by language developments in the motherland and by its language policy (e.g., support of a minority school system, see Mattheier 2002: 132). 2. so-called “border minorities” (e.g., the German-speaking minorities along the German-Romance linguistic border and the Hungarian minorities located in various countries bordering on Hungary, Romania, Croatia). They are defined as non-local and adjoining and have constant contact with the motherland. As a result, they can establish communicative networks across borders. While there is a clear difference between language islands and ethnic minorities without linguistic motherland, the distinction between language islands and border minorities is less clear-cut. Mattheier (1996) suggests using the concept of Sprachhalbinsel ‘language peninsula’ for this type of minority. Moreover, the factor of “geographical isolation”, which had been valid for centuries, is no longer crucial for non-adjoining minorities (see section 5). At this point, another important issue in the definition of language islands comes into play: the relationship between language and space. In this case, one has to differentiate between “minority” and “regional” languages. Linguistic minorities can be described in terms of ethnic and religious membership, whereas regional languages are defined by the region in which they are spoken (cf. Wirrer 2000). Therefore, speakers of regional languages are a part of the ethnic majority and adopt an additional regional identity (e.g., speakers of the Occitan or Gaelic language). Inhabitants of language enclaves usually describe themselves as constituting a distinct ethnic group, belonging to the ethnos of the linguistic motherland. As a consequence, their ethnic origin becomes a fixed component of the community’s name. They call themselves, for example, Ungarndeutsche, Rumäniendeutsche, Teuto-Brazileiros or Texas Germans. This pattern does not, however, hold for language islands that have lost their visible connection to the motherland. An example here would be the so-called Cimbri in Venetia (see Hornung 1994; Rowley 1996). In cases such as these, where mixed language varieties are emerging, the distinction between language islands and ethnic minorities becomes exceptionally fuzzy. In conclusion, two main characteristics crystallize out of the definition criteria of language islands: (i) the relationship to (or dependence on) a linguistic motherland. This can be seen as a form of spatial isolation from the linguistic motherland which entails a discontinuous language space; and (ii) the socio-psychological disposition of language island communities or in other words their collective “awareness of being different” (Mattheier 1996: 815). A special situation arises when language islands find themselves within the language space of other linguistic minorities. This, for example, is the case with the Walser colonies in the Val’ d’Aoste region in North-West Italy. The region in question is an autonomous French-speaking province within an Italian-speaking country (see Zürrer 1999). An even more striking example is that of the Landler communities, an Austrian dialect enclave within the Low German based language island of Transylvania (Schabus 1994: 1996). In most of these cases, we find a complex polyglossic situation, where the language island

19. Discontinuous language spaces population alternates between using its enclave variety, the regional minority varieties and the majority language. Another special case is the so-called “dialect islands” (i. e., a dialect enclaves within a dialect space where a different variety of the same genetic origin is spoken). Two examples of this type are the Palatine dialect enclave located within the dialect area of Eastphalia (Karch 1978) and the Ligurian dialect enclaves situated in Sicily (see section 3.2).

2. The history o language island research (Sprachinselorschung) The sociolinguistic phenomenon of a Sprachinsel is by no means a typical German phenomenon. The term Sprachinsel was, however, coined in the German research tradition. It was later borrowed into English where it emerged as language enclave and language pocket. Language island studies have remained a well-established field of research within German dialectology, including dialect geography and areal-linguistic methods, as well as the neogrammarian tradition (see Murray in this volume) and Kulturraumforschung (cf. Knobloch in this volume). The latter concentrated mainly on Middle and Eastern Europe (cf. Rosenberg 2003: 200). The concept of Sprachinsel is therefore strongly connected with the German research tradition. By contrast, non-German approaches generally subsume this type of minority groups within the category “linguistic minorities” (cf. Mattheier 1996: 812; Edwards 1990). The notion of Sprachinsel first appeared in 1847 in order to describe a Slavonic speech community within a German speaking area near Königsberg (Eastern Prussia). At this time, the term Sprachinsel competed with the term Sprachkolonie (cf. Mattheier 1996: 812) which refers to the historical context in which language island phenomena were first discussed, i. e., the colonization of the East (Ostkolonisation). Since the beginning of the twentieth century, Sprachinsel has become more popular than the notion Sprachkolonie, which is still in use in the Romanophone research tradition (see Kattenbusch 1998). (For other terms such as Enklaven, Sporaden, Außensiedlungen, etc., see Geyer 1999). The main focus of language island research was initially on the origin and history of these enclaves, aspects of areal linguistics, processes of language merger and analysis of the dialects used in the respective communities (see Hutterer 1982). A major objective at that time was to describe language change. On account of their isolation from their respective native countries, these communities were also supposed to have preserved linguistic phenomena that had either receded or completely disappeared in the German inland area, as Gehl (2000: 292⫺293; translation by Iman Laversuch) asserts: The existence of language islands provide a detailed demonstration of the developmental pathway taken by language. Similarly, language island variants often help address difficult questions concerning the historical development of minority languages located within Germany. Thanks to their uncommon resilience, these sociolinguistic communities have a strong tendency to retain for much longer periods of time antiquated linguistic forms which were long since lost in German minority languages, due to their own penchant for linguistic innovation and change.

Studies carried out during the first period of research were conducted by German dialectologists. One of the pioneers of dialectology, Andreas Schmeller, included the dialects

335

336

III. Structure and dynamics of a language space of language islands in Venetia (the so-called Sette communi) in his dictionary of the Bavarian language. Research concentrated on old language islands like Transylvania, the Zips (former Hungary) and Venetia (Italy), the first descriptions of which go back to the sixteenth and seventeenth centuries (cf. Kuhn 1934). In the nineteenth century, language islands which had developed more recently in Slovenia and Hungary/Romania (Gottschee, Banat) became subjects of investigation; and from 1870 on, language islands in Pennsylvania and Australia were included (see Rosenberg 2003: 200). Since the beginning of the twentieth century, especially in Russia, so-called Arbeitsstellen für Sprachinselforschung were founded by, among others, Schirmunski, Dinges and Dulson (see Berend and Jedig 1991). These research units aimed to explore the laws governing dialect merger, the chronology of language change and dialect contact. In addition, they were interested in identifying reasons for the emergence of new dialect borders and the influence of the Russian language (see Berend and Jedig 1991; Domaschnew 2000). A similar tradition, the so-called Donauschwäbische Dialektologie, was established by Gideon Petz in Hungary at the end of the nineteenth century and was carried on by Claus Hutterer and Karl Manherz (see Schwob 1971; Gehl 2000). Between the World Wars, the number of studies on German language island varieties increased rapidly ⫺ most of the research, however, was highly ideologically biased (cf. Knobloch in this handbook, for some reasons). Early studies tried to reconstruct the Urheimat, or origin, of the respective varieties and considered the dialects spoken in language islands as “pure”, “uncontaminated” and “homogenous”. Researchers of the Russian Arbeitsstellen für Sprachinselforschung, however, found that most of the linguistic spaces under investigation not only used a mixture of different dialects, but also developed their own koine´, especially in the socalled “daughter” colonies of the old settlements. It is against this background that Schirmunski (1930) introduced the discussion of “dialect levelling” (Dialektausgleich) and the distinction between primary and secondary dialect features. Primary features are defined as being the most salient and those which are most often avoided when communicating with speakers of other dialects. In dialect contact, they are also the first to disappear. By comparison, secondary features are preserved in communication among speakers of different dialects and do not disappear in contact situations. The Schirmunski distinction is somewhat problematic, given the fact that salience is not defined by linguistic criteria, but by the self-evaluation of the speakers (see Rosenberg 2003). Sometimes speakers also retain emblematic elements to mark their distinctness (see Asfandiarova 1999). Moreover, if Schirmunski refers to the difference which exists between the speaker’s “way of speaking and the norm of the written language” (Rosenberg 2005: 225), that means that the salience of features is defined with in the context of the written code. However, at the time of dialect merger in the Russian German colonies, standard (written) German was rarely available. The main geographical focus of dialect-based research was on language islands in Russia, Hungary, Romania, Northern Italy and Slovenia. These studies are mainly indebted to traditional dialectology research which relied on questionnaires, word-geography and dialect atlases (e.g., Berend 1997; Naiditsch 2006; Gehl 2000; and almost 20 volumes in the series Beiträge zur Sprachinselforschung edited by Maria Hornung). However, this kind of research now faces considerable challenges given the rapid decrease of German speakers in the island areas (see section 5). Recently, there has been an emergence of studies focusing on urban vernaculars (Gadeanu 1998; Gerner, Glauninger and Wild 2002).

19. Discontinuous language spaces As mentioned above, the paradigms of language island research are strongly connected with historical developments. In the nineteenth century many language islands were dialectal enclaves that had minimal contact with a standard variety (whether genetically related or not) and a low rate of literacy. In modern times, there has been a steady increase not only in the accessibility of the genetically related standard (in this case New High German), but also the contact with the surrounding majority language. This development has led to diglossic or triglossic situations in which the majority language has increasingly replaced Standard German as a high-prestige variety (see section 4.2). More recent research on language islands therefore must take multiple factors into account: linguistic and sociolinguistic changes of language island dialects; language mixture and language contact with relevant surrounding and contact languages; communicative structures within the island communities; and language attitudes of the speakers. This was particularly true for language island research on German in the last decades of the twentieth century in the United States (Louden 1994, 2003; Keel 2003), in Latin America (Kaufmann 2004, Auer, Arnhold and Bueno-Aniola 2007) and in Australia (Clyne 1994; Kipp 2006). More recent studies on language islands in Eastern Europe and Upper Italy include Blankenhorn (2003), Deminger (2004), Gerner (2006), KnipfKomlo´si (2003), Zürrer (1999) and Boas (to appear). Furthermore, new research paradigms have also combined sociolinguistic parameters (identity) and linguistic development features (Auer, Arnhold and Bueno-Aniola 2007; Kaufmann 2004). Recent attempts to compare historical, linguistic and sociolinguistic developments and language attitudes across different island constellations are Rosenberg (2003) on Russia and Brazil, Kaufmann (1997, 2004) on Mennonite groups in the Americas (see also Kaufmann in this volume) and Eichinger, Plewnia and Riehl (2008) on the Germanspeaking islands in Middle and Eastern Europe. In contrast to the German research tradition, studies on other non-local minorities have not primarily focused on dialectological issues, but on language contact and sociolinguistic issues (language maintenance and language shift). Dialectological research on rural communities that originated in the nineteenth century (e.g., Pugliese in Crimea, Friulian dialects in Romania, Venetian dialects in Mexico) only emerged in the 1970s and has been more sporadic than systematic (see Corra` and Ursini 1998; Wössner 2002).

3. The history o language islands The fact that there are German-speaking settlements in so many parts of the world remains a unique phenomenon of this particular ethnic group. The reasons are mainly historical: in Europe there has been nothing comparable to the German colonization of the East. While German settlers emigrated in the Middle Ages mainly as a reaction to overpopulation not experienced in any other European country, there was a second major reason for modern German based emigration to Russia and Eastern Europe ⫺ a ban imposed on the recruitment of settlers by many other European countries (Gottas 1995). Also the foundation of rural settlements overseas remains a mainly German phenomenon (with the exception of Italian settlers, see below). An explanation might be that other major European nations (e.g., France, England, Spain) were colonial powers which had colonies at their disposal to which settlers could be sent.

337

338

III. Structure and dynamics of a language space Nevertheless, there are colonies all over the world. The first examples date back to antiquity and include, for example, the Greek colonies in Asia Minor. Many settlements have naturally disappeared over time and traces of their existence can be found sporadically in select lexemes of the majority language such as place names. Such is the case with many medieval settlements like those of the Mozarabian Christians in Northern Spain (see Kattenbusch 1998). Other interesting groups in the context of European migration include the Albanians who from the thirteenth century immigrated to Greece (i. e., the so-called “Arvanites”, see Sasse 1998) and to Southern Italy (Calabria, Sicily, cf. Breu 2005) or the Hungarians who settled in Slovakia, Romania and on the Balkans. Moreover, there are language colonies outside Europe that meet the criteria of discontinuous language spaces, e.g., enclaves of Portuguese in West Africa, of French in Southern Africa, Urdu-speaking communities among Hindi minorities in Northern India (see Rosenberg 2005: 222). However, as mentioned above, these communities are mainly discussed as “linguistic minorities” and there is no common research tradition comparable to the German Sprachinselforschung (see section 2). Moreover, many of these groups only partly meet the criteria defining language islands (e.g., they have lost contact to the motherland and developed their own language and identity). Thus, in this survey, the focus will be on German-speaking islands, with some supplementary examples from Romance-speaking minorities (see section 3.2).

3.1. The history o German-speaking language islands Aside from missionary activities or famines, the major push factors for emigration in the Middle Ages are of social origin. The restructuring of the feudal and economic system brought many farmers to the verge of ruin (Göllner 1979: 19⫺20). In the twelfth and thirteenth centuries, almost 400,000 settlers relocated to the east of the Elbe and Saale Rivers, and moved on to Bohemia, Poland and the Baltic. Smaller groups settled in Transylvania and the Zips, following the invitation of the Hungarian kings. The socalled “colonization of the East” was fostered by lords of the manor who wanted to take advantage of the progressive agrarian techniques and economic organization of the German settlers. In exchange they offered certain social privileges (Gottas 1995: 16). In Transylvania, the Siebenbürger Sachsen (Transylvanian Saxons) were even granted the status of a nation with autonomous economic, clerical and cultural institutions. On the contrary, medieval emigration to upper Italy (Val d’Aosta, Venetia ⫺ Sette e tredici Communi) was only partly initiated by the bestowal of privileges by, for example, the bishops of Vicenza. It rather consisted in a trans-border movement to the Southern edge of the Alps. The immigrant populations mainly settled in completely isolated, unpopulated regions in the alpine valleys or plateaus. Due to this extreme isolation, most of the population, mainly females, remained monolingual in the settler variety up to the twentieth century (see Hornung 1994; Pohl 2005). A second wave of emigration started at the end of the seventeenth century and continued until the end of the nineteenth century, when many settlements were founded in Southeastern and Eastern Europe (e.g., in Hungary, Northern Romania, the Banat and Bacˇka; in Russia in the Volga region and the regions around St. Petersburg and the Black Sea). This movement has been characterized as a population movement that was planned and guided from above (Gottas 1995: 19). It was initiated by the Austro-Hung-

19. Discontinuous language spaces arian monarchs and the Russian empress Catherine the Great. As settlers came from very different parts of Germany, these areas became home to a unique diversity of dialects that triggered different convergence processes. At the same time, people emigrated overseas to the Americas, Australia and South Africa. The majority of them settled in towns and assimilated very quickly (Elitenwanderung and Arbeitswanderung, cf. Rosenberg 1998: 263). But there was still the so-called Siedlungswanderung (Rosenberg 1998), as in previous centuries: In South America, settlers were offered land and privileges by the big landowners, especially in Brazil and Argentina, in order to facilitate the development of the tropical forest, the establishment of military boundaries and the forging of ties between Europe and America. Emigration into rural settlements in the USA was additionally motivated by the desire for religious freedom. Such was the case, for example, of the Amish communities who moved to Pennsylvania in 1730 (for a general overview, see Born and Dieckgießer 1989; on Latin America, Rosenberg 1998). Further interesting developments originated from the migration of inhabitants of language islands to other environments. In 1877 and 1878, German settlers from the Black Sea (especially the Mennonites) moved to Argentina. At the end of the nineteenth century, Russian Germans from the Volga region also came to Argentina and Kansas (Berend 2003). Mennonites who had settled in Ukraine, first moved to Canada and then to Mexico. This translocation was motivated by the worsening political conditions that impeded the open expression of their religious ideas and obliged them to do military service. These spiritual communities are of particular interest, first because of the ways in which they have managed to maintain an archaic variety of Standard German for liturgical purposes; and second because they have adopted elements from different contact languages and have thus developed different attitudes toward the majority society (see Kaufmann 1997, 2004). Depending on the age of the settlement, the separation from the motherland and the heterogeneity of linguistic varieties spoken language islands caused different diatopic varieties to develop: from local vernaculars (colonies in Upper Italy, Slovenia) to merged or leveled dialects (Hungarian varieties, e.g., different Donauschwäbisch dialects) to koine´s (Hunsrückisch in Brazil, Volga German).

3.2. Migration o Romance-origin settlers During the Middle Ages, we find population movements in many places of the Romance world, but primarily in Italy. People from Piedmont and Liguria moved to Southern Italy and Sicily in the twelfth and thirteenth centuries and settled in dialect enclaves among a majority which spoke a different dialect. This migration wave was triggered by the social and economic disaster in the Ligurian hinterland on the one hand and exemptions and other privileges provided by the Almerian kings on the other (see Trovato 1998: 556). As Kuhn (1934: 388) mentions, some French language islands even preceded the German settlements in Eastern European trading towns (Bratislava, Prague etc.), but these islands assimilated quite quickly. A religious group of Waldensian origin which spoke Provenc¸al varieties immigrated to Calabria in the fourteenth century. This community was able, despite a ban imposed by the Inquisition, to retain its language for a considerable period of time thanks to

339

340

III. Structure and dynamics of a language space geographical and social isolation. The Franco-Provenc¸al language islands of Faeto and Celle in Apulia also go back to the fourteenth century. Their specific origin is, however, unclear (see Kattenbusch 1998: 532⫺534). The difference between the Romance and the German migration in the Middle Ages does not only lie in different push and pull factors, but also in the fact that all of these settlers moved to another Romanophone region. As a result, the genetic relationship between the minority and majority languages may have accelerated the interference or even shifting processes. In this case, it is also difficult to differentiate between language islands and dialect enclaves (cf. for instance the Ligurian settlers in Sicily). Exceptions are the immigration of Romanian shepherds into the Balkan region (Kattenbusch 1998: 536) and religiously motivated immigrants such as the French Huguenots in other European Protestant countries (who immigrated later to the USA and South Africa). Although they normally assimilated very quickly, some Huguenot communities in Germany were able to preserve their language and traditions until the Prussian invasion of 1866 (Corra` and Ursini 1998: 561). The large wave of emigration from Romance-speaking countries started at the beginning of the nineteenth century. It especially involved movement from Southern Europe into the Americas. The immigration to Middle and South America is comparable to the German one: recruited by agents, settlers mainly populated rural areas and were largely composed of people from Northern Italy (see Lo Cascio 1987; Corra` and Ursini 1998). In Rio Grande do Sul (Brazil), Italians even developed a Venetian-based dialectal koine´ (Corra` and Ursini 1998). Most of these rural settlements were founded in virgin environments in complete isolation and have been able to preserve their local dialects up to today (see Wössner 2002 on a Venetian language island in Mexico). Nevertheless, the vast majority of migrants, especially those from the beginning of the twentieth century on, did not settle in rural but in urban environments (see section 6).

4. The current state o language islands 4.1. Sociolinguistic perspectives Today the original dialectal varieties spoken in the island communities are likely to become extinct. This holds true to a great extent for the German-speaking language islands in Eastern Europe, where, in the aftermath of the Second World War, more than 90 percent of the German-speaking population was expelled. This dramatic development had three main consequences (see Riehl 2006): 1. Due to the emigration of German-speaking settlers and the immigration of majority language speakers, the number of communicative contexts for second-language usage increased, thereby facilitating language contact and code-switching. 2. Speakers from villages using different dialectal varieties came into contact, creating new contexts of variety contact and dialect convergence. 3. The significant increase in the number of interethnic marriages in which children were raised monolingually in the majority language and were consequently only able to develop a passive command of the minority dialect.

19. Discontinuous language spaces In many Eastern European language islands, language use among the generations is as follows. The oldest generation (born before 1932) has at its disposal the German dialect and (regional) standard German as L1-system and a relative command of L2. The second generation (born between 1932 and 1952) speaks both German dialect and the majority language as their L1, but Standard German is maintained only in rudimentary form. The following generation (born between 1952 and 1975) uses the majority language as its L1 and some interlanguage version of Standard German as L2. Those born after 1975, who are also majority language L1-speakers, might (depending on the respective access at school) have a very good, sometimes near-native command of Standard German, but only as a L2 high-prestige variety that is not usually used in in-group conversations (see Knipf-Komlosı´ 2003 and others). The exception in the Eastern European context is Romania, where schooling in the mother tongue was permitted despite major difficulties, during the communist regime. Here, the dialect is being increasingly replaced by the German Standard. The reason for this development is the balanced bilingualism (German⫺Romanian) or trilingualism (German⫺Romanian⫺Hungarian) of the minority speakers that has caused the reduction of the complex pluriglossic system Standard German⫺Saxon dialects⫺Romanian (⫺Hungarian) (see Rein 1999; Szabo´ 2000). Nevertheless, schooling in L1 and minority rights and protection seem to guarantee language maintenance in these particular language spaces (see below section 5). In the language islands in Upper Italy the lack of access to Standard German or even the complete loss of ties with the German motherland, has resulted in the near loss of the island varieties among the younger generations. Only the Walser dialect seems to be more vital (cf. Zürrer 1999: 2000). Within the Anglophone context (the USA and Australia), the German language islands have also become nearly extinct: the combined social and economic restructuring of American society during the Great Depression and World War II led to a massive rural exodus and a general breakdown of previously established communicative networks. As a consequence, there was accommodation to English and a gradual attrition of the native language (Keel 2003). These developments were accelerated by anti-German sentiments and emerging Anglo-American national identity. In Australia as early as the 1970s English replaced the German dialect as a means of in-group communication among almost all of the speakers of the language islands, excluding of course the very oldest members (see Clyne 1994; Kipp 2006). Exceptions here are the various religious communities such as Amish and Mennonites. These groups continue to use Standard German in their worship services and the respective dialect varieties in everyday life. Social isolation has been crucial to maintaining the minority language (“maintenance by inertia”, Louden 2003 and below). Although all members are bilingual and English is exclusively used as the written variety and the primary language of instruction, it is deemed unsuitable for intra-group domains. In contrast to many other minority groups, Pennsylvania German has not only managed to survive since the eighteenth century, due to high birth rates it is even “experiencing dramatic growth” (Louden 2006: 105). While the situation in Middle and South America is similar to that of North America (with the aforementioned exception of religious groups, see Kaufmann 1997, 2004), a somewhat different situation is found in Brazil. Here, the relative stability of the German variety mainly in Rio Grande do Sul is due to the economic and administrative isolation

341

342

III. Structure and dynamics of a language space of the speakers and the extremely rural character of the German population (Damke 1997). Additionally, these colonies enjoyed a comparatively long period of revitalization thanks to the steady influx of newcomers until 1924. Suspension of educational and minority rights in 1940 has progressively led to the de-standardization of the standard variety and a movement toward an acrolectal form of a supra-regional variety, the socalled Hunsrückisch (see Auer 2005; section 4.2).

4.2. Linguistic developments As mentioned above, one of the most frequently discussed linguistic phenomena relevant for the development of language island varieties is linguistic convergence. There are three main types: (i) convergence between dialects; (ii) convergence of dialects toward a (genetically related) standard variety; and (iii) convergence of settler varieties toward the surrounding majority language. While the first two types were the focus of traditional language island research (see section 2), more recent studies have begun to concentrate on convergence toward the contact language. Convergence between dialects can be seen as a process of leveling. In older language islands where there was no roofing standard, a complex struggle took place between salient and non-salient features. According to the Schirmunski criteria, those forms more distant from the standard sometimes were able to win out over less distant features (e.g., [y:] over [u:] for the standard diphthong [au] in some Mennonite varieties of the Ural). The reason for this development is sociolinguistic in origin. Features which are more distant are often perceived as being more authentic. Consequently, they are better suited to serving the role of a group symbol (cf. Rosenberg 2005: 226⫺227). At the morphological level, we can observe a different pattern of leveling. In some cases, it is the form which simplifies the morphological system that is selected. However, in most cases, the form that is the closest to the standard is chosen (examples from the Hunsrückisch koine´ see Auer 2005). One of the processes involved in convergence toward the contact language is case merger (i. e., the replacement of the dative by the accusative case) (examples from Russian German, see Berend and Riehl 2008): (1) a. Auf den Platz sin Leit [instead of: Auf dem Platz] ‘There are people in the square.’ b. Haben wir Deutsch gesprochen mit die Mama [instead of: mit der Mama] ‘We spoke German with our Mummy.’ c. In die Felder habn se geschaffen [instead of: In den Feldern] ‘They have worked in the fields.’ While the results of many initial studies on German-American varieties (Huffines 1994; Louden 1994; and others) seem to indicate that convergence (toward English) was highly probable, the results of more recent investigations involving Russian German and other Eastern European varieties (see Eichinger, Plewnia and Riehl 2008) with a rich case system have suggested other explanations which seem far more plausible. As Rosenberg

19. Discontinuous language spaces (2005: 233) points out, case reduction in German language islands is to a great extent internally (i. e., typologically) motivated. A comparison made between different island varieties also reveals that ⫺ despite variation due to geographic location, age and contact with Standard German ⫺ reduction processes follow a clear path, starting with feminine nouns and including the plural forms in noun groups with definite articles. Subsequently, the indefinite articles, all genders and possessive pronouns are affected: (2) Der woar bei sei Leit [instead of: seinen] ‘He visited his people.’ (Pennsylvania German, cf. Born 2003: 155)

The personal and demonstrative pronouns are retained the best and only undergo reduction in some of the American varieties: (3) … wann ich sie Blumme bringe deet [instead of: ihr] ‘… when I would bring her some flowers’ (Pennsylvania German, cf. Huffines 1994: 51) The comparatively high retention of case in the pronominal system is also typical of most other Germanic languages and may be motivated, among others, by the type of storage (full listed vs. decomposition), frequency and the degree of animacy (cf. Rosenberg 2005). Other processes of convergence can be seen at the syntactic level. One phenomenon is the deconstruction of the so-called “brace construction”. Typical for German, this construction is marked by the finite and the infinite parts of the verb framing or encapsulating other components of the clause (e.g., Hans hat heute das Haus geputzt, lit. ‘Hans has today the house cleaned’). In language island varieties, components are often placed after the infinite verb part (so-called “extraposition”): (4) a. Am besten hat gesprochen die Ältste [instead of: Am besten hat die Ältste gesprochen.] ‘The oldest one spoke best.’ (Russian German, see Berend and Riehl 2008) b. Ohren haben gehört etwas. [instead of: (Die) Ohren haben etwas gehört.] ‘(My) ears have heard something.’ (Russian German, see Berend and Riehl 2008) This development may be motivated by a cognitive principle. Called “the principle of proximity”, this phenomenon was identified by Givo´n (1990: 970⫺971). According to Givo´n, there is a general tendency to keep grammatical operators near their operands. This principle is also evident in the fact that the relative proximity of grammatical morphemes to the stem can be used to indicate conceptual scope relations. Another instance of convergence concerns the gradual removal of verb final position in the German subordinate clause. This leads to a generalization of the verb second position in the affirmative clause:

343

344

III. Structure and dynamics of a language space (5) a. Vorher schon, weil ich hab gelernt auf Kaufmann. [instead of: … weil ich auf Kaufmann gelernt habe.] ‘Even before, because I have been trained to be a merchant.’ (Ukrainian German, unpublished data) b. das hilft … die Fremdsprachen viel besser zu behalten oder erlernen, dass sie können die deutsche Sprache [instead of: … dass sie die deutsche Sprache können.] ‘That helps to keep foreign languages better in mind, so that they know the German language’ (Hungarian German, unpublished data) These developments are generally interpreted as being instances of simplification and reduction (see Gilbert and Fuller 2003: 175). This is due to the destabilization of norm certainty and norm loyalty which in turn accelerates language change (Rosenberg 2003: 217). However, there are also instances where complex structures emerge. One example includes the grammaticalization of optional native constructions. In Pennsylvania German, the copula ⫹ am ⫹ infinitive is used for aspect marking (marking of the progressive aspect): (6) Er is in die Schtadt an gehe nau ‘He is going to town now.’

(Pennsylvania German, see Louden 1994: 85)

In this case, a latent semantic category (that is presently undergoing grammaticalization in substandard varieties of German) is expanded and the optional construction becomes grammaticalized (cf. Heine and Kuteva 2005). This process might be interpreted as being a true instance of convergence toward the contact language. Generally speaking, the external linguistic context has an impact on certain internal tendencies. It accelerates language development following a particular typological direction and according to certain cognitive principles (see Riehl 2004: 95⫺96, Riehl to appear). Thus, as Clyne (1991: 179) points out: The speech of bilinguals will diverge from that of monolinguals in the heartland of the immigrant language not only because of the effects of the dominant language but also because in the relative isolation of the immigrant situation, changes in accordance with the dominant typology of the language are accelerated in the language.

In this context, the lack of a roofing standard for the same diasystem is responsible not only for the retention of archaic features, but also for the acceleration of internal change. This process, however, need not necessarily only include simplification processes (as the example of the aspect construction shows). These observations seem applicable to all language contact situations. For language island research, this commonality facilitates drawing a comparison between the island variety with not only varieties of the so-called “motherland”, but also with other varieties of language islands. Such comparisons are useful in demonstrating the interdependency which exists between the degree of transference and the amount of cultural pressure. As predicted by Thomason and Kaufman (1988: 74⫺76), typological changes in syntax and morphology only take place in language contact situations where the pressure exerted by the majority language is comparatively high. Consequently, it is common to find that typological changes such as the generalization of V2-position are more ad-

19. Discontinuous language spaces vanced in settings where contact with the surrounding language(s) is greater in frequency and/or intensity. In this sense, comparative language island research would seem to be a promising field of research which is useful for obtaining critical insights into the complex interconnections which exist between typological, sociolinguistic and psycholinguistic aspects of language change (see Rosenberg 2003; Riehl 2004).

5. Factors determining language maintenance in discontinuous language spaces One of the major factors for language maintenance in the history of language islands is the type of the settlement, e.g., closed, monolingual settlements such as the Volga Germans or the settlements in the Banat, or mixed communities as found in the Bukovina. In settlements of the latter type, there was neither a common regional identity, nor was there a leveling between different varieties, but rather coexistence (Protze 1995). This also holds true for many urban settlements. Towns play an important role in the emergence of regional standard varieties. The vernaculars in these areas gained a certain cultural prestige and were also acquired by non-native speakers (e.g., the urban vernaculars of Czernovic and Timisoara in the nineteenth century, see Gadeanu 1998). More recently, one major criterion for language maintenance (also mentioned by Kloss 1966) is the number of speakers: the smaller the community, the more it is necessary to speak other languages. Although language shift is fostered by multilingualism, it is not necessarily a consequence of it. Other factors, such as language awareness can also contribute to language retention (see Riehl 2005). Similar importance is ascribed to the prestige of the minority language in comparison to the surrounding majority language: Mainly in the Eastern European context, some languages are commonly considered to be less prestigious than others (e.g., Czech and Romanian, which have low prestige as opposed to Hungarian, which has comparatively high prestige). Normally, the reputation of a language coincides with the social prestige of its language communities. The change of alliances and shift in influence of a certain group may lead to a subsequent change in prestige level. Such was the case with the plummeting reputation of Germany after World War II. Another factor that has to be considered is the level of institutionalization of the minority language and its juridical status. In areas where the minority language functions as a language of administration, it also frequently serves as the language of instruction, public life and media. When no minority status is granted, developments such as shifts in the local control of schools, parishes and newspapers toward more supra-regional control (as has been seen in the USA) might also cause shifts toward the majority language (Louden 2003: 135). There are two other important factors to mention here: endogamy and religious commitment to the language. When a high percentage of the marriages in minority communities are intra-ethnic, the language can be used in the family domain. The use of the native language in the religious domain plays an important role, because in liturgical contexts a standard variety will mainly be employed. Although in most language islands, speakers will have only a passive competence in this variety, the use of the minority language in church can also have a considerable impact on the prestige of the language.

345

346

III. Structure and dynamics of a language space The factors mentioned so far also apply to other types of minorities. The most crucial factor for discontinuous language spaces is the contact with the linguistic motherland. This contact can be established either via the settlement of new speakers from the motherland or the relocation of young people to the motherland for education. The first pertains not only to new urban migratory contexts, but also for language islands (e.g., Bukovina and the Brazilian colonies, see above); the latter is historically pertinent for Transylvania and is relevant for more recent language islands, e.g., in South America. In both cases, language island speakers are confronted with new developments of the standard variety in the motherland and have a literary tradition at their disposal. Today, contact can also be maintained by television transmitted via satellite. Many language islands were able to survive for comparatively long periods of time because speakers were able to use their dialect variety as a medium of communication almost exclusively. Only a few members of the community (so-called “gate keepers” responsible for out-group communication, Mattheier 2003: 24) had an active command of the majority language. The increase of literacy and the introduction of compulsory schooling eventually made it necessary, however, to acquire and use the majority standard variety. In most of the German-speaking communities, Standard German could be used as a language of schooling until the first few decades of the twentieth century. In this case, a stable diglossia with German dialect or regional koine´ and German standard was established. Consequently, the contact with the linguistic motherland was also very important. Only when formal instruction in German had stopped and the second language became important in more and more domains, did a relatively unstable triglossic situation between German dialect, Standard German, a second language and in some cases additional varieties or languages, emerge. As evident in Romania, the diasystem was frequently reduced in such cases, i. e., the use of German was reduced to the standard variety only. This was true for most cases, excluding those where the dialect had a strong in-group prestige or functioned as marker of identity (Pennsylvania Dutch, Venetian dialects in Mexico, Plaudietsch varieties of Mennonites). In conclusion, there are three types of settlements which have proved to be conducive to language maintenance: 1. Rural, agrarian settlements that have existed in isolation over many centuries due to geography and infrastructure (e.g., settlements in remote mountain valleys such as the Walser in Val d’Aoste, or the Sette e tredici communi, or in isolated unpopulated regions like the Provenc¸al speakers in Lucania in Southern Italy and the Tyroleans in Pozuso, Peru). 2. Rural settlements which have been isolated due to their religious beliefs (e.g., the Waldenser population, the Amish and Mennonite communities in the Americas and Russia). 3. Settlements which have been granted a high degree of autonomy with own administration and school system. In these cases, speakers have had access to a standard variety of their language of origin and were able to use it in many different domains (e.g., German-speaking communities in Transylvania; settlers of the Volga region before 1941; German-speaking communities in Rio Grande do Sul). The first two situations coincide with factors already mentioned by Kloss (1966): rural isolation, endogamy, limited social and demographic mobility. When these conditions changed, some effort was needed to maintain the language further (i. e., scholastic education and institutional support by clubs, religious organizations and media).

19. Discontinuous language spaces However, all these efforts failed if the speakers themselves lacked the socio-psychological predisposition to maintain their “otherness” and a positive attitude toward their native language. In such cases, it is important to take into account the kind of variety speakers consider to be their “language”. When dialectal varieties are not roofed by a standard vernacular, the restricted communication radius and a negative attitude of the majority group toward dialects might affect the speakers’ attitude toward the minority language. Exceptions are those religious groups that define themselves through their outsider status.

6. Other types o discontinuous language spaces As we have seen, the factors involved in language maintenance of language islands are mostly defined by the relative isolation of the group over an extended period of time. Historically, these isolated groups have lived in rural settlements that had little or no access to higher education and the accompanying institutional use of a high-prestige variety (excluding that is the formal language used for religious purposes and local administration). Changes in immigration patterns from the nineteenth century onward stimulate the question whether the new emerging urban settlements meet the criteria of language islands, especially the criterion of delayed assimilation (Mattheier 1994, see section 1). Historically, this can be said of urban ghettos in the Americas (e.g., Little Italies, s. Corra` and Ursini 1998: 571⫺573) and currently for Turkish-speaking communities that concentrate in certain quarters of Middle-European cities (Brussels, Amsterdam, Hamburg, Berlin). Such speech communities to a great extent follow the life circle paradigm of language islands as suggested by Mattheier (2003). This paradigm predicts criteria that are valid for language maintenance in the second generation: closed networks, endogamy, productive exchange with the motherland via newcomers who join the community, long visits in the motherland during the holidays, the presence of well-established media (newspapers, television, broadcasting), most of which apply to urban settlements as well, especially to the Turkish-speaking community. Despite its relatively low reputation among majority language speakers, the Turkish language has a conspicuous covert prestige among native speakers and other migrant groups (see Dirim and Auer 2004, and for more details, Krefeld in this volume). Moreover, as there is no institutionalization of the minority language, most second and third generation speakers have no command of a written variety. This development is comparable to the loss of a standard roof and the use of dialectal or regional varieties in language island settings. In addition, we also can observe convergence processes toward the German language (cf. Rehbein 2001; Rehbein and Karakoc¸ 2004). Taking all this into account, it could be of considerable interest to apply methods of Sprachinsel research on this kind of urban settlements as well.

7. Conclusion Discontinuous language spaces are ⫺ despite the obsolescence of the metaphor Sprachinseln ⫺ still an interesting field of research. The most important criteria in defining

347

348

III. Structure and dynamics of a language space Sprachinsel involve the relationship of the speakers toward their linguistic motherland and its standard variety and the awareness of ethnic difference among the minority group. In the first case, seen from a linguistic point of view, the typological development of a particular language can be observed under different circumstances. In the second case, observed from a sociolinguistic point of view, it is possible to explore the impact of language and space on language attitude and ethnic identity. Modern language island research has to adopt general methods of other minority research traditions. Interesting insights could be gained not only by comparing German language islands in different constellations, but also by comparing them to minority islands of different linguistic origins. Furthermore, as most of the older language islands are facing an imminent death (see Mattheier 2003), new emerging communities such as urban settlements should be included in research paradigms of discontinuous language spaces.

8. Reerences Asfandiarova, Dania 1999 ‘Mir verzähle doch lutherisch und die katholisch’. Dialektmischung im Vokalismus der deutschen Sprachinsel. Prischib/Aleksejevka (Baschkortostan/Russische Föderation). In: Gisela Brandt (ed.), Historische Soziolinguistik des Deutschen IV, 241⫺262. Stuttgart: Stuttgarter Arbeiten zur Germanistik, vol. 372. Auer, Peter 2005 A(nother) scenario for new dialect formation. In: Björn Melander (ed.), Spra˚k i Tid. Studier Tillägnade Mats Thelander pa˚ 60-a˚rsdagen, 57⫺70. Uppsala: Institutionen för nordiska spra˚k. Auer, Peter, Jacinta Arnhold and Cintia Bueno-Aniola 2007 Being a colono and being daitsch in Rio Grande do Sul. In: Peter Auer and Werner Kallmeyer (eds.), Style and Social Identities. Alternative Approaches to Linguistic Heterogeneity, 85⫺120. Berlin: Mouton de Gruyter. Auer, Peter, Frans Hinskens and Paul Kerswill (eds.) 2005 Dialect Change. The Convergence and Divergence in European Languages. Cambridge: Cambridge University Press. Berend, Nina 1997 Wolgadeutscher Sprachatlas (WDSA). Aufgrund der von Georg Dinges 1925⫺1929 gesammelten Materialien unter Mitarbeit von Rudolf Post. Tübingen: Franke. Berend, Nina 2003 Zur Vergleichbarkeit von Sprachkontakten: Erfahrungen aus wolgadeutschen Sprachinseln in den USA und Russland. In: Keel and Mattheier (eds.), 239⫺268. Berend, Nina and Hugo Jedig 1991 Deutsche Mundarten in der Sowjetunion. Geschichte der Forschung und Bibliographie. Marburg: Elwert. Berend, Nina and Elisabeth Knipf-Komlo´si (eds.) 2006 Sprachinselwelten. Entwicklung und Beschreibung der deutschen Sprachinseln am Anfang des 21. Jahrhunderts. Frankfurt am Main: Lang. Berend, Nina and Klaus J. Mattheier (eds.) 1994 Sprachinselforschung. Eine Gedenkschrift für Hugo Jedig. Frankfurt am Main: Lang.

19. Discontinuous language spaces Berend, Nina and Claudia Riehl 2008 Russland. In: Eichinger, Plewnia and Riehl (eds.), 17⫺58. Blankenhorn, Renate 2003 Pragmatische Spezifika der Kommunikation von Russlanddeutschen in Sibirien. Entlehnung von Diskursmarkern und Modifikatoren sowie Code-switching. Frankfurt etc.: Lang. Boas, Hans C. to appear The Life and Death of Texas German. A Study in New-Dialect Formation, Language Contact, Language Change, and Language Death. Durham: Duke University Press. Born, Joachim and Sylvia Dickgießer 1989 Deutschsprachige Minderheiten. Ein Überblick über den Stand der Forschung für 27 Länder. Mannheim: Institut für deutsche Sprache. Born, Renate 2003 Regression, convergence, internal development: The loss of the dative case in GermanAmerican dialects. In: Keel and Mattheier (eds.), 151⫺164. Breu, Walter (ed.) 2005 L’influsso dell’italiano sulla grammatica delle lingue minoritarie. Problemi di morfologia e sintassi. Atti del Convegno Internationale ⫺ Costanza, 8⫺11 ottobre 2003. (Studi e testi di Albanistica 17.) Rende: Universita` della Calabria. Clyne, Michael G. 1991 Community Languages: The Australian Experience. Cambridge: Cambridge University Press. Clyne, Michael G. 1994 What can we learn from Sprachinseln? Some observations on “Australian German”. In: Berend and Mattheier (eds.), 105⫺122. Frankfurt am Main: Lang. Corra`, Loredana and Flavia Ursini 1998 Romanische Migranten in der Neuzeit. I migranti romanzi in eta` moderna. In: Günter Holtus, Michael Metzeltin and Christian Schmitt (eds.), Lexikon der Romanistischen Linguistik, vol. 7, Kontakt. Migration und Kunstsprachen, 559⫺585. Tübingen: Niemeyer. Damke, Ciro 1997 Sprachgebrauch und Sprachkontakt in der deutschen Sprachinsel in Südbrasilien. Frankfurt am Main: Lang. Deminger, Szilvia 2004 Spracherhalt und Sprachverlust in einer Sprachinselsituation: Sprache und Identität bei der deutschen Minderheit in Ungarn. Frankfurt am Main: Lang. Dirim, Inci and Peter Auer 2004 Türkisch sprechen nicht nur die Türken: Über die Unschärfebeziehung zwischen Sprache und Ethnie in Deutschland. Berlin: de Gruyter. Domaschnew, Anatoli 2000 Die Marburger Schule und die Entwicklung der deutschen Sprachinselforschung in Russland. In: Stellmacher (ed.): 387⫺398. Edwards, John 1990 Notes for a minority-language typology: Procedures and justification. Journal of Multilingual and Multicultural Development 11: 137⫺151. Eichinger, Ludwig M. 2003 Island Hopping: Vom Nutzen und Vergnügen des Vergleichens von Sprachinseln. In: Jannis Androutsopoulos and Evelyn Ziegler (eds.), “Standardfragen”. Soziolinguistische Perspektiven auf Sprachgeschichte, Sprachkontakt und Sprachvariation, 83⫺107. Frankfurt am Main: Lang. Eichinger, Ludwig M., Albrecht Plewnia and Claudia M. Riehl (eds.) 2008 Handbuch der deutschen Sprachminderheiten in Mittel- und Osteuropa. Tübingen: Narr.

349

350

III. Structure and dynamics of a language space Földes, Csaba 1996 Mehrsprachigkeit, Sprachenkontakt und Sprachenmischung. Flensburger Papiere zur Mehrsprachigkeit und Kulturenvielfalt im Unterricht 14/15: 5⫺79. Földes, Csaba 2006 “Sprachinsel”-Paradigma auf dem Prüfstand: Konzept, Terminologie und Forschungsmethodologie. Zeitschrift für Deutsche Philologie 125: 321⫺341. Ga˘deanu, Sorin 1998 Sprache auf der Suche. Zur Identitätsfrage des Deutschen in Rumänien am Beispiel der Temeswarer Stadtsprache. Regensburg: S. Roderer. Gehl, Hans 2000 Aktuelle Erkenntnisse in der donauschwäbischen Dialektologie. In: Stellmacher (ed.): 288⫺333. Gerner, Zsuzsanna 2006 Identität ⫺ soziales Netzwerk ⫺ nationale Stereotype. Zur Identitätsbildung und Identitätsforschung in den deutschen Sprachinseln in Ungarn. In: Berend and Knipf-Komlo´si (eds.), 149⫺171. Gerner, Zsuzsanna, Manfred Glauninger and Katharina Wild (eds.) 2002 Gesprochene und geschriebene deutsche Stadtsprachen in Südosteuropa und ihr Einfluss auf die regionalen deutschen Dialekte: Internationale Tagung in Pe´cs, 30. 3.⫺2. 4. 2000. Wien: Praesens. Geyer, Ingeborg 1999 Sprachinseln. Anmerkungen zu Definition und Forschungstradition. In: Peter Wiesinger, Werner Bauer and Peter Ernst (eds.), Probleme der oberdeutschen Dialektologie und Namenskunde. Vorträge des Symposiums zum 100. Geburtstag von Eberhard Kranzmayer, 152⫺170. Wien: Praesens. Gilbert, Glenn and Janet Fuller 2003 The linguistic atlas of Texas German revisited. In: Keel and Mattheier (eds.), 165⫺176. Giro´n, Talmy 1990 Syntax. Amsterdam: Benjamins. Göllner, Carl 1979 Geschichte der Deutschen auf dem Gebiete Rumäniens, vol. 1: 12. Jahrhundert bis 1848. Bucharest: Kriterion. Gottas, Friedrich 1995 Die Deutschen in Südosteuropa. Von den Ansiedlungen im Mittelalter und im 18. Jahrhundert zur Rückbewegung im 20. Jahrhundert. In: Grimm and Zach (eds.), 13⫺30. Grimm, Gerhard and Krista Zach (eds.) 1995 Die Deutschen in Ostmittel- und Südosteuropa. Munich: Verlag Südostdeutsches Kulturwerk. Heine, Bernd and Tania Kuteva 2005 Language Contact and Grammatical Change. Cambridge: Cambridge University Press. Hinskens, Frans, Peter Auer and Paul Kerswill 2005 The study of dialect convergence and divergence: conceptual and methodological considerations. In: Auer, Hinskens and Kerswill (eds.), 1⫺48. Holtus, Günter, Michael Metzeltin and Christian Schmitt (eds.) 1998 Lexikon der Romanistischen Linguistik, vol. 7: Kontakt. Migration und Kunstsprachen. Tübingen: Niemeyer. Hornung, Maria 1994 Die sogenannten zimbrischen Mundarten der Sieben und Dreizehn Gemeinden in Oberitalien. In: Maria Hornung (ed.): 19⫺43. Hornung, Maria (ed.) 1994 Studien zur Dialektologie, vol. 3, Die deutschen Sprachinseln in den Südalpen: Mundarten und Volkstum. Hildesheim: Olms.

19. Discontinuous language spaces Huffines, Marion 1994 Directionality of language influence: The case of Pennsylvania German and English. In: Berend and Mattheier (eds.), 47⫺58. Hutterer, Claus 1982 Sprachinselforschung als Prüfstand für dialektologische Arbeitsprinzipien. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke, Herbert E. Wiegand (eds.), Dialektologie. Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 1, 178⫺189. Berlin: de Gruyter. Hutterer, Claus 1994 Sprachinseldialektologie. In: Klaus J. Mattheier and Peter Wiesinger (eds.), Dialektologie des Deutschen. Entwicklungsstand und Entwicklungstendenzen, 93⫺101. Tübingen: Niemeyer. Karch, Dieter 1978 Braunschweig-Veltenhof: Pfälzische Sprachinsel im Ostfälischen. Tübingen: Niemeyer. Kattenbusch, Doris 1998 Romanische Migranten im Mittelalter. Überblick. In: Holtus, Metzeltin and Schmitt (eds.), 526⫺538. Kaufmann, Göz 1997 Varietätendynamik in Sprachkontaktsituationen. Attitüden und Sprachverhalten rußlanddeutscher Mennoniten in Mexiko und den USA. Frankfurt am Main: Lang. Kaufmann, Göz 2004 Eine Gruppe ⫺ zwei Geschichten ⫺ drei Sprachen. Zeitschrift für Dialektologie und Linguistik 71: 257⫺306. Keel, William D. 2003 Patterns of shift in Midwestern German Speech Islands. In: Keel and Mattheier (eds.), 303⫺325. Keel, William D. and Klaus J. Mattheier (eds.) 2003 German Language Varieties Worldwide: Internal and External Perspectives. Frankfurt am Main: Lang. Kipp, Sandra 2002 German⫺English bilingualism in the Western district of Victoria. Unpublished PhD disseration, Department of Linguistics and Applied Linguistics, University of Melbourne. Kipp, Sandra 2006 Reconstructing the social networks of a nineteenth century Sprachinsel. In: Berend and Knipf-Komlo´si (eds.), 109⫺124. Kloss, Heinz 1966 German-American language maintenance efforts. In: Joshua Fishman (ed.), Language Loyalty in the United States, 206⫺252. The Hague: Mouton. Knipf-Komlo´si, Elisabeth 2003 Sprachwahl und kommunikative Handlungsformen der deutschen Minderheit in Ungarn. In: Keel and Mattheier (eds.), 269⫺281. Knipf-Komlo´si, Elisabeth 2006 Sprachliche Muster bei Sprachinselsprechern am Beispiel der Ungarndeutschen. In: Berend and Knipf-Komlo´si (eds.), 39⫺56. Kuhn, Walter 1934 Deutsche Sprachinsel-Forschung. Geschichte, Aufgaben, Verfahren. Plauen i. Vogtland: Wolff. Lo Cascio, Vincenzo (ed.) 1987 L’italiano in America latina. Firenze: Le Monnier. Louden, Mark L. 1994 Syntactic change in multilingual speech islands. In: Berend and Mattheier (eds.), 73⫺91.

351

352

III. Structure and dynamics of a language space Louden, Mark L. 2003 An eighteenth-century view of Pennsylvania German and its speakers. In: Keel and Mattheier (eds.), 69⫺85. Louden, Mark L. 2006 Patterns of language maintenance in German American speech islands. In: Linda L. Thornburg and Janet M. Fuller (eds.), Studies in Contact Linguistics. Essays in Honor of Glenn G. Gilbert, 127⫺145. New York: Lang. Maitz, Pe´ter 1998 Die Sprachinsel als Forschungsgegenstand: Ein geschichtlich-thematischer Überblick am Beispiel der ungarndeutschen Sprachinselforschung. Sprachtheorie und germanistische Linguistik 8(2): 205⫺219. Mattheier, Klaus J. 1994 Theorie der Sprachinsel. Voraussetzungen und Strukturierungen. In: Berend and Mattheier (eds.), 333⫺348. Mattheier, Klaus J. 1996 Methoden der Sprachinselforschung. In: Hans Goebl, Peter H. Nelde, Zdenek Stary and Wolfgang Wölck (eds.), Contact Linguistics. An International Handbook of Contemporary Research, vol. 1, 812⫺819. (Handbücher zur Sprach- und Kommunikationswissenschaft 12.1.) Berlin: de Gruyter. Mattheier, Klaus J. 2002 Sprachinseln als Arbeitsfelder. Zu den zentralen Forschungsdimensionen der Erforschung deutscher Sprachinseln. In: Maria Erb, Elisabeth Knipf, Magdolna Orosz and La´szlo´ Tarno´i (eds.), “Und thut ein Gnügen seinem Ambt”: Festschrift für Karl Manherz zum 60. Geburtstag, 135⫺144. Budapest: ELTE. Mattheier, Klaus J. 2003 Sprachinseltod: Überlegungen zur Entwicklungsdynamik von Sprachinseln. In: Keel and Mattheier (eds.), 13⫺31 Meier, Jörg 2000 Die deutschen Dialekte in der Zips/Spis: Anmerkungen zur Sprachinsel- und Sprachkontaktforschung. In: Stellmacher (ed.): 362⫺386. Naiditsch, Larissa 1996 Dialekte der Russlanddeutschen. Muttersprache 106: 250⫺262. Naiditsch, Larissa 2006 Zur phonologischen Beschreibung der Entwicklungsprozesse in einer Inselmundart: Das Mennonitenplatt. In: Berend and Knipf-Komlo´si (eds.), 227⫺236. Pohl, Heinz 2005 Sprachen und Sprachinseln im südalpinen Raum: Ein Überblick. Europa Ethnica 62(3/ 4): 91⫺100. Protze, Helmut 1995 Zur Erforschung deutscher Sprachinseln in Südost- und Osteuropa. Siedlung, Sprache, Geschichte und Wechselwirkungen. In: Grimm and Zach (eds.), 55⫺84. Pusch, Claus D. 2005 Normativer Hiat und Skripturalität: Die Problematik des Oralitätsprimats im Kontext der Standardisierung von Minderheitenidiomen, dargestellt an galloromanischen Beispielen. In: Carsten Sinner (ed.), Norm und Normkonflikt in der Romania, 21⫺38. Munich: Peniope. Rehbein, Jochen 2001 Turkish in European Societies. Lingua e Stile 36: 317⫺334. Rehbein, Jochen and Birsel Karakoc¸ 2004 On contact-induced language change of Turkish aspects: Languaging in bilingual discourse. In: Christine B. Dabelsteen and J. Normann Jørgensen (eds.), Languaging and

19. Discontinuous language spaces Language Practices, 125⫺149. (Copenhagen Studies in Bilingualism 36.) Copenhagen: Faculty of the Humanities, University of Copenhagen. Rein, Kurt 1999 Diglossie und Bilingualismus bei den Deutschen in Rumänien und Ungarn sowie den GUS-Staaten. In: Thomas Stehl (ed.), Dialektgenerationen, Dialektfunktionen und Sprachwandel, 37⫺53. Tübingen: Narr. Riehl, Claudia M. 2004 Sprachkontaktforschung. Eine Einführung. Tübingen: Narr. Riehl, Claudia M. 2005 Code-switching in bilinguals: Impacts of mental processes and language awareness. In: James Cohen, Kara T. McAlister, Kellie Rolstad and Jeff MacSwan (eds.), Proceedings of the 4th International Symposium on Bilingualism, 1945⫺1957. Somerville, MA: Cascadilla Press. Riehl, Claudia M. 2006 Sprachwechselprozesse in deutschen Sprachinseln Mittel- und Osteuropas. Varietätenkontakt und Varietätenwandel am Beispiel Transkarpatiens. In: Berend and Knipf-Komlo´si (eds.), 189⫺204. Riehl, Claudia M. 2008 Die deutschen Sprachgebiete in Mittel- und Osteuropa. In: Eichinger, Plewnia and Riehl (eds.), 1⫺16. Riehl, Claudia M. to appear Norm and variation in language minority settings. In: Alexandra Lenz and Albrecht Plewnia (eds.), Grammar Between Norm and Variation. Rosenberg, Peter 1994 Varietätenkontakt und Varietätenausgleich bei den Russlanddeutschen: Orientierungen für eine moderne Sprachinselforschung. In: Berend and Mattheier (eds.), 123⫺164. Rosenberg, Peter 1998 Deutsche Minderheiten in Lateinamerika. In: Theo Harden and Elke Hentschel (eds.), Particulae particularum. Festschrift zum 60. Geburtstag von H. Weydt, 261⫺291. Tübingen: Stauffenburg. Rosenberg, Peter 2003 Comparative speech island research: some results from studies in Russia and Brazil. In: Keel and Mattheier (eds.), 199⫺238. Rosenberg, Peter 2005 Dialect convergence in the German language islands (Sprachinseln). In: Auer, Hinskens and Kerswill (eds.), 221⫺235. Rowley, Anthony 1996 Die Sprachinseln der Fersentaler und Zimbern. In: Robert Hinderling and Ludwig M. Eichinger (eds.), Handbuch der mitteleuropäischen Sprachminderheiten, 263⫺285. Tübingen: Niemeyer. Sasse, Hans-Jürgen 1998 Arvanitika: The long Hellenic centuries of an Albanian variety. International Journal of the Sociology of Language 134: 39⫺68. Salmons, Joseph 1994 Naturalness and morphological change in Texas German. In: Berend and Mattheier (eds.), 59⫺72. Schabus, Wilfried 1994 Beobachtungen zu Sprachkontakt, Varietätenausgleich, Sprachloyalität und Sprachwechsel in Pozuzo (Peru) und bei den “Landlern” in Siebenbürgen. In: Berend and Mattheier (eds.), 221⫺262. Schabus, Wilfried 1996 Die Landler. Sprach- und Kulturkontakt in einer alt-österreichischen Enklave in Siebenbürgen (Rumänien). Wien: Praesens.

353

354

III. Structure and dynamics of a language space Schirmunski, Viktor M. 1930 Sprachgeschichte und Siedlungsmundarten. Germanisch-Romanische Monatsschrift 18, 113⫺122 and 171⫺188. Schwob, Anton 1971 Wege und Formen des Sprachausgleichs in neuzeitlichen ost- und südostdeutschen Sprachinseln. Munich: Oldenbourg. Siebenhaar, Beat 2004 Die deutschen Sprachinseln auf den Jurahöhen der französischsprachigen Schweiz. Zeitschrift für Dialektologie und Linguistik 71: 180⫺212. Stellmacher, Dieter (ed.) 2000 Dialektologie zwischen Tradition und Neuansätzen. Beiträge der Internationalen Dialektologentagung, Göttingen, 19⫺21 Oktober 1998. Stuttgart: Steiner. Szabo´, Csilla-Anna 2000 Aspekte des heutigen Gebrauchs der Sprache im Sathmarer Gebiet, Nordwestrumänien. In: Stellmacher (ed.): 347⫺361. Thomason, Sarah and Terrence Kaufman 1988 Language Contact, Creolization, and Genetic Linguistics. Oxford: University of California Press. Trovato, Salvatore C. 1998 Galloitalienische Sprachkolonien. I dialetti galloitalici di Sicilia. In: Holtus, Metzeltin and Schmitt (eds.), 538⫺559. Wiesinger, Peter 1983 Deutsche Dialektgebiete außerhalb des deutschen Sprachgebiets: Mittel-, Südost- und Osteuropa. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke and Herbert E. Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 2, 900⫺929. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.2.) Berlin: de Gruyter. Wirrer, Jan 2000 Vorwort. In: Jan Wirrer (ed.), Minderheiten- und Regionalsprachen in Europa, 7⫺10. Opladen: Westdeutscher Verlag. Wössner, Christiane 2002 Qua parlo´n fa noantri! Spracherhalt und ethnische Identität in Chipilo ⫺ einer Sprachinsel des Veneto in Mexiko. Frankfurt am Main: Lang. Zürrer, Peter 1999 Sprachinseldialekte. Walserdeutsch im Aosta-Tal (Italien). Frankfurt am Main: Sauerländer. Zürrer, Peter 2000 Kontaktlinguistische Variation in Sprachinseldialekten. Europa Ethnica 57(3/4): 148⫺165.

Claudia Maria Riehl, Köln (Germany)

20. Linguistic stability in a language space

20. Linguistic stability in a language space 1. 2. 3. 4.

Linguistic stability: Some preliminary considerations Stability in European languages Conclusions References

1. Linguistic stability: Some preliminary considerations In general, linguistic stability tends to be(come) a more and more marked situation in Europe. That this is not merely an impression becomes clear when we consider a number of important recent publications on language change, such as Auer, Hinskens and Mattheier (1997), Kallen, Hinskens and Taeldeman (2000), Auer, Hinskens and Kerswill (2005), Delbecque, van der Auwera and Geeraerts (2005) and Hinskens (2006). All these books report on the results of recent conferences in which linguistic variation, as an instantaneous reflection of language change (and, more particularly, of processes of convergence and divergence) formed the central theme. Panta rhei! It is not difficult to indicate the main sources of the increasing marginalization of linguistic stability: factors such as an immense geographical and social mobility, the ever growing impact of the mass media, a high level of education and the overall informalization of public life have brought about spectacular shifts in the European dialect-standard language constellations (see Auer 2005). More particularly, almost everywhere in Europe, these changes have resulted in the replacement of older, relatively stable situations of diglossia (in which the standard languages and the dialects were at a considerable structural distance and displayed an almost complementary functionality) by diaglossic constellations, which are mainly characterized by verbal continua from dialect (henceforth: DIA) to standard (henceforth: SL) and hence by the development of intermediate varieties with overlapping functions (type C in Auer’s typology). Linguistic (in)stability should be dealt with from different angles. I distinguish two distinct dichotomies: (1) functional vs. structural (in)stability and (2) horizontal vs. vertical (in)stability. ⫺ Functional stability: in this situation the functions of a variety remain stable throughout a certain period. ⫺ Structural stability: by this we mean that no structural changes take place within the linguistic system. ⫺ Horizontal stability: this implies that there is stability among the varieties of the same horizontal layer on the dialect⫺standard dimension (e.g., no shifts or leveling processes among the dialects of the same region). ⫺ Vertical stability: there are no changes in the place that linguistic elements occupy on the vertical dialect⫺standard axis. We may expect that these types of (in)stability will interact in many ways, depending on a number of intralinguistic and extralinguistic factors. Exploring the huge amount of literature on linguistic (in)stability, we find that these factors and their impact may differ from variety to variety (depending on their place on the vertical dialect⫺standard axis).

355

356

III. Structure and dynamics of a language space For this reason we have decided to organize our exploration of linguistic (in)stability by vertical kind of linguistic variety (see section 2): 2.1 Stability in standard varieties; 2.2 Stability in dialects; 2.3 Stability in intermediate varieties. In every subsection a number of hypotheses will be formulated and evaluated against linguistic facts and situations from across Europe. Most attention, however, will be paid to the linguistic community where I live and which I have studied in depth for a number of decades, viz. Dutch-speaking Belgium (Flanders).

2. Stability in European languages 2.1. Stability in standard varieties In general, and certainly with respect to diaglossic language communities (type C in the typology of Auer 2005), we may hypothesize that spoken standard varieties are subject to loss of homogeneity and structural stability. Everywhere it is possible to observe an increase of variation as a reflection of ongoing processes of change. In order to find out which factors promote or reduce stability in standard varieties, we will compare the two national standard varieties of Dutch: the northern or Netherlandic one and the southern or Flemish one. Especially in oral usage, standard Flemish Dutch is (much) more stable and displays less variation than standard Netherlandic Dutch. In the domain of phonology, this has been convincingly demonstrated by Van de Velde (1996): whereas pronunciation in standard Flemish Dutch remained quite stable during the period between 1935 and 1990, the northern standard was subject to dramatic phonological changes (e.g., devoicing of word-initial fricatives, development of a retroflex postvocalic r (the socalled Gooise r), lowering of the diphthongs ei, ui and ou, diphthongization of the long mid vowels ee, eu and oo, etc.). In my view, there are at least four factors that may explain this diverging development in the two national standard language (SL) varieties. 1. In Flanders the oral usage of the SL has remained (even up to the present day) a kind of a privilege of a cultural elite (see also Willemyns 2003: 299⫺338). This is mainly due to the fact that in Flanders the standardization process started relatively late (the end of the nineteenth century). Before that (and even up until the 1930s) French functioned as the (exoglossic) SL of the economic, social and cultural elite. Moreover, standardization in Flanders was not the result of a real process but a matter of imposition. After a fierce ideological struggle between the “particularists” (who aimed at creating a Flemish standard) and the “Groot-Nederlanders” (who wanted to adopt the SL-norm from the Netherlands in order to erect a strong barrier against French), the latter group won and started imposing (to a certain extent) the northern SL, which differed greatly both structurally and psychologically from the Flemish regiolects and dialects. This is the main reason why this Netherlandic norm never reached the lower social strata of the Flemish community. In the Netherlands on the other hand, larger and more diverse segments of the population use the SL daily and in a wider variety of situations. This even started a process of destandardization to the extent that Stroop (1998) even in the title of his book referred to it as “the decline of the standard language”. This leads us to the first hypothesis

20. Linguistic stability (I)

357

If a variety (also a standard variety) obtains a symbolic value for a specific group (in the social and/or geographical and/or ethnic sense), variation will tend to be more limited.

To put it differently: if more people with different social, regional or ethnic profiles use the SL, an increase of variation and change is proportionally more likely. 2. In the Netherlands, the intermediate (substandard) varieties are structurally much closer to the SL than in Flanders (see for instance Geeraerts 1999). This also applies to the dialects, but here we should add that in both areas the upward processes of dialect and regiolect leveling do not take place at the same speed. In both the Netherlands (see for instance Hagen 1990) and in Flanders (see Willemyns 1997) we have to make a distinction between (relatively) central and (relatively) peripheral areas: in the central areas the leveling towards the SL has already progressed (much) further than in the peripheral ones. Table 20.1 illustrates the situation. Observations such as these lead us to a second hypothesis: (II) If the structural distance between the SL and the intermediate/substandard varieties is smaller, an upward transfer of [⫺SL] elements is proportionally easier and, hence, the SL will be less stable.

Tab. 20.1: Central and peripheral zones in the Dutch language area central Netherlands Flanders

the ‘Randstad’ (Amsterdam, The Hague) Brabant (including Antwerp)

peripheral northeast southeast Limburg (to the east) West Flanders (to the west)

3. In the Netherlands, there is more functional overlap of standard and nonstandard varieties than in Flanders. In the former area, speakers switch from one variety (SL) to the other (substandard) more often. In Flanders, there is a stronger tendency to restrict the use of SL to formal situations, whereas in (relatively) informal situations a “lower” variety will be used: an intermediate variety (called tussentaal) by younger Flemings and dialect or regiolect by older ones. Normally speaking, younger Flemings should be able to switch to SL rather easily (but they refuse to do so, unless the situation is very formal); among older Flemings (especially those who do not belong to the cultural “elite”) the (very) restricted use of SL is mainly due to a limited command of this variety. However this may be, in Flanders there is less functional overlap of [⫹SL] and [⫺SL] varieties. This results in a third hypothesis: (III) If [⫹SL] and [⫺SL] varieties display more functional overlap, the standard variety will be less stable. 4. In Flanders the normative control of the SL has always been much stronger than in the Netherlands. Whereas there is no prescriptive authority in the Netherlands, several Flemish institutions have taken on a normative task. For example, there is a corps of language advisors in the government of the Flemish community in Belgium, and an

358

III. Structure and dynamics of a language space expert in Dutch monitors correct usage of the standard variety at the official broadcasting station (the VRT). This more rigid normative control in Flanders obviously has its roots in linguistic history: it can be related to the exoglossic threat of French until the late 1930s and the fact that an imposed standard had to be spread throughout the Flemish community (which was considered a social duty).These observations engender a fourth hypothesis with respect to the (in)stability of standard varieties. (IV) If there is stronger normative control of the SL, this variety will be more stable and homogeneous. If we broaden our horizon to other language areas in Europe (see Auer 2005 for an interesting synopsis), we can conclude that the Flemish situation (with a rather high degree of stability) is the more marked one. The four factors which we discussed in the preceding paragraphs and which in our view account for the relative stability of the SL in Flanders, do not seem to cooperate so strongly and in such a combined way anywhere else. Surveying the wealth of relevant reports (especially the books referred to in the first paragraph of this article), we can state that: (1) in most language areas (in Europe) the SL has already percolated further; (2) on average, the structural distance between [⫹SL] and [⫺SL] varieties has become smaller; (3) there is more functional overlap between [⫹SL] and [⫺SL] varieties; and (4) the normative control of the Flemish type is becoming something of an anachronism, even in countries such as France, Spain and Sweden, which have a language academy. There is a more tolerant attitude towards variation in standard and substandard varieties everywhere.

2.2. Stability in dialects On the basis of the literature on processes of language change in almost all of the linguistic communities of Europe (see for example Auer 2005 on the shifts in Europe’s standard⫺dialect constellations), we may say that the local basilects are or will be the big victims or losers of changes in progress. Even in a “dialect paradise” such as the German-speaking part of Switzerland, there are recent reports on dialect leveling (see, for example, Christen 1998). Dialects everywhere are liable to two kinds of leveling processes, although their speed and intensity may differ from region to region. Eventually these processes will result in the (gradual) loss of the basilects. First of all, there is horizontal dialect leveling. This means that dialect A loses features by borrowing the corresponding features from dialect B. Second, vertical dialect leveling may take place as well: in this case dialects lose features by the downward infiltration of features from a “higher” variety. Both leveling processes may result in convergence in the sense that they reduce structural differences within the verbal repertoire of a language. There are lots of reports on dialect leveling and the effects on the (in)stability of dialects and dialect features (see for example Auer 1998; Bellman 1997; Berruto 2005; Hinskens 1998; Hoppenbrouwers 1990; Kerswill 2003; Pedersen 2005; Thelander 1980; Tsiplakou et al. 2006; Villena Ponsoda 1996). Time and time again, the same observations can be made: (1) one dialect is more resistant to both kinds of leveling processes than another; (2) one dialect feature is more resistant to change by leveling than another; (3) stability of a dialect (feature) is increasingly becoming a marked phenomenon.

20. Linguistic stability

2.2.1. Re (1): Dialect A vs. dialect B Many reports on dialect leveling and (in)stability show that the degree of stability of a dialect may vary considerably for each language area, but also ⫺ within the same language area ⫺ from region to region. Whereas the Swedish and the Danish dialects, for example, are subject to spectacular leveling processes (see for instance Thelander 1980; Kristensen and Thelander 1984 for Sweden, and Pedersen 2005 for Denmark), the dialects of the German-speaking part of Switzerland display a relative stability (see for example Haas 1992), which only recently has started to break down (cf. Christen 1998 and Siebenhaar 2000). In the Netherlands, the situation tends to resemble the Scandinavian rather than the Swiss German one (see for example Hoppenbrouwers 1990 and Hagen 1991), although Willemyns (1997) correctly points out that there are considerable differences with respect to leveling processes in the more “central” and in the more “peripheral” dialects. In the more “peripheral” regions, the structural distance to the SL is bigger but leveling proceeds along two “lines”: leveling towards the SL and leveling by the development of regional koine´s (see Hinskens 1993 for a transparent description of the two co-occurring types of leveling in southeast Limburg). In Flanders, too, the dialects display quite different degrees of stability, but here the difference between (relatively) “central” versus (relatively) “peripheral” regions is less relevant. I shall illustrate this on the basis of three dialect groups: the Brabantine, the East Flemish and the West Flemish ones (see Map 20.1 for the geographical situation of the three groups). Among these three groups, the West Flemish dialects display the highest degree of stability (see for example Vandekerckhove 1993), although there are recent reports on processes of horizontal leveling (cf. Vandekerckhove 2002). These may result in internal/ inward regional koine´ization, but also in external/outward leveling, in the sense that traditional dialect variation, at least partially, is eliminated and replaced by borrowing some features from the Brabantine dialect, which in Flanders functions as a kind of “superdialect” (see section 2.2.1.3 for this notion). The Brabantine dialects (and more particularly the Antwerp dialect) and the East Flemish dialects are to a much higher degree subject to several kinds of leveling processes. Yet there are also remarkable differences between the two regions: whereas in the

Map 20.1: Main dialect areas in Dutch-speaking Belgium

359

360

III. Structure and dynamics of a language space province of Antwerp the leveling processes are almost exclusively of the horizontal kind (leveling towards the urban Antwerp dialect), the East Flemish dialects are subject to vertical (i. e., towards the SL) as well as horizontal leveling (i. e., reduction of the internal variation by leveling out of local phenomena while elements of the Brabantine/Antwerp “superdialect” are adopted). We have now reached the point at which we have to raise the question of which factors trigger the different degrees of (in)stability among dialects. In my view, at least three factors (which may interact to a considerable extent) are of crucial importance in this respect. These will be discussed in the next subsections.

2.2.1.1. The unctional strength o (a group o) dialects Both the Swiss German and the West Flemish dialects still possess a very strong functionality. Dialect is still the medium of primary socialization and dialect is still spoken in all relatively less formal situations. In order to compare the functional strength of the West Flemish and the East Flemish dialects, I will mainly refer to two Master’s theses from the University of Ghent (prepared under my supervision): Strijkers (1995) and Sabbe (2005). Both (former) students carried out similar research projects in two comparable places, viz. in Melle (a suburban residential location near Ghent in East Flanders with 10,000 inhabitants) and Zedelgem (a similar kind of place near Bruges in West Flanders, 8,400 inhabitants), respectively. See Table 20.2. Tab. 20.2: The use of dialect in two suburban Flemish places

Average functionality of dialect in ten less formal situations Use of dialect among parents and children

Melle

Zedelgem

58,5 % 21,4 %

80,2 % 75,0 %

It is clear that in Zedelgem (and in West Flanders in general) the local dialect is still transmitted to most children and that the entire local community still uses DIA in almost all less formal situations. In Melle (and in fact in the entire East Flemish dialect area), the local dialect has lost much of its functionality. On the basis of several other East Flemish Master’s theses (Jacobs 1990; Dermaut 1999; De Vogelaer 1999; Van Lerberghe 1999; Coessens 1999; De Bruycker 2001) we may conclude that in East Flanders barely fifteen percent of the children are still brought up in dialect. In West Flanders this number is at least four times as high. Keeping this in mind we formulate our first proposition: dialects with a strong functionality are more likely to display a high(er) degree of stability.

2.2.1.2. The symbolic value o a dialect In regions and places where the local dialect still has a high symbolic value (i. e., where it still functions as a marker of local or regional identity), DIA is more likely to maintain a high degree of stability. In this case the local or regional dialect may be(come) the medium for manifesting his or her “Ortsloyalität” for a speaker (see Mattheier 1985 for this notion). To a very high degree, this is still the case in West Flanders and (the province of) Antwerp, but much less in East Flanders. It is not an accident that in the latter area leveling processes often result in vertical leveling (in the direction of the SL). In Antwerp, horizontal leveling towards a regional koine´ is much more likely.

20. Linguistic stability In my own research on the (in)stability of the Flemish dialect landscape (see Taeldeman 1986, 2000 and 2006), I systematically distinguished between so-called “central areas” (German: Kerngebiete) and transitional zones (at the border of two central areas). Areas of the second type are characterized by an intensive spectrum of local variation (in the structural sense), rather abrupt transitions, a relatively high degree of awareness of the spatial oppositions and a strong association of a particular DIA-variety with a particular place. In such areas it is quite possible that the local dialect becomes the favored medium with which to express one’s “Ortsloyalität”. This may have a stabilizing effect and may even lead to the polarization of dialect oppositions. We now formulate our second proposition: If a regional or local dialect becomes a marker of membership of the local community (“Ortsangehörigkeit”), the chance of relative stability will become higher.

2.2.1.3. The social-economic-cultural basis o a dialect (group) It seems quite logical that dialects of regions with a strong social, economic, or cultural basis are more resistant to leveling processes from outside, both horizontal (proceeding from other dialects) and vertical (proceeding from “higher” varieties). It is much more probable that they will play an agentive part on both axes: they may be the innovators of horizontal leveling towards dialects of regions with a weaker social-economic and cultural basis, and they may also pass on particular features to “higher” varieties. This was, for example, the case of Paris and the Iˆle-de-France in France, the London area in England, Copenhagen in Denmark (see Pedersen 2005), the Stockholm area in Sweden (also Pederson 2005), Sofia in Bulgaria (see Videnov 1998), the “Randstad” region in the Netherlands (see for example Hagen 1991) and in recent times the city of Luxemburg in the Grand Duchy (cf. Gilles 2000). In more extreme cases, such regions or cities may even have a destabilizing effect on existing SL varieties (i. e., destandardization). In Flanders, Brabant (and more particularly the city of Antwerp and its immediate surroundings) forms the region with the strongest socioeconomic and cultural position. The urban area of Antwerp has some 650,000 inhabitants. Its harbor district is the economical heart of Flanders and the city functions as the cultural capital of Flanders. Antwerp is also the trendsetter in political developments. The Antwerp dialect hardly ever plays a recipient role, not even vis-a`-vis the standard variety of Belgian Dutch. It takes a dominant position in the formation of new Flemish intermediate varieties (see also below, section 2.3.) and is the epicenter of considerable horizontal dialect leveling (not only in the province of Antwerp, but also elsewhere in Flanders, see above). In this context I refer to the gradual replacement (in the “Kempen” region) of the long mid vowels [ε:] and [œ:] (the equivalents of Standard Dutch ei/ij and ui) by the “heavy” diphthongs of the Antwerp city dialect (e.g., [aes] ⫽ ijs ‘ice’ and [c.es] ⫽ huis ‘house’). It is fascinating to compare the role of the Antwerp city dialect with that of the dialects of the capitals of the old County of Flanders: Ghent (in East Flanders) and Bruges (in West Flanders). Ghent (225,000 inhabitants) is Flanders’ second largest city and occupies a strong industrial and cultural position, but in general its socioeconomic strength is rather well balanced with the rest of the province. Its dialect has remained quite vital within the city boundaries, but, unlike Antwerp and the Antwerp dialect, it exports very few of its urban features to the rest of the East Flemish dialect area. In fact it holds a very insular position in the East Flemish dialect landscape. This insular position is so strong that Wondelgem, a suburban district which was integrated into the

361

362

III. Structure and dynamics of a language space Ghent conurbation in the 1950s, only took over the city dialect features for which the Ghent dialect is closer to the SL (see Oosterlinck 1992). So the changes in the (more rural) Wondelgem dialect (and in the other East Flemish dialects outside Ghent) must be attributed to convergence to the Belgian Dutch SL or to a Flemish intermediate variety (so-called tussentaal) with Brabantine characteristics, much more than to the influence of the Ghent city dialect (although the latter variety is a bit more resistant to influence of the SL). Bruges (110,000 inhabitants) is now a smaller regional centre, whose major asset is its great attraction to tourists. The town’s socioeconomic impact has become very limited since all industrial activities have been transferred to zones outside the town and especially to the port of Zeebrugge. The typical features of the traditional city dialect (e.g., diphthongization of long mid ee, eu and oo J eei, eui and oou) are being leveled out (even within the city walls) in the direction of a general rural West Flemish koine´ (bearing witness to horizontal leveling from outside). Observations of this kind generate a third thesis: the stronger the social/economic/ cultural basis of a place or a region, the more resistant the related dialect will be against leveling processes from outside. It will develop a considerable degree of “agentivity”: it will impose its characteristics horizontally (i. e., in dialect leveling) and even vertically (by contributing to the creation of intermediate varieties (cf. the role of Antwerp in Flanders) and even sometimes in destandardization processes).

2.2.2. Re (2): Dialect eature a vs. dialect eature b Much recent variationist research focuses on the fact that within a particular dialect (area) one feature can display a (much) higher degree of stability than others. In the relevant literature we see three important lines of investigation that may shed light on this phenomenon. These will be discussed in the following three paragraphs.

2.2.2.1. Primary vs. secondary vs. tertiary DIA-eatures In his (now famous) article on German Siedlungsmundarten ‘settlement dialects’ in the Soviet Union, Schirmunski (1930) made a distinction between dialect features that are quite liable to loss (which he called “primary”) and dialect features that more strongly resist any kind of accommodation (so-called “secondary” features). In fact, he considered these two types as conglomerates of characteristics. The so-called “primary” features are characterized by (1) a limited spatial distribution, (2) a remarkable structural distance to the competing elements and (hence) (3) a high degree of salience and awareness. Schirmunski concluded that primary features are (much) more liable to change than secondary ones. More than fifty years later, a new generation of variation linguists rediscovered this issue and confronted Schirmunski’s insights with contemporary linguistic facts. Thelander (1982) and Hinskens (1986 and 1993) offered convincing evidence for the proposition that DIA-features with a small-scale spatial diffusion are more liable to change than features with a wide diffusion. In this context, we should also mention Trudgill (1986) because of his introduction of the notion of “salience” and the discussion this aroused (see Hinskens, Auer and Kerswill 2005: 43⫺45). In an article on this matter (Taeldeman 2006), which was based on observations in the Flemish dialect landscape, I tried to extend and refine Schirmunski’s basic proposal in two ways: (1) next to “pri-

20. Linguistic stability

363

Tab. 20.3: Characteristics of tertiary, secondary and primary features TERTIARY (accent) ⫺ ⫺ ⫺ ⫺

a relatively big area gradual transitions very stable if any changes, then very slowly, gradually in all respects (distance, lexically, intermediate forms) ⫺ a very low degree of awareness (if any) ⫺ no attitudinal engagement with respect to the home form ⫺ no Ortsloyalität B difficult to suppress

SECONDARY ⫺ ⫺ ⫺ ⫺

intermediate intermediate changes are possible changes are more gradual in nature

⫺ an intermediate degree of awareness ⫺ no attitudinal engagement with respect to the home form ⫺ (almost) no Ortsloyalität B not difficult to suppress

PRIMARY ⫺ ⫺ ⫺ ⫺

a relatively small area sharp transitions strong liability to change changes are absolute in all respects (distance may be big, no intermediate forms)

⫺ a very high degree of awareness ⫺ a high attitudinal markedness ⫺ high (chance of) Ortsloyalität B easy to suppress

mary” and “secondary” features, I also distinguished “tertiary” DIA-features. Features of the latter type constitute what is commonly called a “regional accent” (cf. Walsh and Diller 1981, “low order processes such as pronunciation [which, J. T.] are dependent on the early maturing and less adaptive macro-neural circuits, which make a foreign accent [e.g., a standard accent, J. T.] difficult to overcome after childhood)”; (2) for the three types I designed a more extended conglomerate of characteristics. See Table 20.3 (taken from Taeldeman 2006: 247). From this table we can draw some conclusions with respect to the degree of stability of DIA-features: 1. Tertiary features are very difficult to suppress and are as such marked by a high degree of stability. Normally speaking, they will also be transferred to “higher” varieties (intermediate, substandard or even standard, see, e.g., Rys and Taeldeman 2007). 2. Secondary features are not difficult to suppress; hence their liability to change depends on a variety of mainly extralinguistic factors. Usually they will not be transferred to “higher” varieties. 3. Primary features are easy to suppress, but whether or not they are suppressed depends mainly on socio-psychological factors: in case of strong external pressure or a negative attitude they may or will disappear quickly (cf. Schirmunski 1930), but in case of a positive attitude they are very likely to survive (but will, normally, not be transferred to “higher” varieties). Especially when a primary feature becomes strongly associated with Ortsloyalität, it may result in polarization (the central theme in Taeldeman 2000 and 2006). When a dialect opposition is polarized, the following four phenomena co-occur: 1. Polarization and its spatial correlate, i. e., an abrupt transition between dialects, go hand in hand with at least two other polarization phenomena on the linguistic level: a. the linguistic distance (between two opposing features) tends to be either stabilized or maximized and no intermediate (compromise) forms arise;

364

III. Structure and dynamics of a language space b. the linguistic scope (i. e., the linguistic material that takes part in the opposition) also tends to be maximized, so no exceptions and no lexical diffusion would be expected. 2. On the spatial level, at least one of the participating variants has a small-scale diffusion. 3. On the mental level, there is a high degree of awareness of the opposition. 4. On the socio-psychological level, people take a strong attitudinal position towards the variant of their homeplace. The positive association of one or more linguistic forms with the home location may become so strong that a kind of “Ortsloyalität” may maintain the opposition (at least over a certain period). I admit that all the clear cases of polarization that I discovered in the Flemish dialect landscape are situated in relatively rural areas, where the overall conditions for a linguistic panta rhei (see the first paragraph of section 1) are not adequately fulfilled. As a kind of illustration, let me briefly outline a remarkable case of polarization in Flanders, and more particularly in the transition zone between the Brabantine and the East Flemish dialect area (see Map 20.2 for a more detailed localization).

Map 20.2: Short front vowels in the Flemish and Brabantine dialects

20. Linguistic stability In this transition zone, there are two villages, Vlekkem (300 inhabitants, Vl on Map 20.2) and Ottergem (500 inhabitants, Ot on Map 20.2), which are not more than two kilometers apart. One of the “deep” phonological oppositions between the East Flemish and the Brabantine dialects relates to the aperture of the short front vowels: these are much higher on the Brabantine side than in the old County of Flanders (to the west). Within the Flemish area, however, the short front vowels have their most open realization in Vlekkem and a few neighboring East Flemish settlements (including the small town of Wetteren): so [vis] ‘fish’ and [pyt] ‘pit’ in the Brabantine dialect of Ottergem contrast with [ves] and [pœt] in Vlekkem. In most of this transition zone, the transition is more gradual (with intermediate forms such as [ves] and [pøt]), but in and around the two villages mentioned the phonetic distance is extremely large. Precisely at this point there are five already mentioned dimensions of polarization: an abrupt (spatial) transition, extreme phonetic contrasts, no exceptions, an extremely high degree of awareness and of “Ortsloyalität”.

2.2.2.2. Structural stability and structural embedding There are many reports on the fact that the (degree of) stability of DIA features strongly correlates with (and depends on) the degree of (structural) embedding within the linguistic system. To a lesser degree, this is true for varieties which are also written, since written (SL) varieties have a strong conservative effect on their spoken equivalent. Dialects only make use of spoken registers and in the absence of a socially imposed norm they are more open to natural changes that act on “structural imbalances in the language system” (Van Marle 1997: 24). Van Marle quite clearly demonstrated that in (the Dutch) dialects exceptions in the mainly rule-governed components (syntax, morphology and postlexical phonology) are the first candidates to undergo “regularization processes” (mainly by analogy). Most of his examples are taken from the morphological component and related domains (morphophonology and morphosyntax) of the Dutch dialects. For example, he points out that verb paradigms which display (among other things) vowel alternation, are “regularized” much more rapidly in the dialects than in the SL (e.g., in some modal auxiliaries: standard ik mag ⫺ wij mogen becomes dialect ik mag ⫺ wij magge ‘I am ⫺ we are allowed’). The same applies to the so-called “strong” verbs (e.g., ik las J ik leesde ‘I read (past)’, ik ving J ik vangde ‘I caught’). With respect to the latter case, I can add an interesting observation for the Flemish dialects: in this area this tendency is by far the strongest in the East Flemish dialects (where forms such as ik leesde ‘I read’, ik slaapte ‘I slept’, ik vergeette ‘I forgot’, ik zwemde ‘I swam’ and ik zoekte ‘I sought’ occur very frequently), which is, not by accident, the dialect group with the weakest functional strength (see section 2.2.1.1). Another example from Van Marle (1997) deals with regularization processes in the system of the personal pronouns and more particularly with the tendency (at least in the plural) to equate the subject forms and the object forms, see Table 20.4. In general, we can say that linguistic elements which are rule-governed to a high degree also display a high degree of stability. This became very clear to me when I studied the phonology of my own dialect (the dialect of Kleit, a village between Bruges and Ghent): postlexical processes appeared to be much more stable than lexical phonological aspects (see Taeldeman 1993). In this perspective I also refer to some very interesting findings by Rys (2007), who investigated the acquisition of phonological aspects of a dialect by children and adolescents who had not been brought up in that dialect (at

365

366

III. Structure and dynamics of a language space Tab. 20.4: Personal pronouns in standard Dutch and in Netherlandic dialects Subject form(s)

Object form(s)

Standard

1st person 2nd person 3rd person

wij, we jullie zij, ze

ons jullie hun, hen

Dialect

1st person 2nd person 3rd person

ons jullie hun

ons jullie hun

home they spoke a “higher” variety of Belgian Dutch). One of her findings was that, on average, rule-governed aspects of this dialect (as a second variety) were acquired better than lexically determined elements. I presume that the same speakers, if they are confronted less intensively with this dialect at some later stage, will better retain the rulegoverned aspects of the dialect. In connection with this, the notion of “rule-governed”ness deserves some refinement: rules have a context and one context may be more complex than the other. In general, we may state that linguistic elements which are governed by a rule with a relatively simple context are more stable than elements which are generated by rules with a more complex context. In the latter case, language change by rule simplification is all but implausible (see, for instance, Taeldeman 2003 for several examples in the phonology of the Flemish dialects). Yet, we should remain cautious with respect to this. In the Flemish dialects I discovered a number of morphophonological rules with complex conditions which nevertheless displayed a remarkable degree of stability. I shall briefly discuss one example. In the Flemish dialects, adnominal words (articles, demonstrative and possessive pronouns and adjectives) end in -en before masculine nouns, but the inflectional -n only surfaces if the next word starts with a vowel, t-, d- or b- (see De Wulf and Taeldeman 2001 for more details). This is not an “elegant” rule since t/d/b do not constitute a natural class. However, t/d/p/b form the natural class of anterior stops, t/d form the natural class of coronal stops and d/b constitute the natural class of voiced anterior stops. One would expect that rule (context) simplification might take place, either to t/d/p/b, to t/d or to d/b. In reality, however, the unnatural condition holds out, across both space (cf. De Wulf and Taeldeman 2001) and time. The oldest dialect inquiry in Flanders and large parts of the Netherlands dates from the 1880s. In this material we also found the same unnatural context. The big problem with such regularization or simplification processes is that it is impossible to predict whether they will take place and, if they do, where and when this will happen. One factor seems to have universal validity: a (very) high frequency (both with respect to token frequency and type frequency) may protect irregularities or structural imbalances against regularization for quite some time (consider for example the paradigms of ‘to be’, ‘to have’ and ‘to go’, which in many languages even display suppletion).

2.2.2.3. Stability in the various components o a dialect In the literature of the last 25 years on language change and linguistic (in)stability, we often find rather casual statements in which the degree of (in)stability is related to a particular language component (e.g., syntax vs. morphology vs. phonology vs. [parts of] the lexicon). Yet, at least in my perception, there has not been massive research into this

20. Linguistic stability matter. An interesting exception to this is Van Bree (1990), who thoroughly investigated the stability (within the dialect ⫺ standard continuum) of syntactic, lexical, morphological and phonological dialect features in Twente, in the northeast of the Netherlandic area. He focused mainly on syntactic and lexical elements. I summarize his most important findings: (1) Syntax: a. It is impossible to situate the entire syntactic component en bloc at one particular point of a stability scale. b. Special regional constructions (e.g., in his case ik heb de band lek ~ SL mijn band is lek ‘I got a puncture’) are more stable than more straightforward aspects of word order (e.g., … dat hij gekomen is vs. … dat hij is gekomen ‘… that he has come’). (2) Lexicon: a. It is impossible to situate the entire lexicon at one particular point on a stability scale. b. Function words (e.g., articles and subordinating conjunctions) are more stable than adverbs of time and place and these adverbs in turn are more stable than full content words (e.g., concrete nouns). Discussing his findings, Van Bree refers to the degree of awareness and (with reference to Trudgill 1986) to the connected notion of “salience” as the most important stabilizing factor(s): if dialect speakers are more aware of (the “dialecticity” of) a linguistic element (and the more salient it is), there is a bigger chance that it will be leveled out by the “higher” varieties. Secondly, he also refers to the degree of spatial diffusion, frequency and the structural distance (between DIA and SL): if these factors are stronger, they will tend to promote dialect stability more. Van Bree’s article revealed that the four main components of a linguistic system cannot be situated en bloc on a stability hierarchy. In the preceding section we also demonstrated this point with respect to the phonological component: tertiary features appeared to be more stable than secondary ones, secondary features are more stable than primary ones and the latter usually display the lowest degree of stability, unless they are backed up by a (very) positive attitude and a high symbolic value, which may turn them into markers of Ortsloyalität. In general, it is my opinion that time and again we arrive at the same factors that determine the degree of stability: the degree of awareness or salience, the scope of spatial diffusion, the connection with positive or negative attitudes, frequency and ⫺ in the structural field ⫺ the degree of embedding in the system and the structural distance between DIA and SL and between DIA1 and DIA2. For the sake of completeness, we might also add the degree of abstractness: on average, content aspects are more stable than form aspects.

2.3. Stability in intermediate varieties (?) Auer (2005) rightly states that in most of the European language areas a kind of diaglossia (implying a verbal continuum from SL to DIA) has been created. The idea of a continuum goes hand in hand with the idea that within the intermediate field between the two poles (DIA and SL) there is no more or less stationary intermediate layer: no

367

368

III. Structure and dynamics of a language space commonly distinguishable “interlect” (Dutch: tussentaal). The notion of diaglossia seems more reconcilable with a situation in which every member of a particular language community selects his or her own mix of dialectical and more or less standardized elements. (This is the reason why the title of this section is followed by a question mark.) Yet, in the literature of the last few decades there are indications that also in this intermediate field between local dialects and (sub)standard varieties intermediate lects may ⫺ at least to a certain extent ⫺ come into existence. In the direction of the dialectical/regiolectical pole there are quite a few reports on koine´ization processes (see e.g., Sobrero 1996 and Berruto 2005 for Italy, Kerswill 1994, 2002 and Kerswill and Williams 2000 for Britain, etc. and Kerswill and Trudgill 2005 for a multilingual overview). In Auer and Hinskens (1996: 3; italics are mine, J. T.), koine´ization is defined as “the development (through dialect mixing, simplification and reduction) of a regional lingua franca”. The question now is whether in the “higher” part of that continuum (but clearly below SL) similar processes of stabilization and acceptance may come about, which in fact result in the development of a substandard koine´, which contains not just SL elements and DIA elements but also new elements (mostly compromise forms). Almost everywhere in Europe there appear to be symptoms of such developments, but to my knowledge one of the clearest examples of such a development in the direction of a substandard koine´ can now be observed in Flanders (Dutch-speaking Belgium). In what follows I briefly examine the ingredients of this new variety (see section 2.3.1) and discuss the factors that determine and explain this process (see section 2.3.2).

2.3.1. Ingredients o the Belgian-Dutch substandard koine´ Most ingredients of this Flemish tussentaal are borrowed from the Belgian variant of the Dutch SL, but there is also a surprisingly strong interference of the dialects and regiolects and especially of the Brabantine “superdialect” (see also section 2.2.1). If we analyze the “dialectogene” ingredients in more detail, it is striking that for some reason they were all designated as very stable and vital (cf. the previous section). They have much in common with what we have termed tertiary dialect features: a rather low degree of awareness and salience, a broad spatial diffusion and the absence of an attitudinal marking. A brief survey of the phonological and grammatical characteristics follows. 1. Phonology. Here we can distinguish two categories of features: (a) tertiary features in the sense of regional “accents” and (b) phonological features that are present in (almost) all of the dialects of Dutch-speaking Belgium. Referring to Rys and Taeldeman (2007) I mention (b.1) phonological alternations in the short function words which historically (and also in the Dutch SL) end in -t (dat, wat, niet, met), (b.2) deletion of final -e if the following word starts with a vowel (e.g., onz’ arbeiders ‘our workers’), (b.3) the tendency in …V(C)C #V… sequences to transfer the final consonant to (the onset of) the next syllable (resyllabification) (e.g., ont-erven ‘to disinherit’ J on-terven) and (b.4) the assimilation of t ⫹ d to [t] (e.g., uitdoen ‘to put off’ J ui[t]oen). Whereas the features under (1) cause regional diversity, the features under (2) have a leveling effect. 2. Grammar. Here I restrict my overview to the five dialectisms or regionalisms that obtained the highest score in an extensive corpus of spontaneous spoken Belgian

20. Linguistic stability Dutch by 38 Flemish politicians (see Van Laere 2003 for more details): (2.1) the inflection of adnominal words (e.g., ne langen dag ⫽ SL een lange dag ‘a long day’), (2.2) expletive dat as an intensifier after a subordinating word (e.g., ’k weet nie hoe da(t) ’k da(t) moet doen ⫽ SL ’k weet niet hoe ik dat moet doen ‘I don’t know how to do it’), (2.3) the generalization of gaan as an auxiliary of the future (e.g., da(t) gaa(t) nie(t) meer gebeuren ⫽ SL dat zal niet meer gebeuren ‘this won’t happen again’), (2.4) the intrusion of nonverbal constituents in the final verb group (e.g., … da(t) we moeten bezig zijn ⫽ SL … dat we bezig moeten zijn ‘…that we have to be busy’) and (2.5) the use of the pronominal ge-system (e.g., ge moet u realiseren dat … ⫽ SL je moet je realiseren dat … or u moet zich realiseren dat … ‘you have to realize that …’). Again these (and other) grammatical features are quite common throughout the Flemish dialect landscape. In general we may say that in Flanders an intermediate substandard variety (between the Belgian variant of Standard Dutch and the Flemish dialects) is developing and is becoming stable. The peculiarity of this process lies in the fact that it comprises all components of the language. In most other language areas of Europe the development of such an interlect (Flemish: tussentaal) seems to be related predominantly to the domain of phonology (regional “accents”) and much less to the grammar and the lexicon.

2.3.2. The (socio)linguistic context It looks as if in Flanders a rather exceptional standard ⫺ dialect constellation is being shaped in that between the two polar varieties a clearly recognizable, rather stable substandard variety with a kind of roofing function is developing. In general, we may consider this as a trans-componential conglomerate of (a) Belgian SL elements and (b) elements which are rooted in the Flemish dialects but for quite some time have been spreading all over Flanders. This tussentaal mainly results from three (socio)linguistic motives: 1. A deliberate dissociation from the “lower” varieties with too local or regional a flavor. These dialects or regiolects are increasingly considered to be incompatible with social prestige and, with the exception of a regional “accent” (tertiary phonological features), this will result in the loss of the traditional dialects. 2. A deliberate dissociation from the Belgian variant of the Dutch SL, which never could percolate through to the “lower” social strata. For this reason, the use of SL has always been felt as too rigid, too unnatural for everyday informal contacts. This negative attitude has been enhanced by a faulty kind of “linguistic engineering”: the Flemish language ideologists have been propagating the northern/Netherlandic SL as the only correct norm too rigorously and over too long a time period (see for example Jaspers 2001). 3. The need for a (more) natural Flemish variety which is fit for daily contact in all parts of Flanders and which is not felt to be dialect. This new and stabilizing interlect results from a spontaneous process without any linguistic engineering: this is the reason why so many Flemings feel “at home” when using it.

369

370

III. Structure and dynamics of a language space

3. Conclusions This quest for various kinds of linguistic stability provides some general conclusions: 1. Generally speaking, linguistic stability is becoming more and more a marked phenomenon. More and more we find a panta rhei situation. 2. Nevertheless, in all varieties on the vertical standard ⫺ dialect axis we have been able to discover several cases of relative linguistic stability, even in language communities which are fundamentally characterized by a situation of diaglossia. Under certain circumstances (see for example Flanders in section 2.3), stabilization processes may be at work within the linguistic continuum resulting in rather stable intermediate varieties. 3. The factors which may have a stabilizing effect appear to be of very different kinds. Some of these factors are more relevant for one type of language variety than for others. 4. I give a brief survey: a. Intralingual (structural) factors: in this context we elaborated on the factor of “structural distance” between two varieties: if this distance is bigger, the chance that this factor will have a positive effect on linguistic stability will be higher. Moreover, the degree to which a linguistic element is embedded in the system will contribute to greater stability. b. Extralingual factors: in the course of our explanation several kinds of extralinguistic factors turned out to be capable of exerting a positive effect on linguistic stability (under certain circumstances). i. Social: if the social, economical or cultural basis of “lower” varieties is stronger, they appear to resist leveling processes from outside better (both vertical and horizontal). ii. Functional: if the functional strength of [⫺SL] and more particularly [⫹DIA] varieties is greater, the chance that they will display some kind of linguistic stability will be higher. iii. Ideological: if normative control (in this case of the [⫹SL] variety) is stronger, the chance of relative stability will be higher (see for example the Belgian variety of the Dutch SL in Flanders). iv. Socio-psychological: in this respect we pointed out that some varieties or features may attain a high symbolic value for a particular (socially or geographically defined) group of language users. In this case this will contribute considerably to their linguistic identity and create positive attitudes. If this symbolic value is higher, some kind of stability of the varieties or features in question will be more likely. v. Mental: the factor mentioned above interferes with the degree of awareness (and salience) of a particular feature. Especially with respect to the “lower” varieties this factor appeared to act on the degree of stability in two ways: (1) in case of “upward” leveling processes, the degree of awareness turned out be inversely proportional to the degree of stability and (2) if a high degree of awareness coincides with positive attitudes, a strong stabilizing effect is to be expected and polarization is far from excluded.

20. Linguistic stability Our findings with respect to linguistic stability must certainly be confronted with more material from more language areas and from more language varieties. This will constitute a fascinating challenge for today’s and tomorrow’s variationist linguistics.

4. Reerences Auer, Peter 1998 Dialect levelling and the standard varieties in Europe. Folia Linguistica 32: 1⫺9. Auer, Peter 2005 Europe’s sociolinguistic unity, or: A typology of European dialect/standard constellations. In: Delbecque, van der Auwera and Geeraerts (eds.), 7⫺42. Auer, Peter and Frans Hinskens 1996 The convergence and divergence of dialects in Europe. New and not so new developments in an old area. In: Ammon, Mattheier, and Nelde (eds.), 1⫺30. Auer, Peter, Frans Hinskens and Paul Kerswill (eds.) 2005 Dialect Change: Convergence and Divergence in European Languages. Cambridge: Cambridge University Press. Auer, Peter, Frans Hinskens and Klaus J. Mattheier (eds.) 1997 Konvergenz und Divergenz von Dialekten in Europa. (Sociolinguistica 10.) Tübingen: Niemeyer. Bellmann, Günter 1997 Between base dialect and standard language. Folia Linguistica 32: 23⫺34. Berruto, Gaetano 2005 Dialect/standard convergence, mixing, and models of language contact: The case of Italy. In: Auer, Hinskens and Kerswill (eds.), 81⫺95. Christen, Helen 1998 Dialekt im Alltag: Eine empirische Untersuchung zur lokalen Komponente heutiger schweizerdeutscher Varietäten. Tübingen: Niemeyer. Coessens, Jan 1999 Taalgebruik, taalattitudes en lexicale standaardisering in Zottegem. Unpublished master’s thesis, University of Ghent. De Bruycker, Heleen 2001 Taalattitudes en lexicale voorkeuren in Zuid-Oost-Vlaanderen. Unpublished Master’s thesis, University of Ghent. Delbecque, Nicole, Johan van der Auwera and Dirk Geeraerts (eds.) 2005 Perspectives on Variation: Sociolinguistic, Historical, Comparative. Berlin/New York: Mouton de Gruyter. Dermaut, Joren 1999 Taalattitudes, taalgebruik en lexicale standaardisering in drie verkavelingen in Merelbeke. Unpublished Master’s thesis, University of Ghent. De Vogelaer, Gunter 1999 Taalgebruik, taalattitudes en lexicale standaardisering in Nieuwkerken en Sint-Niklaas. Unpublished Master’s thesis, University of Ghent. De Wulf, Chris and Johan Taeldeman 2001 Apocope en insertie van -n na sjwa in de zuidelijke Nederlandse dialecten: Conditionering en geografie. Taal en Tongval 14: 7⫺51. Geeraerts, Dirk 1999 Hoe “gans het volk” is de taal? Over Taal 38: 30⫺34. Gilles, Peter 2000 Dialektausgleich im Le¨tzebuergeschen. (Phonai 44.) Tübingen: Niemeyer.

371

372

III. Structure and dynamics of a language space Haas, Walter 1992 Mundart und Standardsprache in der deutschen Schweiz. In: Jan B. Berns and Jan A. van Leuvensteijn (eds.), Dialect and Standard Language, 312⫺336. Amsterdam: Noord Hollandsche. Hagen, A. M. 1990 Groepsportret van het Nederlands. Onze Taal 59: 32⫺39. Hagen, A. M. 1991 Waar is de regenboog gebleven? In: Herman Crompvoets and Ad Dams (eds.), Het Dialectenboek, 9⫺17. Waalre: Stichting Nederlandse Dialecten. Hinskens, Frans 1986 Primaire en secundaire dialectkenmerken: Een onderzoek naar de bruikbaarheid van een vergeten (?) onderscheid. In: Jon Creten, Guido Geerts and Koen Jaspaert (eds.), Werk in Uitvoering: Momentopname van de Sociolinguı¨stiek in Belgie¨ en Nederland, 135⫺158. Leuven/Amersfoort: ACCO. Hinskens, Frans 1993 Dialectnivellering en regiolectvorming: Bevindingen en beschouwingen. Taal en Tongval 6: 40⫺61. Hinskens, Frans 1998 Dialect levelling: A two-dimensional process. Folia Linguistica 32: 35⫺51. Hinskens, Frans (ed.) 2006 Language Variation ⫺ European Perspectives. (Studies in Language Variation 1.) Amsterdam: Benjamins. Hinskens, Frans, Peter Auer and Paul Kerswill 2005 The study of dialect convergence and divergence: Conceptual and methodological considerations. In: Auer, Hinskens and Kerswill (eds.), 1⫺48. Hoppenbrouwers, Cor 1990 Het Regiolect: Van Dialect tot Algemeen Nederlands. Muiderberg: Coutinho. Jacobs, Liesbet 1990 Kennis en gebruik van Nederlandse taalvarie¨teiten bij R.U.G.-studenten. Unpublished Master’s thesis, University of Ghent. Jaspers, Jürgen 2001 Het Vlaamse stigma. Over tussentaal en normativiteit. Taal en Tongval 53: 129⫺153. Kallen, Jeffrey L., Frans Hinskens and Johan Taeldeman (eds.) 2000 Dialect convergence and divergence across European borders. International Journal of the Sociology of Language 145. Kerswill, Paul 1994 A New Dialect in a New City: Children’s and Adults’ Speech in Milton Keynes. Final report. Reading: Dept. of Linguistic Science. Kerswill, Paul 2002 Koineization and accomodation. In: J. K. Chambers, Peter Trudgill, and Natalie Schilling-Estes (eds.), The Handbook of Language Variation and Change, 669⫺702. Oxford: Blackwell. Kerswill, Paul 2003 Dialect levelling and geographic diffusion in British English. In: David Britain and Jenny Cheshire (eds.), Social Dialectology: In Honour of Peter Trudgill, 223⫺243. Amsterdam: Benjamins. Kerswill, Paul and Ann Williams 2000 Creating a new town koine: Children and language change in Milton Keynes. Language in Society 29: 65⫺115. Kerswill, Paul and Peter Trudgill 2005 The birth of new dialects. In: Auer Hinskens, and Kerswill (eds.), 196⫺220.

20. Linguistic stability Kristensen, Kjeld and Mats Thelander 1984 On dialect levelling in Denmark and Sweden. Folia Linguistica 18: 223⫺246. Mattheier, Klaus J. 1985 Ortsloyalität als Steuerungsfaktor von Sprachgebrauch in örtlichen Sprachgemeinschaften. In: Klaus Mattheier and Werner Besch (eds.), Ortssprachenforschung, 139⫺157. Berlin: Schmidt. Oosterlinck, Carolien 1992 Fonologische dialectvariatie in Wondelgem. Taal en Tongval 5: 114⫺128. Pedersen, I. L. 2005 Processes of standardisation in Scandinavia. In: Auer, Hinskens and Kerswill (eds.), 171⫺195. Rys, Kathy 2007 Dialect as a second language. Linguistic and non-linguistic factors in secondary dialect acquisition by children and adolescents. Unpublished dissertation, University of Ghent. Rys, Kathy and Johan Taeldeman 2007 Fonologische ingredie¨nten van Vlaamse tussentaal. In: D. Sandra, R. Rymenans, P. Cuvelier and P. Van Petegem (eds.), Tussen taal, spelling en onderwijs, 1⫺9. Ghent: Academia Press. Sabbe, S. H. 2005 Dialecten in Zedelgem: Kennis, attitudes, functionaliteit en resistentie. Unpublished Master’s thesis, University of Ghent. Schirmunski, Viktor 1930 Sprachgeschichte und Siedelungsmundarten. Germanisch⫺Romanische Monatschrift 18, 113⫺122 (Part I), 171⫺188 (Part II). Siebenhaar, Beat 2000 Sprachvariation, Sprachwandel und Einstellung: Der Dialect der Stadt Aarau in der Labilitätszone zwischen Zürcher und Berner Mundart. Stuttgart: Steiner. Sobrero, Alberto 1996 Italianization and variations in the repertoire: the koinai. In: Auer Hinskens, and Mattheier (eds.), 104⫺111. Strijkers, Heidi 1995 Taalgebruik, taalattitudes en lexicale standaardisering in Melle. Unpublished Master’s thesis, University of Ghent. Stroop, Jan 1998 Poldernederlands: Waardoor het ABN Verdwijnt. Amsterdam: Bert Bakker. Taeldeman, Johan 1986 Dialekt versus Soziolekt in Übergangsgebieten. In: Peter von Polenz, Johannes Erben and Jan Goossens (eds.), Kontroversen, alte und neue, vol. 4, Dialektologie und Soziolinguistik: Die Kontroverse um die Mundartforschung, 263⫺272. Tübingen: Niemeyer. Taeldeman, Johan 1993 Dialectresistentie en dialectverlies op fonologisch gebied. In: Taal en Tongval 6: 102⫺119. Taeldeman, Johan 2000 Polarisering. Taal en Tongval 52: 229⫺244. Taeldeman, Johann 2003 Vlaamse klankfeiten en fonologische theoriee¨n. In: Georges De Schutter and Steven Gillis (eds.), Fonologische kruispunten, 71⫺84. Ghent: KANTL. Taeldeman, Johan 2006 Polarization revisited. In: Hinskens (ed.): 233⫺248. Thelander, Mats 1980 De-dialectalization in Sweden. Uppsala: Uppsala University. FUMS rapport, 86. Trudgill, Peter 1986 Dialects in Contact. Oxford, Blackwell.

373

374

III. Structure and dynamics of a language space Tsiplakou, Stavroula, Andreas N. Papapavlou, Pavlos Pavlou and Marianna Katsoyannou. 2006 Levelling, koineization and their implications for bidialectism. In: Hinskens (ed.): 265⫺ 276. Van Bree, Cor 1990 De stabiliteit van de syntaxis en andere taalsectoren. Taal en Tongval 3: 186⫺210. Vandekerckhove, Reinhild 1993 Dialectverlies in West-Vlaanderen? De vitaliteit van het Deerlijkse dialect. Taal en Tongval 6: 120⫺135. Vandekerckhove, Reinhild 2002 Vervlakking op de West-Vlaamse dialectkaart. Diminutiefvorming vroeger en nu. Taal en Tongval 54: 24⫺46. Van de Velde, Hans 1996 Variatie en verandering in het gesproken Standaard-Nederlands (1935⫺1993). Unpublished PhD dissertation, Katholieke Universiteit Nijmegen. Van Laere, Annelies 2003 Tussentaalelementen in de taal van Vlaamse politici. Unpublished Master’s thesis, University of Ghent. Van Lerberghe, Sabine 1999 Taalgebruik, taalattitudes en lexicale standaardisering in De Pinte en Gijzenzele. Unpublished Master’s thesis, University of Ghent. Van Marle, Jaap 1997 Dialect versus standard language: Nature versus culture. In: Jenny Cheshire and Dieter Stein (eds.), Taming the Vernacular: From Dialect to Written Standard Language, 13⫺34. London/New York: Longman. Videnov, Michail 1998 The present-day Bulgarian language situation: Trends and prospects. International Journal of the Sociology of Language 135⫺137: 11⫺36. Villena-Ponsoda, Juan A. 1996 Convergence and divergence in a standard-dialect continuum: Networks and individuals in Malaga. Sociolinguistica 10: 112⫺137. Walsh, Terence M. and Karl C. Diller 1981 Neurolinguistic considerations on the optimum age for second language learning. In: Karl C. Diller (ed.), Individual Differences and Universals in Language Learning Aptitude, 3⫺21. Rowley MA: Newbury House. Willemyns, Roland 1997 Dialektverlust im niederländischen Sprachraum. Zeitschrift für Dialektologie und Linguistik 64: 129⫺154. Willemyns, Roland 2003 Het verhaal van het Vlaams. Antwerp/Utrecht: Standaard Uitgeverij/Het Spectrum.

Johan Taeldeman, Ghent (Belgium)

21. Old minorities within a language space

375

21. Old minorities within a language space 1. 2. 3. 4. 5. 6.

Preliminary remarks How to define (ethno)linguistic minorities? Typology and areal patterns of established (ethno)linguistic minorities Minority language rights, legislation and language space: The territorial principle Minorities and language spaces: A glance at three case studies References

Ce n’est pas l’espace qui de´finit la langue, mais la langue qui de´finit son espace. (Pailhe´ 2007: 70)

1. Preliminary remarks Within the humanities, the notion of space is a central concept and a recurrent object of analysis mainly in the area of geography (cf. Johnstone in this volume), a science for which spatial reference ⫺ be it to areas, to structural or to distributional patterns ⫺ is fundamental and constitutes the key element of the way that this discipline defines itself. Space, in its traditional geographic definition, is understood as a three-dimensional entity, the extension of which may vary according to the structures and patterns to be taken into account. This primarily territorial view of space is due to the fact that the first and foremost objects of study in geography are “natural” spaces, which are delimited by boundaries considered as given by nature or, at least, reconstructible as traced primarily along physical, topographical units and not “manufactured” by humans (Harnisch in this volume). In linguistics, this geographically-based notion of territorial space has been taken up in (traditional) dialectology, which has been describing language variation within spatial entities similar to those that constitute the object of study of geography, i. e., primarily territorial ones. At the same time, however, dialectology has shown the insufficiency of territoriality to circumscribe language space(s): with the rare exception of zones of spatial contact between typologically distant languages, language borders have turned out to be extremely difficult to trace, as isoglosses are rarely bundled up in a clear-cut way and large zones of dialect transition are far more widespread than straightforward lines of separation. This echoes the insight gained in geography that many physical dividing lines considered to constitute “natural” boundaries at first sight have in reality not led to separation between cultures or populations at all, or, if they did, did so to a far lesser degree than one might have expected. As a reaction to the difficulty of defining spatial entities through “naturally given”, physically circumscribed territories, both geography and areally-oriented linguistics (including linguistic geography and geolinguistics; see Pailhe´ 2007: 67⫺70 on this distinction) adopted a concept of reference space based on man-made, “manufactured” boundaries, namely legal and administrative ones. This perspective had the advantage of being in accordance with the nation-state, the dominant model of socio-political organization in Europe and the parts of the world under European influence. In dialectology, this

376

III. Structure and dynamics of a language space opened the era of “national” language surveys and atlases, and language space, though still defined in a predominantly territorial way, became increasingly aligned to political space (cf. Gal in this volume). In geography, this perspective led to a wealth of studies focusing on individual regions, states, countries or conglomerates of countries. In a more recent step, geographers ⫺ mainly from the area of human and social geography ⫺ turned away from these rather static physical and territorial definitions and emphasized the dynamic character of geographic space, which in current reference work is defined as a functional space of action and interaction emerging through the pursuance of basic necessities of day-to-day life. In the realm of language studies, this re-perspectivization of language space was favored by the pragmatic turn in linguistics, which put strong emphasis on the situational and social circumstances that guided and, to a certain extent, determined language use. As in geography, the notion of space in areally-oriented linguistics became more dynamic and floating, thus allowing for the accommodation of increasing spatial and social mobility that is typical for contemporary society in industrialized countries. With language space(s) being increasingly determined by functional needs and interactional practices instead of fixed territorial entities, the attitudinal and representational dimension has gained importance, and the question that is now mostly addressed ⫺ namely in sociolinguistics ⫺ is how language space is constructed or imagined. And while some radical approaches in geography suggest to abandon the territorial dimension of the definition of geographic space altogether (Werlen 1993) in favor of a vision of space in permanent, interactionally-based emergence, in linguistics the understanding of communicative space has almost entirely shifted away from traditional dialects and is now based on a complex network of variational dimensions and languagedetermining variables subsumed under concepts such as the architecture of language (Berruto in this volume and Viaut 2007). It is important to bear in mind these different definitions and evolving conceptualizations of space when the issue of (ethno-)linguistic minorities and their relation to language spaces is addressed, and even more so when “old”, i. e., territorialized (established/ indigenous/autochthonous) minorities and their language(s) are concerned. It may legitimately be questioned whether it is adequate to speak of linguistic minorities within a language space, as the language through which such a minority is defined constitutes its own language space, both in a territorial and in a functional sense. On the other, it is obvious that the increase in bilingualization ⫺ a frequent phenomenon in minority contexts that has gained strength namely within the last century ⫺ is perhaps the most striking result of an overarching standard or national language intruding into the linguistic minorities’ “own” linguistic space and their language territory (cf. Viaut 2007 for the discussion of this distinction). This affects significantly the status of the minority, which, as will be shown in the following section, may be defined in very divergent terms anyway.

2. How to deine (ethno)linguistic minorities? At first sight, the most obvious approach to define linguistic minorities seems to be a quantitative one: a linguistic minority within a given space would be a part of the population making use of a distinct language or language variety, which has a lower number of members and/or occupies a smaller part of this territory than another or the other linguistically defined group of the population, which in turn would be designated as the

21. Old minorities within a language space majority. The socio-territorial criteria (how much territory of a given space is occupied by which part of the population) is rarely applied, contrary to the socio-demographic criteria. However, the socio-demographic attributes of minority/majority may shift according to the reference space that is taken into account: for instance, speakers of French are a minority within North America (as a geographic space) and within Canada (a political space; 22.3 percent of the population in 2006), with speakers of English as the majority. This still holds true, and even more saliently, in the Province of Ontario, where French speakers make up some 4.4 percent of the population. However, when the Eastern Ontarian County of Russell is considered as reference space, French speakers, with 62.5 percent of the county’s population, are no longer minoritarian; and in the ⫺ again administrative ⫺ space of the Village of Casselman, located within Russell County, French speakers ⫺ 86.5 percent of the townspeople ⫺ are the uncontested majority (figures from LeTouze´ 2004). Only the apriorism of considering the provincial or national level as the relevant reference space for a quantitative analysis allows describing the French speakers of Casselman as belonging to a linguistic minority. However, a purely quantitative approach to linguistic minority status is unable to show why the nation-state of Canada or the province of Ontario should be a more important communicative space for Casselman French speakers than the county or the village. It becomes obvious that the identification of the relevant and adequate reference space is frequently delicate and arbitrary when the case of French speakers in Canada is compared to that of speakers of French in Switzerland: whereas the former are generally awarded linguistic minority status in the relevant literature and such a classification does not normally raise much controversy, most scholars would feel uneasy to attribute the same status to the latter, although from a quantitative point of view, Swiss Francophones (with twenty percent of this country’s total population) are even more “minority-prone” than Canadian speakers of French. This is so because the Canadian French speakers tend to feel to belong to a linguistic minority (even if a speaker from above-mentioned Ontario probably has stronger feelings of this kind than his French-speaking neighbor in the province of Que´bec), whereas the Swiss French speakers generally do not feel as belonging to a linguistic minority or do so only in particular situations or in specific small-scale geographic border areas. This is something that numbers will not tell. Since a purely quantitative definition of minority status quickly turns out to be insufficient to capture reality, various more elaborate definitions have been proposed. These definitions take into account qualitative aspects of minority status, namely those alluded to somehow loosely in the previous paragraph as “feelings”, i. e., attitudinal, perceptional and other psychological factors, but also quantifiable criteria that go beyond socio-demographic data and concern the socio-economic or socio-political stand of communities within a given space. According to these definitions, minority groups are characterized by their limited access to key positions in the economic, political and cultural sphere in the administrative space in question ⫺ typically the nation-state they live in ⫺ and by their relatively lower position on the average income scale or in relation to the level of education that the typical member of the minority group receives. Economic and political marginalization of this kind was an important aspect of the so-called “Quiet Revolution” that took place in French-speaking Canada, mainly in the province of Que´bec, in the 1960s, where French Canadians ⫺ who in Que´bec, at that time, were, and still are, the socio-demographic majority ⫺ realized that not only was their mother tongue rarely heard in political institutions but that the average income of a francophone person was also significantly below that of an English-speaking citizen. Furthermore,

377

378

III. Structure and dynamics of a language space speakers of French were blatantly under-represented in top positions in private enterprises as well as in public administration. Educational marginalization does not normally mean, at least in developed countries, that no access to schooling is provided but that schooling is available only in the majority or dominant language, so that members of linguistic minorities receive education in a language other than their primary language, which may lead to less satisfactory results and increased educational failure. However, the most crucial aspect of qualitative definitions of minority status seems to be the sociopsychological one that concerns attitudes (cf. Gallois, Watson and Brabant 2007): an (ethno-)linguistic community may only be regarded as a “minority” if it is considered as such by the community itself and by the other group(s) in the reference space. That is, despite the fact that many attributes and features associated with minorities are without doubt externally motivated and therefore “given”, they may only affect the community in question if members of the minority (and members of the majority) are aware of these features and conscious of their relevance. Thus, minorities are always created to a certain extent, either by “the others”, i. e., the members of the majority, through an act of heterocategorization (Krefeld in this volume) or by themselves through autocategorization as a different, culturally, linguistically or otherwise deviant, less powerful or less “important” part of the whole population of a nation-state or another reference space. Theoretically, both auto- and heterocategorization as a minority group may have positive or negative outcomes. In reality, though, the case of a (socio-demographic) majority which holds in high esteem and perhaps even admires an ethno-linguistic minority is much rarer then the contrary constellation where (socio-demographic) minorities are considered as inferior, as “underdogs” or political trouble-makers by the majority group. An example for the seldom-found former situation might be the attitude that the German-speaking majority in Switzerland has towards the Rhaeto-Romance-speaking minority in their country, an attitude that, as sociolinguistic surveys have repeatedly indicated, is generally favorable and even very positive. Self-awareness of minority status, on the other hand, may lead to ambiguous and contradictory statements on this issue, too. It is well-known that members of ethno-linguistic minorities may show very negative attitudes towards their own community, language and culture when they become aware of its minority status and of the fact that the majority considers them as minoritarian, a phenomenon labeled as self-hatred. This may lead to extreme cases of self-induced alienation and loss of one’s own identity. For instance, among heritage speakers of socalled Islen˜o or Brule´ Spanish in the New Orleans region of Lousiana, in the southern United States, it is reported that members of the community not only change their linguistic behavior, but also eliminate all ethno-linguistic attributes as they were aware that, as such, they were considered as uneducated, uncultivated, poor and “dirty” by members of the English-speaking majority group (Holloway 1997).

3. Typology and areal patterns o established (ethno)linguistic minorities Among the different typologies and ensuing terminological proposals for minorities and situations of minorization, Edwards’ (1995; 2004: 463⫺467; cf. also Turell 2001) model is particularly suited to shed light on the recurrent patterns of spatial localization and organization of minorities. Edwards’ grid involves three fundamental distinctions:

21. Old minorities within a language space The first is among minority languages that are unique to one state, those that are non-unique but that are still minorities in all contexts in which they occur, and those that are minorities in one setting but majority varieties elsewhere. This initial distinction gives rise to the terms unique, non-unique, and local-only minority. The second categorization deals with the type of connection between speakers of the same minority language in different states; are they adjoining or non-adjoining? Finally, what degree of spatial cohesion exists among speakers within a given state? Here, the terms cohesive and non-cohesive can be used. (Edwards 2004: 465)

As can be deduced from the quote, Edwards follows the general practice and considers the legal entity of the (nation-)state as the relevant reference space for situating linguistic minorities. However, the model clearly favors a trans-national perspective in that the minorities’ areal patterns are analyzed, not only within the political space and the concomitant language space of the national or “official” language(s), but also within neighboring political or national entities, by the primarily geographic criteria of spatial adjunction and cohesion. As a matter of fact, when nation-states and their territorialized established minorities are graphically represented ⫺ a language map of Europe being the prime example (cf., e.g., Plasseraud 2005: 12⫺13) ⫺ it is a remarkably recurrent fact that the areas where minority languages are spoken are located at the fringes of the states. Synchronically, therefore, many ethnolinguistic minorities are border minorities, whereas inland minorities are less common. This distributional pattern is obviously the result of case-specific historical events but may also be linked to more general regularities of horizontal diffusion of innovations that concern, among others, social structures and language use: such innovations and changes usually irradiate from a powerful center point or area, frequently ⫺ although not necessarily ⫺ the national capital or capital region. This leads to processes of increased language contact, bilingualism and diglossic functional hierarchization, and perhaps, ultimately, to language shift in areas close to the national diffusion center but not necessarily reaching, or having the same impact on, more peripheral zones. Such an historical development is exemplified by the language distribution on the Iberian Peninsula between the beginning of Christian re-conquest of Arab-dominated areas from the eighth century onwards and the present, where, among other reasons, the establishment of a powerful administrative and cultural center in the form of a capital region in the Toledo-Madrid area favored a phase of lateral expansion of the Castilian language. It is tempting to link the socio-economic, political and cultural marginalization that characterizes linguistic minorities in the quantitative approaches mentioned above to their spatial marginalization, with the latter conditioning the former. The example of the linguistic landscape on the Iberian Peninsula, however, clearly shows that geographic marginalization within a nation-state is quite often a secondary process and a consequence of incipient or even entrenched political domination and socio-economic marginalization, and that the relationship between peripheral geographic location and socio-political subordination is a complex one, especially when the diachronic axis is duly taken into account. Another recurrent phenomenon in Europe’s nation-states which explains many cases of border minorities is that political frontiers frequently cut through traditional language zones. Some political borders of modern nation-states go back to older legal separation lines that had been drawn in a period where the ideal of linguistic homogeneity ⫺ along the “one nation ⫺ one language” line of thinking typical for many modern states (Do-

379

380

III. Structure and dynamics of a language space rian 2004: 441⫺443) ⫺ did not have much importance. In other cases, borderlines are the results of battles, conquests and diplomacy. But on many occasions in the process of nation-building, the myth of “natural frontiers” has been put forward, with political powers seeking to establish ⫺ through war and violence if necessary ⫺ borders along geographical guiding lines such as mountain ranges or rivers. The border between France and Germany in the Upper Rhine region is a point in case, where the French considered the river as a “natural frontier”, whereas the Germans ⫺ with some additional linguistic arguments at hand ⫺ found the ridges of the Vosges Mountains to constitute a “natural” borderline. On closer examination, it turns out that both sides were wrong as neither the mountains nor the river really had lead to a hermetic separation of ethnic groups, languages and cultures. An blatant example of a mismatch between nation-state borders traced along “natural” guidelines and ethnolinguistic zones is provided by the Francoprovenc¸al language group, which stretches over a geographical area that embraces the highest peaks of Europe but which is now split up through national borders among France, Italy and Switzerland. Cases like this correspond to Edwards’ unique adjoining minorities; whereas the French-German borderline that follows the Upper Rhine has led the Alsatian dialect of German to become a local-only adjoining minority language. It is hardly surprising that the latter type of minority, with a large and adjacent hinterland sharing the same language and culture, is, generally speaking, in a much more favorable socio-linguistic situation than the former type, with German in eastern Belgium or northern Italy (South Tyrol) ⫺ but less so in Alsace ⫺ being prime examples. Although the importance of geographic adjacency of border minorities belonging to the same ethno-linguistic group is beyond doubt, immediate geographic contact is not necessary for contiguity effects to arise; in Pailhe´’s (2007: 66) wording, connectivity is more decisive than proximity. For instance, neither the Danish speakers in northern Germany (Schleswig-Holstein) nor the speakers of Swedish on Finland’s south and west coast are in direct and uninterrupted spatial contact with their language neighbors “on the other side” but still profit from their linguistic hinterland in about the same way as German speakers in South Tyrol or Flemish speakers in northern France do. If the linguistic and cultural hinterland is itself in a minority situation and lacks political or socio-economic vigor, no strengthening contiguity effects show up, as, e.g., in the case of Occitan speakers in northern Spain (Val d’Aran) and northwestern Italy (Piemontese Valadas Occitanas) who do not get any an advantage from the (geographically) large Occitan-speaking territory in southern France (cf. section 5.1). Edwards’ third distinction between cohesive and non-cohesive minorities also is a primarily geographic one, but again geography and spatial distribution patterns will not tell the whole truth. The fact that ethnolinguistic communities do not occupy a continuous area but are scattered may be due to the fact that the minority is non-sedentary, but then it would not count as a territorialized minority anyway; or it may have purely geographical reasons, as for example in the case of Rhaeto-Romance speakers in the Swiss canton of Grisons (Graubünden) and around the Sella Massif in northern Italy, where the sub-groups (and their different language varieties) are separated by high mountain areas that do not permit permanent settlement. The geographic separation of the Swiss and the northern Italian Rhaeto-Romance communities, or the scattering of Acadian French speakers in eastern Canada (cf. section 5.3), are due to the immigration of people speaking other languages into a once more cohesive (minority) language space, with the consequence of an areal retreat of the latter, as in the case of the Rhaeto-

21. Old minorities within a language space Romance communities, or geographical displacement, as in the Acadian example. Migration of this kind may also lead to very isolated, small ethnolinguistic groups known as Sprachinseln in the German terminology (Riehl in this volume). Whether different non-contiguous sub-groups of linguistic minorities have strong internal ties or not is also conditioned by psychological, attitudinal and political factors. Where such ties exist, for instance, in the Acadian French communities (including Louisiana’s Cajun-French minority), common roots and a common heritage have been a central element of the cultural memory of the community. On the other hand, attitudinal factors and political action may also undermine common heritage and psycho-linguistic ties and thus lead to a loss of internal cohesion, even in cases where there is no spatial separation at all. For instance, the decrease of internal cohesion between the Catalonian and the Valencian sub-group of the Spanish mainland Catalan minority is largely, if not exclusively, ideology-driven and the result of political instrumentalization (cf. section 5.1), despite the fact that both communities are territorially contiguous and not separated by any geographic dividing line whatsoever.

4. Minority language rights, legislation and language space: The territorial principle A key topic in the literature on ethnolinguistic minorities and the future of their languages is the juridical situation of these communities and, more specifically, the legislation concerning the use of the minorized language(s). The status of language rights, namely in relation to more general human rights, is the subject of a vast body of literature (cf. de Varennes 1996; May 2001: 128⫺197). Among the debated points is the question whether language rights are individual or collective rights, but also whether the non-discrimination or protection of ethnolinguistic minorities can and should be formulated in the form of legal rights or if they are better looked upon as moral rights (Edwards 2007: 453⫺455). As a derivation or “particularized application” (Edwards 2007: 453) of basic human rights, the right to use a specific language for a specific purpose in a specific social setting is a priori an individual right, but since language as a communication system necessarily involves at least two interlocutors and, in order to really function as such, a speech community, language rights are de facto individual and collective/social rights at the same time. If language rights are considered legal rights in some way, and some political action in the form of language planning and language policy is taken to enforce them (cf. Skutnabb-Kangas 2007), they are unavoidably territorialized, since planning actions and policies can only be implemented in an area over which the social actors in question have power and jurisdiction (Viaut 2007: 48). This antagonism may be captured, following Grin (1994), in the contrast between two principles, the personality principle and the territorial principle; both play an important role when established minorities are considered, but the latter is intrinsically linked to the question of minority languages and space. The territorial principle “traditionally means that each language should correspond to a specific area, in order to ensure the latter’s linguistic homogeneity; the language rights enjoyed by individuals are then conditional on their geographic position” (Grin 1994: 35). Although said to be successfully applied in various nation-states, most prominently perhaps in Switzerland, the territorial

381

382

III. Structure and dynamics of a language space principle raises the crucial question of how the language area(s) are determined. This can be done on synchronic, and predominantly socio-demographic, or historical grounds, or a fusion of both. Since legal action in order to ensure the territorial integrity of a language is frequently taken only when this integrity (and the minority language itself) is endangered, a demographic definition can be problematic since at this stage the internal cohesion of the linguistic minority will already be low, bi- or multilingualism prevalent and allophone immigration into the area widespread. Therefore, a critical threshold value is frequently applied: on this basis, a zone is considered to belong to the area covered by the territorial principle if the linguistic minority has a certain demographic weight there, even if it is below fifty percent. This may still be insufficient to prevent “holes” in and around urban centers (if there are any within the minority’s area). As a purely synchronic definition of language territories would not lead to results that are in accordance with the very aims of the territorial principle ⫺ namely continuous and stable areas ⫺ some kind of diachronic argument is also usually invoked. For instance, the territory of a language may be defined as an area where a language was traditionally spoken and historically dominant. But which period or point in history should be taken as a reference point in order to decide where a language was “traditionally spoken”? The Italian Fascist activist Ettore Tolomei justified the forced assimilation or expulsion of the German-speaking minority in South Tyrol by arguing that their area of settlement belonged to the “traditional” zone of Italian or, put in more cautious a way, of the ItaloRomance varieties, because the Romans had colonized this area some two millennia earlier and the spread of Germanic was but the result of illegitimate posterior “intrusion”. A similar argument has been made by certain representatives of the Flemish community at the peak of the Belgian language battle when they claimed ⫺ in the name of the territorial principle! ⫺ that Brussels, which had been a Flemish-speaking town for many centuries but had undergone a process of language shift in the late eighteenth century, should be included into the monolingual territory of Flanders, although currently speakers of French clearly dominate with around 85 percent of the city’s population (Nelde 1982). Theoretically, toponymy might help to determine the area where a language had been “traditionally spoken”, but toponyms are sometimes difficult to interpret and to assign to precise historical layers (and languages). They would in most cases lead to overlapping language territories, i. e., to the contrary of what the territorial principle aims at. As a consequence of these problems, in most cases areas to which the territorial principle has been applied are the result of a compromise between historical, sociodemographic, more strictly linguistic and political factors. Boundaries between territories have been determined during the twentieth century mostly with regard to the contemporary situation, but in borderline cases evidence from administrative documents and linguistic sources (namely dialectological surveys that started to be available from the late nineteenth century on) have been taken into account as well. Therefore, minority (and, conversely, majority) language territories mostly reflect the extension of the language at the moment when language-political action was taken, or just before. Where neither reliable historical nor contemporary linguistic evidence was available and the question of the boundary remained unsettled, political considerations prevailed and the alignment of language territories with already established administrative entities was given preference. In some cases, small zones along the territories’ border have been “swapped”. An emblematic case is the Belgian municipality of Fourons/Veuren, a traditionally multilin-

21. Old minorities within a language space gual zone with a French-speaking majority which used to be part to the Francophone province of Lie`ge but joined the Dutch-speaking Flemish province of Limburg during the 1960s as a compensation, as Flanders had lost some municipalities to the Walloon region. It is mainly in such areas along the border that conflicts arise later on and that the territorial principle is challenged, usually without success. A particularly delicate issue, which frequently occurs but is not restricted to states and regions where the territorial principle applies, are so-called included minorities. Sometimes, such minority groups within the territory of a politically or demographically superordinate ethnolinguistic minority receive special attention and political support; the Occitan-speaking (sub-)minority of the Val d’Aran in Catalonia is a point in case, as well as the Rhaeto-Romance-speaking Ladin population living within the territory of the German-speaking minority in South Tyrol/Northern Italy. The contrary may be true when these sub-minorities are made up of speakers of the dominant (national) language on the state level; as examples, Anglophone inhabitants of Montre´al and other areas in Que´bec/Canada or Castilian speakers in Barcelona and other parts of Catalonia come to mind. According to Grin (1994: 38) such an unequal treatment of languages in an overall clearly asymmetric sociolinguistic situation is in accordance with the territorial principle when applied to minority languages that are threatened languages at the moment when legal action is taken. However, this author is aware of the inherent dangers for social stability if language rights are systematically withheld from included minorities, and particularly local-only minorities, on all levels. In order to achieve a balance of rights, he therefore develops the concept of “territorial multilingualism” (see Grin 1994: 41⫺45 for details).

5.

Minorities and language spaces: A glance at three case studies

5.1. Catalan and Occitan speakers in Spain and France The Catalan- and the Occitan-speaking communities (cf. the maps in Plasseraud 2005: 24⫺27) are typical “old” minorities that have not moved from their territories since their languages came into being, as a result of the Roman colonization of Western and Southern Europe between the third and the first century b.c. Up to the early twentieth century, Catalan and Occitan were treated as one language (or diasystem) by many scholars in Romance philology due to internal criteria, although this ⫺ nowadays contested ⫺ view might have been biased by the fact that written Occitan has remained the dominant literary language in the Catalan-speaking area for a longer period than anywhere else. Typologically, they share basic Gallo-Romance features and features that link them to the realm of Ibero-Romance, although these features are much more visible in Catalan than in Occitan (and restricted, in the latter case, to the Southwestern dialects of Gascony). Historically, both languages also have much in common. After a period of literary splendor, associated with high political and economic influence, a period of decay followed in the Middle Ages (for Occitan) and the Early Modern Times (for Catalan), which came to an end with the rise of European Romanticism in the nineteenth century only. The current situation of both languages, however, is totally different, and consciousness of common origin or specific links between the two communities is absent, with the

383

384

III. Structure and dynamics of a language space exception of a small (mostly academic) elite. On European language maps, both languages seem to occupy rather large and coherent territories. Occitan covers the southern third part of France and Catalan two-thirds of the eastern coastal fringe of the Spanish state (plus isolated island areas). But these maps are misleading (Pailhe´ 2007: 76; cf. for instance Plasseraud 2005: 25 for a more differentiated cartography of Occitan). Catalan is much more vivid than Occitan and has a significantly better sociolinguistic standing. The Catalan-speaking community may be estimated at 7 to 8 million, whereas the Occitan-speaking community is impossible to quantify, due to the absence of corresponding census data, but estimates range between 100.000 and several million speakers. According to Edwards’ topological typology, both speakers of Occitan and Catalan would most readily be qualified as non-unique adjoining cohesive minorities. For the Catalan community, however, such a classification is not entirely correct, not only because of the presence of the language on the Balearic Islands and a small area in Northwestern Sardinia, but also because Catalan is not considered as a minority idiom in the Principality of Andorra, where it is the only official language. However, apart from some symbolic acts (e.g., Catalan songs due to Andorran participation in the recent editions of the European Song Contest) little effect is derived from this status. Today Catalan is a de facto minority language even in Andorra as autochthonous Andorrans are now outnumbered by immigrants (cf. Sinner 2004). In the far more populous Spanish part of the Catalan territory, the language enjoys a relatively good sociolinguistic and legal status and is co-official with the dominant national language Castilian, a status that it shares with the languages of the other “historical” autonomous communities within the Spanish state such as Basque in the Basque Country and Galician in Galicia. Furthermore, Catalan has a strong position as the language of administration and schooling, especially in the autonomous regions of Catalonia and the Balearic Islands (less so in the Valencian Community region), and a remarkable presence in the media, in science and in commerce. These and other features make the Catalan-speaking community a rather untypical example of an ethno-linguistic minority and totally distinct from the Occitan-speaking community. Occitan has official status in a small area only, i. e., in the previously mentioned Val d’Aran, which administratively belongs to Catalonia and, thus, to Spain. In the case of Occitan, the lack of a generally accepted standardized language (Kremnitz 1974; Sumien 2006: 15⫺110) has led to weak psychological cohesion. Dialectal divergence is significant, enforced by some ⫺ albeit minoritarian ⫺ community activists and even linguists who challenge the linguistic unity of Occitan (cf. Blanchet and Schiffman 2004 for samples of this position). Many speakers believe that they do not speak a language of their own at all, but a patois, i. e., some kind of “bad French”. The case of the Catalan community is different in this respect: internal psychological cohesion is much stronger, and the consciousness to speak the same language is high. However, challenges to the ⫺ linguistically entirely uncontroversial ⫺ unity of the Catalan language space have arisen in recent decades through an increasing political antagonism between the two large mainland regions where Catalan is spoken, i. e., Catalonia and the Valencian Community (Doppelbauer 2006; Triano Lo´pez 2007). This so-called “language conflict” in the Valencian community with respect to Catalonia is in fact more readily explained in terms of a political and economic competition between the regions and a result of battles between political parties and social groups with divergent ideological orientations. Still, this conflict undermines significantly the position of Catalan and the

21. Old minorities within a language space entire Catalan-speaking community within the Spanish state and its national, Castiliandominated language space. In legal terms, in post-Franquist Spain the latent conflict between individual and social language rights is tamed by the co-officiality of the national language ⫺ Castilian ⫺ and the regional languages ⫺ such as Catalan ⫺ within the “traditional” territories of the latter. This legitimates and, eventually, promotes bi- or multilingualism within the minority languages’ territories and therefore departs significantly from some basic ideas behind the territorial principle. In Finland a completely different approach to the reconciliation of individual and collective language rights and linguistic territoriality has been chosen.

5.2. Swedish speakers in Finland Finland (Suomi) is an officially bilingual state (Latomaa and Nuolijärvi 2005 [2002]) that came into being in 1917, after more than a century as an autonomous Grand Duchy of Russia and ⫺ more important for the language issue ⫺ after having been under Swedish rule between the early fourteenth century and the early nineteenth century. 500 years of Swedish presence left their traces on the linguistic landscape (Tandefeldt and Finnäs 2007): until independence was gained, Swedish was not only the dominant language of politics, administration and commerce but was also considered the language of the elite, so that a typical situation of diglossia was in place, when in the nineteenth century Finnish nationalism developed and the use of Finnish began to be promoted also in formal registers. Geographically, speakers of Swedish concentrated in the cities and along the southern and western coasts of the country, areas that had experienced important waves of colonization and settlement during previous centuries; the Swedishspeaking community of Finland therefore is a significantly “younger” minority than the Occitan- and Catalan-speaking communities presented above. Today, this geographical pattern is still valid (cf. the cartographic presentation in Plasseraud 2005: 35), with Swedish-speaking Finns living on small coastal stretches and on nearby islands in the south ˚ land Islands, which are closer to the Swedish than to the and in the west and on the A Finnish mainland. In 2001, 92.3 percent of the Finnish population had Finnish as their first language, while 5.7 percent spoke Swedish as a first language; in the 1920s at the moment of political independence, their share had been at 11 percent (figures from Latomaa and Nuolijärvi 2005: 128 and 135). According to the (recently reformed and updated) National Language Act from 1922, no territoriality principle applies in communicative exchanges with the national government, where a reply in either language may be expected. A strict territorial principle, comparable to Switzerland or Belgium, is only observed on ˚ land archipelago, where Finnish is almost non-existent in daily life, with 93.5 perthe A cent of the population indicating Swedish as their mother tongue (Svenska Finlands folkting 2004: 19). On the mainland, however, the language status is determined on the administrative level on the basis of rather “generous”, i. e., minority-friendly numerical criteria: a municipality is considered bilingual when the minority group ⫺ be it Finnishor Swedish-speaking ⫺ makes up at least eight percent of the population or totals up to a minimum of 3,000 speakers. In bilingual municipalities, both languages may be used with local authorities and schooling is provided for both groups. Currently (2002), there

385

386

III. Structure and dynamics of a language space are 385 monolingual Finnish-speaking municipalities in Finland, as compared to 21 monolingual Swedish-speaking municipalities. Out of 42 officially bilingual municipalities, 22 are predominantly Swedish- and 20 are predominantly Finnish-speaking; among the latter are large cities in the South, where the percentages of Swedish-speaking Finns are low but absolute numbers exceed 3,000, such as the national capital Helsinki (Helsingfors in Swedish; 6.4 percent/35,800 Swedish-speakers) and the cultural centre of the ˚ bo; 5.2 percent/9,000). The highest ratios of Swedish-speakers Swedish Finns, Turku (A on the mainland are found on the western coast, in the Ostrobothnia region, where several municipalities boast Swedish majorities between 90 and 97 percent. The Finnish system does not normally lead to coherent language territories as favored by the territorial principle. Still Finnish-Swedish bilingualism is generally characterized as “harmonic” and mostly free of conflicts (Saari 2000: 5⫺6), which may be attributed to two facts: first, Swedish is not an endangered language and the Swedish-speaking community of Finland is, according to Edwards’ typology, a local-only minority which receives significant institutional and medial backing from Sweden. And, second, the members of the community generally do not feel as Swedes living in another nationstate but as Finns speaking another language, and this language is considered as an element of the national heritage not only by Swedish-, but also by (many) Finnishspeakers. Therefore, it would seem inadequate to designate the Swedish-speaking community in Finland as an ethno-linguistic minority; despite much cultural specificity, it is best described as a linguistic-only minority. The linguistic harmony, however, seems to be less assured in recent years than before: although the demographic weight of the Swedish-speaking group has remained stable in absolute figures between the late nineteenth and the twenty-first century, the Finnishspeaking population has grown more rapidly and the relative weight of Swedish-firstlanguage speakers has gone down by two thirds between 1880 and the present (Latomaa and Nuolijärvi 2005: 135). Opposition to compulsory teaching of the second national language surfaces regularly and most vigorously from the Finnish-speakers, many of whom no longer see much utility in learning Swedish and advocate an earlier and longer inclusion of international languages ⫺ namely English ⫺ in the schools’ language curriculum (comparable to current moves within certain German-speaking cantons of Switzerland). In the end, such a readjustment would lead to a lack of persons among the Finnish-speaking majority group with a sufficient proficiency in the minority language to comply with the requisites of the official bilingualism. Such an evolution would, in the long run, destabilize the language situation in Finland and weaken the position of the Swedish-speaking community within the national Finnish language space.

5.3. Acadian French speakers in Canada The recognition and protection of a clearly circumscribed and stable language territory is aimed at by most “old”, established ethno-linguistic minorities and considered a requirement for their survival. In view of this fact, it is rather surprising to find an “old” minority for which the non-existence of a language territory of its own fulfils this identificational and emblematic function; speakers of Acadian French in the eastern provinces of Canada are such a minority group. In the discourse of Acadian-French speakers, the reference to a common territory called “Acadia” (Acadie) is omnipresent,

21. Old minorities within a language space although this territory is virtual, not visible on maps and atlases and not delimitable as a geographic reality. Obviously, speaking of “established minorities” in the Canadian context is ambiguous and controversial; strictly speaking, only the Indian and Inuit first nations would legitimately qualify as “old”, autochthonous minorities. However, among the European migrant groups that settled on Canadian territory, the Acadian-French or, more briefly: the Acadians were indeed among the first to arrive: After the foundation of Port-Royal (today’s Annapolis-Royal) in 1604 by the discoverer and colonizer Samuel de Champlain, French migrants mainly from western and central parts of France settled on the A(r)c(c)adie island known today as Nova Scotia (Nouvelle-Ecosse) (cf. Griffiths 2005; Kolboom 2005 for historical details). Since they wanted to maintain a neutral position vis-a`-vis both competing colonial powers in the region, France and England, in the mideighteenth century they refused to swear an oath of allegiance on the British Crown, which led to their expulsion from Nova Scotia and their deportation by the British to New England and other British colonies on the American East coast, to England and back to France. A number of Acadians arrived, sometimes on very roundabout ways, on the marshes of the lower Mississippi River, in the future U.S. state of Louisiana (then under Spanish rule), where they formed the francophone community known today as “Cadiens” or Cajuns. Other Acadians managed to escape from deportation and took refuge, with the help of the befriended Micmac Indians, in inaccessible areas of the nearby mainland province of New Brunswick. During the late eighteenth and early nineteenth century, many Acadians who had been brought to European destinations returned to Canada, but, since their former territory in Nova Scotia was no longer available to them, settled elsewhere in the Atlantic provinces, the French islands of Saint-Pierre and Miquelon and in Newfoundland. Due to this varied history, the Acadian minority, which in eastern Canada outside Que´bec amounts today to some 255,000 persons, is scattered over the entire Atlantic region (cf. Rossillon 1995: 73 for maps), but the highest concentration of AcadianFrench speakers is found in northern, eastern and southeastern New-Brunswick (Nouveau-Brunswick), where stretches, e.g., along the Upper Saint-John River around Edmundston, opposite the US state of Maine, and on the so-called Acadian Peninsula east of Bathurst are almost exclusively French-speaking. It is also in New-Brunswick that the Acadian-French minority has the most favorable sociolinguistic status, since the province adopted in 1968 an Official Language Act (updated in 2002), through which official bilingualism was installed (cf. Kolboom 2005: 230⫺246); New-Brunswick is ⫺ up to the present ⫺ the only officially bilingual province of Canada. Official and administrative bilingualism, though far from being considered as “harmonic” and even less so as symmetric, is now actively promoted by the government as a distinctive feature and “brand” for the province. That such a development was possible in New-Brunswick but not elsewhere in the eastern Canadian provinces with an Acadian-French population is partly due to the low (or lower) degree of linguistic assimilation of New-Brunswick Acadians and partly to their demographic weight: whereas francophones account for a third of the population (33 percent in 2006) in that province, they constitute only around four percent of the populations of neighboring Nova Scotia and Prince-Edward-Island (Iledu-Prince-Edouard) and are even more marginalized, quantitatively speaking, in Newfoundland (Terre-Neuve; 0.4 percent) (cf. the contributions in Valdman, Auger and Piston-Hatlen 2005). Language policy in favor of the francophones is almost absent in these

387

388

III. Structure and dynamics of a language space provinces and limited to minimum services in the realm of schooling, as required by federal Canadian law. However, despite these low figures and equally low spatial and administrative cohesion, Acadians are very “present” and visible even outside of New-Brunswick due to a strong socio-cultural and ideological cohesion. Literature, music, theater and other performing arts are the vectors of this manifest presence of the Acadian group, which creates or strengthens internal ties and radiates outside the community. Regular manifestations of common origins, heritage and values, such as the annual celebration of the national Acadian holiday on 14 August, the annual Acadian Festival in Caraquet, an entirely francophone municipality on the Acadian Peninsula in New-Brunswick, regular family meetings and the episodic organization of an Acadian World Congress (the third edition of which will take place in 2009 in the Caraquet region, too) contribute to develop and maintain a communicative space with a dense interactional network, so that territorial space is considered as less prominent. Although geographic concentrations of Acadian-French speakers are obvious and acknowledged as such in the community, no real move to a territorialization of “New Acadia” has been made so far, and speakers do not hesitate to argue that Acadia is where someone who feels as an Acadian has taken residence for at least 24 hours. Such an attitude, which is in strong contrast to the one prevailing among European minority groups, is certainly linked to the migrantcolonial history of the American continent and to a different conceptualization of “homeland” and geographical origin there; but it can also be linked, at least partially, to the fact that the Acadian-French minority members have successfully constructed an “Acadia that has no frontiers”, as expressed in a song by the Que´bec-Acadian singer Carolyne Jomphe (L’Acadie n’a pas de frontie`res; cf. Kolboom 2005: 12). Among the speakers of Acadian French, even more than among Swedish speakers in Finland, language space is replacing language territory, in the sense of Viaut (2007).

6. Reerences Auer, Peter and Li Wei (eds.) 2007 Handbook of Multilingualism and Multilingual Communication. (Handbook of Applied Linguistics 5.) Berlin/New York: Mouton de Gruyter. Blanchet, Philippe and Harold Schiffman (eds.) 2004 The sociolinguistics of southern “Occitan” France revisited. International Journal of the Sociology of Language 169. Cardinal, Linda 2005 The ideological limits of linguistic diversity in Canada. Journal of Multilingual and Multicultural Development 26: 481⫺495. Doppelbauer, Max 2006 Vale`ncia im Sprachenstreit: Sprachlicher Sezessionismus als sozialpsychologisches Phänomen. (Ethnos 67.) Vienna: Braumüller. Dorian, Nancy 2004 Minority and endangered languages. In: Tej K. Bhatia and William C. Ritchie (eds.), The Handbook of Bilingualism, 437⫺459. (Blackwell Handbooks in Linguistics 15.) Malden: Blackwell. Edwards, John 1995 Multilingualism. London: Penguin.

21. Old minorities within a language space Edwards, John 2004 Language minorities. In: Alan Davies and Catherine Elder (eds.), The Handbook of Applied Linguistics, 451⫺475. (Blackwell Handbooks in Linguistics 17.) Malden: Blackwell. Edwards, John 2007 Societal multilingualism: Reality, recognition and response. In: Peter Auer and Li Wei (eds.), Handbook of Multilingualism and Multilingual Communication, 447⫺467. (Handbook of Applied Linguistics 5.) Berlin/New York: Mouton de Gruyter. Gallois, Cindy, Bernadette Watson and Madeleine Brabant 2007 Attitudes in language and communication. In: Marlis Hellinger and Anne Pauwels (eds.), Handbook of Language and Communication: Diversity and Change, 595⫺618. (Handbook of Applied Linguistics 9.) Berlin/New York: Mouton de Gruyter. Griffiths, Naomi E. S. 2005 From Migrant to Acadian: A North American Border People 1604⫺1755. Montre´al/London/Ithaka: McGill-Queen’s University Press. Grin, Franc¸ois 1994 Combining immigrant and autochthonous language rights: A territorial approach to multilingualism. In: Tove Skutnabb-Kangas and Robert Phillipson (eds.), Linguistic Human Rights: Overcoming Linguistic Discrimination, 31⫺48. (Contributions to the Sociology of Language 67.) Berlin/New York: Mouton de Gruyter. Hellinger, Marlis and Anne Pauwels (eds.) 2007 Handbook of Language and Communication: Diversity and Change. (Handbook of Applied Linguistics 9.) Berlin/New York: Mouton de Gruyter. Holloway, Charles E. 1997 Dialect Death: The case of Brule Spanish. (Studies in Bilingualism 13.) Amsterdam/Philadelphia: Benjamins. Kolboom, Ingo 2005 Die Akadier ⫺ Frankreichs vergessene Kinder: Der lange Weg zu einer Nation ohne Grenzen. In: Ingo Kolboom and Roberto Mann (eds.), Akadien, ein französischer Traum in Amerika: Vier Jahrhunderte Geschichte und Literatur der Akadier, 5⫺322. Heidelberg: Synchron. Kremnitz, Georg 1974 Versuche zur Kodifizierung des Okzitanischen seit dem 19. Jahrhundert und ihre Annahme durch die Sprecher. (Tübinger Beiträge zur Linguistik 48.) Tübingen: Narr. Latomaa, Sirkku and Pirkko Nuolijärvi 2005 [2002] The language situation in Finland. In: Robert B. Kaplan and Richard B. Baldauf Jr. (eds.), Language Planning and Policy in Europe: Hungary, Finland and Sweden, vol. 1, 125⫺232. Clevedon/Buffalo/Toronto: Multilingual Matters. [First published in Current Issues in Language Planning 3, 95⫺202.] LeTouze´, Sophie 2004 Les francophones du comte´ de Russell (Ontario): Profil statistique. Ottawa: Centre interdisciplinaire de recherche sur la citoyennete´ et les minorite´s (CIRCEM), Universite´ d’Ottawa. May, Stephen 2001 Language and Minority Rights: Ethnicity, Nationalism and the Politics of Language. Harlow: Longman. Nelde, Peter H. 1982 Conflit ethnoculturel et changement de langue a` Bruxelles. In: Jean Caudmont (ed.), Sprachen in Kontakt: Langues en contact, 37⫺57. (Tübinger Beiträge zur Linguistik 185.) Tübingen: Narr. Pailhe´, Joe¨l 2007 Ge´ographie des langues: proble´matique et application. In: Viaut (ed.), 65⫺85.

389

390

III. Structure and dynamics of a language space Plasseraud, Yves 2005 Atlas des minorite´s en Europe: De l’Atlantique a` l’Oural, diversite´ culturelle. Paris: Autrement. Rossillon, Philippe (ed.) 1995 Atlas de la langue franc¸aise: Histoire, Ge´ographie, Statistiques. Paris: Bordas. Saari, Mirja 2000 Schwedisch als zweite Nationalsprache Finnlands: Soziolinguistische Aspekte. Linguistik online 7(3). Available from . (Accessed 25 February 2008.) Sinner, Carsten 2004 Sprachen und Sprachkontakt in Andorra: Ein Überblick. Zeitschrift für Katalanistik 17: 91⫺110. Skutnabb-Kangas, Tove 2007 Language planning and language rights. In: Marlis Hellinger and Anne Pauwels (eds.), Handbook of Language and Communication: Diversity and Change, 365⫺397. (Handbook of Applied Linguistics 9.) Berlin/New York: Mouton de Gruyter. Sumien, Domergue 2006 La standardisation pluricentrique de l’occitan: Nouvel enjeu sociolinguistique, de´veloppement du lexique et de la morphologie. (Publications de l’Association Internationale d’E´tudes Occitanes 3.) Turnhout: Brepols. Svenska Finlands folkting (ed.) 2004 Swedish in Finland: La Finlande sue´dophone. Helsingfors/Helsinki: Svenska Finlands folkting. Tandefelt, Marika and Fjalar Finnäs 2007 Language and demography: historical development. International Journal for the Sociology of Language 187⫺188: 35⫺54. Triano Lo´pez, Manuel 2007 The Attitudinal Factor in Language Planning: The Valencian Situation. (Lincom Studies in Romance Linguistics 54.) Munich: Lincom. Turell, M. Teresa 2001 Spain’s multilingual make-up: Beyond, within and across Babel. In: M. Teresa Turell (ed.), Multilingualism in Spain: Sociolinguistic and Psycholinguistic Aspects of Linguistic Minority Groups, 1⫺57. (Multilingual Matters 120.) Clevedon/Buffalo/Toronto/Sydney: Multilingual Matters. Valdman, Albert, Julie Auger and Deborah Piston-Hatlen (eds.) 2005 Le franc¸ais en Ame´rique du Nord: Etat pre´sent. Quebec: Presses de l’Universite´ Laval. de Varennes, Fernand 1996 Language, Minorities and Human Rights. Den Haag: Nijhoff. Viaut, Alain 2004 La frontie`re linguistique de la ligne a` l’espace: Ele´ments pour une sche´matisation. Glottopol 4: 6⫺22. Viaut, Alain 2007 Le territoire linguistique et ses limites. In: Viaut (ed.): 47⫺63. Viaut, Alain (ed.) 2007 Variable territoriale et promotion des langues minoritaires. Pessac: Maison des Sciences de l’Homme d’Aquitaine. Werlen, Benno 1993 Gibt es eine Geographie ohne Raum? Zum Verhältnis von traditioneller Geographie und zeitgenössischen Gesellschaften. Erdkunde 47: 241⫺266. Williams, Colin H. (ed.) 1988 Language in Geographic Context. Clevedon/Philadelphia: Multilingual Matters.

Claus D. Pusch, Freiburg im Breisgau (Germany)

IV. Structure and dynamics across language spaces 22. Contact-induced grammatical change: A cautionary tale 1. 2. 3. 4. 5. 6. 7. 8.

Introduction Background Establishing change Applying the comparative variationist framework to the investigation of change Putting contact-induced change to the test Discussion Acknowledgement References

1. Introduction When the space between two languages ⫺ or their speakers ⫺ is reduced enough to result in contact, a variety of outcomes may ensue. Most prominent by far is change. So ubiquitous has the association between contact and change become that it is invoked to account for a wide array of developments that differ from some (typically ill-defined) norm. This is the case even where the candidate development already exists in the recipient language, where no other-language form is visible, and where the transition period is as short as a single generation. And despite the general consensus that grammatical structure is particularly resistant to transfer, reports of dramatic structural changes in languages known to have been in contact are legion. In this article we suggest that much of the evidence brought to bear on contactinduced change ⫺ diachronic as well as synchronic ⫺ either fails to demonstrate that change has occurred, and/or if it has, that it is the product of contact and not internal evolution. These issues, together with the possibility that the inherent variability characteristic of all spoken language may have been mistaken for change, need to be resolved before a contact explanation can be justified. In the ensuing sections we first provide a brief overview of some of the more salient types of change reported to result from contact. We then review the kinds of confounding evidence that plagues the study of change, contact-induced or otherwise, and propose criteria for establishing its existence empirically. In section 3 we detail a method, combining the machinery of variationist sociolinguistics with the comparative method of historical linguistics, capable of identifying change and ascertaining its source. We then present a series of case studies illustrating the utility of this approach and conclude with some observations about the implications of our findings for understanding the existence, nature and extent of contact-induced change. Our focus is on the grammar, but the considerations we present also apply to other types of change, with the exception of lexical borrowing.

392

IV. Structure and dynamics across language spaces

2. Background Contact-induced structural change has long been a subject of debate among linguists, due in no small part to the primacy traditionally accorded the internal evolution of language (e.g., Lass 1980; Martinet 1960). This was traditionally assumed to be regular and rule-governed, while externally-motivated changes could be arbitrary, unpredictable and idiosyncratic. External explanations were thus often invoked as a last resort, when a more satisfactory account was not available (Romaine 1988: 349). Early opinions ranged from outright denial of the possibility of a mixed language (e.g., Müller 1871: 86; Whitney 1881) to the claim that there are no completely unmixed languages (Schuchardt 1884: 5). But most of those who admitted the possibility of contact-induced change did so cautiously and under specific conditions. Haugen’s comment that “[on] those relatively rare occasions when bound morphemes are actually transferred from one language to another, they are such as fit readily into a preexistent category in the recipient language” (1954: 386), is but one example. Similarly, Jakobson (1962: 241) argued that “a language accepts foreign structural elements only when they correspond to its own developmental tendencies”. Notwithstanding such caveats, major structural changes said to result from language contact continue to be reported. Among the more spectacular changes of the past figure the borrowing of Turkish inflectional suffixes into dialects of Asia Minor Greek (Dawkins 1916), the incorporation of the entire Spanish preposition system into Malinche Mexicano (Field 2005: 352), the borrowing of Russian finite verb morphology into Mednyj Aleut (Thomason 2001), the development of Michif from French noun phrases and Cree verb phrases (e.g., Bakker 1997; Bakker and Papen 1997) and the (in)famous convergence of Kupwar Urdu, Marathi and Kannada into “a single surface syntactic structure” (Gumperz and Wilson 1971: 155). After reviewing many such changes, as well as a variety of theories of what constrains them, Thomason was led to reject the idea that categorical constraints against contactinduced change could be enunciated. Instead she proposed that under the right social and linguistic conditions, […] anything goes, including structural borrowing that results in major typological changes in the borrowing language. In phonology, loss or attrition of entire phonetic and/or phonological categories in native words and of all kinds of morphophonemic rules. In syntax, sweeping changes in such features as word order, relative clauses, negation, coordination, subordination, comparison, and quantification. In morphology, typologically disruptive changes such as the replacement of flexional by agglutinative morphology or vice-versa, the addition or loss of morphological categories that do not match in source and borrowing languages, and the wholesale loss or addition of agreement patterns. (Thomason 2001: 71)

At the time of this writing, Thomason’s (and Schuchardt’s) views enjoy wide support. The implicit assumption is that change is an almost inevitable result of language contact (e.g., Appel and Muysken 1987: 154; Bynon 1977: 240; Harris and Campbell 1995: 149, among many others; although see Thomason 2007 for discussion of deliberate resistance to contact-induced change). To be sure, some scholars continue to caution against promiscuous inferences, particularly in the absence of robust sociohistorical and linguistic evidence (e.g., Dorian 1993; Mougeon, Nadasdi and Rehner 2005). And Heine and Kuteva (2005: 8) observe that contact is often invoked without satisfactorily ruling out

22. Contact-induced grammatical change other causal mechanisms. Nonetheless, the onus seems to be on the analyst to justify why change did not occur in an apparently favorable situation: But since [the above] predictions are robust ⫺ that is, they are valid in the great majority of cases that have been described in the literature ⫺ any violation should provide interesting insights into social, and, to a lesser extent, linguistic determinants of contact-induced change. (Thomason 2001: 71, emphasis ours)

And in fact, the kinds of changes enumerated above are reported in contemporary contact situations as well. In the least conspicuous cases, an existing minor-use pattern is claimed to be developing into a major-use pattern via overextension of recipient-language options into new contexts (e.g., Heine and Kuteva 2005: 44⫺62; Klein 1980; SilvaCorvala´n 1994a: 167; Toribio 2004: 167). Thus Toribio (2004: 170) states that when the “bilingual mode” of two Spanish-English bilinguals was experimentally induced, their use of overt subject pronouns, while not ungrammatical in Spanish, reflected the influence of English, a non-pro-drop language. Savic´ (1995: 487, cited in Heine and Kuteva 2005: 69) reaches similar conclusions in relation to slightly elevated uses of optional overt subject and object NPs in the discourse of Serbian-English bilinguals. Heine and Kuteva also discuss an increase of overt pronouns among Russian and Hungarian bilinguals, again on the model of English (2005: 68), as do Otheguy, Zentella and Livert (2007) for New York City Spanish and Backus (2005: 333) for Turkish. The same process has been said to initiate word order changes, as in the shift from the Spanish majority pattern of postposed adjectival placement (la persona ma´s importante) to the pre-nominal adjectival placement (la ma´s importante persona) characteristic of English (Toribio 2004: 167). In another type of change, recipient-language morphemes may encode a completely new grammatical category, as in Tariana-Portuguese bilinguals’ use of interrogative pronouns (e.g., kwana ‘who’) to mark relative and complement clauses on the model of Portuguese, or of Portuguese lexemes to express the Tariana evidentiality marking system (Aikhenvald 2002: 182⫺183, 315⫺316). Other contemporary cases of convergence involve outright violation of recipient-language grammar. Clyne (2003: 132) discusses overgeneralization by Dutch-English and, to a lesser extent, German-English bilinguals of SVO word order in contexts where SOV and VSO are prescribed in the non-contact varieties. Apart from extension of an existing option, addition of a totally new category and violation of recipient-language rules, contact is widely believed to accelerate attrition and loss of grammatical features in one language with no equivalent in the other (e.g., Seliger and Vago 1991). The subjunctive mood in New World Romance varieties is frequently cited as an example of contact-induced erosion (see e.g., Garcı´a 1985; Gutie´rrez 1994; Lantolf 1978; Ocampo 1990; Silva-Corvala´n 1994b for Spanish and Laurier 1989 for Canadian French). We explore this phenomenon in more detail below. Other morphosyntactic elements said to be receding as a result of contact include the BE auxiliary (sein, zijn, essere) in Australian German, Dutch and Italian respectively, in favor of haben, hebben, avere ‘have’, on the model of English, which admits only have in such contexts (Clyne 2003: 118⫺119). Clyne does acknowledge regional variation in auxiliary usage, even in some non-contact varieties of these languages, but follows Thomason and Kaufman (1988: 58⫺59), in arguing that contact reinforces language-internal tendencies. And these examples could be multiplied ad libitum.

393

394

IV. Structure and dynamics across language spaces Much of the evidence for contact-induced change may lend itself equally well to alternate explanations. Sometimes inherent variability is equated with change (as in the aforementioned Australian example). Elsewhere, perceived differences from a standard are designated changes (as in the increase of overt subject pronouns). And it is often unclear whether the change in question is idiosyncratic or community-wide. The next section summarizes these and other pitfalls and details some of the considerations that must be taken into account in the empirical establishment of change.

3. Establishing change 3.1. Variability is not change The hallmark of speech is inherent variability, alternate ways of saying the same thing (Labov 1969). Variability is a necessary condition for change, but is not, in and of itself, coterminous with it. In many areas of the grammar, alternations among variant forms may persist for centuries, but linguists who believe that language is invariant often interpret them as signs of change. This inference is particularly prevalent when the forms in question are detected among speakers or groups considered vulnerable to external influence (e.g., bilinguals, residents of minority-language communities in intense contact with a majority language, etc.). Change is also routinely invoked when one of the variants is salient or stigmatized, especially when a null form is involved. Thus, the widespread variability between subjunctive and indicative morphology in (prescribed) subjunctiveselecting contexts cited above is usually characterized in contact situations as loss of the subjunctive, despite the fact that this alternation has been attested for centuries (Poplack 2003; St-Amand 2002). So a first requirement is to determine whether variability is involved in change. Variant choice is never random, but is instead highly constrained by features of the (linguistic and extra-linguistic) environment. Measuring the trajectory of variant use over time, coupled with analysis of the fine conditioning of variant choice, provide a sensitive and refined diagnostic of change, in ways we illustrate in section 5 below.

3.2. Locating an earlier stage At the most trivial level, change shows up as a difference (qualitative or quantitative) between an outcome and an earlier stage, but in much of the current literature, the structure of the earlier stage is either not appealed to, or at best, only inferred (e.g., Winford 2003: 64). Absent systematic comparison with a relevant precursor, any claims about what has changed must remain speculative. (Indeed, lack of an appropriate diachronic baseline is at least partially responsible for the widespread assumption that many salient features of contemporary vernaculars, both contact and non-contact, are recent innovations; see also Zwicky 2005). The first step in establishing the existence of change is comparison over time. This may not be simple or straightforward, given the often fragmentary nature of surviving diachronic evidence. This abiding (and at times insoluble) problem in historical linguis-

22. Contact-induced grammatical change tics has given rise to an entire methodology devoted to reconstruction (e.g., Campbell 1999, 2003; Hoenigswald 1960; Lehman 1992; Ringe 2003, to name but a few). Synchronically, however, with a modicum of ingenuity, earlier stages or surrogates for them can often be found.

3.3. The nature o the reerence point Not every earlier stage is equally useful in assessing change, however. In most reported cases, the comparison point (diachronic or synchronic) is some poorly defined “standard”. Constituted only of forms that have been prescriptively ratified, the standard rarely acknowledges variants that cannot be associated (rightly or wrongly) with variant functions or readings (Poplack et al. 2002). As a result, the language ratified as standard may diverge considerably from the language actually spoken, which often develops independently of normative injunction (Poplack and Dion to appear; Leroux 2003; Poplack and Malvar 2007; Willis 2000). The variability most characteristic of vernacular speech involves competition among standard and nonstandard features. When compared with an idealized invariant standard, the latter may appear to be changes. A nice illustration of the importance of an appropriate reference, or baseline, variety comes from a recent study of Spanish in contact with English in New York City (Otheguy, Zentella and Livert 2007). As noted above in connection with the extension of minor-use patterns, rates of overt pronouns are thought to be increasing among secondgeneration Spanish speakers, on the model of English, in which the subject is assumed to be categorically expressed (but see Leroux and Jarmasz [2006] for conditions governing variable subject pro-drop in English). Of course, some overt pronoun expression is tolerated in Spanish, but the standard baseline does not specify how much. What then constitutes an “increase”? Vis-a`-vis what? Otheguy, Zentella and Livert’s comparison with one surrogate pre-contact stage of Spanish (spoken by newcomers to New York) revealed, contrary to the widely espoused convergence hypothesis enunciated above, an even higher rate of overt pronouns than that displayed by their New York born-andraised counterparts (2007: 786). Only once the appropriate comparison variety was identified (i. e. by distinguishing two newcomer dialects, each with a different initial rate of overt pronouns), were they able to demonstrate that pronoun expression had in fact increased in the second generation. Invidious comparisons can only be avoided if the reference variety is commensurate with the variety hosting the candidate for change. If it is a spoken vernacular, the baseline should also be a spoken vernacular (or an appropriate surrogate thereof); if it is a contact variety, the baseline should be a pre-contact variety, and so on. As we will show below, when the inference of change is pursued scientifically via systematic comparisons (vertical and horizontal) of appropriate reference varieties, potential candidates often turn out either not to be changes, or not to be contact-induced, but rather cases of garden-variety inherent variability firmly rooted in the internal structure of language (Leroux and Jarmasz 2006; Poplack et al. 2006). Efforts invested in locating and mining appropriate earlier stages have paid off handsomely (Ayres-Bennett 2000; Gordon et al. 2004; Martineau and Mougeon 2003; Poplack and St-Amand 2007). The procedure is a good deal more laborious than simply inferring what the precursor was like, but in view of the potential for misidentification, we see this as a requirement for any claim of change based on synchronic data.

395

396

IV. Structure and dynamics across language spaces

3.4. The social embedding o change Another major impediment to detecting change resides in the nature, quantity and quality of the evidence. Manifestations of change are relatively rare in running discourse. This may be why much of the published evidence comes from data collected in experimental or quasi-experimental conditions (e.g., informant elicitation, subjective reaction tests, even “forced” interviews with semi-speakers [see, e.g., Altenberg 1991; Huffines 1991; Kaufman and Aronoff 1991; Montrul 2002; Schmidt 1991]). Atypical situations may entail atypical behavior, with little or no bearing on what transpires in real-life bilingual settings (Poplack 1983: 116). Without knowing how well, if at all, these experiments reflect language use in the community, their results cannot be interpreted. This brings us to the distinction between innovation and change. Every change involves an innovation in the speech of one or more individuals and its subsequent diffusion across the rest of the community (Croft 2000; Joseph and Janda 2003: 13; Milroy 1993: 223; Shapiro 1995: 105, fn. 1; Thomason 2007). For some linguists, the innovation is the change, and it is assumed to occur during first language acquisition. Theoretically, we would argue (along with Labov 1994: 310⫺311 and most other sociolinguists), that “an innovation in a speaker’s output is not a linguistic change until it has been agreed on and adopted by some community of speakers” (Milroy 1992: 221), a fortiori because the vast majority of changes do not “catch on” (e.g., Thomason 2001: 130). From an empirical perspective, however, this is a moot point: in all but the rarest of cases, the moment of innovation is simply unobservable. We can rephrase the issue in terms of the relationship between the individual and the community. Claims of change are often based on exiguous speaker samples (e.g., Toribio’s 2004 study of two individuals), if any. The linguistic behavior of the individual may be ephemeral, situation-bound and/or simply deviant from the norms of the remainder of the community (Labov 1972a; Poplack 1983: 119), especially when this behavior is elicited under laboratory conditions. Alternatively, it may be a predictable outcome of one or more explanatory factors, linguistic and/or extra-linguistic. Only once the speaker’s linguistic behavior is situated with respect to that of the appropriate reference variety (other speakers of the same age, bilingual ability, educational level, etc.) is it possible to decide. If change has occurred, it should be observable (at least to some degree) in the speech of the relevant members of the community (e.g., the younger cohort, if spearheaded by youth; the highly bilingual, if induced by contact, etc.). In other words, in the absence of diffusion, change cannot be confirmed.

3.5. Type o bilingual population The locus of the candidate for contact-induced change is also important. As Bullock and Toribio (2004: 91) observe, researchers have drawn on an array of bilingual populations, including heritage language speakers, incomplete acquirers and those undergoing language loss. Shifting populations have garnered most attention (see, e.g., Dorian 1981, 1989; Gal 1979; Mougeon and Beniak 1991; Otheguy, Zentella and Livert 2007; SilvaCorvala´n 1991), but the linguistic behavior they display is not necessarily applicable to stable bilingual communities (Poplack 1993; Winford 2003: 64). In the widely studied

22. Contact-induced grammatical change contexts of language restriction and obsolescence, structural differences (from the mainstream or non-contact variety) may in fact be inevitable, given that many of the informants may not (or no longer) use it with any regularity. In effect, then, although ethnically minority-group members, they may be L2 speakers of the minority language (Poplack 1997: 305). L2 speakers can expected to make all manner of acquisition errors, depending on their level of proficiency at the time of data collection. Though these may be transitory, in contact situations, they too are often construed as changes.

3.6. The linguistic embedding o change In addition to a certain level of diffusion across the community, the innovation should also occur (to some degree) in relevant linguistic contexts. Relevant contexts are determined by situating the innovation in the larger host linguistic system⫺synchronic and diachronic. Among the questions to be answered are: To what extent has the innovation gained a foothold? What is its current role? Has it replaced a native form in one or more functions? Did it introduce a new function into recipient-language grammar? Isolated, anecdotal or exceptional examples, so often cited in the literature, reveal nothing about the regularity of the innovation, its productivity, or the extent to which it is entrenched in the language (Weinreich, Labov and Herzog 1968: 185). This can only come from studying the sustained discourse of speakers in context. Because changes examined synchronically are likely to still be in progress, and thus continue to feature variability, these questions can only be addressed by establishing the structure of the variability, as it emerges from the constraints conditioning variant choice.

3.7. Recognizing change Most reported changes are tangible: a novel lexical element (e.g., a borrowing), unwarranted presence of an existing form (e.g., increased pronoun expression in pro-drop languages), absence of a prescribed form (e.g., the French subjunctive), or choice of the (prescriptively) “wrong” form (e.g., auxiliary haben for sein in Australian German or avoir for eˆtre in Canadian French). But observed changes tend to be far more subtle, especially during the sometimes very lengthy transition period from one language state to another. Linguists who focus on inherent variability recognize, in ascending order of importance, change in frequency of one or more competing forms, change in statistical significance of one or more factors contributing to variant choice and change in linguistic structure. With rare exceptions, these elude casual observation, but can be detected with the aid of the methods described in section 4.

3.8. Distinguishing contact-induced change rom internal evolution Once it has been established that a change has indeed taken place, it remains to demonstrate that it is contact-induced and not the product of drift. This step is usually skipped in the literature, much of which is predicated on the aforementioned (widespread but

397

398

IV. Structure and dynamics across language spaces unfounded) assumption that linguistic differences occurring in bilingual contexts are necessarily 1) changes and 2) contact-induced. Thomason (2007: 42) defines contact-induced change as follows: “a particular linguistic change is caused at least in part by language (or dialect) contact if it would have been less likely to occur outside a particular contact situation”. Given how little we know about the likelihood of particular changes ⫺ within or outside contact situations ⫺ we offer alternative criteria: A candidate for contact-induced change in a contact variety is present in the presumed source variety and either 1) absent in the pre-contact or non-contact variety, or 2) if present (e.g., through interlingual coincidence), is not conditioned in the same way as in the source, and 3) can also be shown to parallel in some non-trivial way the behavior of a counterpart feature in the source.

Compliance with these conditions can only be ascertained through systematic quantitative comparisons, of a diagnostic linguistic feature, with an earlier or pre-contact stage (or a surrogate thereof), with a non-contact variety, and most important, with the presumed source. As we shall see in what follows, these criteria have not been met by most extant studies, with the result that the existence, frequency and extent of contact-induced change remain unclear. Summarizing, sufficient understanding of the outcomes of contact appears beyond the reach of anything but systematic corpus-based research carried out within the bilingual community. Application of an empirically accountable quantitative methodology to large and representative bodies of bilingual speech can contribute significantly to resolving the difficulties associated with ascertaining whether contact-induced change has authentically occurred. It is to such a methodological framework that we now turn.

4. Applying the comparative variationist ramework to the investigation o change 4.1. The variationist paradigm The approach that we exemplify in the remainder of this article draws on the combined resources of variationist sociolinguistics (e.g., Labov 1972a, Labov 1972b; Labov et al. 1968; Weinreich, Labov and Herzog 1968) and the Comparative Method of historical linguistics (e.g., Baldi 1990; Campbell 1998; Hoenigswald 1960; Meillet 1967). A central tenet of the variationist paradigm is that the recurrent choices made by speakers in discourse, though not completely predictable, are not random, but are subject to constraints imposed by “the phonological environment, the syntactic context, the discursive function of the utterance, topic, style, situation and personal and/or sociodemographic characteristics of the speaker or other participants” (Sankoff 1988a: 151). The key theoretical construct of this framework is the linguistic variable (Labov 1966/1982), which comprises a set of variants among which speakers alternate in expressing a given meaning or function. This results in a heterogeneous but structured linguistic system. A foundational working principle is that the structure ⫺ grammatical and social ⫺ underlying the heterogeneity can be inferred from the distribution and conditioning of competing variants in discourse.

22. Contact-induced grammatical change

4.2. Communities and speakers We noted above that change proceeds from diffusion, and extent of diffusion can only be assessed in reference to a community of speakers. A crucial step, though often neglected, is to locate an appropriate (here, bilingual) community in which to carry out the study. The nature and makeup of the community allow us to test hypotheses about the extra-linguistic factors often claimed to be relevant to contact-induced change. These include intensity of contact (the more intense, the greater the likelihood of structural interference), length of contact (the greater the time-depth, the greater the likelihood of change), status of the languages in the community (minority or majority) and size of speaker population (the fewer and more marginal the speakers, the more likely that they will borrow from the language of the dominant group) (e.g., Romaine 1995; Thomason 2001; Winford 2003). Study sites can be selected to represent these parameters and to test whether these factors are or are not operating. Speakers constituting these communities can be sampled to represent individual dimensions of the contact axis. Sampling protocols are also guided by hypotheses about how contact-induced change works. According to one, individual level of bilingualism is inversely correlated with mastery of the minority-language grammar, such that the greater the proficiency in the other language, the more likely the possibility of simplification or loss of minority-language linguistic structure. Another proposes that the incorporation on the local level of majority-language material (via code-switching or borrowing) into otherwise minority-language discourse may bring with it associated grammatical properties, and these in turn may lead to structural convergence (e.g., Backus 2005; King 2005; Muysken 2000; Thomason 2001; Winford 2003). According to this hypothesis, a speaker’s individual propensity to code-switch would be a predictor of contact-induced change. Individual attitudes toward, and relative prestige of, each of the languages could also affect the extent and direction of change (e.g., Appel and Muysken 1987: 158; Joseph 2007; Romaine 1995: 66; Thomason 2001). The case studies in section 5 demonstrate how these and other hypotheses can be tested on linguistic production data provided by speakers representing different combinations of these conditions, thus enabling us to ascertain which are most explanatory of contact-induced change and which community members are most likely to lead it.

4.3. Sampling bilingual speech Once the appropriate speakers have been identified, we can record samples of their spontaneous speech (bilingual and monolingual) in quantities sufficient to permit accountable analysis. The speech of choice for such an endeavor is the vernacular. Not only does the vernacular offer “the most systematic data for linguistic analysis” (Labov 1984: 29), it is also the register in which linguistic manifestations of language contact are most likely to occur. Variationists mine vernacular corpora to uncover the regular patterns that characterize natural exchanges in the speech community and to distinguish them from isolated or idiosyncratic occurrences. A pattern may be defined as “a series of parallel occurrences (established according to structural and/or functional criteria) occur-

399

400

IV. Structure and dynamics across language spaces ring at a non-negligible rate of use” (Poplack and Meechan 1998: 129). Patterns can only be discerned by systematic quantitative analysis of the data in accordance with the principle of accountability (Labov 1972b: 72) . This requires not only that all the relevant occurrences of a feature be incorporated into the analysis, but additionally all the contexts in which it could have occurred but did not. Adherence to the principle of accountability allows the analyst to identify anomalous occurrences which are neither representative nor recurrent (and thus unlikely to trigger change), while identifying regular tendencies, which may (or may not) herald change.

4.4. Variable rule analysis Certain linguistic and extra-linguistic conditions can be expected to promote change, though none is fully determinant. Claims that even favorable linguistic factors can be overridden by the right constellation of social factors (e.g., Thomason 2001: 77; Winford 2003: 25) raise the issue of the relative weights each contributes to a particular outcome and the relationship of such weights to different types of contact situations. This is a problem that lends itself well to multivariate analysis, the major analytical tool in the variationist framework. Variable rule analysis (Rand and Sankoff 1990; Sankoff, Tagliamonte and Smith 2005) facilitates extraction of regularities from the welter of data making up running discourse. The multiple regression procedure incorporated within it helps determine which aspects of the linguistic (e.g., phonological, morphological, syntactic, etc.) and extra-linguistic (e.g., minority status, bilingual ability, age, etc.) contexts contribute statistically significant effects to variant choice when all are considered simultaneously (e.g., Sankoff 1988b). These aspects are hypotheses about the structure of variant choice. Variable rule analysis enables us to operationalize them (as factors) and determine which give the best account of the data. The goals are to ascertain whether change has taken place and if so, whether it is the product of contact or internal development. Much empirical work on contact-induced change in progress revolves around frequencies. As detailed by Poplack and Tagliamonte (2001: 92), however, differences in rates of variant selection must be used with caution to infer change, contact-induced or otherwise. The key diagnostic in assessing the relationship and provenance of forms is the constraint hierarchy, the configuration of environmental factors affecting the probability that a given variant form will be selected, along with the direction of their effects. We construe these as a portion of the grammar underlying the variability. To the extent that constraint hierarchies are language-specific, they are powerful indicators of change (see section 5). Similarities in surface form in contact and source varieties can be misleading. They may result from borrowing or transfer, justifying the inference of change, but may also be due to interlingual coincidence or to linguistic universals. To rule out alternative explanations, we rely on conflict sites (Poplack and Meechan 1998: 132): functional, structural and/or quantitative differences, typically manifested as a conflict in constraint hierarchies. The conflict site, together with detailed cross-variety comparison, plays a crucial role in detecting change and identifying its ultimate provenance. If the hierarchy of constraints conditioning the variable occurrence of a candidate for change (e.g., Ø subjunctive, overt subject pronoun, preposition stranding, null relative marker, periphrastic future) in a contact variety is the same as that of its pre-contact precursor, while

22. Contact-induced grammatical change differing from that of its presumed source, no structural change has taken place. If it features a constraint hierarchy different from those of both its pre-contact precursor and the presumed source, we can infer that change has occurred, but not one that is contactinduced. Only when a candidate for change in a contact variety features a constraint hierarchy different from that of its pre-contact precursor, but parallel to that of its presumed source, can we conclude in favor of contact-induced change.

4.5. Cross-variety comparison Comparisons are at the heart of inferring change, and as such, figure, at least implicitly, in all studies of contact-induced change. But many have been carried out without reference to the principle of accountability, or involve an idealized comparison variety. Still others are simply anecdotal. Innovative in the research program we describe here is the extent of the scientific apparatus brought to bear on the comparison. The comparative variationist enterprise (Poplack and Meechan 1998; Poplack and Tagliamonte 2001; Tagliamonte 2002) contributes not only a methodological aspect ⫺ the construction, statistical analysis and linguistic interpretation of suitable comparison varieties (pre-contact, non-contact and post-contact), but also the critical capacity to decide among competing hypotheses. The focus is on identifying empirical criteria capable of both detecting change and ascertaining its source, and testing hypotheses empirically to determine the goodness of fit with the data. This involves principled data collection, enunciation and operationalization of hypotheses and their statistical evaluation in large-scale corpora. In section 5, we illustrate this approach by considering a number of case studies which attest to the versatility of quantitative variationist methodology for identifying contactinduced change.

5. Putting contact-induced change to the test The studies we present below were carried out in a context of long-term intense contact between French and English in Canada, a situation that meets all the criteria considered favorable to structural convergence (e.g., Thomason and Kaufmann 1988; Winford 2003: 90). Depending on locality, each of the languages has a different official status: English is the majority language in Ontario, while in Quebec, French assumes this role. This configuration allows for comparison of each of the languages in minority and majority guise. The 284 speakers involved in these studies are distributed among these four conditions. Other characteristics relevant to the contact situation include density of contact on the local level (as expressed by the ratio of anglophones to francophones in the community), individual bilingual ability (as measured by a bilingual proficiency index [Poplack 1989, Poplack, Walker and Malcolmson 2006]) and relative propensity to borrow or code-switch (as inferred by comparing individual rates of each across sample members). Standard extra-linguistic factors like social class, education, etc. were also taken into account. The inference of change over time was investigated by comparing the speech of younger and older generations.

401

402

IV. Structure and dynamics across language spaces The linguistic data come from two massive corpora of spontaneous speech collected from these speakers: the Ottawa-Hull French Corpus (Poplack 1989) and the Quebec English Corpus (Poplack, Walker and Malcolmson 2006). The typological similarity of French and English should render them particularly hospitable to contact-induced change. In addition, superficially similar structures were deliberately targeted, as these are considered to be prime candidates for grammatical convergence. Indeed, at least two of the cases described below had been reported to result from this process.

5.1. Attrition o the subjunctive in Ottawa-Hull French In the most straightforward case, we locate, in community speech, a linguistic feature that is used differently from the way it is prescribed and determine which individuals are deviating the most. If the difference is the result of change, and if that change is contact-induced, its leaders should figure among the younger, the more bilingual and those in most intense contact with the majority language. The French subjunctive is an excellent candidate. It is (prescriptively) obligatory after the class of “subjunctive-selecting” main verbs, like falloir in example (1), but in Canadian French it has been found to be variable ((1); Poplack 1990). (This and all ensuing examples are reproduced verbatim from audio recordings from one of the corpora described above. The codes in parentheses refer to corpus [QEC: Corpus of Quebec English; OH: Corpus of Ottawa-Hull French], speaker and line number. The examples cited throughout this paper are instantiations of patterns that emerged from the large-scale quantitative analyses presented in sections 5.1⫺5.3) qu’ (1) Fallait Necessary (past) that ge´ant”. giants Fallait qu’ Necessary (past) that ‘She had to say “yes, sentence.’

elle re´pond (ind) “oui, tu peux faire trois pas de she answer (ind) yes you can make three steps of

elle re´ponde (subj) la phrase comple`te. she answer (subj) the sentence complete. you may take three giant steps”. She had to say the whole (OH.025.2186)

Because the subjunctive is no longer used productively in English, the (variable) occurrence of the indicative in contexts like (1) is inferred to be a stage in the loss of this form, a development which Laurier (1989) characterized as the result of contact. Table 22.1 displays the results of a multivariate analysis of the extralinguistic factors contributing to choice of subjunctive. In this and the ensuing tables, the higher the figure, the greater the probability the “application” variant will be selected in the environment listed on the left. Table 22.1 reveals a small, but statistically significant, tendency for the most bilingual speakers to use least subjunctive. This result supports the hypothesis that high levels of bilingualism lead to linguistic simplification and loss. Most studies would stop at this demonstration (if they provided any quantitative information at all). But if this is a contact-induced change, effects of the other factors should point in the same direction. This not the case. Instead, the two French-majority-language communities in Que´bec, Vieux Hull and Mont Bleu, which should favor the subjunctive most under this

22. Contact-induced grammatical change

403

Tab. 22.1: Variable rule analysis of the contribution of extralinguistic factors to the choice of subjunctive under verbal matrices Corrected mean Total N

0.715 2694

SEX Female Male

.52 .46

AGE 45⫺54 55⫺64 15⫺24 25⫺34 65⫹ 35⫺44 NEIGHBOURHOOD OF RESIDENCE Vieux Hull, Quebec (FR majority working class) West End, Ontario (FR minority working class) Vanier, Ontario (FR minority upper-working class) Basse-Ville, Ontario (FR minority working class) Mont Bleu, Quebec (FR majority upper-middle class) ENGLISH PROFICIENCY mid-low low mid-high high

.55 .53 .52 .50 .48 .44 .57 .51 .50 .50 .43 .54 .53 .48 .42

hypothesis, instead use most and least, depending on social class. (Working-class speakers, even more counter-intuitively, apparently use it the most.) Also militating against the inference of change is the fact that the age distribution is not monotonic: the youngest speakers are as likely to select the subjunctive as their oldest counterparts. Further analysis revealed that these apparent extra-linguistic effects are in fact epiphenomena of a strong lexical effect (Poplack 1990, 1992: 2001). Subjunctive morphology is highly associated with only a few main verbs, and because those verbs happened to be unevenly distributed across the subgroups represented in Table 22.1, this showed up as a discrepTab. 22.2: Variable rule analysis of the contribution of English proficiency to the choice of subjunctive in verbs embedded under four classes of verbal matrix English Proficiency

VERBS EMBEDDED UNDER high frequency/high subjunctive matrix verb: falloir high frequency/high subjunctive matrix verbs: vouloir / aimer low frequency/low subjunctive matrix verbs low frequency/variable subjunctive matrix verbs

low

mid-low

mid-high

high

.55

.58

.40

.42

.46 .68 .34

.58 .47 .60

.42 .49 .61

.60 .31 .39

404

IV. Structure and dynamics across language spaces Tab. 22.3: Variable rule analysis of the contribution of socioeconomic class to the choice of subjunctive in verbs embedded under four classes of verbal matrix Socioeconomic class

VERBS EMBEDDED UNDER high frequency/high subjunctive matrix verb: falloir high frequency/high subjunctive matrix verbs: vouloir / aimer low frequency/low subjunctive matrix verbs low frequency/variable subjunctive matrix verbs

unskilled

skilled

sales / service

professional

.45

.53

.47

.69

.36 .56

.62 .47

.64 .42

.52 .58

.50

.42

.42

.68

ancy in subjunctive usage. When the lexical identity of the main verb is factored out, as in Table 22.2, bilingual proficiency no longer shows a systematic effect. The real explanation for differential use of the subjunctive is class-based, and is not relevant to the contact situation at all, as illustrated in Table 22.3 (see Poplack 1997 for detail). This result could not have been intuited from casual comparisons with either the idealized standard benchmark (which prescribes the subjunctive categorically), or amongst proficiency groups. This is but one illustration of our earlier contention that differences in overall rates of variant occurrence need not be indicative of change, contact-induced or otherwise. Indeed, real-time analysis over a time span of 120 years (StAmand and Poplack 2002) showed that the current situation of the subjunctive is actually a retention.

5.2. Preposition stranding in Ottawa-Hull French It has been argued that a local factor, the extent to which individual speakers draw on the two languages in discourse, is a better predictor of contact-induced change than macro-level factors like minority status or density of contact. Under this hypothesis, speakers who code-switch the most would lead contact-induced change (Backus 2005; Bullock and Toribio 2004; King 2005; Thomason 2001; Winford 2003). Zentz (2006) was the first to test this hypothesis empirically, using as a diagnostic the syntactic phenomenon of preposition stranding in Canadian French, as in (2). (2) Comme le gars que je sors avec, lui il parle- bien il est franc¸ais Like the guy that I go out with him he speaks- well he is French ‘Like the guy I’m going out with, he speaks- well he is French.’ (OH.013.2122) Because preposition stranding is normatively proscribed, examples like (2) have been ascribed to contact, if not convergence, with English (e.g., King 2000; Roberge and Rosen 1999), in which stranding is presumed to occur freely.

22. Contact-induced grammatical change

405

But the phenomenon of stranding cannot be fully understood in isolation from the behavior of the remainder of the preposition system. In particular, French does admit phrase-final prepositions in another, superficially similar, construction, known as orphaning (Zribi-Hertz 1984). This is illustrated in (3). In contrast to stranding, orphaning is not admissible in English, as can be seen from the gloss of (3). (3) Si tu veux l’ avoir avant, il faut tu payes pour. If you want it to.have before it have you pay for *‘If you want it before, you have to pay for.’

(OH.013.258)

This raises the question of whether preposition stranding in French is an extension (to relative clauses) of the native orphaning, or a change induced by contact with English. Reasoning that if the latter were the case, and the above claims were correct, speakers who code-switched frequently to that language should engage in more stranding, Zentz (2006) contrasted the stranding behavior of “high” and “low” code-switchers. Using constraint hierarchies to first determine and then compare the structure of the choice mechanism, Zentz showed (Table 22.4) that the strongest predictor of preposition orphaning in French is lexical. “Strong” prepositions (a category made up, for the most part, of four: avec ‘with’, pour ‘for’, dedans ‘in’, dessus ‘on’) favor orphaning, while “weak” prepositions (here, a` ‘to’ and de ‘from’) disfavor it almost categorically. This, pattern, which is perfectly consistent with normative injunctions (Grevisse 1980: 1509), is shared by all speakers, regardless of propensity to code-switch, as can be seen by the parallel constraint hierarchies. In comparison, stranding, the candidate for change, not only occurs at the same low rate as orphaning (approximately ten percent of all eligible contexts), but is constrained in the same way. Here too, the most important conditioning factor is the lexical identity of the preposition. As with orphaning, stranding is also eschewed with weak prepositions and highly promoted by strong prepositions. And the strong prepositions in question are precisely those that favor orphaning. This effect is again as true of high as of low code-switchers. Once stranding ⫺ the superficially English structure ⫺ is examined in the context of the recipient-language system, three independent results counter the notion that it was borrowed from English. Stranding occurs at the same (low) rate and more important, obeys the same variable constraints as native orphaning. Like orphaning, stranding is restricted to like orphaning, standing is basically restuded to the same four strong prepositions. In English, on the other hand, stranding would seem to be equally productive with a wide variety of lexical prepositions, strong and weak. Finally, there is no distinction between high and low code-switchers, refuting claims that the former are agents of contact-induced change (see also Torres-Cacoullos and Travis to appear). The great merit of the Zentz study is that it marshals accountable scientific methodology to demonstrate that, contrary to appearances, a syntactic structure apparently identical to a counterpart in the majority language, and which is moreover (prescriptively) proscribed in the minority language, is nonetheless unlikely to be a contact-induced change. This is because the conditions giving rise to stranding in French parallel those constraining native orphaning, while differing from those assumed to be operating in English. The drawback of this study (and most others) is that it compares the situation

406

IV. Structure and dynamics across language spaces Tab. 22.4: Independent variable rule analyses of the contribution of linguistic factors to preposition orphaning and stranding for low and high code-switchers (brackets indicate that the factor was not statistically significant)

Corrected mean Total N SYNTACTIC FACTORS Place of topic Discourse referent Intra-sentential referent Type of construction Cleft Object precedes relative Pseudo-cleft Presence of another complement None Additional Post-V complement Clitic SEMANTIC FACTORS Semantic content of preposition “Weak” “Strong” Obligatoriness of preposition Obligatory Non-obligatory Humanness of noun complement Non-human Human

Orphaning

Stranding

“Low” “High” Code-switchers Code-switchers

“Low” “High” Code-switchers Code-switchers

.09 982

.02 662

.01 224

.03 116

[

]

[

]

N/A

N/A

[

]

[

]

N/A

N/A

N/A

N/A

.35

[

]

N/A N/A

N/A N/A

.40 .99

[ [

] ]

[

]

[

]

[

]

[

]

[ [

] ]

[ [

] ]

[ [

] ]

[ [

] ]

0

.19 .96

.42 .61

[ [

] ]

.57 .36

[ [

] ]

.19 .99

[ [

.19 .95

] ]

[ [

] ]

.46 .82

[ [

] ]

in the contact language with what is presumed to be the case in the source. Absent any accountable study of preposition stranding in the contact variety of English, we cannot be fully confident about exactly what was transferred. In additional, there is no evidence (in this or other studies), that the current state of affairs represents a change. Demonstration of contact-induced change would minimally include a vertical comparison (preferably with a pre-contact stage, but at the very least with an earlier stage), a horizontal comparison with a non-contact variety and comparison with the structure of the source.

22. Contact-induced grammatical change

5.3. Relative marker expression in Quebec English A study by Poplack et al. (2006) illustrates the unexpected results to emerge from such an approach. It involves the variable use of overt and null relative markers in accusative relative clauses in Quebec English, a minority language in intense contact with French. This is illustrated in example (4). Quebec French shows the same (surface) alternation in the same context (5). (4) a. I said “Oh yeah, I remember I was wearing this disgustingly ugly green suit that you made me wear”, or something. (QEC.031.1135) b. I was always known as Carrot Head and Freckle Face and a few other very vulgar names Ø they called me. (QEC.006.602) (5) a. Bien je comprenais le franc¸ais, mais c’ est les mots que je comprenais Well I understood the French but it is the words that I understood pas. not ‘Well, I understood the French, but it’s the words that I didn’t understand.’ (OH.059.812) b. Les maisons Ø ils faisaient dans ce temps la` le bois e´tait tre`sThe houses Ø they made in that time there the wood was veryil e´tait pas cher pantoute. it was not expensive at all ‘The houses they made in those days, the wood was very- it wasn’t expensive at all.’ (OH.050.698) In particular, both languages permit a null variant (4b, 5b), although this is prescriptively unacceptable in French. How is it generated? Table 22.5 displays the factors constraining choice of null variant across mainstream varieties and cohorts of contact-variety speakers. A first important observation is that the structure of variable relative marker expression differs in the two non-contact varieties. In English (column 2), the strongest predictor of a null relative marker is the presence of intervening material between the subject of the relative clause and its antecedent; in French (column 1), the major effect is phonological. Because these effects are language-specific, we can use them as diagnostics of contact-induced change. We may now compare the contact variety to its non-contact counterpart (columns 2 and 3). For all factor groups, the constraint hierarchies are parallel. This means that, in the aggregate, the contact variety of English is no different from its mainstream counterpart. But according to the hypotheses enunciated earlier, certain subgroups of the population could be expected to spearhead changes. So we will want to establish whether all of the minority-language speakers behave in the same fashion, by comparing the conditioning of variant choice according to speaker age and bilingual ability. Predictably, the behavior of the older (column 5) and least bilingual (column 7) speakers is very similar to that of speakers of noncontact English (column 2) in terms of constraint hierarchies. But the younger (column 4) and more bilingual speakers (column 6) share less with the main-

407

408

IV. Structure and dynamics across language spaces Tab. 22.5: Independent variable rule analyses of the contribution of linguistic factors to the selection of the null relative marker: across mainstream varieties

Corrected mean: Total N

1

2

3

4

5

6

7

Mainstream French

Mainstream English

Contact English

Younger speakers

Older speakers

Most bilingual speakers

Least bilingual speakers

.019 50/460

.367 109/288

.532 150/281

.400 131/316

.502 128/253

.503 75/146

.568 75/135

.52 .17

.54 .24

.55 .19

Intervening material None Present

[ [

Antecedent type Pronoun Noun

.71 .44

.73 .46

[ [

Type of subject of the relative clause Pronoun Noun

0

.52 .28

.51 .32

Following phonological environment Sonorants Obstruents

.13 .94

[ [

] ]

[ [

] ]

.54 .36

Definiteness of antecedent Definite Indefinite

[ [

[ [

] ]

[ [

] ]

[ [

] ]

.59 .36

Sentence type Cleft Other

.61 .44

[ [

] ]

[ [

] ]

[ [

] ]

[ [

] ]

] ]

] ]

[ [

] ]

.57 .14

[ [

[ [

] ]

.76 .45

[ [

] ]

.89 .42

[ [

] ]

.53 .15

[ [

] ]

.53 .18

[ [

] ]

[ [

[ [

] ]

.60 .31

[ [

] ]

[ [

[ [

] ]

] ]

] ]

] ]

] ]

stream in terms of the fine conditioning of variant choice. This kind of result is consistent with their predicted role as leaders of change. Are these in fact changes? Comparison with older speakers suggests that they are. But are they contact-induced? As we noted earlier, to sustain any claim in this respect, the involvement of the source language (here, French) must be established. The three major “departures” of the younger and more bilingual cohorts with respect to their older and less bilingual counterparts involve the effects of antecedent type, type of subject of the relative and following phonological environment. The first two effects have been neutralized in the younger generation, while in French, as shown in column 1, they are strong and significant. The effect most relevant to the contact situation is phonological. Recall that this is the major predictor of variant choice in French, and the younger and most bilingual anglophones are the only ones who display it. This result is consistent with the inference of contact-induced change. But comparison of constraint hierarchies shows that the French effect goes in the opposite direction. (The apparent English phonological effect is an epiphenomenon of a developing tendency for certain pronominal subjects which are coincidentally obstruent-initial (he, she, they) to disfavor deletion.) Here again, differences among minority-language cohorts, suggestive as they appear on the surface, cannot be attributed to convergence

22. Contact-induced grammatical change

409

Tab. 22.6: Independent variable rule analyses of the contribution of linguistic factors to the selection of the null relative marker: younger contact vs. benchmark speakers Younger contact speakers

Younger mainstream speakers

.482 83/168

.299 48/148

Intervening material None Present

.56 .18

.55 .10

Antecedent type Pronoun Noun

[ [

] ]

.76 .45

Type of subject of the relative clause Pronoun Noun

[ [

] ]

[ [

Following phonological environment Sonorants Obstruents

[.53] [.38]

[.54] [.34]

Definiteness of antecedent Definite Indefinite

.59 .40

[ [

Corrected mean: Total N

] ]

] ]

with French. To what then are they due? Comparison with the non-contact benchmark provides the answer (Table 22.6). When speakers of mainstream English are analyzed according to age, we find the same apparent changes in progress. In this case our method has enabled us to detect change ⫺ the young minoritylanguage anglophones are diverging from their elders (column 5 of Table 22.5) ⫺ and to attribute it to the correct source: participation in a change in progress in mainstream English. These examples, along with many others (e.g., Dion 2003; Jarmasz 2005; Leroux and Jarmasz 2006; Petrik 2005; Poplack and St-Amand 2007; St-Amand 2002) demonstrate that when the inference of contact-induced change is pursued systematically, it becomes increasingly difficult to justify. Particularly instructive is the fact that even in quantitative studies, differences between cohorts or overall rates may be masking effects that are independent of the contact situation.

6. Discussion Change is widely considered a predictable, if not inevitable, outcome of language contact. The literature is rife with accounts of radical alterations of the grammars of contact languages, whether by virtue of incorporating forms or functions with no antecedent in the recipient language, or losing one or more of its core grammatical elements. Many of these changes would appear to be very rapid, occurring even in the space of a single

410

IV. Structure and dynamics across language spaces generation. This is consistent with another widespread, if unproven, idea: that contact accelerates change. On both of these counts, contact-induced change appears suspiciously different from internal evolution, which we know to be gradual, moderate and vastly more conservative. Observing that scientific demonstration of change, contact-induced or other, was often missing from this research, we reviewed some of the phenomena that could lead the analyst down the garden path of inferring change, even where such an inference is not warranted. Chief among them are the inherent variability characteristic of spontaneous speech and the dearth of accountable corpus-based studies of the languages in contact. When confronted with an idealized standard variety, but lacking systematic comparison with the more relevant pre- and non-contact varieties, or with the supposed source variety of the claimed change, we detailed how the (variable) occurrence of non-standard variants could easily be confused with change. We also outlined a number of conditions necessary to establish the existence of contact-induced change. Many had already been enunciated elsewhere (Poplack 1993; Poplack and Meechan 1998); some were recently reiterated (e.g., Backus 2005: 310). The following are paraphrased from Thomason (2001: 93⫺94): 1. 2. 3. 4.

Situate the proposed change with respect to its host linguistic system Identify a presumed source of the change Locate structural features shared by the source and recipient languages Prove that the proposed interference features were not present in the pre-contact variety 5. Prove that the proposed interference features were present in the source variety prior to contact 6. Rule out (or situate) internal motivations Thomason herself concedes that “in many, possibly even most contact situations around the world we cannot at present satisfy requirements 4 and 5”, going so far as to caution that “if we can’t satisfy all the requirements, then we can’t make a solid case for contactinduced change” (Thomason 2001: 94). This raises the issue, which has preoccupied us throughout this article, of just how many of the contact-induced changes currently treated as facts satisfy any of these conditions, let alone all. This is a very important question, since these “facts” are the basis for our theories about the nature, extent and mechanisms of contact-induced change. We observed earlier that for many diachronic changes, the crucial evidence would now be impossible to reconstruct. But as we have been at pains to illustrate, this is certainly not the case for synchronic situations. It is thus all the more puzzling that the requisite information is missing from all but a handful of contemporary studies. Why should this be? The growing, but patently false, perception among linguists working in other paradigms that it is too difficult, if not impossible, to access the relevant evidence may play a role. The treatment of frequency is a case in point. Frequency is at the root of most inferences of change, whether manifested as raw rates of occurrence of a form, or its diffusion across a community or linguistic system. Among the few who have addressed this issue at all, a certain skepticism about the feasibility of ascertaining frequency empirically is detectible. Thus with reference to the appearance of foreign elements in bilingual speakers’ discourse, Thomason observes, “determining the frequency of appearance is a hopeless task” (2001: 134), and goes on to query: “how can one

22. Contact-induced grammatical change determine whether a speaker used [an element] once, occasionally or frequently?” Likewise, while Backus (2005) admits that most changes are likely of the “frequency-changing kind”, he contends that these are precisely the changes that “are virtually impossible to demonstrate empirically” (2005: 334). Such assertions betray a surprising lack of familiarity with corpus-based studies, and fly in the face of the important results of largescale quantitative analyses that have been performed on languages in contact. Despite recent concerns about the absence of an empirically well-founded framework for investigating bilingual speech (Van Coetsem 2000: 39⫺40; Winford 2007: 22) and the alleged limitations of variation-based techniques for illuminating contact phenomena (Muysken 2000: 250), we submit that the rigorous comparative methodology presented here is capable of detecting change and distinguishing external from internal motivations. It thus discriminates among competing hypotheses about the nature of these processes. Agreeing with Thomason (2001: 88) that the crucial similarity between the processes of contact-induced and internal change has to do with competition (i. e. variability) between an innovative feature and an older feature, we have outlined a procedure for assessing contact-induced change within a framework that takes account of its symbiotic relationship with linguistic variation. We hope to have shown that reliance on the structure of variant choice affords a depth of analysis that transcends impressionistic observations based on perceived similarities between surface forms and rates of occurrence. Indeed, the very strength of the variation-based approach resides in its capacity to discern even minor perturbations in constraint hierarchies, which may be harbingers of change, contact-induced or otherwise. Otheguy, Livert and Zentella (2007) provide the most recent confirmation that it can detect even the smallest adjustments in constraint ranking across generations of bilingual speakers. Illustrating with studies of three well-documented candidates, we found no evidence of change in the first two, and the change documented in the third was not contactinduced. Moreover, in line with Thomason’s (2001: 91⫺92) caveat that a “case for contact origin can’t be made on the basis of single feature”, additional research situating a wide variety of Canadian French and English vernacular variants in historical, social and linguistic context corroborates the findings presented here: none of the contemporary non-standard variants typically ascribed to either linguistic change or language contact are in fact innovations (Poplack and Dion to appear; Elsig and Poplack 2006; Jarmasz 2005; Leroux 2007; Poplack 1993). Rather, virtually all were not only present in the pre-contact variety, but, more important, they were largely constrained in the same way. These results are especially striking when we consider that all the popularly invoked prerequisites for contact-induced change ⫺ social (protracted contact, extensive bilingualism; high levels of proficiency in the respective source languages, etc.) and linguistic (i. e. typological congruence [Thomason 2001: 71], interlingual equivalence [Backus 2005: 326]; grammatical “weak points” [Bullock and Toribio 2004: 91], attractiveness [Johanson 2002: 2]) ⫺ are met. Why then did contact-induced change fail to occur? Given the present state of our knowledge, we can only speculate. The length of contact may have been too brief to result in change, although a time span of two centuries is a good deal longer than those on which many recent claims have been based. Perhaps the morpho-syntactic features we examined are particularly resistant to change, although they were specifically selected on the basis of typological similarities thought to facilitate it. We cannot rule out the possibility that our analytical tools may have misidentified or excluded changes; we

411

412

IV. Structure and dynamics across language spaces reserve judgment until a demonstration to that effect is offered. Certainly the circumstances ⫺ linguistic, social and attitudinal ⫺ could not have been more propitious. But while favorable circumstances may encourage change, we know that they do not determine it. A remaining possibility is that contact-induced change is a good deal less common than the literature suggests. This is our best guess at the moment. Of course, the analyses we have presented, despite being based on corpora of bilingual speech with sizes ranging into the millions of words and large representative samples of individuals in contact situations, cannot begin to do justice to the full trajectory of change over time and across all speakers of a language. But opting for the alternative ⫺ anecdotal observation ⫺ simply because “it is much easier to show a qualitative change” (Backus 2005: 334) ⫺ ignores the important scientific advances in empirical linguistics made over the last half century. Contact-induced change is not an inevitable, nor possibly even a common, outcome of language contact. Only more accountable analyses of more contact situations will tell. In the interim, the burden of proof is on those who claim that it has occurred.

7. Acknowledgement The research reported here was generously funded by grants to Poplack from the Social Sciences and Humanities Research Council of Canada and the Killam Foundation. Poplack holds the Canada Research Chair in Linguistics at the University of Ottawa.

8. Reerences Aikhenvald, Alexandra 2002 Language contact in Amazonia. New York: Oxford University Press. Altenberg, Evelyn P. 1991 Assessing first language vulnerability to attrition. In: Seliger and Vago (eds.), 189⫺206. Appel, Rene´ and Pieter Muysken 1987 Language Contact and Bilingualism. London: Edward Arnold. Ayres-Bennett, Wendy 2000 Voices from the past: Sources of spoken seventeenth-century French. Romanische Forschungen 112(3): 328⫺348. Backus, Ad 2005 Codeswitching and language change: One thing leads to another? International Journal of Bilingualism 9(3/4): 307⫺340. Baldi, Philip 1990 Introduction: The Comparative Method. In: Philip Baldi (ed.), Linguistic Change and Reconstruction Methodology, 1⫺13. Berlin: Mouton de Gruyter. Bakker, Peter 1997 “A Language of Our Own”: The Genesis of Michif, the Mixed Cree-French Language of the Canadian Me´tis. (Oxford Studies in Anthropological Linguistics 10.) Oxford: Oxford University Press. Bakker, Peter and Robert A. Papen 1997 Michif: A mixed language based on Cree and French. In: Sarah G. Thomason (ed.), Contact Languages, 295⫺363. Amsterdam: John Benjamins.

22. Contact-induced grammatical change Bullock, Barbara E. and Almeida Jacqueline Toribio 2004 Introduction: Convergence as an emergent property in bilingual speech. Bilingualism: Language and Cognition 7(2): 91⫺94. Bynon, Theodora 1977 Historical Linguistics. Cambridge: Cambridge University Press. Campbell, Lyle 1998 Historical Linguistics: An Introduction. Edinburgh: Edinburgh University Press. Campbell, Lyle 1999 Historical Linguistics: An Introduction. Cambridge. MA: MIT Press. Campbell, Lyle 2003 How to show languages are related: Methods for distant genetic relationship. In: Joseph and Janda (eds.), 262⫺282. Clyne, Michael 2003 Dynamics of Language Contact. Cambridge: Cambridge University Press. Croft, William 2000 Explaining Language Change: An Evolutionary Approach. Harlow/New York: Longman. Dawkins, Richard MacGillivray 1916 Modern Greek in Asia Minor. Cambridge: Cambridge University Press. Dion, Nathalie 2003 L’effacement du que en franc¸ais canadien: Une e´tude en temps re´el. Unpublished MA thesis (me´moire), Department of Linguistics, University of Ottawa. Dorian, Nancy C. 1981 Language Death: The Life Cycle of a Scottish Gaelic Dialect. Philadelphia, PA: University of Pennsylvania Press. Dorian, Nancy C. (ed.) 1989 Investigating Obsolescence: Studies in Language Contraction and Death. Cambridge: Cambridge University Press. Dorian, Nancy 1993 Internally and externally motivated change in language contact settings: Doubts about dichotomy. In: Charles Jones (ed.), Historical Linguistics. Problems and Perspectives, 131⫺154. London: Longman. Elsig, Martin and Shana Poplack 2006 Transplanted dialects and language change: Question formation in Que´bec. University of Pennsylvania Working Papers in Linguistics 10(2): 77⫺90. [Selected Papers from NWAV 34.] Field, Frederic 2005 Long-term effects of CS: Clues to structural borrowing. International Journal of Bilingualism 9(3/4): 341⫺360. Gal, Susan 1979 Language Shift: Social Determinants of Linguistic Change in Bilingual Austria. New York: Academic Press. Garcı´a, Erica C. 1985 Shifting variation. Lingua 67: 189⫺224. Gordon, Elizabeth M., Lyle Campbell, Jennifer Hay, Margaret Maclagan, Andrea Sudbury and Peter Trudgill 2004 New Zealand English: Its Origins and Evolution. Cambridge: Cambridge University Press. Grevisse, Maurice 1980 Le Bon Usage. Duculot: Paris. Gumperz, John J. and Robert Wilson 1971 Convergence and creolization: A case from the Indo-Aryan/Dravidian border. In: Dell Hymes (ed.), Pidginization and Creolization of Languages, 151⫺167. Cambridge: Cambridge University Press.

413

414

IV. Structure and dynamics across language spaces Gutie´rrez, Manuel 1994 Simplification, transfer and convergence in Chicano Spanish. The Bilingual Review/La Revista Bilingüe: 19: 111⫺121. Harris, Alice C. and Lyle Campbell 1995 Historical Syntax in Cross-Linguistic Perspective. Cambridge: Cambridge University Press. Haugen, Einar 1954 Review of Weinreich (1953). Language 30: 380⫺388. Heine, Bernd and Tania Kuteva 2005 Language Contact and Grammatical Change. Cambridge: Cambridge University Press. Hoenigswald, Henry 1960 Language Change and Linguistic Reconstruction. Chicago: University of Chicago Press. Huffines, Marion Lois 1991 Pennsylvania German: Convergence and change as strategies of discourse. In: Seliger and Vago (eds.), 125⫺137. Jakobson, Roman 1962 Sur la the´orie des affinite´s phonologiques entre des langues. In: Stephen Rudy (ed.), Roman Jakobson: Selected Writings, vol. 1, 234⫺246. The Hague: Mouton. Jarmasz, Lidia-Gabriela 2005 “I think I know what I’m thinking”: The alternation between simple and progressive forms in stative verbs. Unpublished MA thesis (me´moire), Department of Linguistics, University of Ottawa. Johanson, Lars 2002 Structural Factors in Turkic Language Contact. London: Curzon. Joseph, Brian D. 2007 Attitudes and ideology in contact-induced change: Trumping all psychological and structural factors. Paper presented at the 18th International Conference on Historical Linguistics (ICHL). Universite´ du Que´bec a` Montre´al, August. Joseph, Brian D. and Richard D. Janda (eds.) 2003 The Handbook of Historical Linguistics. Oxford: Blackwell. Kaufman, Dorit and Mark Aronoff 1991 Morphological disintegration and reconstruction. In: Seliger and Vago (eds.), 175⫺188. King, Ruth 2000 The Lexical Basis of Grammatical Borrowing. Philadelphia/Amsterdam: John Benjamins. King, Ruth 2005 Crossing grammatical borders: Tracing the path of contact-induced linguistic change. In: Markku Filppula, Juhani Klemola, Marjatta Palander and Esa Pentillä (eds.), Dialects Across Borders: Selected Papers from the 11th International Conference on Methods in Dialectology (Methods XI), Joensuu, August 2002, 233⫺251. Amsterdam/Philadelphia: John Benjamins. Klein, Flora 1980 A quantitative study of syntactic and pragmatic indications of change in the Spanish of bilinguals in the U.S. In: William Labov (ed.), Locating Language in Time and Space, 69⫺82. New York: Academic Press. Labov, William [1966] 1982 The Social Stratification of English in New York City. Washington: Center for Applied Linguistics. Labov, William 1969 Contraction, deletion and inherent variability of the English copula. Language 45(4): 715⫺762. Labov, William 1972a Language in the Inner City. Philadelphia: University of Pennsylvania Press.

22. Contact-induced grammatical change Labov, William 1972b Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Labov, William 1984 Field methods of the project on language change and variation. In: John Baugh and Joel Schezer (eds.), Language in Use, 28⫺53. Englewood Cliffs, NJ: Prentice Hall. Labov, William 1994 Principles of Linguistic Change: Internal Factors. Oxford: Blackwell. Labov, William, Paul Cohen, Clarence Robins and John Lewis 1968 A study of the non-standard English of Negro and Puerto Rican speakers in New York City. Final report, Co-operative Research Report 3288, vol. I, US Regional Survey. Lantolf, James 1978 The variable constraints on mood in Puerto Rican Spanish. In: Margarita Sun˜er (ed.), Contemporary Studies in Romance Linguistics, 193⫺217. Washington, DC: Georgetown University Press. Lass, Roger 1980 On Explaining Language Change. Cambridge: Cambridge University Press. Laurier, Michel 1989 Le subjonctif dans le parler franco-ontarien: Un mode en voie de disparition? In: Raymon Mougeon and E´douard Beniak (eds.), Le Franc¸ais Canadien Parle´ Hors Que´bec: Aperc¸u Sociolinguistique, 105⫺126. Quebec: Presses de l’Universite´ Laval. Lehmann, Winfred P. 1992 Historical Linguistics. 3rd ed. London/New York: Routledge. Leroux, Martine 2003 Le passe´ recompose´: Analyse diachronique et multivarie´e de la valeur temporelle du passe´ en franc¸ais oral de Hull (Que´bec). Unpublished MA thesis (me´moire), Department of Linguistics, University of Ottawa. Leroux, Martine 2007 Something old, something new, something borrowed, something true: What are null subjects in French? Paper presented at New Ways of Analyzing Variation 36, University of Pennsylvania, Philadelphia, October. Leroux, Martine and Lidia-Gabriela Jarmasz 2006 A study about nothing: Null subjects as a diagnostic of convergence between English and French. University of Pennsylvania Working Papers in Linguistics 12(2): 1⫺14. Martineau, France and Raymond Mougeon 2003 A sociolinguistic study of the origins of ne deletion in European and Quebec French. Language 79(1): 118⫺152. Martinet, Andre´ 1960 E´le´ments de Linguistique Ge´ne´rale. Paris: Colin. Meillet, Antoine 1967 The Comparative Method in Historical Linguistics. Paris: Librairie Honore´ Champion. Milroy, James 1992 Linguistic Variation and Change: On the Historical Sociolinguistics of English. Oxford: Blackwell. Milroy, James 1993 On the social origins of language change. In: Charles Jones (ed.), Historical Linguistics: Problems and Perspectives, 215⫺236. London: Longman. Montrul, Silvana 2002 Incomplete acquisition and attrition of Spanish tense/aspect distinctions in adult bilinguals. Bilingualism: Language and Cognition 5: 39⫺68. Mougeon, Raymond and E´douard Beniak 1991 Linguistic Consequences of Language Contact and Restriction: The case of French in Ontario. Oxford: Oxford University Press.

415

416

IV. Structure and dynamics across language spaces Mougeon, Raymond, Terry Nadasdi and Katherine Rehner 2005 Contact-induced linguistic innovations on the continuum of language use: The case of French in Ontario. Bilingualism: Language and Cognition 8 (2): 99⫺115. Müller, Max 1871 Lectures on the Science of Language. 1st ed. New York: Scribner Muysken, Pieter 2000 Bilingual Speech: A Typology of Code-Mixing. Cambridge: Cambridge University Press. Ocampo, Francisco 1990 The pragmatics of word order in constructions with a verb and a subject. Hispanic Linguistics 4(1): 87⫺127. Otheguy, Ricardo, Ana Celia Zentella and David Livert 2007 Language and dialect contact in Spanish in New York: Towards the formation of a speech community. Language 83: 1⫺33. Petrik, Katrina 2005 Deontic modality in Quebec English: “Everything you need to know”. Unpublished MA thesis (me´moire), Department of Linguistics, University of Ottawa. Poplack, Shana 1983 Bilingual competence: Linguistic interference or grammatical integrity? In: Lucia Elı´asOlivares (ed.), Spanish in the U.S. Setting: Beyond the Southwest, 107⫺131. Arlington: National Clearinghouse for Bilingual Education. Poplack, Shana 1989 The care and handling of a mega-corpus. In: Ralph Fasold and Deborah Schiffrin (eds.), Language Change and Variation, 411⫺451. Amsterdam: Benjamins. Poplack, Shana 1990 Prescription, intuition et usage: le subjonctif franc¸ais et la variabilite´ inhe´rente. Langage et Socie´te´ 54: 5⫺33. Poplack, Shana 1992 The inherent variability of the French subjunctive. In: Christiane Laufer and Terrell Morgan (eds.), Theoretical Analyses in Romance Linguistics, 235⫺263. Amsterdam: Benjamins. Poplack, Shana 1993 Variation theory and language contact. In: Dennis Preston (ed.) American Dialect Research: An Anthology Celebrating the 100th Anniversary of the American Dialect Society, 251⫺286. Amsterdam: Benjamins. Poplack, Shana 1997 The sociolinguistic dynamics of apparent convergence. In: Gregory Guy, John Baugh and Deborah Schiffrin (eds.), Towards a Social Science of Language, 285⫺309. Amsterdam: Benjamins. Poplack, Shana 2001 Variability, frequency and productivity in the irrealis domain of French. In: Joan Bybee and Paul Hopper (eds.), Frequency and the Emergence of Linguistic Structure, 405⫺428. Amsterdam: Benjamins. Poplack, Shana 2003 Language contact and the evolution of French. Paper presented at Linguistic Symposium on the Romance Languages (LSRL) 33 and Conference on French in North America, Indiana University, Bloomington, April. Poplack, Shana and Nathalie Dion to appear “Prescription versus praxis: The evolution of future temporal reference in French”. Language. Poplack, Shana, Nathalie Dion, Lidia-Gabriela Jarmasz, Carmen LeBlanc and Nicole Rosen 2002 Re´pertoire Historique des Grammaires du Franc¸ais. Corpus and documentation. Sociolinguistics Laboratory, University of Ottawa.

22. Contact-induced grammatical change Poplack, Shana, Adrienne Jones, Allison V. Lealess, Martine Leroux, Chelsea T. Smith, Yukiko Yoshizumi, Lauren Zentz and Nathalie Dion 2006 Assessing convergence in contact languages. Paper presented at New Ways of Analyzing Variation 35, The Ohio State University, Columbus, November. Poplack, Shana and Elisabete Malvar 2007 Elucidating the transition period in linguistic change. Probus 19(1): 121⫺169. Poplack, Shana and Marjory Meechan 1998 How languages fit together in code-mixing. Special issue of the International Journal of Bilingualism 2(2): Instant Loans, Easy Conditions: the Productivity of Bilingual Borrowing: 127⫺138. Poplack, Shana and Anne St-Amand 2007 A real-time window on 19th century vernacular French: The Re´cits du franc¸ais que´be´cois d’autrefois. Language in Society 36(5): 707⫺734. Poplack, Shana and Sali Tagliamonte 2001 African American English in the Diaspora. Oxford: Blackwell Publishers. Poplack, Shana, James Walker and Rebecca Malcolmson 2006 An English “like no other”?: Language contact and change in Quebec. Canadian Journal of Linguistics, 51(2/3): 185⫺213. Rand, David and David Sankoff 1990 Goldvarb 2.1: A variable rule application for the Macintosh. Montreal. Centre de Recherches Mathe´matiques, University of Montreal. Version 2. http://www.crm.umontreal. ca/~sankoff/GoldVarb_Eng.html. Ringe, Don 2003 Internal reconstruction. In: Brian D. Joseph and Richard D. Janda (eds.), The Handbook of Historical Linguistics, 244⫺261. Oxford: Blackwell. Roberge, Yves and Nicole Rosen 1999 Preposition stranding and que-deletion in varieties of North-American French. Linguistics Atlantica 21: 153⫺168. Romaine, Suzanne 1988 Pidgin and Creole Languages. London: Longman. Romaine, Suzanne 1995 Bilingualism. 2nd ed. Malden, MA: Blackwell. Sankoff, David 1988a Sociolinguistics and syntactic variation. In: Frederick J. Newmeyer (ed.), Linguistics: The Cambridge Survey, vol. 4, Language: The Socio-Cultural Context, 140⫺161. Cambridge: Cambridge University Press. Sankoff, David 1988b Variable rules. In: Ulrich Ammon, Norbert Dittmar and Klaus J. Mattheier (eds.), Sociolinguistics: An International Handbook of the Science of Language and Society, vol. 1, 984⫺997. (Handbücher zur Sprach- und Komunikationswissenschaft 3.1.) Berlin: de Gruyter. Sankoff, David, Sali Tagliamonte and Eric Smith 2005 Goldvarb X: A Variable Rule Application for Macintosh and Windows. Toronto: Department of Linguistics, University of Toronto. Savic´, Jelena M. 1995 Structural convergence and language change: Evidence from Serbian/English codeswitching. Language in Society 24(4): 475⫺492. Schmidt, Annette 1991 Language attrition in Boumaa Fijian and Dyirbal. In: Seliger and Vago (eds.), 112⫺124. Schuchardt, Hugo 1884 Slawo-deutches und Slawo-italienisches. Graz: Leuschner und Lubensky.

417

418

IV. Structure and dynamics across language spaces Seliger, Herbert and Robert M. Vago (eds.) 1991 First Language Attrition. Cambridge: Cambridge University Press. Shapiro, Michael  1995 A case of distant assimilation: /str/ J / tr/. American Speech 70: 101⫺107. Silva-Corvala´n, Carmen 1991 Spanish language attrition in a contact situation with English. In: Seliger and Vago (eds.), 151⫺171. Silva-Corvala´n, Carmen 1994a Language Contact and Change. Oxford: Clarendon Press. Silva-Corvala´n, Carmen 1994b The gradual loss of mood distinctions in Los Angeles Spanish. Language Variation and Change 6(3): 255⫺272. St-Amand, Anne 2002 Le subjonctif suivant une expression non-verbale. Unpublished MA thesis (me´moire), Department of Linguistics, University of Ottawa. St-Amand, Anne and Shana Poplack 2002 A real-time window on 19th century vernacular French: The Re´cits du franc¸ais que´be´cois d’autrefois. Paper presented at the annual meeting of the Canadian Linguistics Association, University of Toronto, Toronto. Tagliamonte, Sali 2002 Comparative sociolinguistics. In: Jack Chambers, Peter Trudgill and Natalie SchillingEstes (eds.), Handbook of Language Variation and Change, 729⫺763. Malden/Oxford: Blackwell. Thomason, Sarah G. 2001 Language Contact: An Introduction. Washington, DC: Georgetown University Press. Thomason, Sarah G. 2007 Language contact and deliberate change. Journal of Language Contact 1: 41⫺62. Thomason, Sarah G. and Terrence Kaufman 1988 Language Contact, Creolization, and Genetic Linguistics. Berkeley/Los Angeles: University of California Press. Torres-Cacoullos, Rena and Catherine E. Travis to appear Variable yo expression in New Mexico: English influence? In: Susana V. Rivera-Mills and Daniel J. Villa (eds.), Spanish of the Southwest: A Language in Transition. Madrid: Iberoamericana/Vervuert. Toribio, Almeida Jacqueline 2004 Convergence as an optimization strategy in bilingual speech: Evidence from codeswitching. Bilingualism: Language and Cognition, 7(2): 165⫺173. Van Coetsem, Frans 2000 A General and Unified Theory of the Transmission Process in Language Contact. Heidelberg: Winter. Weinreich, Uriel, William Labov and Marvin Herzog 1968 Empirical foundations for a theory of language change. In: Winfred P. Lehmann and Yakov Malkiel (eds.), Directions for Historical Linguistics, 97⫺195. Austin: University of Texas Press. Whitney, William Dwight 1881 On mixture in language. Transactions of the American Philosophical Association 12: 1⫺26. Willis, Lauren 2000 Eˆtre ou ne plus eˆtre: Auxiliary alternation in Ottawa-Hull French. Unpublished MA thesis (me´moire), Department of Linguistics, University of Ottawa. Winford, Donald 2003 An Introduction to Contact Linguistics. Oxford: Blackwell.

23. Areal language typology

419

Winford, Donald 2007 Some issues in the study of language contact. Journal of Language Contact⫺Thema 1 (2007). Available online from . Zentz, Lauren 2006 “C’est c¸a je travaille dessus”: Orphaned prepositions and relativization in Canadian French. Unpublished MA thesis (me´moire), Department of Linguistics, University of Ottawa. Zribi-Hertz, Anne 1984 Orphan Prepositions in French and the Concept of “Null Pronoun”. Indiana University Linguistics Club: Bloomington, IN. Zwicky, A. 2005 More illusions. Available at . Accessed 29 April 2008.

Shana Poplack and Stephen Levey, Ottawa (Canada)

23. Areal language typology 1. Introduction: Setting the stage 2. Language typology 3. Language change: The social and linguistic dynamics of contact-induced convergence and areality 4. Consequences for Linguistic Typology and its generalizations 5. Conclusion and outlook 6. References

1. Introduction: Setting the stage Linguistic typology deals with linguistic variation and the discovery of cross-linguistic universal patterns reflected in that variation. The aim of the present article is to show that the robustness and the reliability of typological generalizations crucially depends on the appropriate assessment of areality, i. e., the influence of geographic closeness and language contact on structural properties of languages. Thus, the analysis of structures and their overall statistical distribution in the population of the world’s languages does not guarantee any direct access to valid generalizations about language universals and about properties of the human cognition. This can be illustrated by an example from basic word order. In his 1986 volume, Tomlin looked at the overall percentage of the six logically possible combinations of S (Subject), O (Object) and V (Verb) in a statistically balanced sample of 402 languages. As it turned out, the majority of these languages were either SOV (180 languages, 44.78 percent) or SVO (168 languages, 41.79 percent). Only 37 languages (9.20 percent) were VSO. The sequences in which O preceded S were only represented by a small minority of twelve languages (2.99 percent) with VOS, five languages in the Amazonian region of

420

IV. Structure and dynamics across language spaces Brazil (1.24 percent) with OVS and no language with OSV. Tomlin’s (1986) data are statistically representative but they do not allow conclusions of what may be typical or normal properties of languages. This can be shown by a rather extreme thought experiment discussed by Newmeyer (1998): [S]uppose that a nuclear war wiped out most of humankind and its written history, but spared the Amazonia region of Brazil. Some centuries later, a carefully constructed sample of the world’s languages would in all probability show those with OVS order to be relatively common. (Newmeyer 1998: 307)

The general problem that underlies Newmeyer’s thought experiment is the contact-induced diffusion of linguistic structures across geographic areas to the detriment of other cognitively possible structures. In this context, Dryer (1989) points out that some 71 percent of all the VSO languages belong to the Austronesian family. Without the expansion of the speakers of these languages, VSO would be attested in some three percent of the world’s languages only. Similarly, the percentage of SVO languages would be considerably lower without the Niger-Congo family, which makes up about 40 percent of the languages belonging to this word-order type (Dryer 1989). The above example gives an idea of the dependence of typological findings on language contact and its areal consequences. A more detailed account will be given in the following sections. Section 2 will deal with language typology, its methods and some of its generalizations and functional motivations. The same section will also show how Integrative Functionalism (Croft 1995: 2000) provides a basis for linking typological research with social factors and their impact on the contact-induced diffusion of linguistic structures. Section 3 will discuss the social dynamics of contact and language change, its areal consequences, the possible correlation between contact and complexity and the new resources for finding areal correlations provided by the World Atlas of Language Structures (Haspelmath et al. 2005). Section 4 considers some suggestions of how to implement social aspects into typological methods and the problems they raise. Finally, section 5 will present an outlook on the role of areality for assessing the overall robustness of typological findings and for the arbitrariness of linguistic structures.

2. Language typology 2.1. Methods, universals and motivations Language typologists try to discover universal patterns inductively in a representative sample of the world’s languages. They start from a certain semantically or pragmatically defined concept and look at how this concept is expressed morphologically and syntactically in that sample. Croft (2003: 14) characterizes the typological method in terms of the following three standard research strategies: i. ii.

Determine the particular semantic(-pragmatic) structure or situation type that one is interested in studying. Examine the morphosyntactic construction(s) or strategies used to encode that situation type.

23. Areal language typology

421

iii. Search for dependencies between the construction(s) used for that situation and other linguistic factors: other structural features, other external functions expressed by the construction in question, or both. The application of these strategies will be illustrated by a brief look at the genitive construction and the relative clause construction. Cross-linguistic comparison crucially depends on a common basis (tertium comparationis) that is universally applicable. The definition of a semantic or pragmatic structure as outlined in strategy (i) serves exactly that purpose. In our example, the genitive construction can be roughly defined in terms of two entities, a possessor and a thing possessed, and the possessional relation between them. Relative clause constructions are characterized by an entity that is semantically modified by a state of affairs (for a more detailed definition, cf. Lehmann 1984). On the basis of these definitions, it is now possible to apply strategy (ii) and to examine how these situation types are formally expressed. For the sake of brevity, I will only look at word order, i. e., at the position of the possessor (Gen ⫽ genitive) relative to the thing possessed (N) and at the position of the state of affairs (Rel ⫽ relative clause) relative to the entity it modifies (N). A look at English shows that both word orders are possible in the case of possession (GenN: my friend’s house; NGen: the house of my friend; on genitive variation in English, cf. Rosenbach 2002), while states of affairs have to follow the entity they modify (NRel: the house [I want to buy]). Each of these expression formats (GenN, NGen, NRel) represent individual types. The search for cross-linguistic dependencies between the genitive construction and the relative-clause construction in terms of strategy (iii) reveals an interesting pattern concerning the combination of types. There are two parameters in our example. Each of them consists of two types. The parameter concerning the position of the genitive relative to the head noun consists of GenN and NGen. In the case of the relative-clause parameter, we have the types of RelN and NRel. From a logical perspective, each of the four mathematical combinations should be equally possible. As is illustrated in Table 23.1, the combination [NGen & RelN] is not attested ([⫹] means “attested” in the world’s languages, [⫺] means “unattested”): Tab. 23.1: Positions of Gen and Rel relative to N

NRel RelN

NGen

GenN

⫹ ⫺

⫹ ⫹

Since the above pattern is parallel to implications in propositional logic and since it holds universally in the world’s languages, it is called an implicational universal. Greenberg (1966) was the first who discovered implicational universals. The implicational correlation between the parameters NGen/GenN and NRel/RelN is defined as follows by Hawkins (1983; also cf. 1994: 2004): (1) Hawkins (1983): Universal (IX’): NG 傻 NRel If in a language the genitive follows the noun, then the relative clause follows likewise.

422

IV. Structure and dynamics across language spaces Implicational universals can be combined into chains or hierarchies. The two parameters based on the positions of the genitive (Gen) and the relative clause (Rel) within the noun phrase are part of a wider framework which additionally includes the positions of demonstratives (Dem), numerals (Num) and adjectives (Adj) and the likelihood with which these modifiers deviate from word-order patterns with prenominal modifiers in prepositional languages. As can be seen from (2), the adjective can only precede the noun if the demonstratives and the numerals do likewise, etc. This hierarchical correlation was discovered by Hawkins (1983; also cf. Hawkins 1994, 2004) who called it the Prepositional Noun Modifier Hierarchy: (2) Prepositional Noun Modifier Hierarchy (PrNMH): If a language is prepositional, then if RelN then GenN, if GenN then AdjN, and if AdjN then DemN/NumN. Prep 傻 ((NDem/NNum 傻 NAdj) & (NAdj 傻 NG) & (NG 傻 NRel)) There is a number of well-known hierarchies in linguistic typology. I only mention the animacy hierarchy (Silverstein 1976, Dixon 1979) as another example. This hierarchy includes animacy and person and covers a wide area of grammatical processes involving word order, choice of grammatical relation and case marking (e.g., split ergativity, direct/ inverse marking, etc.). It shows up in various versions (cf. Croft 2003: 165⫺175, 178⫺ 182). The version in (3) is from Dixon (1979): (3) Animacy hierarchy (Dixon 1979: 85): 1st Person > 2nd Person > 3rd Person > Proper Noun > Human Common Noun > Animate Common Noun > Inanimate Common Noun For linguistic typologists, universal patterns of the above type are not arbitrary. They are motivated by properties of human cognition and by the fact that language is used for communication. This leads to the following two types of motivations: ⫺ Cognitive motivations: parsing, iconicity, economy ⫺ Motivations from the speech situation: discourse, pragmatics The cognitive motivation of parsing is based on the properties of the human parser and the assumption that there are grammatical structures that can be more easily processed than others. Hawkins (1983, 1994, 2004) gives the most impressive processing-based account of implicational universals. One of his efficiency principles, the Minimize Domains Principle, states that the parser prefers shorter processing domains to longer ones in combinatorial and/or dependency relations (Hawkins 2004: 31). The Prenominal Noun Modifier Hierarchy can be explained in terms of this principle. The most extensive discussion of iconicity and economy in linguistic typology is found in Haiman (1978, 1980, 1983, 1985). While iconicity implies a certain isomorphism between a concept and the way in which it is expressed, economy reflects the desire of speakers and hearers to strive for “the least effort” or “to do things in the simplest way”. The two principles of iconicity and economy lead to opposite structural consequences. Iconicity supports overt marking of semantic distinctions, while economy tends to reduce them. From such a perspective, typological universals are seen as the result of a competition between these two motivations.

23. Areal language typology Motivations from discourse have been discussed from various perspectives. A prominent approach understands discourse as the starting point of a grammaticalization cline from which more rigid syntactic and morphological structures are developed (Givo´n 1979: 209). Thus, typological variation in basic word-order patterns (subject/object/verb) are derived from topic structures. Similarly, nominative/accusative patterns versus ergative/absolutive patterns are motivated by the properties of topic and focus structures, respectively (DuBois 1987). For usage-based approaches, grammar is crucially related to frequency and relevance in discourse (cf. the contributions in Bybee and Hopper 2001). Finally, pragmatic inference is seen as an important factor in various approaches to grammaticalization (Bybee et al. 1994; Hopper and Traugott 2003).

2.2. What about social actors?  Integrative Functionalism The above description of linguistic typology does not say anything about social factors. And indeed, there does not seem to be any space for social factors at first glance if typology is about semantic structures and patterns of their morphosyntactic expression. But this is not the whole story. Social factors play a considerable role if one looks at the geographical diffusion of linguistic structures. How social factors matter for typology is shown by Integrative Functionalism as defined by Croft (1995, 2000). Integrative Functionalism is one of the three types of linguistic theories discussed by Croft (1995, 2000). The other two are Formal Linguistics and External Functionalism. The two criteria for distinguishing these approaches are (i) the assumption of an innate syntax-oriented Universal Grammar and (ii) the assumption that grammar is self-contained. The following list provides a survey of the three types of theories and the assumptions that define them: ⫺ Formal Linguistics:

Existence of an innate syntax-oriented language capacity (Universal Grammar, UG) ⫺ External Functionalism: Nonexistence of an innate UG, but syntax and other aspects of grammar are self-contained ⫺ Integrative Functionalism: Syntax and other aspects of grammar are not self-contained, they are open to language-external factors. Formal Linguistics claims the existence of an innate Universal Grammar that is selfcontained. External Functionalism denies innateness and has a broader concept of grammar which goes beyond syntax. What it shares with Formal Linguistics is the assumption that grammar is in some way self-contained. A typical representative of External Functionalism is Haiman (1983) for whom grammar can be seen as the outcome of a selfcontained system of iconicity and economy (cf. section 2.1). The theoretical approach that is crucial for dealing with areality, i. e., Integrative Functionalism, assumes neither the existence of Universal Grammar nor the self-containedness of grammar ⫺ grammar is open to social factors. Integrative Functionalism is an evolutionary account that models the successful diffusion (“propagation” in terms of Croft 2000) of structural innovations in a language. The environment that is responsible for the selection and propagation of linguistic structures is formed by the individual members of a language community who bring along their cognitive equipment and their social status in a given

423

424

IV. Structure and dynamics across language spaces speech situation. If this is true, grammatical structures as we find them in a language are not only the result of the cognitive properties of the human brain (cf. parsing, iconicity, economy in section 2.1), they are also due to social factors. Croft (2000) even claims that social factors alone are responsible for the propagation of innovations. In my view, this claim disregards the fact that the environment responsible for selection is not primary, but secondary. At a primary level, evolution produced humans and their brains in interaction with the species-external environment. At a secondary level, the very same human brains with their cognitive structure become a secondary environment which selects linguistic structures. Thus, the cognitive properties of this brain matter in addition to the social environment (Bisang 2001, 2004, 2006a). However, social factors may ultimately be stronger than cognitive factors. Thus, even a cognitively odd structure can win if there are social factors that support it. The social factors and their manifestation in the areal distribution of linguistic types will be described in the following section.

3. Language change: The social and linguistic dynamics o contact-induced convergence and areality 3.1. Social models o language change This section starts with a general account of the interdependence of social factors and contact-induced structural change as presented by Thomason and Kaufman (1988) and Thomason (2001). The second part of this section will briefly deal with approaches that take the perspective of individual actors, their role within their social environment and the structural consequences of their behavior. The approaches to be discussed are the social-network model with its distinction between weak and strong ties (Milroy and Milroy 1985), the relevance of leaders of language change (Labov 2001) and the invisiblehand model (Keller 1990, 1994). Thomason and Kaufman (1988: 50) distinguish between two fundamental processes of language change, i. e., maintenance (borrowing) and shift, which are triggered by general social factors such as duration of contact, number of bilingual speakers, population size and socioeconomic dominance. In the case of maintenance, native speakers of a language A keep their language but they integrate elements from another language, the target language B, into it. Under low intensity of contact and if the number of bilingual speakers in a given population is small, borrowing will be limited to non-basic vocabulary. With increasing contact and with an increasing number of bilinguals, there will be more lexical borrowing and even moderate or heavy structural borrowing. In the case of language shift, speakers of A abandon their language in favor of the target language B. The social factors determining its structural outcome are socioeconomic and political relations of domination, population size and the degree to which the target language B is accessible for speakers of language A. The effect of the first factor on language shift is straightforward ⫺ the dominating culture of B speakers promotes the disappearance of language A in favor of B. Typical processes in modern history that bring about shift in this context are urbanization and industrialization (Thomason 2001: 22). If the number of A speakers is significantly smaller than the number of B speakers, the target language B will be adopted perfectly without leaving any traces of A. The

23. Areal language typology percentage of bilinguals in a population of shifters is relevant because it determines the degree of imperfect learning, i. e., the extent to which structural properties of A can be found in B. If the number of shifters from A to B is large and the number of bilingual speakers of A and B is relatively small the new B variety emerging from that situation will preserve certain properties of A. In the case of maintenance with very intensive contact and in the case of shift with a large shifting group and imperfect learning, the structural consequences seem to be basically the same. Otherwise, maintenance and borrowing produce different structural effects. Maintenance is primarily associated with lexical change, while shift primarily triggers changes in phonology and syntax. No matter what the structural changes are, Thomason (2001: 21) clearly points out “that social factors are the only ones that need to be considered in assessing stability: linguistic factors (such as overall structural similarity of the languages in contact) seem to be totally irrelevant”. In the social-network model, actors are graphically represented as points, while the relations between them are marked by lines. Of particular importance for the diffusion of changes is the quality of the relations between the actors in terms of weak and strong ties (Granovetter 1973). Strong ties are defined by frequent and reciprocal contacts for a number of different reasons (e.g., sports, friendship, shopping, child care, etc.). Weak ties are based on only one reason, they are not necessarily reciprocal and they are more frequent because they can be established more easily. Linguistic innovations first spread between actors who are involved in a number of weak networks and serve as liaisons. As soon as they are disseminated in a large enough number of weak networks, they will be integrated into strong networks by local influentials, i. e., individuals who have a special status in their community and want to keep it by taking up innovations. Labov (2001) points out that the network model is not sufficient for a full understanding of how innovations spread within a community. In his view, it is necessary to look at the leaders of language change. As he showed in his extensive sociolinguistic study of the city of Philadelphia, leaders of language change have the following properties: (4) Properties of the leaders of language change (Labov 2001: 360): a. The leaders are women; men play no significant role. b. The highest concentration of leaders is in the groups centrally located in the socioeconomic hierarchy; that is, leadership forms a curvilinear pattern. c. The leaders are people with intimate contacts throughout their local groups, who influence first people most like themselves. d. The leaders are people who are not limited to their local networks, but have intimate friends in the wider neighborhood. e. These wider contacts include people of different social statuses, so that influence spreads downward and upward from the central group. It is certainly true that the discovery and the description of individuals who are more important than others in the diffusion of linguistic innovations is vital for understanding the social mechanisms of language change. The question that remains is whether the properties presented by Labov (2001) are universal or whether they are subject to cultural variation. For instance, one may wonder whether women always play a more central role than men. The invisible-hand model deals with processes of purposeful individual actions at the microlevel and the emergence of new structures at a macrolevel. Keller (1990, 1994) has

425

426

IV. Structure and dynamics across language spaces applied this model to language change. On a microlevel, individual speakers act according to certain maxims in order to reach certain purposes. If a large number of individuals follow at least partially similar purposes, the causal consequence of these individual purposeful actions on a macrolevel is a new phenomenon which cannot be predicted from the purposes and the maxim(s) involved. Linguistic structures and their change are such invisible-hand phenomena. There are different possible structures for communicating a certain concept in a given speech situation and the speaker selects the most suitable structure according to certain maxims such as the following: (5) LePage and Tabouret-Keller (1985): Talk in such a way that you are recognized as a member of the group you wish to identify with. Maxims of the above type are of considerable importance for contact-induced change (cf. section 4). Depending on the group a speaker wants to be associated with, s/he will select structures from one or the other of the languages involved. If enough speakers of a community share the same behavior this will have its impact on whether language change rather tends towards maintenance or towards shift.

3.2. Mechanisms o contact-induced change The integration of linguistic structures of one language into another is the result of various mechanisms which may operate individually or in combination. Thomason (2001: 129⫺156) discusses code-switching, code alternation, passive familiarity, negotiation, second-language acquisition strategies, bilingual first-language acquisition and deliberate decision. The present subsection will be limited to code-switching and language acquisition strategies. Since many mechanisms have hardly been taken into consideration in linguistics, the following quotation from Thomason (2001) describes the current situation very adequately: A full account of the specific processes through which foreign material gets into a language would require attention to innumerable social and psychological details, and linguists are nowhere near any comprehensive understanding of all the relevant processes. (Thomason 2001: 129)

Code-switching is discussed in numerous publications under various definitions (for a survey, cf. Muysken 1995). The following definition given by Thomason (2001: 132) is relatively general: “Code switching is the use of material from two (or more) languages by a single speaker in the same conversation”. In terms of Muysken (2000: 1), codeswitching is “the rapid succession of several languages in a speech event”. A more specific definition in terms of the Matrix Language-Frame Model describes code-switching as “the selection by bilinguals or multilinguals of forms from an embedded variety (or varieties) in utterances of a matrix variety during the same conversation” (Myers-Scotton 1993: 3; for a later model cf. Myers-Scotton and Jake 2000). Whatever its exact definition may be, code-switching does not seem to be a necessary element of contact-induced change although it is an important one. There are at least

23. Areal language typology two reasons for this (Thomason 2001: 132⫺134): The first reason is that code-switching is not observed in every bilingual speech community, although contact-induced change takes place. The second reason is that the assumption that there is always a development from code-switching to maintenance/borrowing is problematic. The difference between the cognitive processes involved in the use of a code-switched word and a stable loan word is hardly understood. Given the fuzzy borderline between these two types of words it is often hard to decide whether a word of language A in language B is due to interference or code-switching. Language acquisition plays a considerable role in contact-induced change. The role of bilingual speakers has been discussed in subsection 3.1. Silva-Corvala´n (1994), who analyzed the speech of English/Spanish bilinguals in Los Angeles, discusses the following structural consequences: (6) Structural consequences of bilingual first-language acquisition (Silva-Corvala´n 1994: 6): a. Simplification of grammatical categories and lexical oppositions b. Overgeneralization of forms, frequently following a regularizing pattern c. Development of periphrastic constructions either to achieve paradigmatic regularity or to replace less semantically transparent bound morphemes d. Direct and indirect transfer of forms from the superordinate language e. Code-switching Each of these structural consequences is a result of Silva-Corvala´n’s overall hypothesis “that, in language contact situations, bilinguals develop strategies aimed at lightening the cognitive load of having to remember and use two different linguistic systems” (SilvaCorvala´n 1994: 6). In addition to first-language acquisition, adult second-language learners use a number of specific strategies that can have an impact on language structure under certain social conditions such as the ones underlying language shift (see subsection 3.2). The following strategies are listed by Thomason (2001: 146⫺148): Second-language learners insert material from their mother tongue into the target language. They transfer structures from their first language into the second language as in the case of speakers of English who use SVO word order in German subordinate clauses. And finally, they do not acquire distinctions which do not exist in their mother tongue.

3.3. The problems o deining a linguistic area and the relevance o zones o convergence The concept of a linguistic area (Sprachbund) was coined by Trubetzkoy in 1928 (published in Trubetzkoy 1930) and has experienced a lot of different and often divergent definitions since that time. Generally speaking, linguistic areas are characterized by structural similarities across a set of geographically adjacent languages which are not due to genetic relatedness. The following definition by Thomason (2001) is relatively neutral and covers various other definitions (for a survey and discussion of definitions, cf. Stolz 2002):

427

428

IV. Structure and dynamics across language spaces A linguistic area is a geographical region containing a group of three or more languages that share some structural features as a result of contact rather than as a result of accident or inheritance from a common ancestor. (Thomason 2001: 99)

Since the presentation of the Balkan Sprachbund by Sandfeld (1930), several other linguistic areas have been discussed, among them the Circum-Baltic area (Dahl and Koptjevskaja-Tamm 2001), the Ethiopian area (Tosco 1994, 2000; Bisang 2006b), South Asia (Masica 1976) and East and mainland Southeast Asia (cf. Bisang 1996). In Bisang (2006b), I have tried to show the systematic impossibility of providing an exact definition of a linguistic area because such a definition would have to be based either on arbitrary criteria or on criteria that are difficult to apply. The most straightforward instances of arbitrariness have to do with quantification (for other criteria, cf. Bisang 2006b: 87⫺88). Thus, some definitions restrict the number of languages or language families involved in a linguistic area, others opt for a minimal number of features. Such criteria are meaningless as long as one cannot prove that there is a significant correlation between the exact quantification of languages, families or features and either a certain linguistic structure or a certain social pattern present in several linguistic areas. An example of a criterion that is difficult to apply is exact cross-linguistic structural identity, another one is social symmetry. The criterion of exact identity does not only raise practical problems, it also gets right to the heart of formal as well as functional theories. In my view, none of these theoretical frameworks provides a safe ground for the exact evaluation of the comparability of languages. A brief look at the diametrically divergent approaches of Chomskyan Formal Linguistics and Croft’s (2001) Radical Construction Grammar may give an idea of why there is no basis for determining exact identity (also see Bisang 2006b). Formal approaches deductively stipulate universal categories and relations which are reflected in the patterns of syntactic distribution across the world’s languages. As a consequence, structures that look identical at first glance may be generated by different properties of Universal Grammar. In most cases, the data on situations of language contact are not detailed enough to show whether structures from two languages are exactly identical. In addition, Croft (2001: 34) argues that “[p]ositing universal categories would imply identical behaviour across languages, which is empirically false”. In his view, categories and relations are language particular. If that is true, cross-linguistic comparability is only possible across constructions and the markedness patterns that can be derived from cross-linguistic analysis. Thus, a linguist who tries to prove the exact identity of two structures from two different languages will find her/ himself in the unbridgeable gulf between stipulated universal categories (claimed to be false by Croft 2001) and the comparison of markedness patterns within constructions (which are only surface phenomena that may have a divergent universal explanation). The criterion of social symmetry was brought about by Aikhenvald and Dixon (2001), who claim that linguistic areas emerge exclusively out of contact with multilateral diffusion without any relationship of dominance. As was pointed out by Dahl (2001), there is almost no case of language contact without at least some social asymmetries. Thus, Aikhenvald and Dixon’s (2001) claim is difficult to apply to real situations of contact. The above problems with the exact definition of a linguistic area do not exclude the relevance of contact-induced change within a certain geographic area. The question is whether it makes sense to strive for a rigid definition of these phenomena in terms of a linguistic area. In fact, there may be two prospects for why finding such a definition could be theoretically relevant (Bisang 2006b):

23. Areal language typology i. ii.

The prospect of drawing further typological conclusions from the existence of a linguistic area. “If x is a linguistic area it must have certain structural properties”. The prospect of finding some concrete conclusions concerning social and/or historical properties of a speech community. “If x is a linguistic area, this is due to the social or historical facts x, y and z”.

Since the features that are discussed for each individual linguistic area are extremely divergent, it seems rather unlikely to find further typological conclusions in the sense of prospect (i). There are only some rather general observations such as the preference of periphrastic structures over morphological structures. However, such generalizations are too general to be associated with the degree of homogeneity associated with linguistic areas. Prospect (ii) applies at a more general level such as maintenance vs. shift (see subsection 3.1). But since the correlation between phenomena of structural convergence and their concrete historical or social background is normally rather general and abstract, too clear-cut a definition does not look promising for finding potentially existing correlations. Since a rigid definition of a linguistic area is unlikely to lead to concrete typological generalizations and only allows rather general conclusions about the interaction of social/historical facts with properties of linguistic structures, I argue in Bisang (2006b) that a less rigid concept in terms of zones of convergence is needed in linguistic typology. This concept explicitly tries to avoid unnecessary reification. Thus, zones of convergence are basically geographical areas in which individual languages develop certain structural similarities through contact among their speakers. While a rigid definition of a linguistic area is problematic, a looser concept in terms of zones of convergence is necessary for at least the following two reasons: Zones of convergence can be seen as generators of typologically new or rare structures (cf. the case of Ethiopia presented in subsection 3.4) and they avoid language samples that are statistically biased because of contact-induced convergence among the languages selected (section 4).

3.4. Language contact and the emergence o new or rarely attested types In some cases, typologically rare structures can be accounted for in terms of language contact. This is illustrated by the positions of the genitive and the adjective relative to the head noun in Bayso, a Cushitic OV-language (South Lowland East Cushitic: OmoTana) spoken in the Ethiopian zone of convergence (Banti 1988). According to Greenberg (1966), word order for Gen and Adj in OV languages is subject to the following universal: (7) Greenberg’s Universal 5 (1966: 110, also cf. Hawkins 1994: 320): If a language has dominant SOV order and the genitive follows the governing noun, then the adjective likewise follows the noun. Bayso however has NGen & AdjN, a word order that is excluded for OV languages by (7). Tosco (1994: 416⫺417) convincingly argues that Bayso is situated in a contact zone between two groups of languages which he calls Highland East Cushitic (HEC) and

429

430

IV. Structure and dynamics across language spaces Lowland East Cushitic (LEC). Both groups are OV but they differ with regard to the positions of Adj and Gen: In HEC, Adj and Gen consistently precede the noun, while they consistently follow the noun in LEC. Bayso takes up NGen from LEC and AdjN from HEC. While contact seems to be a good explanation for NGen & AdjN in Bayso, this does not apply to the same word order in Tigre (Ethiosemitic). As Tosco (1998) shows, the situation in Tigre can be accounted for in terms of Hawkins (1994) and his performance theory. Tigre has a preposed clitic article which produces optimal constituent-recognition domains with NGen & AdjN. The lesson to be learnt from Bayso and Tigre is that there are different motivations for rare structural patterns. They can be due to languageexternal processes (Bayso) and to language-internal processes (Tigre) (and probably to both processes in mutual interaction). The reason why they are not attested more frequently may thus be a coincidence. Given the right socio-historical constellation, they might have the potential for cross-linguistic diffusion and for becoming statistically unmarked.

3.5. Language contact and grammatical properties speciic to certain areas  an example rom the Sinitic languages Sinitic languages such as Mandarin Chinese are well-known for the asymmetry between their basic VO word order and the head-final structure of their noun phrases. In Dryer’s (1992: 86) extensive data sample of languages, Mandarin is the only instance with VO and RelN, while VO & NRel is attested in 60 genera. In spite of the typological rarity of this word order pattern, it is diachronically very stable. In the literature, the extraordinary situation in Chinese has been explained in two ways, either by claiming that Chinese is SOV and thus naturally takes the relative clause before its head noun or that the RelN sequence is due to contact with Altaic SOV languages. As will be briefly shown, none of these arguments is convincing. Li and Thompson (1975) argue that Chinese is SOV. Their claim is based on the baconstruction in which the object is moved in front of the noun. However, a look at the frequency of this construction reveals that it is not at all frequent enough (Sun and Givo´n 1985) to support the hypothesis of a change from SVO to SOV. The explanation of SVO & RelN in terms of contact with Altaic goes back to Hashimoto (1976). More recently, Dryer (2003, 2008) has taken up the same argument. In his view, Mandarin Chinese has preserved the OV word-order properties of proto-SinoTibetan through the influence of the Altaic languages spoken in the North. The problem with this explanation is that the oldest accessible stage of Chinese represented in the oracle bone inscriptions (fourteenth to eleventh centuries b.c.) is clearly SVO (Djamouri 2001). Thus, it is very unlikely that Chinese word order types such as GenN and RelN are the remnants of an erstwhile consistently head final stage in proto-Sinitic. Whatever the reasons for the emergence of that situation in Chinese may be, its typological asymmetry is characterized by long-time stability ⫺ a time stability which may well be due to social factors. This can be inferred from the fact that not all Sinitic languages were equally resistant to contact-induced change. An important factor that supported the resistance against word-order change may have been the influence of classical Chinese as a prestigious written standard that prevented change towards a typologically unmarked word order pattern.

23. Areal language typology

431

3.6. Spread zones and residual zones Nichols (1992: 13⫺24, 192⫺195) makes a distinction between spread zones and residual zones, each of them referring to areas of subcontinental size which are characterized by relative diversity, center, periphery and internal stability. The two types of zones are not understood as an exhaustive classification of areality defined by the above criteria, but they are “two important and common types of areas” (Nichols 1992: 13) and they form opposition pairs with regard to a number of features (see Table 23.2).

Tab. 23.2: The seven properties of spread zones vs. residual zones

(1) (2) (3) (4) (5) (6)

(7)

Spread zones (Nichols 1992: 16)

Residual zones (Nichols 1992: 21)

Little genetic diversity Low structural diversity The language families present in the spread zones are shallow Rapid spread of languages or language families and consequent language succession Classic dialect-geographical area with innovating center and conservative periphery No net long-term increase in diversity. A spread zone is a long-lasting phenomenon, but it preserves little linguistic evidence of its history The spreading language serves as a lingua franca for the entire area or a large part of it

High genetic diversity High structural diversity The language families, or at least a good number of them, are deep No appreciable spread of languages or families. No language succession No clear center of innovation. Accretion of languages and long-term net increase in diversity. Language isolates and isolate families are likely to be found in residual zones No lingua franca

Typical spread zones are western Europe, central Australia, interior North America, Mesoamerica, the Ancient Near East and central insular Oceania. The residual zones analyzed by Nichols (1992) are Ethiopia and Kenya (as a part of eastern Africa), the Caucasus, the Pacific Coast of Northern Asia, northern Australia and California (representing the Pacific Coast of North America). Even though she found no simple correlation between residual areas and high complexity, or spread zones and low complexity, there was a correlation between high-complexity languages and their occurrence in areas characterized by contact and by high linguistic diversity. Since residual zones are characterized by diversity, “we can expect the languages in residual zones to exceed the averages of their continents in complexity” (Nichols 1992: 193). However, if there are enough contact situations in a spread area that spread area can also score high in complexity. The situations in which this is possible in a spread zone have the following properties: [W]hen the center shifts from group to group so that the spread is not confined to a single language, or when a political or cultural spread is of recent formation so that diverse indigenous languages still survive, we have a spread situation within which there can nonetheless still be contact among diverse languages. (Nichols 1992: 193)

432

IV. Structure and dynamics across language spaces Structural complexity is not only related to differences in contact density within residual zones and spread zones, it also shows up in extreme situations of contact that lead to the development of Creole languages. As is shown in the next subsection, contact triggers a reduction of complexity in contact situations of that type.

3.7. Structural complexity and language contact  the case o Creoles Creole languages are not primarily associated with areality but they are the product of language contact and they are the source for various claims concerning the universal properties of human language. The present subsection will concentrate on basic structural properties of Creoles which can be accounted for in terms of contact and it will conclude with an example of areality. Creole languages differ from “ordinary” languages by the fact that “they came into existence at some point in time” (Muysken and Smith 1997: 4). They roughly are the result of an extreme contact situation in which speakers of different mutually unfamiliar languages are forced to communicate with each other. In such a situation, speakers tend to avoid categories and structures that go beyond what is needed for maintaining a basic conversation. This is the point of departure for McWhorter’s (2001, 2005) claim of the relative lack of complexity in Creole languages: Creoles, in being recently borne of communication vehicles deliberately designed to eschew all but the functionally central (pidgins), are unique examples of natural languages with much less contingent accumulation of “ornamental” elaboration than older grammars drag along with them. (McWhorter 2005: 43)

Many grammatical categories and grammatical distinctions attested in the world’s languages are the result of a long period of development within a more or less stable speech community. Typical phenomena of a later development are inflectional affixation, tone and derivational noncompositionality (McWhorter 2001; 2005: 10). This situation leads McWhorter (2001, 2005) to the conclusion that Creole languages are generally less complex than ordinary languages on the basis of the following definition: (8) Complexity (McWhorter 2005: 45): [A]n area of grammar is more complex than the same area in another grammar to the extent that it encompasses more overt distinctions and/or rules than another grammar. McWhorter (2005: 45⫺46) further specifies complexity in terms of the following four diagnostics: (9) Four diagnostics of complexity (McWhorter 2005: 45⫺46): a. A phonemic inventory is more complex to the extent that it has more marked members. […] [M]arked phonemes are those encountered less frequently in the world’s languages than others conventionally deemed unmarked […]. b. A syntax is more complex than another to the extent that it requires the processing of more rules, such as asymmetries between matrix and subordinate clauses.

23. Areal language typology c. A grammar is more complex than another to the extent that it gives overt and grammaticalized expression to more fine-grained semantic and/or pragmatic distinctions than another. d. Inflectional morphology renders a grammar more complex than another one in most cases. As a consequence of these diagnostics, many grammatical categories that are found in Non-Creole languages are not attested in Creoles. Thus, none of the 19 Creole languages analyzed by McWhorter (2001) had any of the following categories: ergativity, grammaticalized evidential marking, inalienable possessive marking, switch-reference marking, inverse marking, obviative marking, “dummy” verbs, syntactic asymmetries between matrix and subordinate clauses, grammaticalized subjunctive marking, verb-second, clitic movement, any pragmatically neutral word order but SVO, noun class or grammatical gender marking (analytic or affixal) or lexically contrastive morphosyntactic tone beyond a few isolated cases (McWhorter 2001: 163). If the presence of these categories in a language is the result of a historical process their explanation in terms of a combination of cognitive and social factors seems to be more straightforward than an account in terms of Universal Grammar. A particularly interesting case are serial verb constructions, which are claimed to belong to the inventory of the Language Bioprogram by Bickerton (1981). As McWhorter (1997: 39) convincingly shows, extensive use of serial verb constructions “have appeared around the world in precisely the creoles which had serializing substrata”. Serial verb constructions are also an areal property of Kwa/Nigerian languages. Thus, a successful areal structure made its way into another group of languages that contributed to its further diffusion.

3.8. The World Atlas o Language Structures (WALS) On the basis of a sample of 150 to 400 languages (with a core sample of 200 languages), Haspelmath et al. (2005) have mapped the geographic distribution of 140 structural variables in the languages of the world. With this pioneering work it has become possible to analyze and to visualize completely new sets of areal clusters of grammatical structures. In spite of its enormous advantages, one should keep in mind that WALS does not fully reflect areality. It only represents a selection of languages. Thus, areas which look homogeneous with regard to the distribution of a certain feature may turn out to be much more complex when more languages from that area are taken into account. Nevertheless, collecting data in this way is the only way to find out more about the geographic diffusion of patterns, about large linguistic areas and about structural properties which are extremely time-stable (also cf. Bisang 2004: 30).

4. Consequences or Linguistic Typology and its generalizations The robustness of typological generalizations crucially depends on the appropriate assessment of areality (Bisang 2004, 2006a). The reason for this is pointed out very clearly in a programmatic paper by Bickel (2007: 241): “[T]he observable distributions are sub-

433

434

IV. Structure and dynamics across language spaces stantially influenced by population history, and this makes it fundamentally problematic to try and predict them by principles of grammar alone” (Bickel 2007: 241). This raises the question of how linguistic typology integrates social factors into its sampling methods and how reliable these methods are. Since Greenberg’s (1966) seminal paper, establishing a language sample involves the two classical factors of genetic relatedness and geographical closeness of the languages selected. The relevance of genetic relatedness is based on the assumption that there is a certain structural homogeneity in languages that share a common ancestor. Geographic closeness matters because it sets the preconditions for contact-induced convergence. Rijkhoff et al. (1993) and Rijkhoff and Bakker (1998) discuss two types of samples depending on the questions a typological research project wants to address: ⫺ Variety samples aim at discovering the full range of possible variation in the world’s languages. ⫺ Probability samples aim at discovering correlation pairs (e.g., Dryer’s [1992] research on typological features that covary with the position of V and O relative to each other; cf. below). Rijkhoff et al. (1993) and Rijkhoff and Bakker (1998) are only interested in the criteria for setting up a variety sample. Although these authors do not deny the relevance of geographic factors they only integrate genetic factors into their model. They argue that migration and variable geographic distances between languages are difficult to quantify adequately, while there is a certain correlation between the genetic tree structure of a language family and the extent of diversity across its member languages. Based on the number of intermediary layers between the head node of a family and the final node that represents the level of the individual languages, they calculate a diversity value for each of the 17 families of the world distinguished by Ruhlen (1987). The diversity values are then integrated into a matrix which shows how many grammars need to be selected from the 17 families as well as from Pidgin and Creole languages, language isolates and unclassified languages for constructing a sample of x languages, whereby x has to be minimally 30. Dryer (1989, 1992) presents the most elaborate method of constructing a probability sample. The aim of his study is to find word-order patterns that covary with the position of V and O. An instance of such a covariation pair is the position of the relative clause relative to its head noun: The position of N covaries with V and the position of Rel covaries with O. Dryer’s (1989, 1992) basic units are not individual languages but genera, i. e., groups of languages which can be roughly compared to subfamilies of Indo-European (e.g., Germanic, Indo-Aryan, Indo-Iranian, Romance, etc.). Dryer (1992) lists 252 genera. As soon as a particular word-order type is attested in at least one language of a genus, that genus will get the value 1 for that type. To avoid the impact of areality, Dryer (1992) divides the 252 genera into the following six large linguistic areas of continental size: (i) Africa, (ii) Eurasia, (iii) Southeast Asia and Oceania, (iv) Australia and New Guinea, (v) North America and (vi) South America. With this division it is possible to see whether a certain type is a world-wide phenomenon or whether it is local. A pair of parameters is said to be a correlation pair of VO/OV and thus of typological relevance only if the two values covary equally through all six large areas. Considerations of genetic relatedness and geographic closeness are necessary prerequisites for discovering significant typological generalizations but they are not sufficient.

23. Areal language typology As pointed out by Maslova (2000: 307), these methods do not allow to “distinguish between general distributional universals and accidental statistical properties of the current language population”. The two methods discussed are reliable only if the linguistic types have reached stationary distribution, i. e., a random distribution which is not biased by former typological properties of a given language or family. The factors that determine the geographic distribution of linguistic types are language death and language birth. Since there is an enormous lack of information concerning these two factors and the types involved in them, it is simply impossible to say whether the global distribution of types has reached a statistically neutral stage. To overcome this problem, Maslova (2000: 328⫺329) suggests a method which integrates estimates of transition probability, i. e., the probability that a language shifts from one type (Ti) to another type (Tj) within a given period of time: It is clear […] that if we want to estimate the probability of a shift Ti J Tj we must compare two quantities: the number of languages which have undergone this shift and the number of languages which have retained type Ti or shifted to another type within the same time interval. Accordingly, in order to estimate a transition probability p (Ti J Tj) for some time interval t, one would need a sample of languages which can be assumed to have been in a state Ti t years ago. The current frequency of type Tj in this sample would give an estimate (Maslova 2000: 329) of the transition probability p (Ti J Tj).

This method is based on the hypothesis that type shift exclusively depends on language internal factors ⫺ a hypothesis that seems to be implicitly or explicitly supported by a large number of typologists. The following quotation from Nichols (1992) is an explicit commitment to that hypothesis: The rate of spread is determined by nonlinguistic factors such as migrations, political or economic power, etc. These things cannot figure in a linguistic model, except as unknowns. But the rate at which structural features change is a purely linguistic matter and does figure in a linguistic model. (Nichols 1992: 209)

From the perspective of Integrative Functionalism, there is no reason why the above hypothesis should hold. In situations of contact, the question of whether a speech community adopts a new type may be due to social factors such as the maxim of language use presented in (5). If type shifts depend on social factors, the time interval within which they take place is arbitrary in the sense that there is no cognitive or discourse motivation which may allow generalizations about what tends to be a short-term shift and what tends to be a long-term shift (cf. Bisang 2004: 34⫺35). If this is true the possibility of calculating fixed transition probabilities from a type A to a type B becomes highly questionable. Within this framework, the most reliable conclusion may be drawn from looking at those structures that turn out to be diachronically the least stable ones within a sample of languages. But even in these cases, typologists will have to cope with the problem that the number of written sources that document language change only reflects a very small fraction of all the changes that must have taken place in the past. In such a situation, there seem to be natural limits for typologists to assess the impact of social factors on the global distribution of linguistic types.

435

436

IV. Structure and dynamics across language spaces

5. Conclusion and outlook Since a fully reliable method that dissociates cognitive from social factors seems to be impossible, there will always be some degree of uncertainty about the validity of typological generalizations. In spite of this, I share Newmeyer’s (1998: 364) optimism that it is the best strategy “to assume, for want of evidence to the contrary, that the more robust-seeming of them [i. e., linguistic generalizations; W. B.] are valid”. Contact-induced change and areality do not only make the assessment of typological generalizations more complex, they also have important consequences concerning the Saussurean view that grammar is arbitrary. As was pointed out by Newmeyer (2005: 182⫺191), language change in general and contact-induced change in particular support holistic functionalism against atomistic formalism. Atomistic functionalism is based on the assumption that “[t]here is direct linkage between properties of particular grammars and functional motivations for those properties” (Newmeyer 2005: 174), while holistic functionalism does not assume any direct linkage. Since there is considerable evidence from studies on language change that “the forces […] that bring a construction into a language are not necessarily the same ones that keep it there” (Newmeyer 2005: 185), structures that might have been motivated at an initial stage can become unmotivated at a later one. In that sense, findings from language change and areality clearly support holistic functionalism, which makes the discussion of arbitrariness take on a new dimension. If arbitrariness is generally attributed to properties of Universal Grammar in formal linguistics, the assumption of innateness loses at least some of its relevance if arbitrariness can be due to social factors operating in situations of language change. From that perspective, areal typology does not only matter to functional typology, it also has its consequences for formal linguistics.

6. Reerences Aikhenvald, Alexandra Y. and R. M. W. Dixon (eds.) 2001 Areal Diffusion and Genetic Inheritance: Problems in Comparative Linguistics. Oxford: Oxford University Press. Banti, Giorgio 1988 Adjectives in East Cushitic. In: Marianne Bechhaus-Gerst and Fritz Serzisko (eds.), Papers from the International Symposium on Cushitic and Omotic Languages. Cologne, January 6⫺9, 1986, 205⫺259. Hamburg: Buske. Bickel, Balthasar 2007 Typology in the 21st century: Major current developments. Linguistic Typology 11: 239⫺251. Bickerton, Derek 1981 Roots of Language. Ann Arbor: Karoma. Bisang, Walter 1996 Areal typology and grammaticalization: Processes of grammaticalization based on nouns and verbs in East and mainland South East Asian languages. Studies in Language 20: 519⫺597. Bisang, Walter 2001 Areality, grammaticalization and language typology. On the explanatory power of functional criteria and the status of Universal Grammar. In: Walter Bisang (ed.), Language Typology and Universals, 175⫺223. Berlin: Akademie-Verlag.

23. Areal language typology Bisang, Walter 2004 Dialectology and typology⫺an integrative perspective. In: Bernd Kortmann (ed.), Dialectology Meets Typology: Dialect Grammar from a Cross-Linguistic Perspective, 11⫺45. Berlin/New York: Mouton de Gruyter. Bisang, Walter 2006a Contact-induced convergence: Typology and areality. In: Keith Brown (ed.), Encyclopedia of Language and Linguistics, 2nd ed., vol. 3, 88⫺101. Oxford: Elsevier. Bisang, Walter 2006b Linguistic areas, language contact and typology: Some implications from the case of Ethiopia as a linguistic area. In: Yaron Matras, April McMahon and Nigel Vincent (eds.), Linguistic Areas: Convergence in Historical and Typological Perspective, 75⫺98. Hampshire: Palgrave MacMillan. Bybee, Joan and Paul J. Hopper (eds.) 2001 Frequency and the Emergence of Linguistic Structure. Amsterdam/Philadelphia: Benjamins. Bybee, Joan, Revere Perkins and William Pagliuca 1994 The Evolution of Grammar: Tense, Aspect and Modality in the Languages of the World. Chicago/London: The University of Chicago Press. Croft, William A. 1995 Autonomy and functionalist linguistics. Language 71: 490⫺532. Croft, William A. 2000 Explaining Language Change: An Evolutionary Approach. Essex: Pearson Education. Croft, William A. 2001 Radical Construction Grammar: Syntactic Theory in Typological Perspective. Oxford: Oxford University Press. Croft, William A. 2003 Typology and Universals. Cambridge: Cambridge University Press. Dahl, Östen 2001 Principles of areal typology. In: Martin Haspelmath, Ekkehard König, Wulf Oesterreicher and Wolfgang Raible (eds.), Language Typology and Language Universals, vol. 2, 1456⫺1470. (Handbücher zur Sprach- und Kommunikationswissenschaft, 20.2) Berlin/ New York: Mouton de Gruyter. Dahl, Östen and Maria Koptjevskaja-Tamm 2001 Circum-Baltic Languages, vol. 1: Past and Present, vol. 2: Grammar and Typology. Amsterdam/Philadelphia: Benjamins. Dixon, R. M. W. 1979 Ergativity. Language 55: 59⫺138. Djamouri, Redouane 2001 Markers of predication in Shang bone inscriptions. In: Hilary Chappell (ed.), Sinitic Grammar: Synchronic and Diachronic Perspectives, 143⫺171. Oxford: Oxford University Press. Dryer, Matthew S. 1989 Large linguistic areas and language sampling. Studies in Language 13: 257⫺292. Dryer, Matthew S. 1992 The Greenbergian word order correlations. Language 68: 81⫺138. Dryer, Matthew S. 2003 Word order in Sino-Tibetan languages from a typological and geographical perspective. In: Graham Thurgood and Randy LaPolla (eds.), The Sino-Tibetan Languages, 43⫺55. London: Routledge. Dryer, Matthew S. 2008 Word order in Tibeto-Burman languages. Linguistics of the Tibeto-Burman Area 31: 1⫺84.

437

438

IV. Structure and dynamics across language spaces DuBois, John 1987 The discourse basis of ergativity. Language 11: 365⫺399. Givo´n, Talmy 1979 On Understanding Grammar. New York: Academic Press. Granovetter, M. S. 1973 The strength of weak ties. American Journal of Sociology 78: 1360⫺1380. Greenberg, Joseph H. 1966 Some universals of grammar with particular reference to the order of meaningful elements. In: Joseph H. Greenberg (ed.), Universals of Language, 73⫺113. Cambridge, MA/ London: MIT Press. Haiman, John 1978 Conditionals are topics. Language 54: 564⫺589. Haiman, John 1980 The iconicity of grammar: Isomorphism and motivation. Language 56: 515⫺540. Haiman, John 1983 Iconic and economic motivation. Language 59: 781⫺819. Haiman, John 1985 Iconicity in Syntax. Amsterdam/Philadelphia: Benjamins. Hashimoto, Mantaroo 1976 Language diffusion on the Asian continent: Problems of typological diversity in SinoTibetan. Computational Analyses of Asian and African Languages 3: 49⫺65. Haspelmath, Martin, Matthew S. Dryer, David Gil and Bernard Comrie (eds.) 2005 The World Atlas of Language Structures (WALS). Oxford: Oxford University Press. Hawkins, John A. 1983 Word Order Universals. New York: Academic Press. Hawkins, John A. 1994 A Performance Theory of Order and Constituency. Cambridge: Cambridge University Press. Hawkins, John A. 2004 Efficiency and Complexity in Grammars. Oxford: Oxford University Press. Hopper, Paul J. and Elizabeth C. Traugott 2003 Grammaticalization. 2nd ed. Cambridge: Cambridge University Press. Keller, Rudi 1990 Sprachwandel. Tübingen: Frankfurt. Keller, Rudi 1994 On Language Change: The Invisible Hand in Language. London: Routledge. Labov, William 2001 Principles of Linguistic Change: Social Factors. Oxford: Blackwell. Lehman, Christian 1984 Der Relativsatz: Typologie seiner Strukturen, Theorie seiner Funktionen, Kompendium seiner Grammatik. Tübingen: Narr. LePage, Robert B. and Andre´e Tabouret-Keller 1985 Acts of Identity. Cambridge: Cambridge University Press. Li, Charles N. and Sandra A. Thompson 1975 The semantic function of word order in Chinese. In: Charles N. Li (ed.) Word Order and Word-Order Change, 163⫺195. Austin: University of Texas Press. Masica, Colin P. 1976 Defining a Linguistic Area. Chicago/London: University of Chicago Press. Maslova, Elena 2000 A dynamic approach to the verification of distributional universals. Linguistic Typology 4: 307⫺333.

23. Areal language typology McWhorter, John H. 1997 Towards a New Model of Creole Genesis. New York: Peter Lang. McWhorter, John H. 2001 The world’s simplest grammars are creole grammars. Linguistic Typology 5: 125⫺166. McWhorter, John H. 2005 Defining Creole. Oxford: Oxford University Press. Milroy, James and Lesley Milroy 1985 Linguistic change, social network and speaker innovation. Journal of Linguistics 21: 339⫺384. Muysken, Pieter 1995 Code-Switching and grammatical theory. In: Lesley Milroy and Pieter Muysken (eds.), One Speaker, Two Languages: Cross-Disciplinary Perspectives on Code-Switching, 177⫺ 198. Cambridge: Cambridge University Press. Muysken, Pieter 2000 Bilingual Speech: A Typology of Code-Mixing. Cambridge: Cambridge University Press. Muysken, Pieter and Norval Smith 1997 The study of pidgin and creole languages. In: Jacques Arends, Pieter Muysken and Norval Smith (eds.), Pidgins and Creoles: An Introduction, 3⫺14. Amsterdam/Philadelphia: Benjamins. Myers-Scotton, Carol 1993 Duelling Languages: Grammatical Structure and Codeswitching. Oxford: Clarendon. Myers-Scotton, Carol and Janice L. Jake 2000 Four types of morpheme: Evidence from aphasia, code switching, and second-language acquisition. Linguistics 38: 1053⫺1100. Newmeyer, Frederick J. 1998 Language Form and Language Function. Cambridge, MA: MIT Press. Newmeyer, Frederick J. 2005 Possible and Probable Languages: A Generative Perspective on Linguistic Typology. Oxford: Oxford University Press. Nichols, Johanna 1992 Linguistic Diversity in Space and Time. Chicago/London: The University of Chicago Press. Rijkhoff, Jan, Dik Bakker, Kees Hengeveld and Peter Kahrel 1993 A method of language sampling. Studies in Language 17: 169⫺203. Rijkhoff, Jan and Dik Bakker 1998 Language sampling. Linguistic Typology 2: 263⫺314. Rosenbach, Anette 2002 Genitive Variation in English: Conceptual Factors in Synchronic and Diachronic Studies. Berlin/New York: Mouton de Gruyter. Ruhlen, Merrit 1987 A Guide to the World’s Languages, vol. 1, Classification. London: Edward Arnold. Sandfeld, Kristian 1930 Linguistique Balkanique, proble`mes et re´sultats. Paris: Klincksieck. Silva-Corvala´n, Carmen 1994 Language Contact and Change: Spanish in Los Angeles. Oxford: Clarendon Press. Silverstein, Michael 1976 Hierarchy of features and ergativity. In: R. M. W. Dixon (ed.), Grammatical Categories in Australian Languages, 112⫺171. Canberra: Australian Institute of Aboriginal Studies. Stolz, Thomas 2002 No Sprachbund beyond this line! On the age-old discussion of how to define a linguistic area. In: Paolo Ramat and Thomas Stolz (eds.), Mediterranean Languages: Papers from the MEDTYP Workshop, Tirrenia, June 2000. Bochum: Brockmeyer.

439

440

IV. Structure and dynamics across language spaces Sun, Chao-Fen and Talmy Givo´n 1985 On the so-called SOV word order in Mandarin Chinese: A quantified text study and its implications. Language 61: 329⫺351. Thomason, Sarah G. and Terrence Kaufman 1988 Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Thomason, Sarah G. 2001 Language Contact: An Introduction. Edinburgh: Edinburgh University Press. Tomlin, Russell S. 1986 Basic Word Order: Functional Principles. London: Croom Helm. Tosco, Mauro 1994 The historical syntax of East Cushitic: A first sketch. In: Thomas Bearth, Wilhelm J. G. Möhlig, Beat Sottas and Edgar Suter (eds.), Perspektiven afrikanischer Forschung: Beiträge zur Linguistik, Ethnologie, Geschichte, Philosophie und Literatur. X. Afrikanistentag, 415⫺440. Köln: Köppe. Tosco, Mauro 1998 A parsing view on inconsistent word order: Articles in Tigre and its relatives. Linguistic Typology 2: 355⫺380. Tosco, Mauro 2000 Is there an “Ethiopian Language Area”? Anthropological Linguistics 42: 329⫺365. Trubetzkoy, Nikolaj S. 1930 Proposition 16. Actes du Premier Congre`s International de Linguistes a` la Haye, du 10⫺ 15 Avril 1928, 17⫺18. Leiden: Sijthoff.

Walter Bisang, Mainz (Germany)

24. The consequences o migration and colonialism I: Pidgins and creoles 1. 2. 3. 4. 5. 6.

Introduction: Issues of space in the study of “marginal languages” Are creoles a distinct dialectological type? “Creole spaces” in previous research From language through code to symbol: On the recent deterritorialization of some creoles Outlook: Creolophone spaces in the twenty-first century References

1. Introduction: Issues o space in the study o marginal languages The difficulty of placing pidgin and creole languages in geographical space is one of the hidden leitmotifs of pidgin and creole linguistics. An early classic in the field (Reinecke 1937) bears the tellingly ambiguous title Marginal Languages: A Sociological Survey of the Creole Languages and Trade Jargons. If we understand the adjective marginal literally,

24. Pidgins and creoles the places where pidgins and creoles arise are the margins of Empire, the contact zones beyond the confines of regular settler colonialism. But of course it is almost impossible not to think of the many metaphorical meanings of the adjective marginal here: languages at the margins of society, languages of precarious status, which are misrepresented or not recognized in the communities in which they are used, and ⫺ not least ⫺ languages whose status and classification has long been a controversial topic in academic linguistics (cf., e.g., McWhorter 2005; Thomason 1997; Thomason and Kaufman 1988). Before addressing the problem of space in the study of pidgins and creoles in detail, a brief working definition is in order of how these two terms are understood in the present contribution. Pidgins and creoles tend to be lumped together because on one conventional understanding creoles are pidgins which have become natively spoken community languages. Quite apart from the fact that we now tend to assume “degrees of restructuring” (cf., e.g., Neumann-Holzschuh and Schneider 2000) where earlier scholars saw the dramatic working out of the linguistic bio-program in creolization (Bickerton 1981), pidgins and creoles are very different with regard to their spatial distribution. Pidginization of languages is a common, indeed almost ubiquitous phenomenon in spontaneous language-contact situations, whereas creolization is rare and, moreover, as we shall see, largely confined to certain tropical regions with a specific colonial history. There is thus little in common between, say, a presumably transient pidgin such as the type of Gastarbeiterdeutsch ‘guest worker German’ which arose in Germany in the wake of labor migration from the 1950s onwards and a creole language such as Sranan, which arose from a largely English pidgin lexical base, is now the lingua franca of the former Dutch colony Suriname and has a complex history of more than 300 years, including a history of writing that stretches back to the eighteenth century (on which see Arends and Perl 1995). When the term pidgin is used in the present study, reference is thus to the very small proportion of pidgins worldwide which have evolved into creole languages or at least into stabilized and expanded pidgins, such as Tok Pisin (of Papua New Guinea) or Nigerian Pidgin, which ⫺ though not spoken natively by the majority of their habitual users ⫺ have undergone a process of institutionalization and concomitant structural and lexical expansion. How then do this subset of pidgins and the full range of the world’s creole languages pattern with regard to the dimensions of space outlined in the first part of the present Handbook? Maps locating creoles in geographical space (cf. Johnstone in this volume) which are available (e.g., Holm 1988⫺1989, vol. 1: xvi⫺xvii) make clear that with very few exceptions creoles cluster in a circumglobal tropical belt. As for their lexical bases, English, French and Portuguese figure prominently, with other European colonial languages such as Dutch and Spanish following. There are non-European lexifiers (in Holm’s list, for example, for the Nubi Creole Arabic of Uganda or for Hiri Motu, an expanded pidgin of Papua New Guinea), but such cases are generally rare. Another comprehensive and easily accessible source of reference on the global distribution of creoles is compiled in the “Creole” branch in the language-family-trees section of the Summer Institute of Linguistics’ Ethnologue database (http://www.ethnologue.com), which lists 86 creole languages. This is comparable to Holm’s 88 as far as totals are concerned, although there are minor differences of coverage which make a direct comparison difficult. Both lists obviously include central representatives of the Creole type such as Haitian Creole or Jamaican Creole, but differ with regard to how they handle

441

442

IV. Structure and dynamics across language spaces borderline cases. Holm’s list, for example, includes Afrikaans as a semi-creole, whereas the Ethnologue database excludes Afrikaans, but lists two Afrikaans-based creoles, Tsotsitaal and Oorlams. Holm does not list Tsotsitaal in his general survey, but refers to it as “Fly Taal” (1988⫺1989, vol. 2: 350⫺352) in a section devoted to partial creolization of Dutch. As for Hindi, Ethnologue refers to Andaman Creole Hindi, which is not treated by Holm, whereas Holm treats the Pidgin Hindustani of the South Seas (which is absent from the Ethnologue listing). The geographical distribution which becomes apparent in such surveys points to a set of sociohistorical circumstances most conducive to the creolization of languages, namely the massive and frequently forced dislocation of large populations in the wake of the colonial expansion of mainly European powers. This forced migration notoriously took the form of chattel slavery in the plantation colonies of the English, Dutch and French Caribbean, and arrangements which were only marginally more humane than slavery in the Pacific region. It is probably fair to say that the following generalization expressed by Loreto Todd is not just confined to creoles with English as their lexifier: “the more we study English-related pidgins and Creoles, the more they speak to us of the suffering inflicted on one branch of humanity by another” (1984: 251). With regard to movement in space, creoles are languages which owe their existence to the movement of populations, and very frequently they themselves have become languages on the move. This becomes most obvious when looking at the island Caribbean, one of the world’s focal areas of creole genesis in the seventeenth and eighteenth centuries. Here, few members of the original Arawak and Carib populations had survived contact with Europeans by 1650, when the trade in African slaves started booming. Europeans were almost everywhere outnumbered by Africans on the large semi-industrial plantation estates at rates of at least ten to one, which provided the demographic basis for the creolization of English, Dutch and French. The development of the plantation system in relatively small island territories almost inevitably led to ecological disasters of one kind or another and to a degree of overpopulation which has become a shaping influence on the history of these islands to the present day. Notwithstanding later immigration to the island Caribbean from India (designed to compensate for the loss of the labor force after the emancipation of the slaves), China and the Middle East (trade and services), most islands became exporters of people after the plantation phase. Workers from the British West Indies helped build the Panama Canal and were active in various other parts of Central America. When given the opportunity, they provided labor for US agribusiness, and after World War II out-migration from the British West Indies/Commonwealth Caribbean became a global phenomenon, leading to the establishment of thriving diaspora communities in Britain, Canada and the United States (particularly Florida and the New York region). Economic pressure, intensified by political turmoil, led to the relocation of a sizable proportion of Suriname’s Sranan-speaking community to the Netherlands. Haitians, as it happens the numerically largest creolophone community in the world, have fled their country and established themselves in neighboring Caribbean islands (e.g., Jamaica) and in major cities along the US Atlantic seaboard. Elevation of Guadeloupe and Martinique to overseas-department status in France after 1945 has made free movement between France and these two islands possible. Given such a shared history of oppression and degradation, it is clear that the first approach to creoles should thus be made through cultural/social spaces, i. e., the second

24. Pidgins and creoles spatial dimension emphasized in this Handbook (cf. ch. 2). It is probably in reaction to a history of denigration that many creole languages have become focuses of intense and frequently contradictory positive and negative evaluation in the course of their history. Most creole languages lack any overt sociolinguistic prestige and owe their survival to their extremely strong covert prestige. Language attitudes tend to be very negative on the part of outsiders and non-speakers, and “paradoxical” among speakers themselves, as they see their languages both as a “symbol of powerlessness and degeneracy” and a “symbol of solidarity and truth” (Rickford and Traugott 1985: 252). Such attitudes are very often strong enough to thwart technocratically rational language planning measures of undoubted practical usefulness, which are usually designed to capitalize on the status of expanded pidgins and creoles as ethnopolitically neutral lingue franche in their respective territories. Thus, West African Pidgin English, an expanded pidgin routinely used by tens of millions of Nigerians and other West Africans and a natively spoken creole for several millions more, is refused any kind of official recognition and support in multilingual and multiethnic states which could do very well with precisely this kind of neutral link language (Deuber 2005; Schröder 2003). Bislama, a creole language spoken in Vanuatu (Crowley 1990), is probably the only official language anywhere in the world whose use is discouraged in the country’s educational system. As to the dimension of transnational spaces (cf. Jacquemet in this handbook), it should be noted that none of the creole languages involved in these currents of migration are under threat in their home bases, and many of them show more than usual persistence in the diaspora. Creoles can thus with some justification be claimed to have become transnational languages. The combined effect of a century of social and political upheaval (which has seen the rapid decomposition of European colonial empires and unprecedented migration on a global scale) and half a century of booming linguistic research in the field of “creolistics” has been that Reinecke’s “marginal languages” have now moved to the center ⫺ both in terms of global demography and in linguistic theory. Creole languages are now part of the sociolinguistic fabric of the world’s global cities ⫺ from New York, London and Toronto to Lagos, Johannesburg and the urban northeast of Brazil. At the same time the study of pidgins and creoles has come to occupy a central place in contemporary linguistics, both in the empirical description of languages, dialects and varieties and in linguistic theory building. There are now numerous introductory surveys (e.g., Arends, Muysken and Smith 1995; Mühlhäusler [1986] 1997; Sebba 1997; Holm 2000) and ⫺ more importantly ⫺ ambitious theoretical accounts (Escure and Schwegler 2004; McWhorter 2005) and major collaborative projects (e.g., Michaelis et al., to appear: The Atlas of Pidgin and Creole Language Structures).

2. Are creoles a distinct dialectological type? Owing to their history of discontinuous transmission and because they tend to share a number of important structural features in their core grammars (e.g., extreme analyticity; SVO order; preverbal markers for tense, mood, aspect; all-purpose locative preposition covering a wide range of spatial relations), creole languages have been claimed to represent a distinct linguistic type. The most vocal proponent of this position has, of course,

443

444

IV. Structure and dynamics across language spaces been Derek Bickerton (e.g., 1981) who ⫺ inspired by Chomsky’s notion of a genetically wired Universal Grammar ⫺ has argued that the core grammar of creole languages provides the most direct and immediate reflection of this linguistic “bio-program”. Not surprisingly, therefore, “creole exceptionalism” (De Graff 2003) has been a controversial topic in theoretical-linguistic debates on creole genesis and the typological classification of creole languages. However, notwithstanding such disputes about the genesis and status of creole languages, all but a tiny minority of professional linguists are agreed that creoles, in spite of occasionally very strong social stigmatization, are in no way inferior to any other type of natural human language in terms of overall structural complexity and expressive potential. One of the rare dissenting voices is Whinnom (1971: 110), who argues that […] modern linguists may have been dangerously sentimental about creole languages which, with only a few notable exceptions, constitute in most communities a distinct handicap to the social mobility of the individual, and may also constitute a handicap to the creole speaker’s personal intellectual development. […] linguists do not have the evidence to assert with confidence that speakers of these languages are not handicapped by their language, and should not, while any doubt remains, make […] assertions to the contrary.

When it comes to dialectological and sociolinguistic investigations of the use of creole languages, on the other hand (that is, the subject proper of the present Handbook), there is little reason to assume special status for them. The general dynamic of the assignation of prestige and stigma may play itself out in extreme ways in the case of creole languages in continued coexistence with their prestigious lexifiers, but there are no fundamental differences between creoles and other highly stigmatized vernaculars. Similarly, in bilingual and multilingual settings creoles may be found in diglossic constellations or as endpoints of continua linking them to local standard varieties, but there do not seem to be constellations which are unique to creoles. If it exists at all, then, the separate sociolinguistic identity of creoles is to be found in the fact that they tend to locate at the boundaries of traditional dialectological/sociolinguistic categories. For example, especially in cases in which they are used alongside their original lexifiers, the status of creoles tends to be controversial: are they to be seen as varieties or dialects of their lexifiers, or as distinct and independent languages? General attitudes among speakers ⫺ as reflected in names such as patois/dialect (for Jamaican Creole), Pidgin English or broken English (for the stabilized and expanded Englishlexicon pidgin of Nigeria) or Negerhollands (for a Dutch-lexicon Caribbean Creole of the Virgin Islands) ⫺ strongly favor the former point of view. Thus, a speaker of Jamaican Creole might plausibly express his deeply felt conviction that “is English we a taak” (focus.marker English we progressive.marker talk), meaning ‘It is English we are speaking’)⫺in a grammatical structure which is remote from both standard English and nonstandard varieties of English without a history of creolization. Aligning the creole with its more prestigious lexifier, however, is no guarantee against the type of linguistic self-hatred evinced in the following authentic statement by an informant (“Dinah”) in Patrick’s (1999: 275) sociolinguistic study of the urban mesolect of Kingston, Jamaica: (1) mi no riili waan spiik ina patwa langwij I NEG really want speak in patois ‘I don’t really want to speak in Jamaican Creole.’

24. Pidgins and creoles Linguists, on the other hand, have emphasized precisely these basic structural contrasts to the lexifier to claim the status of independent languages for creoles, partly in an attempt to raise speakers’ consciousness and claim recognition for creoles after centuries of neglect and denigration. Another scholarly tradition (e.g., Alleyne 1980: 1996) has similarly argued against the popular perception of creoles as varieties of the superstrate/ lexifier, but asserted that they are not so much languages in their own right as varieties of the West African substrate languages. In the absence of any precisely definable linguistic criteria to settle the question of whether a variety is a language or a dialect, a technical linguistic debate on the issue is pointless and should not be continued. Popular perceptions, however, can have noticeable effects on linguistic behavior, for example the development of implicational scaling which links creole basilects to perceived standards or acrolects. Similarly, the idea that a creole is a dialect of a prestigious standard limits the options for language planning in education. It reduces the motivation to write in the creole while at the same time overestimating active and receptive competence in the acrolect. Focusing on matters of space, creoles stand out as vernacular languages on the move ⫺ both in their genesis and history and through their involvement in contemporary global migrations, and in fact it is most interesting to study how the spread of English, French or Portuguese-based creoles interconnects with the usually more orderly process of the spread of the standard varieties of the respective lexifiers. Again, however, such constellations are not unique. Compare, for example, the case of Hindi, which has undergone a complex history of pluricentric standardization on the Indian subcontinent and has since developed a worldwide diaspora, and the ⫺ historically prior ⫺ globalization of the closely related Bhojpuri, which from its Northern Indian base was spread to Nepal, Mauritius, Suriname (cf. Damsteegt 2002 on the present-day fate of the Sarnami community), South Africa (Mesthrie 1991) and the British West Indies, in particular Trinidad. All things considered, then, the terminological and conceptual framework developed in contemporary sociolinguistics to handle highly stigmatized language varieties with exclusively covert prestige would seem to suffice for the study of stabilized/expanded pidgins and creoles.

3. Creole spaces in previous research In creolophone communities which are spread out over larger territories speakers usually assert that there are regionally distinctive forms (for Jamaica, cf., e.g., Beckford Wassink 1999: 63). For at least one major creole, Haitian, we have now a monumental six-volume dialect atlas (Fattier 1998). Nevertheless, it remains true to say that the systematic search for regionally distinctive usages has never been a priority in research on creole languages. In fact, expert analysts commonly tend to second-guess native speakers’ intuitions about regional variation and assume that it is epiphenomenal and an accidental by-product of more systematic ethnic, social or rural⫺urban variability (Beckford Wassink 1999: 83). Creoles being “young” languages, it is, of course, plausible to argue that there has not been enough time for stable regional usages to emerge ⫺ just as in the case of colonial settler koine´s of European languages such as Canadian English or Australian

445

446

IV. Structure and dynamics across language spaces English there is less regional variability than in the traditional dialects of Britain. This, however, is only a partial explanation. Most creolophone communities are characterized by very sharp ethnic and social divisions, and these easily overshadow regional background as a major determinant of linguistic variation. “Creole space” is, therefore, first and foremost sociolinguistic space. Creoles tend to be mass vernaculars which are used in classic diglossic situations (Haiti, cf. Ferguson 1959) or which are linked to standard-like versions of the original lexifier through gradual and ordered continua of usage (cf. DeCamp 1971 for Jamaica or Rickford 1987 for Guyana). Even when the ties to the lexifier have been cut, as is the case with Sranan in Suriname, they commonly figure as vernacular lingue franche in complex multilingual constellations. Quite justifiably, this has led researchers to emphasize the study of creoles in social (i. e., sociolinguistic and social-psychological attitudinal) space rather than in geographical space narrowly understood. For creoles which are still in contact with their lexifiers, the parsimonious model of creole variational space remains the continuum, which “⫺ when reduced to its essentials ⫺ [is] descriptively accurate and theoretically convenient” (Rickford 1987: 15). The caveat in the quote refers to contested issues such as whether synchronic variation along the creole⫺standard continuum reflects or implies a diachronic process of decreolization. A model providing for implicationally scaled and essentially unidimensional moves between the opposite poles of “creoleness” and “standardness” accounts for intra-speaker stylistic variability and inter-speaker variability associated with social status and education. Critics of this model do not invoke plain regional variation as the extra complication or second dimension which might invalidate the continuum. Rather, they argue that it is not fit to account for the full complexity of the observed social and ethnic variation, which ⫺ they argue ⫺ cannot be mapped along one dimension only. The resulting multidimensional creole space, however, is a metaphorical one and not directly anchored in territorial geography, as the following definition proposed by Carrington (1992: 98) makes clear: 1. Creole space is multidimensional. 2. The individuals who inhabit creole space have multi-systemic repertoires. 3. The density of networks of communication will determine the stability, and consequently the identifiability, of systems within this space. 4. The space coheres because networks of communication overlap. Cross-network communication is possible because of the shared parts of repertoires, but the nature of the overlap may well be neither constant nor systematic. The most original theoretical impulse that general sociolinguistic theory has received from sociolinguistically orientated pidgin and creole studies in this tradition is probably the acts-of-identity model proposed by LePage and Tabouret-Keller (1985). In spite of (or rather because of?) its largely metaphorical use, the collocation “creole space” (and its analogues in other ex-colonial languages, such as e´space creole, espacio criollo, espac¸o crioulo) have become widely used motifs in cultural studies approaches to creole speakers’ experience ⫺ embodied in titles of academic publications such as “Race, space and the poetics of moving” (Philip 1998).

24. Pidgins and creoles

4. From language through code to symbol: On the recent deterritorialization o some creoles The majority of the world’s expanded pidgins and creoles are small languages, and many of them are therefore endangered. In addition to the usual factors making language maintenance difficult for speakers of small languages, creoles suffer from another complication. Not only are they frequently considered corrupt, and therefore expendable, versions of their lexifiers, but they also lack the prestige of traditional indigenous languages as the symbolic correlates of a supposedly authentic indigenous culture. Thus, if in contemporary Nigeria standard English has the overt prestige of modernity and progress, and languages such as Hausa, Yoruba and Igbo connote pride in the respective ethnic traditions, Nigerian Pidgin, as a lingua franca based among the masses of the urban poor living at the margins of progress, loses out against both. However, there are repeated instances in which creoles of the urban poor have come to be identified of with specific subcultural lifestyles which in turn have been globalized by the entertainment industry in more or less diluted fashion. The first, most effective and best-known case of such deterritorialization of a creole language is still the spread of Jamaican Creole, in its specific manifestation of Rastafarian slang, on the back of the success of reggae music and its contemporary successors (dancehall, ragga). A language historically rooted in a strictly localizable speech community has thus become a code, optionally and consciously employed in a far more ill-defined “communities of practice” (Meyerhoff 2003; Eckert 2005). Selected features of Jamaican Creole/Rasta talk have thus become available for appropriation by others beyond the extent made possible by diffusion through more traditional language contact. Patterns of conventional face-to-face contact may explain why, in diaspora communities such as London, first-generation and second-generation Caribbean immigrants from small-island communities such as St. Lucia or Antigua should assimilate to Jamaican norms to some extent. Jamaica is the demographic and cultural centre of the Anglophone Caribbean, an important source of linguistic innovation, and this role persists in the diaspora. Why, on the other hand, Asian and even white-British adolescents should find it attractive to “cross” (Rampton 1995) into Jamaican Creole is a more complicated question to answer. In spite of its strong stigmatization, this expatriate version of Jamaican Creole seems to have come to function as the code of a multiethnic urban street culture in contemporary Britain. As such, it is as likely to be used symbolically, as part of a rhetorical strategy, as it is to be used spontaneously. In addition to popular culture, a further factor promoting the deterritorialization of creoles has been the internet and computer-mediated communication, whose sociolinguistic impact was almost immediate, as is made clear by the following remark Velma Pollard (2000: xiii⫺xv) felt it necessary to add to the second edition of her study of the language of Rastafari: This book focuses on the importance of language to Rastafarians, and any study of this “unofficial” Jamaican religion would be incomplete without a thorough reading of it. This collection of papers, originally published in 1994 […], now includes a new chapter and a bibliography. The new chapter takes into account some recent developments in what can only be termed the “globalization” of Rasta culture, much of it represented on the Internet. […]

447

448

IV. Structure and dynamics across language spaces Chapter 6 is a discussion of the sociological and linguistic implications of Rasta dictionaries available on the Internet, as well as other publications. Here the wordmaking process is revisited and transculturacio´n, a term which reflects another aspect of the globalization of Rasta culture, is introduced. The conclusion is that variations in Dread Talk are inherent in this type of cultural “exchange” and are bound to move the language into new, unanticipated spheres.

On the part of the speakers themselves, the computer seems to have lowered traditional inhibitions with regard to the use of creole in writing. Among the pidgins and creoles without an extensive writing tradition which have recently established a serious presence in the chat rooms of the World-Wide Web are Jamaican Creole, Nigerian Pidgin and Haitian Creole. It is clear that such chat rooms are frequented by members of the public, often based in or close to the diaspora, who are frequently educated and middle class and therefore fully competent in English and French. For them, the creoles are an optional stylistic repertoire, a code available for switching into for purposes of more effective identity and discourse work. As such, this code becomes available to outsiders for “crossing into”, much in the same way that occasional members cross into the various music-centered communities of practice mentioned above.

5. Outlook: Creolophone spaces in the twenty-irst century The future outlook for the world’s known 100 or so expanded pidgins and creoles is mixed. Many of them are still marginal languages, spoken in small communities of disadvantaged speakers in remote and isolated areas. Like almost all micro-languages, they face the threat of attrition and eventual extinction. Several creolophone communities, however, are thriving, often both in their historical home bases and in the diaspora. Haitian, Jamaican and Guinea Bissauan Crioulo are cases in point. These creoles, and others, show no sign of disappearing. All of them, however, are changing very rapidly under the influence of prestigious standard languages they coexist with in diglossic situations or continua. Sociolinguistics and contact linguistics have developed the instruments to describe these changes. (There is, of course, a less conspicuous counter-trend to this “de-creolization”: the modification of the standard varieties of the lexifier languages under the impact of the creole substrate and thus the emergence of new local standards.) In diaspora situations, creoles may transform into new ethnic vernaculars. Patrick, for example, has noted “emerging ethnically distinctive varieties of BrE spoken primarily by Caribbean-origin Britons incorporating various elements of Jamaican Creole-like speech” (2004: 234). What poses a challenge to established linguistic methods is the globalization of some creole usages through their currency in global popular culture. The roots of this process are clear: the strong covert prestige of creoles and the resulting appeal at the level of urban street culture. Contemporary sociolinguistics, however, is still developing the means to cope with usage which is far removed from the unconscious vernacular studied in its speech community. What we are looking at is, rather, often conscious rhetorically styled usage in shifting communities of practice, or even largely decontextualized use of creole linguistic signs which would have to be treated as an instance of a very high order of indexicality in the framework of Silverstein (2003) or as straightforward commodification of language in the terms of Heller (2003).

24. Pidgins and creoles

6. Reerences Alleyne, Mervyn 1980 Comparative Afro-American: An Historical-Comparative Study of English-based AfroAmerican Dialects of the New World. Ann Arbor, MI: Karoma. Alleyne, Mervyn 1996 Syntaxe historique cre´ole. Paris: Ed. Karthala. Arends, Jacques, Pieter Muysken and Norval Smith 1995 Pidgins and Creoles: An Introduction. Amsterdam: Benjamins. Arends, Jacques and Matthias Perl (eds.) 1995 Early Suriname Creole Texts: A Collection of 18th-Century Sranan and Saramaccan Documents. Frankfurt: Vervuert. Beckford Wassink, Alicia 1999 Historic low prestige and seeds of change: Attitudes towards Jamaican creole. Language in Society 28: 57⫺92. Bickerton, Derek 1981 Roots of Language. Ann Arbor, MI: Karoma. Carrington, Lawrence 1992 Images of creole space. Journal of Pidgin and Creole Languages 7: 93⫺99. Carrington, Lawrence 1993 Creole space: A rich sample of competence. Journal of Pidgin and Creole Languages 8: 227⫺236. Crowley, Terry 1990 Beach-la-Mar to Bislama: The Emergence of a National Language in Vanuatu. Oxford: Clarendon. Damsteegt, Theo 2002 Sarnami as an immigrant koine´. In: Eithne B. Carlin and Jacques Arends (eds.), Atlas of the Languages of Suriname, 249⫺264. Leiden: KITLV Press. De Graff, Michel 2003 Against creole exceptionalism. Language 79: 391⫺410. DeCamp, David 1971 Toward a generative analysis of a post-creole continuum. In: Dell Hymes (ed.), Pidginization and Creolization of Languages: Proceedings of a Conference Held at the University of the West Indies, Mona, Jamaica, April 1968, 349⫺370. Cambridge: Cambridge University Press. Deuber, Dagmar 2005 Nigerian Pidgin in Lagos: Language Contact, Variation and Change in an African Urban Setting. London: Battlebridge. Eckert, Penelope 2005 Variation, convention, and social meaning. Paper presented at the annual meeting of the Linguistic Society of America. Available at . Escure, Genevieve and Armin Schwegler (eds.) 2004 Creoles, Contact, and Language Change: Linguistics and Social Implications. Amsterdam: Benjamins. Fattier, Dominique 1998 Contribution a` l’e´tude de la gene`se d’un cre´ole: L’atlas linguistique d’Haı¨ti. 6 vols. Villeneuve d’Ascq: Presses Universitaires du Septentrion. Ferguson, Charles A. 1959 Diglossia. Word 15: 325⫺340.

449

450

IV. Structure and dynamics across language spaces Heller, Monica 2003 Globalization, the new economy, and the commodification of language and identity. Journal of Sociolinguistics 7: 473⫺492. Holm, John 1988⫺1989 Pidgins and Creoles, vol. 1, Theory and Structure; vol. 2, Reference Survey. Cambridge: Cambridge University Press. Holm, John 2000 An Introduction to Pidgins and Creoles. Cambridge: Cambridge University Press. LePage, Robert and Andre´e Tabouret-Keller 1985 Acts of Identity: Creole-based Approaches to Language and Ethnicity. Cambridge: Cambridge University Press. McWhorter, John 2005 Defining Creole. Oxford: Oxford University Press. Mesthrie, Rajend 1991 Language in Indenture: A Sociolinguistic History of Bhojpuri-Hindi in South Africa. Johannesburg: Witwatersrand University Press. Meyerhoff, Miriam 2003 Communities of practice. In: Jack Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.), Handbook of Language Variation and Change, 526⫺548. Oxford: Blackwell. Michaelis, Susanne, Philippe Maurer, Magnus Huber and Martin Haspelmath to appear The Atlas of Pidgin and Creole Language Structures (APiCS). Mühlhäusler, Peter [1986] 1997 Pidgin and Creole Linguistics. Expanded and revised edition. Westminster Creolistics Series. London: University of Westminster Press. Neumann-Holzschuh, Ingrid and Edgar Schneider (eds.) 2000 Degrees of Restructuring in Creole Languages. Amsterdam: Benjamins. Patrick, Peter 1999 Urban Jamaican Creole: Variation in the Mesolect. Amsterdam: Benjamins. Patrick, Peter 2004 British creole: Phonology. In: Bernd Kortmann and Edgar Schneider in collaboration with Kate Burridge, Raj Mesthrie and Clive Upton (eds.), A Handbook of Varieties of English, vol 1: Phonology, 231⫺246. Berlin: Mouton de Gruyter. Philip, Nourbese 1998 Race, space, and the poetics of moving. In: Kathleen M. Balutansky, Marie-Agne`s Sourieau and Yanick Lahens (eds.), Caribbean Creolization: Reflections on the Cultural Dynamics of Language, Literature, and Identity, 129⫺53. Gainesville, FL: University Press of Florida. Pollard, Velma 2000 Dread Talk: The Language of Rastafari. Montreal/Kingston, Jamaica: McGill Queen’s University Press/Canoe Press. Rampton, Ben 1995 Crossing: Language and Ethnicity among Adolescents. London: Longman. Reinecke, John 1937 Marginal languages: A survey of the creole languages and trade jargons. [PhD dissertation, Yale.] Ann Arbor, MI: University Microfilms International. Rickford, John 1987 Dimensions of a Creole Continuum: History, Texts and Linguistic Analysis of Guyanese Creole. Stanford CA: Stanford University Press. Rickford, John and Elizabeth Closs Traugott 1985 Symbol of powerlessness and degeneracy, or symbol of solidarity and truth? Paradoxical attitudes towards pidgins and creoles. In: Sidney Greenbaum (ed.), The English Language Today, 252⫺261. Oxford: Pergamon.

25. Overseas varieties

451

Schröder, Anne 2003 Status, Functions, and Prospects of Pidgin English: An Empirical Approach to Language Dynamics in Cameroon. Tübingen: Narr. Sebba, Mark 1997 Contact Languages: Pidgins and Creoles. London: Macmillan. Silverstein, Michael 2003 Indexical order and the dialectics of sociolinguistic life. Language and Communication 23: 193⫺229. Thomason, Sarah G. (ed.) 1997 Contact Languages: A Wider Perspective. Amsterdam: Benjamins. Thomason, Sarah G. and Terrence Kaufman 1988 Language Contact, Creolization, and Genetic Linguistics. Berkeley, CA: University of California Press. Todd, Loreto 1984 Modern Englishes: Pidgins and Creoles. Oxford: Blackwell. Whinnom, Keith 1971 Linguistic hybridization and the “special case” of pidgins and creoles. In: Dell Hymes (ed.), Pidginization and Creolization of Languages: Proceedings of a Conference Held at the University of the West Indies, Mona, Jamaica, April 1968, 91⫺115. Cambridge: Cambridge University Press.

Christian Mair, Freiburg (Germany)

25. The consequences o migration and colonialism II: Overseas varieties 1. 2. 3. 4. 5. 6.

Introduction: An “explosion of space” A rough typology of overseas varieties Shaping overseas varieties Two illustrative scenarios Conclusion References

1. Introduction: An explosion o space Societies and communities, in Europe and elsewhere, are in a constant process of shift and flux, for a variety of social, political and economic reasons: crusades and war campaigns; climatic changes and natural disasters; famines; the transformation from largely rural and self-sufficient to urbanized and monetary-based economies; etc. The last two hundred years, however, have seen population movements on an unprecedented scale, an abrupt catalyzation of sociodemographic fluctuation, both within Europe and from Europe to other continents. Many of these migratory movements take their roots in the Middle Ages and the Early Modern period: political expansionism following the discov-

452

IV. Structure and dynamics across language spaces eries of seafarers and adventurers (many of them in the service of governments, such as Sir Francis Drake, Ferna˜o de Magalha˜es or Amerigo Vespucci); technological innovations that made it possible to transport ever-increasing masses of personnel and cargo to new and constantly expanding numbers of destinations; the establishment of official bodies to develop international trading schemes (such as the British East India Company or the Dutch Vereenigde Oostindische Compagnie; Lawson 1993); the rivalry of the newly created nation-states for possession of overseas territories, etc. This “explosion of space” was to have a massive impact on all the domains of language and space as discussed in the first division of articles of this handbook. First of all, it stretched and extended the space of European languages, so that French, Spanish, German, Portuguese and English (and other European languages) were exported to numerous locations around the globe, where they came into contact with ⫺ and coexisted alongside ⫺ indigenous languages. Second, the usage of the (non-local) colonizing languages would eventually cut inroads into the local sociolinguistic landscapes and massively affect the local varieties, in many cases leading to unprecedented language endangerment and death. A common scenario is for the number of speakers of English, French and Spanish to increase significantly, not because of population growth in Europe, but due to the subsequent adoption of these languages in the new territories, where they functioned as lingue franche and granted access to power and wealth. The so-called “Scramble for Africa” at the end of the nineteenth century provides a particularly pertinent case in point here (Duignan, Gann and Turner 1975). As late as in the 1880s, most of Africa was a genuine terra incognita. The Portuguese (and also, though not on the same scale, the British) had taken possession of much of the African West Coast (the Slave Coast, in the Bight of Benin; the Gold Coast, present-day Ghana; and the Ivory Coast), present-day Algeria and the Suez Canal in Egypt were under French control and Great Britain held the comparatively small Cape Colony in South Africa. It is estimated that less than ten percent of Africa was colonized when the continent became a primary target of New Imperialism (Curtin 1995). With the onset of World War I in 1914, however, the only countries outside formal European control were Ethiopia and the republic of Liberia (Bennet 1984). As a result, between 1885 and 1914, in less than thirty years, Great Britain took nearly thirty percent of Africa’s population under her control, France fifteen percent, Germany nine percent, Belgium seven percent and Italy one percent. Nigeria alone was to contribute a populace of 15 million to the British Empire, which was more than the whole of French West Africa or the entire German colonial empire combined. The “Scramble for Africa” represents a showcase scenario of the effects of political expansionism and colonialism on language and space. In terms of geographical space, an unprecedented extension into unknown territory (cf. Johnstone in this volume); in terms of cultural/social space, a reshuffling and reorganization of the local socio-ethnic landscape as well as the collapse, merging and consecutive redevelopment of social boundaries and networks in the colonial communities (cf. Mæhlum in this volume); in terms of political space, the creation of new nation states (often based on purely geographic and ethnically ignorant conceptualizations of space, along topographic rather than socio-ethnic boundaries), and a transformation of local economic systems (cf. Gal in this volume); and in terms of migration and colonization, large population movements, accompanied by the establishment of permanent settlements and enduring contacts with the colonial powers (cf. Jacquemet in this volume). Linguistically and sociolinguistically,

25. Overseas varieties the global diffusion and spatial expansion of language(s) ultimately led to the endangerment of local minority languages, often going hand in hand with the development of contact-based varieties such as pidgins and creoles (cf. Mair, this volume). On the other hand, language contact between colonizing languages (and the emergence of indigenous offspring varieties, such as Dutch-based Afrikaans in South Africa; Roberge 2002) and koine´ization due to extensive dialect interaction (see below) were common as well. The “explosion of space” thus triggered a range of social, political, ethnic, economic and linguistic developments, deeply affecting all domains of language and space, and had a profound impact on Europe and the New World territories alike. This article looks into selected aspects of language spread and diffusion and is structured as follows. It begins by providing a rough typological overview of overseas varieties, addressing the historical and societal dimension of migration and colonialism as well as the roots of the varieties under discussion more concisely. Using this as a baseline, it discusses and exemplifies some of the major linguistic and sociolinguistic outcome(s) of language transplantation, namely how extended contact triggers language change and how such contact in turn sees the “birth” of new post-colonial overseas varieties. In particular these are issues related to who is most influential in the formation of these varieties (the widely discussed concept of the “founder principle” [Mufwene 1996], as opposed to (later) dialect swamping [Lass 1990], issues related to social restructuring, settlement patterns, reforming of social networks and communities of practice, identity, etc.), whether linguistic criteria are more important than social ones or not (e.g. total percentage of input varieties), and how interaction between varieties as well as extended contact with the colonial variety affects the overall trajectory of change. A final section discusses two case studies ⫺ from Tristan da Cunha and New Zealand ⫺ and exemplifies some of the processes discussed.

2. A rough typology o overseas varieties Perhaps the most important question in this context is exactly what to include under the label “overseas varieties”. One should at all cost avoid the implication that these varieties are characterized by relative homogeneity (Lass 1987). In fact, the term is so general and all-inclusive that it covers just about every variety that was brought out of Europe and continued to be used elsewhere. To take the spatial extension of German, for instance, overseas varieties have developed in the Americas, Africa and in the Pacific (Timm 2001), notwithstanding the fact that Germany had a limited sphere of colonial involvement and never belonged to the colonial superpowers (Gründer 2004). To name but some offspring varieties of German around the world: Pennsylvania Dutch (a misnomer, Deutsch being misinterpreted as Dutch), certainly the most widely known form, which saw substantial input of dialects from the Palatinate; Amana German in the Amana Colonies in Iowa, founded by Inspirationalists of German origin; and several varieties in South America: Venezuelan Alema´n Coloniero (with prominent input from Low Alemannic varieties), Lagunen-deutsch, which developed around Lake Llanquihue in Chile (and may have undergone extensive mixing with Spanish), and Riograndenser Hunsrückisch, formed by German Brazilians in Rio Grande do Sul, especially in the areas of Santa Catarina, Parana´ and Espı´rito Santo (see Riehl in this handbook as well as the

453

454

IV. Structure and dynamics across language spaces description of these varieties on Ethnologue ). In Africa, there are some 30,000 speakers of German in Namibia (former Deutsch-Südwestafrika), whereas German has not survived in Ruanda, Burundi, Tanzania or Mozambique (former Deutsch-Ostafrika, ceded to the British after World War I). A very special place in the overseas legacy of German is held by Unserdeutsch, also known as Rabaul Creole German, a German-based creole language spoken primarily in Papua New Guinea and the northeast of Australia (Holm 1989), which now is near extinction, with few bilingual speakers left. Consequently, even though the linguistic legacy of “overseas German” is limited, the varieties named display contact-linguistic complexity and have arisen due to dialect contact, language mixing and creolization. What, then, are the major types of overseas varieties and which of them should be discussed here? Probably still one of the most important classificatory tools is the model put forward by Kachru (1985), who argued that varieties (of English, though one could easily generalize to other world languages) could be grouped in three concentric circles, depending on criteria such as historical continuity, function and use. The so-called “inner circle” represents the traditional bases of English: the United Kingdom, the United States of America, (English-speaking) Canada, Australia and New Zealand. The total number of English speakers in the inner circle amounts to approximately 380 million, the majority of whom reside in the United States. The “outer circle” (countries that were part of the British Empire) contains countries where English is not an official language but where it assumes an important institutional role and is important for historical reasons (India, Nigeria, South Africa, the Philippines, Bangladesh, Pakistan, Malaysia, Tanzania, Kenya, etc.). The total number of English speakers in the outer circle is estimated to range from 150 million to 300 million. The “expanding circle” finally, is made up of countries where English plays no historical or governmental role, but where it is nevertheless widely used as a foreign language or lingua franca (China, Russia, Japan, Korea, Egypt, Indonesia, etc.). The total number of speakers of English in this circle is most difficult to estimate, particularly since the usage of English is functionally restricted (e.g., for academic purposes or the tourism industry). Global estimates range from 300 million to more than one billion (Crystal 1997). In Kachru’s view, the inner circle is “norm-providing” so that language norms are developed in these countries, mostly because English is the first (or native) language. The outer circle, by contrast, is “normreceiving” (since many speakers of English are bilingual or multilingual), whereas the expanding circle is seen as “norm-developing”, due to the creative potential of foreign language speakers and language learners and the limited exposure to native-speaker norms. Kachru’s model has not remained uncontroversial (Quirk 1990), but factors such as historical continuity and ancestry, mono-, bi- and multilingualism, the function and usage of a given language are certainly of utmost importance for any classification of overseas varieties. Adopting an approach based on time depth, ancestry and affiliations with the “homeland”, the presence, strength and impact of input varieties other than the colonizing language, the function and communicative role of the languages present, the sociodemographic set-up of the community, etc., then we can distinguish between directly transplanted overseas varieties, where a native-speaker tradition was maintained at all times (e.g., Falklands Islands English, Que´bec French, or Riograndenser Hunsrückisch), second- or foreign language varieties (English in India, Spanish in Paraguay), pidgins and creoles (e.g., English-derived Pijin on the Solomon Islands, French-derived Mauritian

25. Overseas varieties Creole, or Morisyen, on Mauritius, or Italian/Spanish-based Cocoliche in Argentina; Holm 1989), creoloids (varieties with prototypical creole features that have not developed from a pidgin, e.g., Singlish, the basilectal English used in Singapore; Platt 1975) and language varieties shaped via extensive dialect contact and koine´ization (such as Australian English). It must be emphasized, however, that many, if not most, processes of new-variety formation are “messy” and that the development of overseas varieties is never clear-cut. More often than not, it is virtually impossible to delineate and isolate single processes; many of them co-occur and overlap, for instance when language and dialect contact coexist and have a joint impact on the sociolinguistic evolution of a newly forming variety. Consequently, this article focuses on two possible outcomes, dialect contact and creoloidization, to document the complex processes of contact linguistics that underlie the formation of overseas varieties. It excludes offspring varieties of language contact, i. e., pidgins and creoles, since these are discussed by Mair (in this volume).

3. Shaping overseas varieties We can single out several stages in the formation of overseas varieties. First of all, what basically happens during the formation phase of an overseas variety is that the total number of variants present in the contact scenario represents a pool out of which the first native speakers select features (Mufwene 2001). None of the input varieties “wins out” at this stage; were this the case, then all but one variety present in the contact scenario would disappear without a trace and the newly developing variety would represent the equivalent of one of its inputs. Riograndenser Hunsrückisch would merely be a transplanted (and basically unaltered) form of a Rhineland-Palatinate German dialect, Australian English exported Cockney, Que´bec French the equivalent of a regional variety found in France, etc. This is hardly ever the case, for reasons that are obvious. Dialects mix and interact as speakers accommodate to each other, and new dialect formation is particularly strong when children grow up in a multi-dialectal environment. The usual outcome is for the inputs to undergo a stage of mixing, so that the inception phase of a new dialect displays mechanisms of feature selection and retention. Consequently, contact between linguistic systems triggers selection processes of features from several coexisting varieties (Kerswill 1996), and this may be influenced by factors such as total number of features present, salience, stigma and prestige of individual variables, sociodemographic characteristics and social mobility, etc. Fully developed dialects, the endproduct of focusing (LePage and Tabouret-Keller 1985; see below), have thus adopted features from at least two (very often, more) donors. Put differently, a crystallizing contact-based overseas variety combines a mixture of features (phonetic, grammatical, morphological and/or lexical) that derive from some or all the dialects present in the original contact situation. A second process is leveling: the majority of variants found in a diffuse mixture situation gradually disappear as one feature is permanently selected (Trudgill 1986; Siegel 1987; Britain 1997). Tendencies towards regularity and transparency are common in contact situations. Though the precise nature of what determines leveling is still unknown, there is consensus that status (stigma or prestige) and frequency are important

455

456

IV. Structure and dynamics across language spaces criteria. Variants that are regionally or socially marked are usually not maintained (Mesthrie 1993) and those with the widest social and geographical distribution have the highest chances of surviving the selection process (Trudgill 1986). First, the surviving form is usually the one found in the majority of inputs (Mesthrie 1993; Siegel 1987), and this has led Trudgill (2004) to adopt an extreme (and controversial) hypothesis, namely that it should be the frequency of features alone that accounts for adoption. Social factors (prestige, status, social network structures, etc.), as a consequence, should simply be irrelevant. Such an explanatory attempt cannot account for independent developments (see below) or cases of reallocation (Trudgill 1986), when more than one original variant survives the leveling process. Reasons for reallocation are reanalysis and functional redistribution, either as social or stylistic variants, or as phonological variants in complementary environments (Britain 1997). This has been documented by Domingue (1981) in Mauritius Bhojpuri, where the total number of coexisting variants was not reduced due to the fact that former regional variants of Hindi were reinterpreted as indicators of (in)formality and style. New Zealand English (NZE) provides a pertinent case of leveling. Schreier et al. (2003) analyzed the maintenance of voiceless labiovelar /hw-/ fricatives (minimal pairs Wales ~ whales, witch ~ which) in three regions of New Zealand (Otago/Southland, Canterbury and the North Island) and found considerable regional variation in early twentieth-century New Zealand English. Whereas speakers from the North Island and Canterbury were predominantly using /w/ (so that the /hw/ ~ /w/ merger was practically completed by 1950), speakers from the Southland and Otago regions had high levels of /hw-/ well into the second half of the twentieth century. The regional distribution of the variable was linked to population demographics and ancestral effects, reflecting the total input strength of /hw-/ retaining donor dialects. A high overall presence of /hw-/ variants in the inputs thus had an enhancing effect on the adoption and maintenance of the phoneme. The strong presence of /hw-/ in the Otago/Southland dialect region was explained by high input frequency and the disproportionately high input of Scottish settlers, who made a distinction between /hw-/ and /w-/. In the other regions, however, /hw-/ was not adopted, simply because the social configuration and the local contact and mixture situations were different. The inputs were mainly from the South of England, where /hw-/ was a minority feature, and this enhanced leveling-out in the local forms of New Zealand English. A third process involved is what some label simplification (e.g., Trudgill 1986), which however may be preferably considered as a manifestation of regularization or simply analogical language change. I personally favor the latter since the outcome of new-dialect formation is not necessarily “simpler” than any of the corresponding input forms. Rather, a given property of a variant X, no matter if phonetic/phonological, lexical or grammatical, is subject to less variation than it was in the input varieties originally. This process manifests itself in the reduction of the quantitative range of forms. Siegel (1987: 14) points out that simplification is not well understood and argues that it should be investigated quantitatively rather than qualitatively (i. e., to adopt a variationist perspective to trace a putative decrease in variability) and to investigate it diachronically, with reference to earlier evolutionary phases of a variety (or of the input varieties). This problem is also recognized by Britain (1997: 141) who states that simplification represents “an increase in grammatical regularity and decrease in formal complexity”, and Mühlhäusler (1997: 236; emphasis added), who claims that

25. Overseas varieties Simplification only refers to the form of the rules in which a language is encoded, indicating optimalization of existing rules and the development of regularities for formerly irregular aspects, for example, grammaticalization of the lexicon. Simplification is a dynamic concept. It expresses the fact that as one moves along a developmental continuum, more and more regularities appear.

A good example of phonological regularization in overseas varieties is provided by the devoicing of intervocalic alveolar plosives in Afrikaans (Booij 2002). In Dutch, voiceless alveolar plosives (/t/) are voiced in intervocalic environments: /t/ > [d] / V_V Dutch plural formation commonly involves the affixation of an -en marker and changes the phonological environment from /-VC#/ to /-VCV#/. If a singular form ends in /d/, then the corresponding plural form maintains it. A voiceless /t/, however, assimilates to /d/, as in: hoed [hut] ‘hat’, hoeden [1hude]. Afrikaans, in contrast, has regularized this assimilation rule so that the phonological environment has no effect on the alveolar plosive: hoed [hut] ‘hat’, hoeden [1hute], stad [stat] ‘city’, statten [1state]. Afrikaans is thus more regular than its ancestral variety Dutch, in that it has reduced the number of phonological rules and does not make phonological contrasts in this particular environment (Booij 2002). Similarly, a striking case of morphosyntactic regularization is found in Tristan da Cunha English (TdCE), which regularized the past tense paradigm of be (with pivot form was) to the extent that were was categorically absent from the speech of Tristanians born before World War II (Schreier 2002). One notes the interplay of simplification and regularization. A past be paradigm that has undergone regularization (we was, the fishermen was, etc.) is also in a way ‘simpler’. Put differently, there is no process of simplification that does not also entail regularization, which questions whether it is necessary to keep the two terms distinct. Finally, processes of feature selection and retention are not the only ones operating in long-term contact situations. A new dialect affiliates with its inputs by drawing its features from them, but it is erroneous to assume that it is linguistically predetermined by the distinctive properties of the varieties in contact. Contact-derived dialects may develop their own dynamics, namely when “contact between two dialects leads to the development of forms that actually originally occurred in neither dialect” (Trudgill 1986: 62). Some variants are “not actually present in any of the dialects that contribute to the mixture but … arise out of interaction between them” (Trudgill et al. 2000), and this has been amply documented by Britain (1997) in English Fens English. According to Trudgill (1986), “interdialectalisms” represent intermediate variants of original forms (and thus originate in incomplete or faulty accommodation) or overgeneralization and hyper-adaptation (perhaps the most well-known case here is hypercorrection, which occurs when speakers misinterpret and incorrectly generalize rules by applying them to inappropriate contexts; Trudgill 1986: 66). Therefore, mixing and leveling alone cannot be offered as an explanation for all the features of a newly developing variety. Any theory that attempts to arrive at a general outline of contact-induced language change needs to leave room for independent innovation patterns. All these processes contribute to new-dialect formation and are thus part of what LePage and Tabouret-Keller (1985) call focusing. Speakers of a fully focused variety are

457

458

IV. Structure and dynamics across language spaces in agreement about normative structures and have a strong awareness that their dialect differs from others on linguistic grounds. Focused varieties often have a “proper” name and are accompanied by processes of standardization and codification; these reinforce attitudes towards outsiders and foster a strong sense of a local identity in the speech community’s members. In so-called diffuse communities, by contrast, speakers display considerable heterogeneity (resembling the phenomenon of super-diversity), having no (or little) consensus on linguistic norms and the status of the variety. Le Page and Tabouret-Keller (1985) illustrate this with the case of multilingual Belize, where English, Creole and Spanish coexist. There is a continuum between acro- and basilectal varieties, no consensus on the usage of high-standard norms, no clear demarcation between the coexisting varieties and no agreement as to which of the varieties is used in a given context or interaction. The maintenance of multiple varieties in a contact setting is therefore a crucial factor to determine whether a variety focuses or whether it remains diffuse, and this has implications for the development of overseas varieties in general. According to Kerswill and Williams (1992: 13), the initial stages of any dialect mixture situation are characterized by extreme diffuseness. There is no agreement on shared norms in a newly founded community and language usage is characterized by the coexistence of a number of distinct variants. Variability may then be reduced when accommodation patterns between speakers in face-to-face interaction increase and social networks are implemented.

4. Two illustrative scenarios The consequences of spatial extension of language are clear. First, there exist a large number of typologically and structurally distinct overseas varieties, so that the sociopolitical and linguistic landscape of any expanding language becomes richer and more diverse (often at the expense of pre-existing varieties, however). Second, the formation and evolution of overseas varieties depends on the complex interplay of language-internal and extralinguistic criteria; these influence the outcome of new varieties and must be taken into account for any attempt at reconstruction or explanatory modeling. The remainder of the article now illustrates two of these outcomes and looks into the historical and sociolinguistic developments of two varieties in more detail: New Zealand English (NZE, which provides a scenario of dialect contact and koine´ization) and Tristan da Cunha English (TdCE, which arose due to creoloidization by adopting creole features without prior pidginization or creolization).

4.1. Dialect contact and new dialect ormation: New Zealand English New Zealand English is the youngest of the “inner-circle” varieties of English around the world (see above), in the sense that it underwent structural nativization (Schneider 2007) most recently. European expansion into the Pacific began in the late eighteenth century, but by the 1830s, European residents in New Zealand numbered no more than around 2,000. It was only in 1840, after the Treaty of Waitangi was signed and British sovereignty was proclaimed over the country, that the European population of New

25. Overseas varieties Zealand began to grow. This occurred at a remarkable rate, however; by 1872, the European population had reached 256,000 and by 1881 almost half a million (Belich 1996: 278). In the early years of colonization, Auckland became the seat of government and it was to remain the most populous settlement. However, the period between 1840 and 1852 was also characterized by colonizing efforts of bodies in Britain that organized emigration to central and southern New Zealand. Most of the immigrants who arrived in these years came to planned settlements. Although their numbers were not great when compared with later immigration, they nevertheless laid out early patterns in the areas where they settled, particularly in Otago and Canterbury. In a sense, they represent the founders of these settlements and were to have a long-lasting impact on the colonies. The 1871 census indicates that the vast majority of migrants in New Zealand came from the British Isles and that the English formed the largest ethnic group thereof (51 percent). The Scots, who made up ten percent of the population of the British Isles, constituted 27.3 percent of the migrants in New Zealand. They were mainly concentrated in Otago and Southland, but also settled elsewhere throughout the country. The Irish, who in 1871 made up 18.8 percent of the UK population, constituted about 22 percent of the New Zealand migrant population. The Welsh were often conflated with the English in official records but, even allowing for this, the percentage of Welsh migrants was insignificant. By 1871, the North Island provinces generally had more New Zealand born people than the South Island, and Otago and Southland had significantly more immigrants from Scotland than the other provinces. Canterbury led in the numbers from England, though the distinction between this and other provinces in the North Island was not as marked as the predominance of Scots in Otago and Southland. Here, the Scots made up between 60 to 80 percent of the total population, mostly living off agriculture (Olssen 1984: 71). The other large group consisted of miners, mostly from England and Ireland, who made up about 24 percent of the male workforce (Olssen 1984: 71). This is reflected by the settlement patterns of individual Otago towns. The population of Arrowtown, for example, a typical gold-mining town, had roughly similar proportions of settlers from England, Scotland, Ireland and Australia. Milton, on the other hand, a rural agricultural town, had a high percentage of settlers born in Scotland. Canterbury, in contrast, was planned to be “English, Anglo-Catholic, and Conservative” (Sinclair 1991: 92). A study of the population of nineteenth-century Canterbury (Pickens 1977) shows that the “English” stereotype of the Canterbury population goes back to early settlement patterns, where fifty-five percent were of English origin, sixteen percent from Ireland, fourteen percent from Scotland and one percent from Wales. Of the English settlers, the majority came from the South of England, in particular the Southeast. New Zealand’s North Island, on the other hand, witnessed a notably different social history and higher sociodemographic fluctuation. Wellington was the earliest New Zealand Company settlement and one of the first places of European involvement. Places like Dannevirke, in contrast, were planned settlements, set up for government assisted immigrants from Scandinavia in the 1870s, and places such as Rotorua were established as late as in the 1880s. Political insecurities and struggles with the local Maori population hindered colonization of much of the North Island, but the stationing of British soldiers attracted the European population in the 1860s. Numbers rose dramatically in the 1860s, after war broke out between Maori and the British troops. The colonial government launched an ambitious scheme to recruit soldier settlers and to repatri-

459

460

IV. Structure and dynamics across language spaces ate them on land confiscated from the Maori. By the mid 1860s, there were some 12,000 Imperial troops, together with 4,000 local soldiers. According to McGibbon (2000: 325), “most of those who enlisted were young single men born in Great Britain and from the lower stratum of Victorian society, laborers and semi-skilled workers attracted by the promise of a free farm”. On their discharge soldiers were allotted land (the amount determined by rank) though many did not stay long enough to obtain legal ownership. For example, of the 2,056 soldier-settlers who were granted farms in the Waikato province in the 1860s, only 214 still owned farm sections in 1880 (McGibbon 2000: 327), which attests to the high degree of mobility in the North Island at the time. Economically, the provinces of the North Island for a long time lagged behind Otago and Canterbury, where the wool and meat industry yielded high profit margins. The discovery of gold in the South Island in 1861 increased economic growth and extended the population. In the 1870s, a government immigration scheme was set up to boost the population in the wealthier provinces of the south, and there was a massive influx of new immigrants to Otago and Canterbury. The situation was to change from the 1880s on, when the success of refrigerated shipping made the export of dairy produce possible. After 1901, more dense settlement and industrial development gave the Northern provinces an economic advantage, which led to demographic restructuring and a pre-eminent socio-political role of the North Island, which remains to the present day. Sociolinguistically, New Zealand English is a showcase scenario to illustrate the mechanisms of new dialect formation. Though some (e.g., Hammarström 1980) have argued for monogenetic origins, taking the view that Australian English (and, by implication, New Zealand English) basically represent varieties of transplanted Cockney, most agree that New Zealand English originated in a context of dialect transplantation, contact and mixture, classifying it as a contact dialect (or koine´): “While NZE is undoubtedly southern English in origin, it shows features which are found throughout the south of England […] NZE really is a mixed dialect, taking input from throughout Britain” (Bauer 1999: 304; emphasis added). In the words of Trudgill et al. (2000: 302), it “is the result […] of a complex series of processes involving dialect contact between different British Isles varieties of English, followed by dialect mixture, new-dialect formation, and then by subsequent linguistic changes.” The scenario that gave rise to this variety of southern hemisphere English consequently involved several transplanted inputs; their phonological and morphosyntactic properties served as a feature pool from which New Zealand English drew its features. The most eminent of these varieties was southeastern English English, or London, to be more precise. Bauer (1994: 388) states that “phonologically speaking, New Zealand English is a variant of the southeast England system”, a claim supported by the qualitative and quantitative analysis reported in the ONZE (“Origins of New Zealand English”) project (Gordon et al. 2004). Apparent-time data of nineteenth-century New Zealand English (the database used by the ONZE team) is indicative of early rudimentary leveling (of socially and regionally marked variants, such as the /v ~ w/ merger) and extreme variability in the speech of the first locally born generation, whose speech is characterized by unique and often idiosyncratic combinations of features from distinct dialect regions of the British Isles. The impact of language contact was slight, manifesting itself mostly on a lexical level (Bauer 1994; Deverson 1998). Following the course of focusing and new-norm adoption, New Zealand English (like other post-colonial varieties, such as Australian English) underwent regional uniformity and leveling on a nationwide scale, as a result of which regionality does not (yet)

25. Overseas varieties correlate with language variation in present-day New Zealand English: “New Zealand, like Australia, is more remarkable for the absence of regional differences” (Gordon and Deverson 1998: 126). This has also been openly remarked on in earlier American English (“The characters chiefly noted in American speech by all who have discussed it, are, first, its general uniformity throughout the country, so that dialects, properly speaking, are confined to recent immigrants, to the native whites of a few isolated areas and to the negroes of the South”; Mencken 1921) as well as in Australian English (“Australians are for practical reasons almost completely unmarked by region within their speech community”; Bernard 1989). Indeed, this seems to be a highly diagnostic feature of the early evolutionary phase of “inner-circle” overseas varieties (see Schneider 2007 for an in-depth comparison and theoretical analysis of these varieties) and one that sets them apart from others developing in multilingual contexts.

4.2. Dialect and creole contact: Tristan da Cunha English The formation of Tristan da Cunha English (TdCE) in the nineteenth century resembles that of New Zealand English to some extent, namely in that it witnessed dialect contact between varieties of British and American English. On the other hand, it shows how complex and intricately interwoven contact-induced change mechanisms are and how little it takes to alter the trajectory path of overseas varieties, particularly those developing in small, stable and locally confined founding populations. The island of Tristan da Cunha lies in the South Atlantic Ocean, some 2,300 kilometers south of St. Helena, 2,800 kilometers west of Cape Town and about 3,400 kilometers east of Uruguay, with a current population of 285. It was discovered by the Portuguese admiral Trista˜o da Cunha in 1506. The English and Dutch too became aware of the islands, the Dutch being the first to effect a landing (in 1643; Beintema 2000). However, none of the colonial powers developed an interest in establishing a permanent colony on the island. Things changed when, towards the end of the eighteenth century, the American fishing and whaling industry expanded to the South Atlantic Ocean and Tristan da Cunha served as an occasional resort to the sealers and whalers (Brander 1940). The growing economic interest and strategic position along a major sea-route soon attracted discoverers, pirates and adventurers. The island was officially settled in 1816, when the British admiralty formally annexed Ascension Island and Tristan da Cunha, apparently with the intention of blocking a possible escape route for Napoleon Bonaparte, who at the time was exiled on the island of St. Helena (Schreier and Lavarello-Schreier 2003). A military garrison was dispatched to the island, withdrawing after one year only. Some army personnel stayed behind with the intention of settling permanently: two stonemasons from Plymouth (Samuel Burnell and John Nankivel) a non-commissioned officer from Kelso, Scotland, named William Glass, his wife, “the daughter of a Boer Dutchman” (Evans 1994: 245), and their two children. The population increased when shipwrecked sailors and castaways arrived; some of them stayed behind and added to the permanent population. In 1824, apart from the Glass family, the settlers included Richard “Old Dick” Riley (from Wapping, East London), Thomas Swain (born in Hastings, Sussex) and Alexander Cotton (from Hull/Yorkshire), who had arrived in the early 1820s (Earle [1832] 1966). The late 1820s and 1830s saw the arrival of a group of women from St. Helena and three settlers from Denmark

461

462

IV. Structure and dynamics across language spaces and the Netherlands. The population grew rapidly and by 1832 there was a total of 34 people on the island, 22 of whom were young children. The 1830s and 1840s saw a renaissance of the whaling industry and numerous ships called at Tristan da Cunha to barter for fresh water and supplies; this led to the arrival of American whalers, some of whom settled permanently. The second half of the nineteenth century was a period of growing isolation, for a number of political and economic reasons. The whale trade declined quickly, the increasing use of steam ships made bartering unnecessary, and the opening of the Suez Canal in 1869 drastically reduced the number of ships in the South Atlantic Ocean. This affected the influx of settlers and a weaver from Yorkshire (Crawford 1945) and two Italian sailors were the only new arrivals in the second half of the century (Crabb 1980). The state of self-sufficiency and isolation lasted well into the twentieth century. When visiting the island in 1937, the Norwegian sociologist Peter Munch found that the Tristanians had not partaken in the massive changes that had occurred in the western world (Munch 1945) and noted that they basically lived in non-industrialized conditions. This situation came to an abrupt end in April 1942, when the British admiralty ordered the installation of a naval station on Tristan da Cunha. The arrival of the navy corps entailed farreaching economic changes and a South African company obtained exclusive rights to establish a permanent fishing industry on the island, employing practically the entire local workforce. The traditional subsistence economy was replaced by a paid labor force economy, and the traditional way of life was modified as a result of the creation of permanent jobs with regular working hours. Tristan da Cunha was an economic boomtown in the 1950s: the living conditions and housing standards improved and the changes brought about by the development scheme led to a complete transformation of the traditional Tristanian way of life within one generation. These social changes had linguistic consequences as local dialect features eroded via accommodation to and adoption of outside norms (Schreier 2002, 2003). In October 1961, unforeseen volcanic activities forced a wholesale evacuation of the entire population. The Tristanians had to leave the island and were transported to England; virtually all of them returned from their exile to the South Atlantic in 1963. The dramatic evacuation and the two “volcano years” in England affected the islanders more than any other single event in the history of the community. The community underwent quick modernization and adaptation to western culture. A new fishing company provided all the households with electricity; this improved the living conditions considerably and the 1970s and 1980s were a period of economic prosperity again, which led to an increase in mobility (mostly for secondary education and further job training) and a quick opening-up of the community. How did the settlement history and population dynamics shape the development of a local indigenous variety, and exactly what is the sociolinguistic status of Tristan da Cunha English as an overseas variety of English? As Zettersten (1969) points out, it is crucial that there was no indigenous population when the island was colonized. Consequently, the community’s founders found themselves in tabula rasa conditions and did not come into contact with pre-existing language varieties (a major difference to practically all locales in the Caribbean and Africa). The English input varieties to Tristan da Cunha English were dialects from the British Isles (the founders came from the Scottish Lowlands, East Yorkshire, East London and Hastings), the United States (from the New England area; Captain Andrew Hagan, the most influential American resident, was a native of New London, Massachusetts) and from St. Helena. The first languages of the

25. Overseas varieties non-British settlers were Afrikaans, Danish, Dutch and Italian. Several kinds of linguistic contact thus operated during the genesis and formation periods of Tristan da Cunha English, as the following varieties of English were transplanted to the island: British and American English (which led to koine´ization), second language (L2) forms of English (as spoken by the settlers from Denmark, the Netherlands and Italy) and St. Helenian English (StHE), which emerged in context of intense language contact involving English, various African and Asian languages and Malagasy (Schreier 2008). Some (e.g., Hancock 1991) have suggested that StHE may have undergone creolization, so that it is structurally similar to English-based creoles in the North Atlantic. This seems to be a strong claim, but the parallels (particularly in morphosyntax) are obvious; they are of high diagnostic value since there were practically no sociodemographic connections between St. Helena and the Caribbean (see Schreier 2008), excluding the possibility of direct transmission. As for, this variety primarily derives from varieties of British/late eighteenth-century American English and StHE, which served as the most important donors to this variety, and it can be excluded that Tristan da Cunha English emerged in a context of direct language contact. There is first-hand evidence that all the Tristan settlers had at least some knowledge of English (for instance, a visiting captain described the English of the Dutch settler as “excellent”, quoted in Brander [1940: 157]; cf. also Earle 1832; Taylor 1856), which excludes pidginization and creolization effects on Tristan da Cunha. On the other hand, these L2 forms had an impact on Tristan da Cunha English and several non-native features were adopted when the local variety nativized (th-sibilization, i. e., dental fricatives realized as /s/, as in think, throw, etc.). The existence of Creole-type features in Tristan da Cunha English (such as extremely high rates of consonant cluster reduction and absence of -ed past tense marking [Schreier 2005: 152]; /v/ realized as [b]; lack of word-order inversion in questions; copula absence; etc.) can only mean that a creolized form of English was transplanted via (at least some of) the women who cross-migrated in the 1820s from St. Helena. These women had some proficiency in English (Taylor [1856] reports that no other languages were spoken on Tristan in 1851) and they most likely spoke a (perhaps mesolectal) English-based Creole, so that these features were adopted by the first generations of native Tristanians. Zettersten (1969: 134⫺5) reaches a similar conclusion when stating that “the speechhabits of the settlers from St. Helena may have contributed towards simplifying the inflectional structures of the Tristan dialect.” Similarly, Cassidy (1974: 176), in his review of Zettersten (1969), states that Tristan da Cunha English “did not develop its pidgincreole as Pitcairn did, but appears to have imported many common creole features”. Processes of admixture and creole contact have been documented in other varieties of English around the world, and simplification as a result of extensive contact with a regularized or simplified variety is found in Afrikaans (Roberge 2002) or Singapore English (Platt 1975). Accordingly, Tristan da Cunha English is a prominent example of what is most aptly called creoloidization. In Trudgill’s (2000: 182) words, a creoloid demonstrates a certain amount of simplification and admixture, relative to some source language, […] which has never been a pidgin or a creole in the sense that it has always had speakers who spoke a variety which was not subject to reduction.

Thus, Tristan da Cunha English emerged in a context of dialect contact, yet witnessed admixture with an English-derived creole variety and had at the same time some input of ESL and learner varieties of English, which may well represent one of the most complex development histories of all the overseas varieties we know of.

463

464

IV. Structure and dynamics across language spaces

5. Conclusion What general conclusions can we draw from the massive expansion of language(s) on a global scale and how does the emergence and disappearance of overseas varieties challenge us to rethink the relationship between language and space? First of all, the unprecedented amount of language diffusion and spread out of Europe and into the “New World” led to the emergence of countless “new” varieties, shaped in diverse contact settings around the globe. These developments affected all the dimensions of space (geographic, cultural/social, political) and enriched the geo- and sociolinguistic landscape of European languages, as these came into contact with each other (most notably when two colonizing powers vied for the same territory, as in Vanuatu or South Africa) and of course with countless pre-existing local varieties. Regular interaction and subsequent accommodation fostered a variety of sociolinguistic outcomes (trade jargons, pidgins and creoles with European lexifiers, contact dialects, etc.). At the same time, the shaping of overseas varieties had a massive impact on the usage of local languages and is/was instrumental for the alarming rate of language endangerment and death (see the highly emotional debate about European “killer languages”; Skutnabb-Kangas 2000). By the same token, the last decades in turn have witnessed a decrease of many overseas varieties, some of which have already disappeared, for reasons such as renewed social mobility, in- and out-migration, or increasing stigmatization and nationalism (to mention many of the overseas varieties of German in the USA; Bonin/Ogasawara English in Japan, Long 2007; Riograndenser Hunsrückisch in Brazil; Unserdeutsch in Papua New Guinea; Cocoliche in Argentina; and many more). From a theoretical perspective, the search for a general framework to account for these varieties is now at the forefront of linguistics, yet the immensely complex processes that underlie colonization and its sociolinguistic conglomerates mar such efforts considerably. The problem is that so many contact settings are far from straightforward, that practically all colonies witness a steady coming and going, a frequent reshuffling of their sociodemographic set-ups, and that there is an ever-changing coexistence of language varieties (local and indigenous languages, social and regional varieties of the same language, etc.). As a result, all contact scenarios are “messy” and simply do not lend themselves to the neat separation and classification researchers hope to come up with. The scope and importance of overseas varieties for research on language and society is immense, however, affecting disciplines as diverse as contact linguistics, historical linguistics, language variation and change, language endangerment and death, typology and universals and many more. This is the legacy of the spatial expansion of language, and the emergence, evolution and disappearance of overseas varieties now has a central role in linguistic theory.

6. Reerences Bauer, Laurie 1994 Introducing the Wellington corpus of written New Zealand English. Te Reo 37: 21⫺28. Bauer, Laurie 1999 On the origins of the New Zealand accent. English World-Wide 20: 287⫺307.

25. Overseas varieties Beintema, Albert 2000 Early Shipping in Tristan da Cunha Waters. Available at , last accessed: May 6, 2009. Belich, James 1996 Making Peoples: A History of the New Zealanders from Polynesian Settlement to the End of the Nineteenth Century. Auckland: Penguin. Bennet, Norman R. 1984 Africa and Europe: From Roman Times to National Independence. 2nd ed. New York: Holmes & Meier. Bernard, John R. 1989 Regional variation in Australian English: A survey. In: Peter Collins and David Blair (eds.), Australian English: The Language of a New Society, 255⫺259. St Lucia: University of Queensland Press. Booij, Geert 2002 The balance between storage and computation in the language faculty. Science Prestige Lecture given at the University of Canterbury, Christchurch, 30 July. Brander, Jan 1940 Tristan da Cunha 1506⫺1902. London: Allen and Unwin. Britain, David 1997 Dialect contact and phonological reallocation: “Canadian raising” in the English Fens. Language in Society 26: 15⫺46. Cassidy, Fred 1974 Review of Arne Zettersten’s “The English of Tristan da Cunha”. Language 50: 175⫺177. Crabb, George 1980 The history and postal history of Tristan da Cunha. Self-published manuscript. Crawford, Allen 1945 I Went to Tristan. London: Allen and Unwin. Crystal, David 1997 English as a Global Language. Cambridge: Cambridge University Press. Curtin, Philip (ed.) 1995 African History: From Earliest Times to Independence. 2nd ed. Boston: Addison Wesley. Deverson, Tony (ed.) 1998 The New Zealand Oxford Paperback Dictionary. Auckland: Oxford University Press. Domingue, Nicole 1981 Internal change in a transplanted language. Studies in the Linguistic Sciences 4: 151⫺159. Duignan, Peter, Lewis H. Gann and Victor W. Turner (eds.) 1975 Colonialism in Africa, 1870⫺1960. Cambridge: Cambridge University Press. Earle, Augustus [1832] 1966 Narrative of a Residence on the Island of Tristan D’Acunha in the South Atlantic Ocean. Oxford: Clarendon Press. Evans, Dorothy 1994 Schooling in the South Atlantic Islands 1661⫺1992. Oswestry: Anthony Nelson. Gordon, Elizabeth and Tony Deverson 1998 New Zealand English and English in New Zealand. Auckland: New House Publishers. Gordon, Elizabeth, Lyle Campbell, Jennifer Hay, Margaret Maclagan, Andrea Sudbury and Peter Trudgill 2004 New Zealand English: Its Origins and Evolution. Cambridge: Cambridge University Press. Gründer, Horst 2004 Geschichte der deutschen Kolonien. 5th ed. Paderborn: Schöningh. Hammarström, Göran 1980 Australian English. Its Origins and Status. Hamburg: Helmut Buske.

465

466

IV. Structure and dynamics across language spaces Hancock, Ian 1991 St. Helena English. In: Francis Byrne and Thom Huebner (eds.), Development and Structures of Creole Languages. Amsterdam/Philadelphia: Benjamins. Holm, John 1989 Pidgins and Creoles. 2 vols. Cambridge: Cambridge University Press. Kachru, Braj B. 1985 Standards, codification, and sociolinguistic realism: The English language in the outer circle. In: Randolph Quirk and Henry G. Widdowson (eds.), English in the World, 11⫺ 30. Cambridge: Cambridge University Press. Kerswill, Paul 1996 Children, adolescents and language change. Language Variation and Change 8: 177⫺202. Kerswill, Paul and Anne Williams 2000 Creating a new town koine´: Children and language change in Milton Keynes. Language in Society 29: 65⫺115. Lass, Roger 1987 The Shape of English. Structure and History. London: Dent. Lass, Roger 1990 Where do extraterritorial Englishes come from? Dialect input and recodification in transported Englishes. In: Sylvia Adamson, Vivien Law, Nigel Vincent and Susan Wright (eds.), Papers from the 5th International Conference on English Historical Linguistics, 245⫺280. (Current Issues in Linguistic Theory 65.) Amsterdam/Philadelphia: Benjamins. Lawson, Philip 1993 The East India Company: A History. (Studies in Modern History Series.) London: Longman. LePage, Robert and Rene´e Tabouret-Keller 1985 Acts of Identity. Cambridge: Cambridge University Press. Long, Daniel 2007 English on the Bonin (Ogasawara) Islands. (Publication of the American Dialect Society 91.) Durham NC: Duke University Press. McGibbon, Ian (ed.) 2000 The Oxford Companion to New Zealand Military History. Auckland: Oxford University Press. Mencken, Henry Louis 1921 The American Language. New York: Knopf. Mesthrie, Rajend 1993 Koineization in the Bhojpuri-Hindi diaspora ⫺ with special reference to South Africa. International Journal of the Sociology of Language 99: 25⫺44. Mufwene, Salikoko 1996 The founder principle in creole genesis. Diachronica 13: 83⫺134. Mufwene, Salikoko 2001 The Ecology of Language Evolution. Cambridge: Cambridge University Press. Mühlhäusler, Peter 1997 Pidgin and Creole Linguistics. 2nd ed. (Westminster Creolistics Series 2.) London: Battlebridge. Munch, Peter A. 1945 Sociology of Tristan da Cunha. Oslo: Det Norske Videnskaps-Akademi. Olssen, Erik 1984 A History of Otago. Dunedin: McIndoe. Osterhammel, Jürgen 1995 Kolonialismus: Geschichte, Formen, Folgen. Munich: Beck. Pickens, Keith A. 1977 The origins of the population of nineteenth century Canterbury. New Zealand Geographer 33: 60⫺75.

25. Overseas varieties

467

Platt, T. 1975 The Singapore English speech continuum and its basilect “Singlish” as a “creoloid”. Anthropological Linguistics 17(7): 363⫺374. Quirk, Randolph 1990 Language varieties and standard language. English Today 21: 3⫺10. Roberge, Paul T. 2002 Afrikaans ⫺ considering origins. In: Rajend Mesthrie (ed.), Language in South Africa, 79⫺103. Cambridge: Cambridge University Press. Schneider, Edgar W. 2007 Postcolonial English. Varieties around the World. Cambridge: Cambridge University Press. Schreier, Daniel 2002 Past be in Tristan da Cunha: The rise and fall of categoricality in language change. American Speech 77: 70⫺99. Schreier, Daniel 2003 Isolation and Language Change: Sociohistorical and Contemporary Evidence from Tristan da Cunha English. Houndmills, Basingstoke/New York: Palgrave Macmillan. Schreier, Daniel 2005 Consonant Change in English Worldwide: Synchrony Meets Diachrony (Palgrave Studies in Language History and Language Change 3.) Houndmills, Basingstoke/New York: Palgrave Macmillan. Schreier, Daniel 2008 St Helenian English: Origins, Evolution and Variation (Varieties of English around the World Series.) Amsterdam/Philadelphia: Benjamins. Schreier, Daniel and Karen Lavarello-Schreier 2003 Tristan da Cunha. History People Language. London: Battlebridge. Schreier, Daniel, Elizabeth Gordon, Jennifer Hay and Margaret Maclagan 2003 The regional and sociolinguistic dimension of /hw-/ maintenance and loss in early 20th century New Zealand English. English World-Wide 24: 245⫺269. Siegel, Jeff 1987 Language Contact in a Plantation Environment. Cambridge: Cambridge University Press. Sinclair, Keith 1991 A History of New Zealand. 4th ed. Auckland: Penguin. Skutnabb-Kangas, Tove 2000 Linguistic Genocide in Education ⫺ or Worldwide Diversity and Human Rights? Mahwah, NJ/London: Lawrence Erlbaum Associates. Taylor, William F. 1856 Some Account of the Settlement of Tristan d’Acunha in the South Atlantic Ocean. London: Cassell. Timm, Uwe 2001 Deutsche Kolonien. Cologne: Parkland. Trudgill, Peter 1986 Dialects in Contact. Oxford: Blackwell. Trudgill, Peter 2000 Sociolinguistics: An Introduction to Language and Society. 4th ed. London: Penguin. Trudgill, Peter 2004 New-Dialect Formation: The Inevitability of Colonial Englishes. Edinburgh: Edinburgh University Press. Trudgill, Peter, Elizabeth Gordon, Gillian Lewis and Margaret Maclagan 2000 Determinism in new-dialect formation and the genesis of New Zealand English. Journal of Linguistics 36: 299⫺318. Zettersten, Arne 1969 The English of Tristan da Cunha. (Lund Studies in English 37.) Lund: Gleerup.

Daniel Schreier, Zürich (Switzerland)

468

IV. Structure and dynamics across language spaces

26. The consequences o migration and colonialism III: New minorities 1. 2. 3. 4. 5.

“Old” and “new” minorities The static concept of space in traditional dialectology Towards a multidimensional and dynamic model of migrants’ communicative space Conclusion References

1. Old and new minorities The concept of “minority” seems to be based on the sheer quantitative relation of “being less” than the representatives of a corresponding “majority” even though this basic definition is completely inadequate when discussing issues within the context of humanities. While statisticians continuously classify the populations into groups of greater or smaller size, only very few of the smaller groups are actually considered “minorities”. The underlying reason is that this concept implies a specific sociopolitical status which is the product of public communication. Numerous factors contribute to the construction of a minority in a given society. The two extreme opposites are: imposition from above, by means of administrative implementation, without consent or even against the will of the concerned group; or minority status claimed by the group itself without being granted by the state authorities. The middle way is to create a legal framework which allows a certain number of communities to define themselves as a minority or as belonging to the majority. This compromise was chosen by Italy in 1999 when it adopted Legge 482 regarding “norme in materia di tutela delle minoranze linguistiche storiche” (‘norms in the matter of protection of the historical linguistic minorities’; ; cf. Orioles 2003). It should be noted, however, that only “old” minorities are recognized by modern states (Orioles 2003: 50⫺57 and 2006). Often their origins predate the rise of the nationstate (e.g., the Basques, the Welsh); more specifically these minorities have existed long before a majority was established. “New” minorities, on the contrary, may be defined as groups that emerged after the formation of the nation-state, normally as the result of demographic mobility, either voluntary migration or forced displacement (as in the case of the German speaking population of the Volga region or of the Crimean Tartars who resettled in Kazakhstan and in Uzbekistan in 1941 and 1944). Occasionally, the formation of a new state can create “new” minorities without migration, simply by redefining the majority (as experienced by the numerous Russian speaking people residing in the Baltic States after 1990, especially in the large cities like Tallinn). The political privilege of “old” minorities is not surprising; it would be impossible, for pragmatic reasons, to grant the status of a national minority immediately to the numerous groups of immigrants, increasingly attracted to countries of relative economic and/or political stability. The direction of this migration is partly conditioned by colonial history (as exhibited by immigration to France, Belgium, Great Britain or the Netherlands). There are, however, a number of other scenarios as well; some migratory move-

26. New minorities ments have a long tradition, such as the migration from northeastern Italy (Friuli) to southern Germany which started in the middle of the nineteenth century (see Melchior 2006), while others have only developed recently. Some of these cases were not even imaginable twenty years ago. In Europe at least, many nations which for generations represented classical push areas of migration, such as Portugal, Spain or Italy since the nineteenth century, have recently become strong pull regions of mass immigration. A systematic survey of new European minorities is not yet possible due to a lack of basic linguistic information. The three nations mentioned above, however, would suffice in order to exemplify the main linguistic interest in migratory movements, which is to observe processes of language change and language acquisition as in a laboratory. These processes occur simultaneously with the reorganization of the communicative space; they are characteristic of migratory contexts because of their frequency and density in these contexts; but of course, they may occur in other speech communities as well, particularly in regions which are undergoing rapid dialect leveling. Among others, the following two factors must be taken into account: the similarity of the varieties in contact and the circumstances of language acquisition.

1.1. The similarity o the varieties in contact The continuum of contact varieties extends from close cognate languages to completely unrelated ones (such as Turkish and German; see Rehbein 2001). A special kind of language change is decreolization; this change takes place when creole speakers come in close contact with the original lexifier language of their creole as a consequence of immigration into the space of this language. An example is the presence of the Capverdians in Portugal (numbering around 100,000), particularly in certain neighborhoods of Lisbon (see Märzhäuser 2006); decreolization can also be observed in the case of the Haitians in the French-speaking Province of Quebec or in the case of the Caribbeans who speak English-based creoles in London. Obviously, contact phenomena are not restricted to (varieties of) the imported and the autochthonous languages; these phenomena may also lead to convergence between diverse varieties of the imported language, such as in the case of migrants from different parts of southern Italy (Sicily, Campania, Puglia or Sardinia) who live in Germany, Belgium and Australia or those Italian dialect speakers who migrated to North (see Haller 1993) or South America (such as Argentina and Brazil) during the first decades of the last century. In such language contact situations, koine´ization is expected to set in with “new” varieties emerging from diverse diatopic input as has been the case in the past. This is best exemplified by the settlement of “Swabian” Germans into the middle Danube regions of southern Hungary and the northern regions of former Yugoslavia und southwestern Romania.

1.2. The age o L2 acquisition Restricting our definition of a migrant to the literal sense, we consider only those people who pass from a source region into a host region (i. e., the “first generation”); very often these migrants are adults and consequently gain only an elementary L2 competence of

469

470

IV. Structure and dynamics across language spaces the varieties spoken in their host society. Due to the linguistic situation at his or her place of employment, an adult migrant without any educational support or control will acquire a completely different variety of the host language ⫺ a pidgin-like interlanguage which, under certain historical circumstances, may pass from the status of an idiolect to that of a variety used by a social group. Exceptions are cases of rapid acquisition (possibly fossilization) due to the fact that the L2 belongs to the same language family as the L1. A possible example is that of the Romanian immigrants in northern Spain (see Sandu 2000; S¸erban and Grigoras¸ 2000; Radu 2001; Lungescu 2006; Escartı´n and Pinos Quı´lez 2005); this group is particularly interesting as many immigrants have predominantly taken up residence in little villages in the Pyrenees rather than in the large cities (as done by the overwhelming majority of migrants in search of employment). In 2007, 4,389 officially registered Romanians (compared to 2,569 in 2005) were residence only in the province of Huesca. Those Romanian peasants, unlike their compatriots living in the cities, have won the acceptance of the autochthonous population because they contribute to the maintenance of the local rural traditions. The children of immigrants, on the contrary, may acquire two L1s, the language of their parents and that of the host society. Yet even for these bilinguals, bilingual literacy is crucial for the organization of their repertoire and the development of language awareness. Without this support, these speakers will not be able to distinguish standard from non-standard forms, and sometimes it may even be unclear to them to which language a certain word belongs (for examples see Krefeld 2004: 75). The age of L2 acquisition and the structural distance between the contact languages seem to determine the linguistics of immigration. Yet, in contemporary societies, the determination of an individual speaker is never absolute, which means that linguistic data need to be analyzed in close relation to the communicative space, in which the respective speaker is moving.

2. The static concept o space in traditional dialectology At first glance, a spatial approach to new minorities might seem strange, as the concept of space used in traditional dialectology since the nineteenth century is extremely restrictive and naturalistic. In fact, the disregard for migration and new minorities in the dialectological paradigm is in part a reflection of nationalist politics since the notion of space is central for the concept of the European nation-state. The fact that “old” minorities are much more readily accepted than those resulting from recent migration is furthermore grounded in the ideological value of an indigenous and resident population inhabiting its own space. The established political system of a prototypical nineteenth-century nation and its corresponding scientific traditions were founded on exactly the same ideology of space. It is no surprise, therefore, that dialectology ⫺ i. e., the modern linguistic approach to space ⫺ established itself as an academic discipline during the same period which saw the rise of the European nation-state. The conception of linguistic space applied in the research of the day was that of traditional (pre-industrial and rural) landownership. Not even the slightest attention was paid to the consequences of industrialization and urbanization which would not have been possible without mass migration. For instance,

26. New minorities the empirical data of the first linguistic atlas (Wenker’s Sprachatlas von Nord- und Mitteldeutschland auf Grund von systematisch mit Hülfe der Volksschullehrer gesammeltem Material of 1881) was collected right in the area of the Ruhr during its exploding industrialization ⫺ without reflecting upon this demographic revolution (see Krefeld 2002b). The mainstream of traditional dialectology stuck to this static, one-dimensional and highly selective conception of linguistic space. The result was a sterile and closed paradigm of geolinguistic work, which in the long run became isolated in the context of linguistics. To sum up, this kind of dialectology was not an appropriate paradigm for facilitating and encouraging research on more dynamic types of linguistic space or on spatial dynamics itself. For this reason, new minorities have been excluded from traditional spaceoriented linguistics. Nonetheless, the linguistic relevance of space is evident in their case. It may even be said that the linguistic investigation of new minorities generally suggests a new framework of communicative space for the description of linguistic variation as will now be shown.

3. Towards a multidimensional and dynamic model o migrants communicative space The basic challenge is to deal with scenarios of complex linguistic variation in unconventional constellations of varieties. As a typical example consider the German speaking territory of northern and eastern Switzerland, where dialects of different (partly cognate) migrant languages come in contact with one another as well as with various Swiss German dialects and Swiss standard German. Among speakers of southern Italian dialects, Portuguese, Galician (a Portuguese dialect spoken in Spain), Serbian varieties, Greek or Turkish sharing a common workplace with indigenous Swiss citizens evidence shows that not only German, but also more or less simplified varieties of Italian are used as an interethnic lingua franca (cf. Berruto, Moretti and Schmid 1990; Schmid 1994: 26⫺36). Note that all migrant contact varieties depicted thus far are spoken in “roofless” situations, i. e., in situations where the standard variety they belong to (e.g., standard Italian) is not implemented and is scarcely present. On the other hand, the relevance of a lacking standard roof has diminished over the last years: the extensive availability of digital media (internet) in large parts of the world in connection with informal writing (sms), and the transnational reach of satellite TV, has increased the spatially unbounded validity of standard varieties and varieties close to the standard. (And in a certain way it is true that the whole history of the media since the invention of writing can be seen as a process of emancipation of communication from the spatial conditions of interaction.) It is therefore necessary to resort to concepts such as language contact, diatopics, diastratics, pragmatics, orality and media use for the description of these minorities and their language. None of these terms, however, can exhaustively cover the situation due to the extremely heterogeneous (and unpredictable) linguistic behavior in the communities concerned. The term “minority” suggests elementary homogeneity and is thus misleading. Even speakers sharing an ethnic-linguistic background and living in largely similar social and regional contexts may behave in completely different ways. In migratory contexts,

471

472

IV. Structure and dynamics across language spaces the “imported” identity may be weakened or, on the contrary, strengthened in reaction to the acceptance or rejection of the migrants by the host society (see Melchior 2006, 2009). Any attempt to hypothesize a simple match of minority membership and linguistic variation patterns would be naı¨ve. The consequence of this reasoning, however, cannot be to restrict linguistic description to individual migrants’ speech production; the data should rather represent a reliable empirical base for reconstructing patterns or at least prototypes of variation while bearing in mind that the communicative space in which those data were produced must be systematically taken into consideration. Firstly, one will have to consider the data in the narrow context of the informant’s egocentric networks (see Tempesta 2000), secondly in the wider area and territory in which these networks are embedded. In any case, the description must include: (1) the repertoire of the central informants (original space of speaker [⫽ S]); (2) the varieties dominantly used in everyday networks (non-distant speaking space [⫽ N]); (3) a. the varieties used by the indigenous population of the place of residence (space of the autochthonous idiom [⫽ A]); b. the official (standard) language and varieties near to the standard of the nation or region the speaker lives in (“territoriality” or space of the standard language [⫽ T]).

Fig. 26.1: The three basic dimensions of the individual communicative space

These three “objective” factors are necessary but not sufficient neither for classifying the communicative spaces, nor for understanding the particular dynamics they generate, which is the main interest of a linguistic description. It is also important to reconstruct on a second “subjective” level how the speech is perceived by the speakers themselves. With regard to the individual space we have to distinguish: (4) a. the auto-perception of the speaker’s own production and b. the corresponding hetero-perceptions. In other words, we must know whether the migrant speaker is categorized as such or as an autochthonous speaker and whether this categorization is in accordance with his or her own perception. In addition, the question arises of how non-migrant speakers of the

26. New minorities migrant’s L1 perceive their production: does it sound authentic or “strange”, maybe having a specific accent? (For instance, Italians living in Germany are sometimes called germanesi by Italians living in Italy as their Italian pronunciation appears to show a sort of German “accent”.) Above all, however, it is essential to know how the migrant perceives the way in which the autochthonous population speaks as well as the way in which the members of his or her networks speak. Without taking this factor into account, it is impossible to understand in which direction the speakers will adapt their repertoire: by acquiring new varieties or by changing the varieties they already use. This adaptive impact of perception is nevertheless an indirect one; perceptions are filtered by complex mental representations which are partly based on objective language phenomena, and partly reproduce cultural and social stereotypes projected onto linguistic features or varieties (see Pustka 2007: 9⫺11, 2008). Note that a common language does not guarantee social cohesion. This is illustrated in the numerous groups of Italians who live in Munich (about 22,000 people), yet are socially deeply split according to their origins (whether they come from northern and southern Italy). Some of them do not even consider themselves migrants because they think they left their home country on their own accord.

Fig. 26.2: Dynamic frame of communicative space

The empirical application of the model allows us to position each individual migrant’s speech and to profile his or her communicative space. Dimensions 1⫺3 (S-repertory, Nvarieties, TA-varieties) are closely interrelated while the dimension 4 (perception) is key to understanding these interrelations.

473

474

IV. Structure and dynamics across language spaces The communicative spaces of migrants are always more or less dissociated (Krefeld 2002a) as the individual repertoire is normally congruent neither with that of the speaker’s communicative networks nor with that of the indigenous speakers. Dimensions 1 and 2 prove to be crucial in sketching the dissociated migrant spaces. The isolated migrant speaker is below the threshold of a minority: s/he is only loosely, if at all, integrated in networks, in which varieties are used that correspond to badly spoken L2, L3 or L4 varieties in his or her own repertory. In addition, the network varieties do not necessarily belong to the autochthonous language.

Fig. 26.3: Communicative space of an isolated migrant

Even though this specific case of double dissociation is somewhat extreme, it is nevertheless characteristic for contemporary societies, and perhaps not only in the context of a metropolis. Little is known about how communication works in networks of isolated speakers of different linguistic backgrounds. These networks offer good conditions for the formation and stabilization of local pidgin varieties as they are used in certain neighborhoods of the large cities; a common group identity may develop out of these networks and in that case the conscious effort to overcome isolation may lead to an emergent “new minority”. The other extreme of the range is marked by speakers having acquired only a very basic, perhaps only passive, competence of the L1 of their parents. Their communicative networks are largely or exclusively dominated by the L1 varieties of the autochthonous speakers. The awareness of belonging to a minority may outlast the use of the original language and manifest itself in the emergence and maintenance of ethnolectal varieties within the autochthonous dialect and/or the standard language of the respective territory. The prototypical representatives of “new minorities” operate in communicative spaces which are at least partially bilingual; such communicative spaces are based on networks which maintain the allochthonous varieties brought from the countries of origin. Nevertheless, these groups normally consist of individual speakers whose repertoires depend on age (first or later generations), on the professional career, quite often on gender and on biographical and individual contingencies. Even within a single network heterogeneous speakers are often united: i.

monolinguals with their “imported” L1, mostly of the first generation; these are usually women who do not work;

26. New minorities

Fig. 26.4: Communicative space of a migrant in an “ethnic” community

ii.

bilinguals with their “imported” L1 and a more or less advanced oral L2 competence; the spoken L2 variety can also be the local dialect of the place, acquired at the work place; iii. perfect bilinguals with two L1s; iv. bilinguals with the L1 of the country/region of residence and a more or less advanced competence of the “imported” L2 (which can also be dialect); v. quasi-monolinguals with an L1 of the country/region of residence and a certain passive knowledge of the dominant “imported” variety of their parents or siblings; note that the first child of migrant families is very often bilingual whereas the following ones may belong to this latter, quasi-monolingual type. With regard to this complex situation, we can then define minority networks as intersections of different individual communicative spaces. It is important to observe that the members of such networks, apart from those of the balanced type (iii), represent opposite scenarios of language contact. Speakers of the types (i) and (ii) tend to impose phonetic and morphosyntactic structures of the “imported” language on the L2-varieties of the host society. The direction of interference is inverted for speakers of the types (iv) and (v). The preferred borrowing of lexical elements from L2 to L1 will show the same inversion (for the opposition of “imposition” versus “borrowing” see Thomason and Kaufmann 1988; Guy 1990; Thomason 1997, 2001).

4. Conclusion Although the term “new minority” is frequently used, it should have become clear that this term is misleading as it suggests the existence of a consistent social unity. In reality, however, we are dealing with a multitude of more or less instable social groups in a permanent state of restructuring. Most of these groups, and especially the most dynamic ones, are located in the cities. The approach outlined above allows us to integrate urban dialectology and urban sociolinguistics into a common model of multidimensional communicative spaces, which are far from “metaphorical”, considering how migrants organize themselves in their own manner by choosing or at least accepting certain spatial structures. In their area of residence, migrants may constitute the majority within certain neighborhoods (such that cities resemble ethnic mosaics with a Chinatown, a Little Italy, a German quarter, etc.). Even when cities or states take measure to hinder the formation

475

476

IV. Structure and dynamics across language spaces of such ghettos (see Krefeld 2006), local ethnic micro-structures, such as preferred housing areas, marketplaces or special roads with “ethnic” stores (groceries, hairdressers, laundries, restaurants, etc.) are still recognizable. An interesting indicator of the “appropriation” of public space in this respect is the renaming of important and highly frequented points of interest, such as markets, bus stops, stations or churches (Amoruso 2002, 2008; D’Agostino 2006). In conclusion, the spatial distribution and organization of individual migrants and of groups belonging to the “new minorities” are of great diagnostic value. A respective map reveals how society (or at least parts of it) integrates these extremely divergent individual communicative spaces. And this exactly should be the future task of “new minority” linguistics: the representation of complex multilingual areas in an atlas of communicative ecology. A first, purely sociological and very schematic attempt at such a map of “human” ecology was made by Ernest Burgess [1925] of the Chicago School of sociology. Unfortunately, however, this approach did not find any linguistic echo, although it merits reconsideration and systematic elaboration (Krefeld 2006).

5. Reerences Amoruso, Chiara 2002 La comunita` ivoriana a Palermo. Frammenti stranieri di una immagine urbana. In: D’Agostino (ed.): 111⫺133. Amoruso, Chiara 2008 Tunisians in Sicily and migration dynamics. Urban settings in comparison In: Krefeld (ed.): 127⫺144. Berruto, Gaetano, Bruno Moretti and Stephan Schmid 1990 Interlingue italiane nella Svizzera tedesca. Osservazioni generali e note sul sistema dell’articolo. In: Emanuele Banfi and Patrizia Cordin (eds.), Storia dell’italiano e Forme dell’italianizzazione, 203⫺228. (Atti del XXIII Congresso Internazionale di Studi, Trento, Rovereto 18⫺20 maggio 1989.) Rome: Bulzoni. Burgess, Ernest W. 1967 [1925] The growth of the city. In: Robert E. Park, Ernest W. Burgess and Roderick D. McKenzie (eds.), The City, 47⫺62. Chicago/London: University of Chicago Press. D’Agostino, Mari 2006 Segni, parole, nomi. Immagini della Palermo plurietnica. In: De Blasi and Marcato (eds.), 207⫺222. D’Agostino, Mari (ed.) 2002 Percezione dello Spazio e Spazio della Percezione. La Variazione Linguistica fra Nuovi e Vecchi Strumenti di Analisi. Palermo: Centro di Studi Filologici e Linguistici Siciliani. De Blasi, Nicola and Carla Marcato (eds.) 2006 La citta` e le sue lingue. Repertori linguistici urbani. (Domini: Quaderni del bollettino linguistico campano 3.) Naples: Liguori. Escartı´n Sese´, Javier and Manuel Pinos Quı´lez 2005 Encuentro en el Espejo. Inmigrantes y Emigrantes en Arago´n. Historias de Vida. Zaragoza: Fundacio´n Seminario de Investigacio´n para la Paz. Guy, Gregory R. 1990 The sociolinguistic types of language change. Diachronica 7: 47⫺67. Haller, Hermann W. 1993 Una Lingua Perduta e Ritrovata: L’Italiano degli Italo-Americani. (Biblioteca di italiano e oltre 14.) Scandicci, Florence: La Nuova Italia.

26. New minorities Krefeld, Thomas 2002a La dissociazione dello spazio comunicativo in ambito migratorio (e come viene percepita dai parlanti): i meridionali in Baviera. In: D’Agostino (ed.): 157⫺172. Krefeld, Thomas 2002b Migration, Sprachbewußtsein und Wissenschaftsideologie: Über dynamische Räume und ihre statische Beschreibung. In: Konrad Ehlich and Venanz Schubert (eds.), Sprachen und Sprachpolitik in Europa, 145⫺170. Tübingen: Stauffenburg. Krefeld, Thomas 2004 Einführung in die Migrationslinguistik: Von der Germania italiana in die Romania multipla. Tübingen: Narr. Krefeld, Thomas 2006 Baviera di Chicago? ⫺ Repertori, reti e spazio urbano. In: De Blasi and Marcato (eds.), 223⫺243. Krefeld, Thomas (ed.) 2008 Sprachen und Sprechen im städtischen Raum. (Spazi comunicativi ⫺ Kommunikative Räume 4.) Frankfurt am Main: Lang. Lungescu, Oana 2006 A tale of two European villages. Aguaviva and Peretu. Available at ; last accessed 29 December 2006. Märzhäuser, Christina 2006 Entre Kriolu e Portugueˆs. A situac¸a˜o de migrantes Caboverdianos em Grande Lisboa ⫺ Um estudio de caso em Alta Cova da Moura. In: Thomas Krefeld (ed.), Modellando lo spazio in prospettiva linguistica, 173⫺194. (Spazi comunicativi ⫺ Kommunikative Räume 1.) Frankfurt am Main: Lang. Melchior, Luca 2006 Lingua, integrazione e spazi comunicativi in ambito migratorio: Una ricerca sui friulani in Baviera. Lingue e idiomi d’Italia (LIDI) I(2): 67⫺103. Melchior, Luca 2009 “Su` pes Gjermaniis”. Zwischen Dissoziation und Integration: Kommunikationräume friaulischer Einwanderer in Bayern. (Spazi comunicativi ⫺ Kommunikative Räume 6.) Frankfurt: Lang. Orioles, Vincenzo 2003 Le Minoranze Linguistiche. Profili Sociolinguistici e Quadro dei Documenti di Tutela. Rome: Il Calamo. Orioles, Vincenzo 2006 Nuove identita` negli agglomerati urbani: Verso il riconoscimento? In: De Blasi and Marcato (eds.), 69⫺81. Pustka, Elissa 2007 Phonologie et Varie´te´s en Contact. Aveyronnais et Guadeloupe´ens a` Paris. (Romanica Monacensia 75.) Tübingen: Narr. Pustka, Elissa 2008 Accent(s) Parisien(s) ⫺ Auto- und Heterorepräsentationen stadtsprachlicher Merkmale. In: Krefeld (ed.): 213⫺245. Radu, Cosmin 2001 De la Craˆngeni-Teleorman spre Spania: antreprenoriat, adventism s¸i migrat¸ie circolatorie [From Craˆngeni-Teleorman to Spain: Management, adventism and circular migration]. Sociologie Romaˆneasca˘ 1(4): 215⫺231. Rehbein, Jochen 2001 Turkish in European Societies. Arbeiten zur Mehrsprachigkeit ⫽ Working Papers Series B, No. 25 SFB 583, Universität Hamburg. Sandu, Dumitru 2000 Migrat¸ia transnat¸ionala˘ a romaˆnilor din perspectiva unui recensa˘maˆnt comunitar [The Romanians transnational migration from a communitarian survey’s perspective)]. Sociologie Romaˆneasca˘ 3(4): 5⫺52.

477

478

IV. Structure and dynamics across language spaces Schmid, Stephan 1994 L’Italiano degli Spagnoli. Interlingue di Immigrati Nella Svizzera Tedesca. Milan: Angeli. S¸erban, Monica and Grigoras¸, Vlad 2000 Dogenii din Teleorman ˆın t¸ara˘ s¸i ˆın stra˘intate. Un studiu asupra migrat¸iei circulatorii ˆın Spania [The Dogeni from Teleorman in Romania and abroad. A study on circular migration to Spain]. Sociologie Romaˆneasca˘ 2: 30⫺54. Tempesta, Immacolata 2000 Varieta` della Lingua e Rete Sociale. Milan: Franco Angeli. Thomason, Sarah Grey 1997 A typology of contact languages. In: Arthur K. Spears and Donald Winford (eds.), Pidgins and Creoles: Structure and Status, 71⫺88. Amsterdam: Benjamins. Thomason, Sarah Grey 2001 Language Contact. Washington DC: Georgetown University Press. Thomason, Sarah Grey and Terrence Kaufmann 1988 Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Wenker, Georg 1881 Sprachatlas von Nord- und Mitteldeutschland auf Grund von systematisch mit Hu¨lfe der Volksschullehrer gesammeltem Material aus ca. 30 000 Orten. Straßburg: Trübner.

Thomas Krefeld, Munich (Germany)

27. Non-convergence despite language contact 1. 2. 3. 4. 5. 6. 7.

Introduction Contact, language and (non-)convergence Some general considerations Some examples for structural factors in (non-)convergence Some examples for sociolinguistic factors in (non-)convergence Conclusion References

1. Introduction Most research into the linguistic consequences of the contact of two or more languages has been concerned with contact-induced convergence (or advergence; cf. Mattheier 1996: 34). Much less scientific effort has been dedicated to the equally important and theoretically perhaps even more interesting question why convergence does not happen (but cf. Martinet’s [in Weinreich 1953: viii] historically interesting counter-claim that “linguistic research has so far favored the study of divergence [e.g., the Romance languages; G. K.] at the expense of convergence”). At least partly, the reason for this neglect seems to be that convergence implies change, whereas the lack of it often implies stasis, which may be considered less interesting (cf. Backus 2005: 328). Besides this, it is rather

27. Non-convergence despite language contact easy to explain contact-induced convergence, firstly because the target structure of such a change is known and can be easily analyzed, and secondly because the sociolinguistic motivation for contact-induced convergence ⫺ normally a high(er) overt prestige of the source language (cf. Auer and Hinskens 1996: 11⫺12) ⫺ can be directly connected to extralinguistic factors such as the number of its speakers and their material, political and/or cultural power. For non-convergence, on the other hand, there is neither a structural target available for analysis, nor does there exist a similarly obvious connection to extralinguistic factors (see section 2.3). In spite of these difficulties, I will attempt to shed some light on sociolinguistic and structural factors for non-convergence despite language contact. First, however, it is necessary to discuss the central concepts of this article, namely contact, language and (non-)convergence.

2. Contact, language and (non-)convergence 2.1. Contact After fifty years of research in sociolinguistics it is commonplace knowledge that the term language contact is a somewhat misleading metaphor and should be replaced by contact of speakers. But even contact of speakers is a rather superficial description for what is happening when speakers of one language borrow linguistic material, structures or rules from speakers of another language. Because of this, Martinet’s (in Weinreich 1953: viii) conviction that “contact breeds imitation and imitation breeds linguistic convergence” has to be questioned. German and French-speaking people, for example, have been in contact for many centuries, but speakers of German have incorporated much more French vocabulary (not to mention French phonemes, stress patterns and grammatical structures) than vice versa. Usually, this imbalance has been explained by the high overt prestige (speakers of) French enjoyed for several centuries. This line of reasoning seems sensible enough, but it overlooks the important detail that French-speaking people did not even have the chance to borrow more German words or structures simply because they did not know enough German in order to do so. The contact of German and French did not happen somewhere along the Rhine River but in the minds of a substantial number of Germans who felt the need to learn French (cf. Weinreich’s (1953: 6, 67) dictum that “the individual is the ultimate locus of contact”). Obviously, the reason for learning French was its high prestige, but this prestige was only one condition for borrowing, the other condition ⫺ and one may say the more important one ⫺ was bilingualism. On the eastern edge of the German-speaking area, the situation was quite different. Here it was West Slavic speakers who learned German and borrowed German words and rules (e.g., initial stress in Czech words), while these languages hardly affected German. Nowadays, the international prestige of French, German and the West Slavic languages is low, at least lower than that of English. The number of French speakers in Germany has been decreasing and there have never been that many speakers of German in France or of Polish or Czech in Germany. Even the previously large number of speakers of German in Poland and the Czech Republic is diminishing (but cf. the still high numbers in StADaF 2006: 2, 14, footnote 18). The consequence of the shrinking number of speakers of the other languages is the almost complete lack of current mutual borrowing along the French-German and German-Polish/Czech borders. But in all these coun-

479

480

IV. Structure and dynamics across language spaces tries there are many speakers of English and, consequently, there are many English words and structures being borrowed into French, German, Polish and Czech in spite of the fact that none of these countries shares a common border with an English-speaking country. From this, we can conclude that geographical contact of speakers of different languages does not automatically imply language contact (Martinet’s “breeds” should, therefore, be replaced by a more cautious “can breed”) and that the lack of geographical contact does not automatically impede language contact, especially in an increasingly globalized and media-struck world.

2.2. Language I will not dwell long on the rather fruitless attempt to distinguish languages from dialects on structural grounds (but cf. some measures of distance in Bechert and Wildgen 1991: 105⫺108). Suffice it to say that it is not at all clear whether Spanish and Portuguese, two different languages, are linguistically more different than Standard German and Low German, a standard and a dialect of the same language (cf. Weinreich 1953: 105), or whether Low German might not, in some respect, be considered closer to Dutch or English than to Standard German. In spite of doubtful or even absurd groupings (Low German as a German dialect; Catalan as a Castilian dialect under Franco; Nahuatl as a dialect, albeit not a Spanish one, in Mexico), one can expect linguistic consequences from such politically motivated categorizations. For example, had the Catalan people not had such a long and independent political and cultural history and, even more importantly, had they not had such a strong economy, Catalan might not have survived as an autonomous language but converged towards Castilian (cf. the somewhat different case of Galicia [Villena Ponsoda 2006: 1807]). Converging languages can become dialects (e.g., Aragonese and Leonese in the Iberian Peninsular) and diverging dialects can become languages (the current developments separating Serbian and Croatian [Auer, Hinskens and Kerswill 2005b: 7]) or develop into varieties of different languages (the dialects along the Dutch-German border [Auer and Hinskens 1996: 16]). Regardless of these considerations, it should be clear that for the analysis of convergence and non-convergence in language contact it is more important to gauge the difference in linguistic subsystems than the one between languages per se. From this point of view, Low German word order is closer to Dutch and Standard German than to English, whereas its consonantal system and nominal morphology are closer to English and Dutch than to Standard German.

2.3. (Non-)convergence With regard to (non-)convergence in language contact, five points have to be made: Firstly, Salmons (1990: 454) writes that “the now common use of ‘convergence’ goes at least back to Weinreich […], where it is understood as ‘partial similarities increasing at the expense of differences’”. Weinreich sees convergence as a process, but convergence can also describe the result of a (past) language contact. What these two views have in common is that both describe the linguistic changes due to language contact and not the mechanism which causes them, i. e., borrowing. Poplack (1993: 256) defines both,

27. Non-convergence despite language contact convergence and borrowing, as mechanisms of change: “[c]onvergence also involves the process of borrowing, although we reserve this term for the transfer of grammatical structure […]”. The difference she makes between the two concepts lies in their extension: convergence is equivalent to structural borrowing whereas borrowing proper is reduced to the lexicon. For the purpose of this article, I will consider convergence as the consequence of (past/ongoing) change which occurs when speakers of one language borrow linguistic material, structures or rules from another language, regardless of the linguistic level involved. Non-convergence, consequently, is the result of a lack of such a transfer. This can, but need not imply linguistic stability. Secondly, not divergence but non-convergence is the opposition of convergence (cf. Auer and Hinskens 1996: 3). Divergence is just one element of non-convergence and in language contact, it is probably a rather rare one (see section 5.2). Thirdly, from a purely structural point of view, even languages whose speakers never come into any type of contact can converge or diverge. There are/were, for example, tendencies in some Germanic varieties to develop an adpositional case. Kaufmann (2008: 94⫺95, footnote 4) describes the use of the definite article dem in Mennonite Low German in Texas. Dem (etymologically dative) appears more frequently in nominal phrases governed by prepositions than in those governed by verbs. Lass (1992: 112) describes a similar case with regard to the suffix -e on Middle English nouns. Such developments make these varieties converge structurally with Hindi, which, too, has a specific adpositional (postpositional) case. But obviously, there is no connection whatsoever between Mennonite Low German, Middle English and Hindi and, therefore, this exclusively structural convergence (or divergence from other Germanic varieties) may be interesting for typologists, but not for us. From a structural point of view, four different types of non-convergence can be distinguished, namely contact-induced divergence, contactinduced simplification (mostly an indirect consequence of the functional loss of lowprestige languages, cf. Silva-Corvala´n 1991: 165⫺166), linguistic stability (which may be contact induced; cf. Labov’s [2001: 297] sociolinguistically motivated “retreat of lower working class males from a female-dominated change” or Thomason and Kaufman’s [1988: 58] structural conserving influence of contact languages) and non-converging endemic change (often linguistic drift or change towards less marked forms). Fourthly, convergence is not an independent linguistic phenomenon. It is the consequence of and interacts with other contact phenomena, such as code-switching, language attrition, second language acquisition, bilingual priming and linguistic accommodation (cf. Muysken 1995: 188; Backus 2005: 315; Poplack 1993: 255). In section 2.1 it was claimed that bilingualism, the consequence of second language acquisition, is a highly important condition for convergence, be it lexical, semantic or structural. For structural convergence, bilingual priming might be a decisive cognitive factor. Learning a second language creates or activates structures and mental representations hitherto unknown to the learner or less frequently used. Misfiring in language production, i. e., the priming of structures in the speaker’s first language by structures of the newly acquired second language, is bound to happen in such a situation (cf. Loebell and Bock [2003] and Backus’ [2005: 326] similar concept of entrenchment, but also the doubts mentioned in Poplack, Walker and Malcolmson [2006: 208]), especially when code-switching and/or linguistic accommodation to the speakers of the contact language are frequent. Language attrition can lead to negative convergence (different from non-convergence; cf. Auer and Hinsken’s [2005: 354] concept of negative accommodation), because marked

481

482

IV. Structure and dynamics across language spaces structures, which are often language specific, disappear first in language attrition (cf. Silva-Corvala´n [1991, 165] and, for dialect contact, Schirmunski’s [1930] concept of primary dialect features). Negative convergence causes differences between the contact languages to be leveled not due to the addition of a new word or rule in the converging language but due to the loss of an old word or rule (cf. Salmons 1990 with regard to discourse markers in Texas German). In the case of non-convergence, one can assume that most other contact phenomena mentioned do not appear or only appear in a reduced form. One would expect such a constellation either when the contact of the groups is conflictive or when the languages involved are of comparable prestige. The latter case can be currently found on the French-German border. The former one is interesting because it shows that although convergence and divergence are contrary linguistic reactions to language contact, they both occur in comparable sociolinguistic settings; both presuppose the contact with a language of high overt prestige. After all, why should speakers change their verbal behavior because they come into contact with a language they consider inferior? Unless, of course, the speakers of this supposedly inferior language start converging. Therefore, the different linguistic behavior of converging and diverging speech communities, i. e., the existence of language loyalty or the lack of it, must be explained by the characteristics of these communities (cf. section 5.2). From a sociolinguistic point of view, these questions are quite interesting, but they do not tell us much about structural restrictions on convergence or about less obvious sociolinguistic factors. This leads us to the fifth and final point: With regard to linguistic theory, non-convergence in language contact is most interesting when we find non-converging speakers or non-converging structures in a situation where we would normally expect convergence. In the light of this, only two of the four types of non-convergence mentioned in point 3 are of interest to us, namely linguistic stability and contact-induced divergence. Our basic questions are: (1) What are the structural reasons for the fact that some linguistic structures remain (relatively) stable in a situation which is otherwise marked by converging structures (cf. Louden 1994: 73), and (2) what are the sociolinguistic reasons for the fact that some speakers do not converge or even diverge from a prestigious source language in a generally converging group?

3. Some general considerations In section 1, the comparatively small amount of research into non-convergence was mentioned. This does not mean, though, that no efforts at all have been undertaken. Weinreich (1953: 3), for example, meticulously compares the surface forms of languages in contact in order to find possible points of interference, but his approach is necessarily pre-generative and pre-Labovian and thus remains somewhat superficial and lacks a more modern empirical foundation. Most current researchers distinguish between core grammar features (“deep” syntax, inflectional morphology, phonological rules), which are relatively stable both with regard to outside influences and with regard to endemic change, and peripheral grammar features (“superficial” syntax [word order], derivational morphology, pronunciation, lexicon), which are more open to outside influences and change more rapidly (but cf. section 5.2 for a different view). Such general comparisons

27. Non-convergence despite language contact do not have much explanatory power, though. Other research projects are too specific. They represent important linguistic levels with rather exotic phenomena (e.g., non-standard emphatic pronoun tags as major representative for syntax in Cheshire, Kerswill and Williams 2005: 159⫺163), a fact which is bound to undermine the validity of farreaching conclusions (Cheshire, Kerswill and Williams 2005: 166). Abstracting from these more methodological problems, the following structural factors are often mentioned as facilitating or restricting convergence or related phenomena: typological distance of the languages involved (Altenberg 1991: 191; cf. also the related concept of isomorphism [Sanchez 2005: 234]), markedness, linguistic saliency or naturalness of the elements in question (Altenberg 1991: 191; Kristiansen and Jørgensen 2005: 288; Mattheier 1996: 41) and questions of cognitive complexity (either with regard to language processing [Altenberg 1991: 191] or with regard to the multilingual speaker’s cognitive load [Sanchez 2005: 235]). These concepts will only be helpful, though, if they are clearly defined and analyzed in a specific contact situation. Even Thomason and Kaufman’s (1988) ground-breaking work fails with regard to the second point, because they compare very different situations. Within a few pages, for example, they address literary Indic-Dravidian language contact dating back thousands of years (Thomason and Kaufman 1988: 79) and present-day Low German as spoken in Nebraska (Thomason and Kaufman 1988: 81⫺82), categorizing both cases as slight structural borrowing. Thomason and Kaufman (1988: 50, table 3, 74⫺76) try to cover all possible contact situations. They distinguish three basic cases (language maintenance [entailing borrowing in their narrow sense], language shift and pidginization) and five different degrees of contact (from casual contact to very strong cultural pressure). However, if one wants to know more about structural or sociolinguistic restrictions on convergence in a general converging constellation, situations where the speakers of one language maintain their language and borrow linguistic material, structures or rules from a contact language should be the main focus. Concentrating on borrowing in Thomason and Kaufman’s narrow sense is justified by the fact that this is the unmarked case of language contact (cf. Backus 2005: 326). Although language shift is quite a common phenomenon, the linguistic consequences in the target language mostly disappear within one or two generations. In addition, one should not forget that the highly complex psychological and linguistic interaction of language contact, language death and second language acquisition in a shift situation further complicates the picture. Besides this, it also seems sensible to restrict one’s research to intermediate levels of contact in order to learn more about structural and sociolinguistic restrictions on convergence. Thomason and Kaufman’s (1988) extreme scenarios are not very helpful in this regard: category 1 only allows for lexical borrowing (Thomason and Kaufman 1988: 77⫺78), while in category 5, the very strong cultural pressure will override any existing restrictions on convergence (Thomason and Kaufman 1988: 91⫺95). Therefore, in order to answer questions (1) and (2) in section 2.3 we need to concentrate firstly on more fine-grained analyses of several phenomena for each linguistic level, and secondly, we should do this in specific and well-defined contact situations, where the sociolinguistic setting (albeit not the reactions to it) is identical for the entire speech community. In such a situation, the overall pressure for convergence will be identical for all linguistic structures and for all speakers. This means that the different behavior of converging and non-converging structures can be analyzed as a function of their structural characteristics. Likewise, the different behavior of converging and non-converging speakers can be analyzed as a function of their different sociolin-

483

484

IV. Structure and dynamics across language spaces guistic dispositions. Another problem with Thomason and Kaufman is their strong relativization of the importance of structural restrictions on convergence ⫺ “it is the sociolinguistic history of the speakers, and not the structure of their language, that is the primary determinant of the linguistic outcome of language contact” (1988: 35; cf. also 4, 15, 19). This conviction is certainly true for situations with extreme differences in the power relations between the groups in contact; for less extreme situations, however, it might detract from the structurally interesting points with regard to convergence and non-convergence (cf. also the critique in Sanchez 2005: 239⫺240). Obviously, even if we follow all these suggestions, success is not guaranteed: in code-switching, the language contact phenomenon most thoroughly studied, the search for structural restrictions is still far from offering unambiguous results. Muysken (1995: 178) states that “we should aim for universal explanations when looking for grammatical constraints” (cf. also Myers-Scotton 1995: 252), but so far the pursuit of this goal has not yet produced any clear results. Muysken (1995: 177) still wonders “to what extent restrictions on the codeswitching process [are] seen as absolute or relative” (cf. section 6).

4. Some examples or structural actors in (non-)convergence 4.1. (Non-)convergence beneath the surace In sections 1 and 2.3 it was argued that non-convergence can imply linguistic stability. But often a more detailed analysis shows that even complete stasis on all linguistic levels may coincide with an intrinsically dynamic situation. Hamel (1997: 113), for example, writes about the contact of indigenous speakers of Otomı´ with the majority culture in Mexico: “[t]he second kind of shift starts with a transformation of the ethnic group’s interpretative basis, that is, with a change of cultural schemes, of patterns of verbal interaction, and of interpretative procedures, while the indigenous language remains on the surface”. What this study shows is that in the contact between a minority and a majority group, it could be highly deceptive to define the status of the minority language exclusively by analyzing its linguistic system (cf. also Poplack 1993: 257). Hamel (1997: 113) continues: “[o]nce the cultural and pragmatic basis of the indigenous is eroded […], the substitution of the language as such can occur much more easily”. Obviously, restrictions on convergence with regard to cultural schemes, patterns of verbal interactions and interpretative procedures are not grammatical in nature; therefore, they will be dealt with in section 5.2. Another case of a somewhat hidden type of convergence is calquing. Sinner (2005a: 564⫺566) mentions the semantic extension of cada in Spanish as spoken and written in Catalonia by people with or without Catalan as mother tongue. Cada dı´a (in Standard Spanish ‘each day’ in a distributive sense) infringes more and more on the semantics of todos los dı´as (‘every day’ in a generalized sense), and it seems to do so under Catalan influence. With regard to semantic borrowing of prepositions and idioms, Louden (1994: 84) writes about the contact of Pennsylvania German and English in the USA: “in spite of such heavy semantic influence from English on Plain Pennsylvania German, including borrowings, the lexicon itself remains primarily German”. Whether the lexicon can still be called “primarily German” in the case of “heavy semantic influence from English” could be debated; Louden’s conviction is interesting, though, because in a certain way, it coincides with the fact that speakers in a contact

27. Non-convergence despite language contact situation are more aware of (and more sensitive to) borrowing complete words than of just borrowing the meaning of a word. It seems that a certain amount of metalinguistic awareness is necessary in order to refrain from borrowing new meanings (cf. section 5.2), especially if the words in question already share part of their meaning (as in the case of cada and todos los; cf. also Berruto 2005: 82). Another group of words facilitating semantic borrowing are bilingual (near-)homophones in related languages (cf. Weinreich 1953: 48⫺50). For Portuguese spoken in the Brazilian-Uruguayan border region, Blaser (1995: 134) mentions estranhar (Standard Portuguese ‘to find [something] weird’) with the Spanish meaning of extran˜ar (‘to miss’); Kaufmann (2000: 171⫺172) mentions embarac¸ada (Standard Portuguese ‘embarrassed’) with the Spanish meaning of embarazada (‘pregnant’). It is difficult to give structural restrictions on convergence with regard to semantic borrowing; what one can say is that facilitating factors seem to be a phonetic and/or a semantic overlap. A different consequence of calquing, which might lead to dramatic changes, is not connected with the meaning of a particular word but with its subcategorization frame. Kaufmann (2005: 77⫺81) cites an example of Mennonite Low German as spoken in North and South America. Indirect objects in German varieties are normally marked by the dative or a general object case. In contrast, all contact languages of Mennonite Low German, i. e., Spanish, Portuguese and English, at least partially use prepositions to mark this syntactic function, and this characteristic has already had an influence on Low German. In the USA, indirect objects are realized with prepositions in 37.5 percent of the cases (mostly with tu ‘to’ or no [< nach, ‘after’]), whereas in Brazil für ‘for’ and tu are used in 19.6 percent of the cases (cf. Portuguese falar para as crianc¸as, ‘talk for the children’, i. e., ‘talk to the children’). Interestingly, the four colonies in contact with Spanish show less prepositional marking (between 4 and 16.7 percent). The reason for this could be the inconspicuous and frequently contracted preposition a, which Spanish uses in this context. Such light forms seem to function as a restriction on convergence. King (2005: 237) sees in this kind of calquing the only way of acquiring foreign grammatical structures (cf. also Backus 2005: 309); she writes: “I argue that grammatical borrowing has a lexical basis. This approach is compatible with calquing as a process of contact-induced change, given that calquing involves change in the properties of lexical items, rather than the direct importation of grammatical structure […]”. Interestingly, in her example of Prince Edward Island French in Canada, calquing leads to convergence and divergence at the same time. This French variety has borrowed many English prepositions which have caused an increase in the application of a rather rare phenomenon in French, namely preposition stranding (cf. King 2005: 243⫺248). At first sight, it appears that we are dealing with a clear case of lexical convergence towards English and an ensuing structural convergence connected to a specific rule of the borrowed prepositions. At the same time, however, this new rule is generalized for all prepositions in Prince Edward Island French, i. e., restrictions on its application in English have not been borrowed, and this leads to subsequent structural divergence.

4.2. (Non-)convergence on the surace It is rather difficult to name clear structural restrictions on convergence in the parts of language which are not phonetically realized. For phonetically realized parts, this will be somewhat easier. With regard to entire words, the most intensively studied type of

485

486

IV. Structure and dynamics across language spaces borrowing, there is a well-known ban on closed-class items (articles, pronouns, etc.) in comparison to open-class items (especially nouns). A high frequency of a word in the recipient language also seems to make borrowing less probable (cf. Weinreich 1953: 57). As for pronunciation, Louden and Page (2005: 1389) state: “[i]t appears that convergence is suppressed where it would involve contrastive sounds […]” (cf. also Weinreich 1953: 23). The fact that Pennsylvania German has adopted the English retroflex r is no counter-example to this rule, because “the system of phonological contrasts in the lexicon is unaffected by this phonetic substitution” (Louden and Page 2005: 1390; cf. also the same development in Canadian French [Poplack 1993: 261⫺262]). Bullock and Gerfen (2004: 99) mention the substitution of the “only marginally contrastive” French vowels in deux ‘two’ and neuf ‘nine’ by the English vowel of bird in Frenchville, Pennsylvania. This is another example for non-system-affecting phonetic convergence enabled by similarity: “[t]he shared structural properties across French-English that are at the locus of this change are similarities in vowel height, roundness, and position along the front-back articulatory dimension” (Bullock and Gerfen 2004: 102⫺103). Such a substitution will consequently not occur if the elements in question do not share at least some structural properties. The fact that these speakers converge to an English vowel, which is strongly marked and does, therefore, not seem to be an attractive target for borrowing, can be explained by the reduction of overall complexity for this bilingual speech community (one instead of two marked vowels [Bullock and Gerfen 2004: 103]). Another interesting point is the phonetic shape of borrowed words (nonce or established): Poplack, Sankoff and Miller (1988: 72) write that “phonological integration proceeds as a function of the social integration of the loanword”. One could also claim that pronouncing a borrowed word in the way of the recipient language is an act of non-convergence within convergence. As for morphology, one can say that the more integrated/synthetic a linguistic form is, the less probable is its transfer (Weinreich 1953: 35, 41; Backus 2005: 323), i. e., the direct borrowing of affixes is a rather rare event. If affixes are borrowed, they are almost always borrowed as part of words. Sometimes they will then be analyzed, and only in this case can they be used productively with stems from the recipient language (e.g., German -ieren < French -i(e)r in verbs with a Germanic stem like inhaftieren, ‘in custody-ieren’, i. e., ‘to imprison’). What is more frequent in morphology is a functional change or a merger of forms in the recipient language under foreign influence (cf. Louden 1988: 146⫺152). With regard to (morpho)syntax, Louden’s (1988, 1994) research comes pretty close to the desiderata formulated in section 3; he compares different (morpho)syntactic phenomena in a precisely defined contact situation. Because of this, he (1988: 227) can make valid comparative statements: What data we have of change in PPG [Plain Pennsylvania German] word order is limited to minor surface phenomena. These minor changes entail the generalization of certain patterns, such as the placement of infinitival complements in clause-final position. However, despite the intimate contact of PPG, an SOV language, with AE [American English], an SVO language, there is no evidence to indicate a major shift of PPG word order away from an underlying verb-final structure.

This quotation illustrates the importance of analyzing several phenomena of each linguistic level, because ⫺ as Louden shows ⫺ these phenomena can behave differently. The only problem with Louden’s work is that he does not analyze his data quantitatively.

27. Non-convergence despite language contact In the following, I will try to demonstrate with examples from Mennonite Low German (cf. Kaufmann 2005; Kaufmann 2007) what such a quantitative analysis might look like. As Louden writes, German verbs govern their complement to the left, i. e., different from English, Spanish and Portuguese, German is an OV-language (den Hund schlagen, ‘the dog beat’, i. e., ‘to beat the dog’). In two of the five Mennonite colonies, namely in Brazil and the United States, the lexical influence of the majority languages on Low German is so strong that one would also expect some structural borrowing. Nevertheless, like in Plain Pennsylvania German there are hardly any cases where the informants produced Low German embedded clauses with the complement occurring to the right of its governing verb (cf. Kaufmann 2005: 87⫺89; Kaufmann 2007: 196). Only a robust number of such tokens would prove a structural influence on Low German. The sequence of verbs and their complements is apparently so deeply ingrained in the system of the language that it functions as a powerful barrier against convergence (but cf. the contact-induced change of Rimella German from an OV to a VO-language (Louden 1994: 88)). In other subsystems of Low German, however, some cases of word order change can be found. The sequence of nominal attributes and their governing heads, for example, shows some influence of the majority languages (cf. Kaufmann 2005: 84⫺86): in the USA, min Bruder sin Lewen (‘to my brother his life’), with 72.7 percent of the cases the most frequent variant (in Brazil only 23.2 percent), seems to follow the English sequence ‘my brother’s life’, while in Brazil daut Lewe von min Bruder is used in 41.1 percent of the cases, probably a consequence of Portuguese a vida do meu irma˜o (‘the life of my brother’; not a single token in the USA). Two factors might have made this convergence possible: first, the respective variants had already existed in the recipient language, i. e., the influence of the contact language did not create a new variant, but only strengthened an already existing one (cf. Mattheier 1996: 34), and second, the position of a nominal attribute seems to be more superficial than the position of a verbal complement. Looking at the two Paraguayan colonies (Menno and Fernheim), there are more interesting things to learn: in Paraguay, the influence of the linguistically distant majority language Spanish is restricted to lexical borrowing, while the strong presence of less distant Standard German has a significant effect on many structural levels, among them the ordering of verbal elements in clause-final clusters. The Mennonites in Menno and Fernheim use the variant also used in Standard German (… dat hei imma sine Mame helpe mut, ‘… that he always his mother help must,’ i. e., ‘… that he always has to help his mother’) in 92.4 and 93.7 percent of the cases respectively, while the Mennonites in the United States, who have hardly any contact to Standard German, use this variant in just 29.9 percent. Again, two factors can account for the difference between the Paraguayan Mennonites who converge to the Standard German variant and the US-American and Brazilian Mennonites who do not converge to English and Portuguese structures: firstly, the linguistic distance between Standard German and Low German with regard to word order is smaller than the one between English or Portuguese and Low German; secondly and more importantly, the nonstandard sequence of verbal elements in verb clusters (… dat hei imma [sine Mame] mut [sine Mame] helpe) is the result of a rather superficial movement of nonfinite verbal elements to the right (with or without scrambling of the complement; cf. Kaufmann 2007: 156⫺157), whereas the basic ordering of verbs and their complements is a much more fundamental characteristic of Germanic varieties. Granted, right now we are comparing two different contact situations, not following the desiderata in section 3, but there are interesting comparisons within

487

488

IV. Structure and dynamics across language spaces the same situation as well: in verb clusters with three verbal elements, convergence to the Standard German variants in Paraguay is less strong: in Menno, the rate drops from 92.4 to 43.5 percent; in Fernheim from 93.7 to 39.8 percent (USA: from 29.9 to 2.4 percent). The reason for this drop lies in the higher complexity of clusters with three verbal elements. Even Standard German, a language with left-branching sequences in most of its verbal syntax, requires in the context in question (modal verbs in the perfect) more parsing-friendly, partly right-branching structures, i. e., the finite verb appears before the nonfinite verbs (… daß er immer [seiner Mutter] hat [seiner Mutter] helfen müssen, ‘… that he always [his mother] had [his mother] help must,’ i. e., ‘… that he always had to help his mother’). The Mennonites in Menno and Fernheim do not always succeed in converging to these variants. They produce completely right-branching structures (… dat hei imma [sine Mame] hat [sine Mame] mut [sine Mame] helpe) in 50 and 56 percent of the cases respectively. We can therefore conclude that the more complex a structure of the source language is, the less convergence there will be.

5.

Some examples o sociolinguistic actors in (non-)convergence

5.1. Age and gender Auer and Hinskens (1996: 4) write: “[e]mpirically rich, well-documented and quantitative investigations of processes of convergence and divergence, and not just of their outcomes, are rare; investigations into the links between social changes and the linguistic developments they can trigger are even rarer” (cf. also Mattheier 1996: 31; Cheshire, Kerswill and Williams 2005: 141). The interaction between social factors (if not changes) and linguistic developments can be illustrated with regard to the Mennonites in Paraguay: as Standard German is a language with a high overt prestige in these colonies, convergence of Low German towards Standard German variants must be classified as change from above. Labov (2001: 274) characterizes this type of change in the following way: “[i]n linguistic change from above, women [and especially young women, one may add; G. K.] adopt prestige forms at a higher rate than men” (cf. also Cheshire, Kerswill and Williams 2005: 143). Young women in Fernheim converge, as expected, most strongly to the Standard German variant in clusters with two verbal elements (100 percent as opposed to 92.2 percent for the five other subgroups; p ⫽ 0.022). Young women in Menno rank second (94.6 as opposed to 91.9 percent; non-significant). The situation with regard to clusters with three verbal elements is somewhat different. In Fernheim, young women still use the two Standard German variants more often than most of the other subgroups (45.2 as opposed to 38.5 percent), but they only rank third and the difference is not significant any more. Young women in Menno still rank second, (52.5 as opposed to 40.8 percent; non-significant). As the wish to use Standard German variants hardly depends on the number of verbal elements, we must either explain the different behavior of young women in Fernheim, i. e., their relative non-convergence in clusters with three verbal elements, with difficulties identifying the standard variant(s) or again with the complexity of this structure. In Menno, there is another interesting sociolinguistic development. Young men there seem to converge less strongly to the Standard German variant in cases of clusters with two verbal elements. While young women

27. Non-convergence despite language contact did not converge in only 5.4 percent of the tokens (the older women do not converge in 10.8 percent; non-significant), the rate of non-convergence for young men is more than twice as high (11.6 percent; the older men 3.7 percent; p ⫽ 0.049). Although the difference between young men and young women is not significant, it is highly suggestive that the older men use the Standard German variant significantly more often than the young men, while the older women use it less often than the young women, at least with regard to absolute frequency. The difference is even more striking with regard to clusters with three verbal elements. In this context, young women use the two possible standard variants in 52.5 percent of the tokens (older women 29.5 percent; p ⫽ 0.032), whereas young men do so in only 33.3 percent (older men 56 percent; p ⫽ 0.038). The difference between young women and young men shows a statistical tendency (p ⫽ 0.092). For several reasons connected to the history of Menno (cf. Kaufmann 2007: 180⫺182), one must classify the behavior of young men as a “retreat […] from a female-dominated change” (cf. Labov 2001: 297). This divergence from the prestigious Standard German variants is only present in part of the speech community and is caused by a special sociolinguistic disposition. This shows clearly that in studies of convergence and non-convergence neither languages nor speech communities should be analyzed as if they were monolithic blocks (cf. sections 2.2 and 3).

5.2. Language loyalty, identity and types o bilingualism In section 2.3, a high degree of language loyalty was mentioned as an important characteristic for a non-converging speech community. Language loyalty is generally not only linked to a positive attitude towards one’s language but also to a positive attitude towards one’s culture. Especially this cultural loyalty is a condition for non-convergence in the less conspicuous areas of language where cultural schemes, patterns of verbal interaction and interpretative procedures are involved (cf. Hamel 1997 in section 4.1). In the context of language enclaves, Mattheier (1994: 334⫺335) calls such culture loyalty a Sprachinselmentalität, a sociopsychological disposition of the members of a minority group, by which they mark their difference from the majority group. In Canada, both types of loyalty seem to exist because neither English as a minority language in Quebec (cf. Poplack, Walker and Malcolmson 2006: 207) nor French as a minority language in Ontario (cf. Poplack, Sankoff and Miller 1988: 57; Poplack 1993: 261) show much lexical or structural borrowing or much code-switching. Besides this, even if, for example, French speakers code-switch or use borrowed words, they make sure that the listener realizes that this is not their normal behavior: “[…] Ottawa-Hull francophones draw attention to, or ‘flag’, their switches, by different discourse devices: metalinguistic commentary, English bracketing, repetition or translation” (Poplack 1993: 263; cf. for lexical borrowing Poplack 1988: 114). Again we are faced with an act of non-convergence within convergence (cf. section 4.1 and 4.2). The most important restrictive factor on convergence in this situation seems to be social pressure exerted by people with a high socioeconomic status (cf. Poplack, Sankoff and Miller 1988: 81; Poplack 1988: 111). Such a status seems to correlate with high metalinguistic awareness which enables speakers to monitor even more hidden types of convergence like semantic borrowing (cf. section 4.1). Other restrictive, but apparently less important factors on convergence are the status as a minority or a majority group (“linguistic security”; cf. Poplack 1988: 95) and the lack

489

490

IV. Structure and dynamics across language spaces of bilingual proficiency (cf. section 2.1; Poplack, Sankoff and Miller 1988: 97; Poplack 1988: 100, 110⫺111). Theoretically, in more conflictive contact situations ⫺ Poplack, Walker and Malcolmson (2006: 209) describe the almost non-converging situation in Canada “as maximally conductive to convergence” ⫺ one could even imagine diverging tendencies, at least with regard to typologically related languages. One sociolinguistic condition for such a development could be a behavior which Sinner (2005a: 561) calls “linguistic paranoia”. He writes about some speakers of Spanish in Catalonia: “[d]icha paranoia lingüı´stica puede manifestarse, por ejemplo, en la tendencia a ver catalanismos ‘por todos lados […]’”. Once people see borrowed words everywhere, i. e., once they feel their linguistic identity threatened, they might well want to avoid using any word or structure which could possibly belong to the threatening contact language. But different from dialect contact, such cases seem to be rare in language contact. Sinner (2005b: 46⫺ 48) mentions some possible examples for lexical divergence in Catalan. There is one more sociolinguistic factor worth while mentioning: the type of bilingualism dominant in the speech community. Louden and Page (2005: 1391) write: “[t]he lexicon is the most cognitively salient component of the grammar. Therefore lexical items are more easily borrowed than grammatical morphemes, phonemes or syntactic patterns in casual language contact” (cf. section 3; Bechert and Wildgen 1991: 69; Villena Ponsoda 2005: 314). But Louden and Page (2005: 1391⫺1392) continue: Conversely, in stable bilingual situations, the phonological shapes of lexical items are the most salient markers of the code being spoken and are therefore resistant to convergence. Maintenance of the Pennsylvania German lexicon, including Pennsylvania German morphophonemic alternations, serves to mark one’s identity as an Old Order sectarian.

Louden (1994: 74) defines a situation of stable bilingualism as a situation where both languages “are acquired sufficiently early and completely”, both languages have “substantial and productive domains of use”, and both languages “enjoy more or less […] equivalent prestige”. One often finds such a situation (stable bi- or multilingualism, no or little borrowing of phonological features and words, massive convergence of syntax) in Sprachbund contexts (Balkans, Kupwar in India) (cf. Bisang in this volume; Louden 1994: 76⫺79).

6. Conclusion Our partial analysis of Mennonite Low German and some of the case studies mentioned may have convinced the reader that working in specific and well-defined contact situations enables us to compare the restrictive power of structural and sociolinguistic factors on convergence in a meaningful way. The question whether one day we will be able to universally determine absolute values for these factors cannot be answered yet. It may well be that this is not a consequence of incomplete or faulty analyses but of the nature of these factors. They simply might not have any universal value and might be better represented by an approach within the framework of Optimality Theory. Structural and sociolinguistic factors could then be ranked according to their restrictive power for specific contact situations. The ranking of the structural factors would be a function of their interaction within the recipient language and the typological interaction between

27. Non-convergence despite language contact the contact languages. Therefore, the ranking would necessarily be language specific and language pair specific, i. e., not universal. The sociolinguistic factors could then be ranked according to their situational strength. Such an approach would not have much explanatory power with regard to single structural or sociolinguistic factors, but could tell us a lot about the interaction of these factors.

7. Reerences Altenberg, Evelyn P. 1991 Assessing first language vulnerability to attrition. In: Seliger and Vago (eds.), 189⫺206. Auer, Peter and Frans Hinskens 1996 Convergence and divergence of dialects in Europe: New and not so new developments in an old area. Sociolinguistica 10: 1⫺30. Auer, Peter and Frans Hinskens 2005 The role of interpersonal accommodation in a theory of language change. In: Auer, Hinskens and Kerswill (eds.), 335⫺357. Auer, Peter, Frans Hinskens and Paul Kerswill (eds.) 2005a Dialect Change: Convergence and Divergence in European Languages. Cambridge: Cambridge University Press. Auer, Peter, Frans Hinskens and Paul Kerswill 2005b The study of dialect convergence and divergence: conceptual and methodological considerations. In: Auer, Hinskens and Kerswill (eds.), 1⫺48. Backus, Ad 2005 Codeswitching and language change: One thing leads to another? International Journal of Bilingualism 9(3): 307⫺340. Bechert, Johannes and Wolfgang Wildgen 1991 Einführung in die Sprachkontaktforschung. Darmstadt: Wissenschaftliche Buchgemeinschaft. Berend, Nina and Klaus J. Mattheier (eds.) 1994 Sprachinselforschung: Eine Gedenkschrift für Hugo Jedig. Frankfurt am Main: Lang. Berruto, Gaetano 2005 Dialect/standard convergence, mixing, and models of language contact: The case of Italy. In: Auer, Hinskens and Kerswill (eds.), 81⫺95. Blaser, Jutta 1995 Das Spanische in Südbrasilien: Die Zerstörung einer Legende durch mikrodialektologische Feldforschung. Frankfurt am Main: Lang. Bullock, Barbara and Chip Gerfen 2004 Phonological convergence in a contracting language variety. Bilingualism: Language and Cognition 7(2): 95⫺104. Cheshire, Jenny, Paul Kerswill and Ann Williams 2005 Phonology, grammar, and discourse in dialect convergence. In: Auer, Hinskens and Kerswill (eds.), 135⫺167. Hamel, Rainer E. 1997 Language conflict and language shift: A sociolinguistic framework for linguistic human rights. International Journal of the Sociology of Language 127: 105⫺134. Kaufmann, Göz 2000 Measuring the unmeasurable or how to gauge the competence in non-standard language varieties. In: Marlene Brugali and Maria da Grac¸a Gomes Paiva (eds.), Avaliac¸a˜o: Novas tendeˆncias ⫺ Novos Paradigmas, 151⫺175. Porto Alegre: Mercado Aberto.

491

492

IV. Structure and dynamics across language spaces Kaufmann, Göz 2005 Der eigensinnige Informant: Ärgernis bei der Datenerhebung oder Chance zum analytischen Mehrwert? In: Friedrich Lenz and Stefan J. Schierholz (eds.), Corpuslinguistik in Lexik und Grammatik, 61⫺95. Tübingen: Stauffenberg. Kaufmann, Göz 2007 The verb cluster in Mennonite Low German: A new approach to an old topic. Linguistische Berichte 210: 147⫺207. Kaufmann, Göz 2008 Where syntax meets morphology: Varianten des bestimmten Artikels und die Variation satzfinaler Verbcluster im Plattdeutschen texanischer Mennoniten. In: Franz Patocka und Guido Seiler (eds.), Dialektale Morphologie, dialektale Syntax: Beiträge zum 2. Kongress der Internationalen Gesellschaft für Dialektologie des Deutschen, Wien, 20.⫺23. September 2006, 87⫺120 Wien: Praesens Verlag. King, Ruth 2005 Crossing grammatical borders: Tracing the path of contact-induced linguistic change. In: Markku Filppula, Juhani Klemola, Marjatta Palander and Esa Penttilä (eds.). Dialects Across Borders: Selected Papers from the 11th International Conference of Methods in Dialectology (Methods XI), Joensuu, Finland 2002, 233⫺251. Amsterdam/Philadelphia: John Benjamins. Kristiansen, Tore and Jens Jørgensen 2005 Subjective factors in dialect convergence and divergence. In: Auer, Hinskens and Kerswill (eds.), 287⫺303. Labov, William 2001 Principles of Linguistic Change, vol. 2: Social Factors. (Language in Society 29.) Oxford: Blackwell. Lass, Roger 1992 Phonology and morphology. In: Norman Blake (ed.), The Cambridge History of the English Language, vol. 2, 1066⫺1476, 23⫺155. Cambridge: Cambridge University Press. Loebell, Helga and Kathryn Bock 2003 Structural priming across languages. Linguistics 41(5): 791⫺824. Louden, Mark L. 1988 Bilingualism and syntactic change in Pennsylvania German. Unpublished PhD dissertation, Cornell University. Louden Mark L. 1994 Syntactic change in multilingual speech islands. In: Berend and Mattheier (eds.), 73⫺91. Louden, Mark L. and B. Richard Page 2005 Stable bilingualism and phonological (non)convergence in Pennsylvania German. In: James Cohen, Kara T. McAlister, Kellie Rolstad and Jeff MacSwan (eds.), Proceedings of the 4th International Symposium on Bilingualism, 1384⫺1392. Somerville, MA: Cascadilla Press. Mattheier, Klaus J. 1994 Theorie der Sprachinsel. Voraussetzungen und Strukturierungen. In: Berend and Mattheier (eds.), 333⫺348. Mattheier, Klaus J. 1996 Varietätenkonvergenz: Überlegungen zu einem Baustein einer Theorie der Sprachvariation. Sociolinguistica 10: 31⫺52. Milroy, Lesley and Peter Muysken (eds.) 1995 One Speaker, Two Languages: Cross-Disciplinary Perspectives on Code-Switching. Cambridge: Cambridge University Press. Muysken, Peter 1995 Code-switching and grammatical theory. In: Milroy and Muysken, (eds.), 177⫺198. Myers-Scotton, Carol 1995 A lexically based model of code-switching. In: Milroy and Muysken (eds.), 233⫺256.

27. Non-convergence despite language contact

493

Poplack, Shana 1988 Language status and language accommodation along a linguistic border. In: Peter H. Lowenberg (ed.), Language Spread and Language Policy: Issues, Implications, and Case Studies, 90⫺118. Washington DC: Georgetown University Press. Poplack, Shana 1993 Variation theory and language contact. In: Dennis Preston (ed.), American Dialect Research: An Anthology Celebrating the 100th Anniversary of the American Dialect Society, 251⫺286. Amsterdam: Benjamins. Poplack, Shana, David Sankoff and Chris Miller 1988 The social correlates and linguistic processes of lexical borrowing and assimilation. Linguistics 26: 47⫺104. Poplack, Shana, James Walker and Rebecca Malcolmson 2006 An English “like no other”?: Language contact and change in Quebec. Canadian Journal of Linguistics 51(2/3): 185⫺213. Salmons, Joe 1990 Bilingual discourse marking: Code switching, borrowing, and convergence in some German⫺American dialects. Linguistics 28: 453⫺480. Sanchez, Tara 2005 Constraints on structural borrowing in a multilingual contact situation. PhD dissertation, University of Pennsylvania. Schirmunski, Viktor M. 1930 Sprachgeschichte und Siedelungsmundarten. Germanisch-Romanische Monatszeitschrift 28, 113⫺122 and 171⫺188. Seliger, Herbert W. and Robert M. Vago (eds.) 1991 First Language Attrition. Cambridge: Cambridge University Press. Silva-Corvala´n, Carmen 1991 Spanish language attrition in a contact situation with English. In: Seliger and Vago (eds.), 151⫺171. Sinner, Carsten 2005a Estudio longitudinal de cambios lingüı´sticos inducidos por contacto lingüı´stico. In: Claus D. Pusch, Johannes Kabatek and Wolfgang Raible (eds.), Romance Corpus Linguistics II: Corpora and Diachronic Linguistics, 559⫺570. Tübingen: Narr. Sinner, Carsten 2005b Evolucio´ de les normes d’u´s als paı¨sos de parla catalana i el problema del concepte de norma. In: Ba`rbara Roviro´, Aina Torrent-Lenzen and Andreas Wesch (eds.), Normen und Identitäten: Sprachwissenschaftliche Beiträge des 19. Deutschen Katalanistentags, Köln 2003, 35⫺57. Titz: Axel Lenzen. StADaF ⫽ Ständige Arbeitsgruppe Deutsch als Fremdsprache (ed.) 2006 Deutsch als Fremdsprache weltweit: Datenerhebung 2005. Berlin/Bonn/Cologne/Munich: StADaF. Thomason, Sarah Grey and Terrence Kaufman 1988 Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Villena Ponsoda, Juan Andre´s 2005 How similar are people who speak alike? An interpretive way of using social networks in social dialectology research. In: Auer, Hinskens and Kerswill (eds.), 303⫺334. Villena Ponsoda, Juan Andre´s 2006 The Iberian Peninsula. In: Ulrich Ammon, Norbert Dittmar, Klaus J. Mattheier and Peter Trudgill (eds.), Sociolinguistics: An International Handbook of the Science of Language and Society, 1802⫺1810. Berlin: de Gruyter. Weinreich, Uriel 1953 Languages in Contact: Findings and Problems. The Hague: London.

Göz Kaufmann, Freiburg (Germany)

V. Data collection and corpus-building 28. Investigating language in space: Methods and empirical standards 1. 2. 3. 4. 5. 6. 7. 8. 9.

Introduction: Defining the question Objectivity Reliability Validity Representativity Quality criteria and areal linguistic methods of data acquisition Existing corpora: Corpus linguistics Quality criteria and attitudinal research References

1. Introduction: Deining the question Language is constituted through the relatively uniform, cooperative and mostly unreflecting actions of many. The relative uniformity is a precondition for language to be able to serve its basic purpose (communication); the existing variability serves a number of other functions. Individual linguistic variants correlate with parameters defined within variation linguistics by stylistics, sociolinguistics or pragmatics, for example, but also with non-linguistic parameters. These can include, for example, gender, age, education, class, milieu, religion, occupation, origin, place of socialization, speech context, intent or perceived role. Some of the terms just listed are drawn from different descriptive traditions, which attempt to capture and describe the existing correlations between speech performance and extralinguistic conditions within a terminological system. Factors like gender and age can easily be operationalized and linked to linguistic data. A more challenging case is the concept of “milieu”, because this is a factor which can only be measured in terms of a larger number of parameters that are, in part, quite complex constructs themselves (e.g., education). We refer to this complex ⫺ whatever its constituents in particular cases ⫺ using the collective term “sociopragmatic factors”. In the investigation of recent, regionally influenced spoken language, the utterances analyzed have usually been produced for a particular survey. Only so was (and is) it possible to examine linguistic issues economically and within a manageable time frame. Observations of “natural” speech play a minor role ⫺ except in the development of hypotheses. At the outset, linguists may notice a regional peculiarity in their own speech or when observing a conversation. They then begin to explore the sociopragmatic conditions under which such forms occur. If they suspect there might be a connection with the speaker’s geographic background, they can test this hypothesis empirically. The linguist simply needs to evoke the appropriate situation in different regions and observe which linguistic form is used. Theoretically, it should not be a problem to build such a corpus: the experimental design for evoking such utterances only needs to allow the

28. Methods and empirical standards parameter of space to vary while any other factors influencing speech are held constant. The problem is, however, that it is extremely difficult to hold all these other factors that affect speech constant and stable over any length of time. Linguistic acts are always social acts as well, and hence they cannot simply be repeated as often as required under the standardized laboratory conditions of a science experiment. There are always more factors affecting speech than one would wish, and most of these cannot be isolated. Some (e.g., role perception) are barely subject to external influence and are therefore virtually impossible to eliminate as disturbing factors. Dialectology has these problems in common with any discipline concerned with human behavior, e.g., the social sciences or psychology, that makes use of experiments or surveys to investigate this behavior. (A survey or an interview can also be seen as an experimental situation.) As a general rule, there are four requirements for any such data collection procedures; these act as quality control and evaluative criteria for academic studies (cf., e.g., Lamnek 2005: 148⫺193; Schnell et al. 2005: 149⫺166): (1) (2) (3) (4)

objectivity reliability validity representativity

These quality criteria are closely interconnected. Insofar as they are relevant for geolinguists, they will be described in the following.

2. Objectivity This requirement specifies that statements about facts should be independent of the observer or describer. Since one can assume that every observer will judge in a “subjective” way, that every observation and every finding is initially tied to the subjectivity of a human, it would be better to speak of intersubjectivity. We are inclined to describe things that everyone sees and talks about in a similar way as objective, things that an individual or a few (in contrast to the others) see in a particular way as subjective. These terms have been communicatively and quantitatively defined by linguists and it is safe to assume that this definition reflects current academic usage. Research findings should be described as intersubjective if they fulfill the criteria (2), (3) and (4) mentioned above. If an academic endeavor complies with these criteria, and if nobody is reasonably able to deny its findings, then they can be regarded as intersubjectively founded, i. e., as objective in the sense described above (on this problem cf., e.g., Kromrey 2006: 43⫺58).

3. Reliability A research finding is reliable, or consistent, if the results are free of random errors and if replication of the experiment generates commensurate results. The more widely the values obtained for a variable differ when measured in an identical experiment or observation context, the less reliable any finding is. The less “internal” variation within a set

495

496

V. Data collection and corpus-building of experiments there is, i. e., the better difficult-to-control factors (subject to limited external influence) can be eliminated, the higher the degree of reliability. Reliability measures repeatability, the stability of research results across identical experimental conditions. In the field dealt with in this article, there are two kinds of reliability problem: (1) Errors of measurement by the person measuring. These can include: ⫺ errors in reading text ⫺ phonetic transcription errors ⫺ using dialectology’s indirect method, informants become measurers; alphabetic transcription, thoroughly inadequate for certain types of inquiry, is the primary source of error here ⫺ parsing errors in the assignment of linguistic measures/categories to text segments The “faulty measurements” described here may arise accidentally or systematically; researchers are generally aware of them as sources of error. Close phonetic transcription in particular poses a major problem since it is here that the replication of results is hardest to achieve, depending on the quality demands imposed. Progress has been made on this front thanks to the electronic presentation of data possible with analysis systems like Praat (cf. and ). It can generally be said that the reliability of a transcription always depends on the goals of a research project and that any deviations from the “objective” transcription can be regarded as harmless as long as they do not exceed the “natural” variation range of a speaker (König 1988) and are not themselves the topic of investigation. For instance, even if I have phonetically transcribed a text incorrectly, I can still use it to analyze syntax. (2) The impact of confounding variables. From a purely theoretical perspective, research into language in space can be described as follows: the parameter of “space” is varied in an experiment or observation while all of the other factors which have an impact on speech are held constant. The resulting linguistic differences can be considered as geographical variants of a language. Their depiction on a map represents an ideal, abstract model of how language changes across space. But such contexts, which have to be re-created from scratch several times, can never be held completely constant. A range of confounding variables of sociopragmatic nature can have an influence and thus distort the results. (For more information on confounding variables, see section 6.) Confounding variables reduce reliability ⫺ the same external experimental setup gives two different results. If an investigator recognizes such a “deviation”, he/ she can repeat the “experiment”, for example by inquiring, “Wasn’t there another way of saying that?” This may even go as far as a suggestion, like “Didn’t people also say XY?” Needless to say, such “intrusions” into the experimental context must be recorded in the protocol. This presupposes that the investigator knows the object of study very well and is able to intervene when she/he has the justified suspicion that a variant other than the one obtained is the valid target usage for a particular location. Reliability is a necessary precondition for validity. Validity can never exceed reliability.

28. Methods and empirical standards

4. Validity A research result can be regarded as valid if the methods used measure what they are supposed to measure. The main question is how resilient, how strong a correlation between two variables is, and to what extent a causal connection between them is probable. In the case of language variation in space, the following problems arise in particular: (1) To what extent do the linguistic parameters identified in our grammatical/pragmatic terminology correspond to factors that really exist in language? Are the constructs we use actually relevant to the research goal? Is, for example, syllable cut [Silbenschnitt] a relevant construct for the phonology of German? One can certainly argue about that, but a category like “plural” is most unlikely to be regarded as irrelevant for German. (2) How suitable are sociopragmatic parameters, as constructs, for representing correlations between linguistic data and themselves? The pragmatic parameters (e.g., intention X correlates with the linguistic form Y) are normally defined by linguists themselves and are thus subject to the validity problems described above under (1). Speaker-related parameters are generally adopted from the social sciences, unless the features are very simple to operationalize, such as age or gender, for example. In contrast, complex factors like speech context, education, social class, and milieu are hard to operationalize, since they are characterized by a multitude of features. Nevertheless, we need to work with such factors as well, in that they should remain constant across speakers in order to make comparable what we hope to compare geographically. However, they mainly crop up as “confounding variables”, because it is virtually impossible to render two speech situations identical, since there are so few speakers for whom all of these variables have the same values. This argument implies that it is possible to specify conditions for the use of every linguistic variant that exists, i. e., that the origin of every variant can be described. Conversely, that in turn requires that there can be no “free” variation. We cannot discuss this question any further here. For the time being, it is probably better to assume that there is no such thing as free variation and thus avoid abandoning the search for determining parameters too early. Free variants in such a situation would then be those whose determining factors are (still) unknown. (3) To what extent is the data appropriate to the research focus? As a general rule, economic constraints mean that dialectological research is based on corpora specifically compiled for that purpose (though cf. section 6.3.5), using methods and designs comparable to those of the social sciences and experimental psychology (cf., e.g., Kromrey 2006; Schulz, Muthig and Koeppler 1981). This raises the question of the extent to which the intended purpose is met by the collected data, i.e., the extent to which the linguistic forms evoked match the research goals. In other words: can the experimental method be used to measure what we hope to measure? We also need to ask how the speech data relate to speech production as a whole.

497

498

V. Data collection and corpus-building To what extent does it support generalizations? To what extent can it predict the form utterances will take? Such questions can be seen as a validity problem, but also as a representativity issue.

5. Representativity A sample is representative if it reflects the characteristics of the population to be investigated, if it represents a miniature image of the whole that is capable of supporting general conclusions about it. Representativity is always of that population. A representative sample drawn from a population is always representative of that population. There is a large body of literature in the social sciences on how to create representative random samples (cf., e.g., Friedrichs 1990: 123⫺147; Kromrey 2006: 265⫺316). Primarily, it deals with the selection of individuals. In areal linguistics, the question of representativity arises in many regards, namely the selection of ⫺ ⫺ ⫺ ⫺

locations individuals, speakers speech contexts, text types, speech styles, register linguistic problems.

At any rate, the goal is to obtain a corpus that is representative for a particular research field and adequate for particular investigations. In general, the greater the variation within such a population ⫺ and in the case of an established national language with a high number of speakers it is very large ⫺ the larger the number of speakers, the amount of text investigated, and the sample have to be in order to be regarded as representative. The less internal variation there is within a language, speech form or register, i. e., the more homogeneous it is, the smaller the sample needs to be to provide representative results. Where little is known about a population, one can only take a random sample. In the case of a completely unknown language, just one informant is sufficient to be able to make many general statements about this language at the beginning. Where the object of study is well known, targeted samples directed at specific, narrowly defined research questions are both possible and necessary. It is practically impossible to create a representative sample of a population like “the German language” ⫺ whatever it is understood to be. However, it is perfectly possible to do so for particular parts of this whole ⫺ always with regard to specific research questions. In investigating base dialects we are generally dealing with smaller, less differentiated social units, i. e., villages. When investigating language forms that lie closer to the standard, we are confronted with barely manageable factors, multiple subdivided units. The requisite sample size and the experimental design also depend on a number of additional factors, e.g., the research objectives. Since the individual quality criteria are closely interconnected, they are not considered separately in the following, and are only discussed where they may pose a problem.

28. Methods and empirical standards

499

6. Quality criteria and areal linguistic methods o data acquisition 6.1. Locations Language in space is determined by language at particular locations. Linguistic forms may vary between two neighboring locations but they may also remain stable across larger areas. They may conform to strict boundaries but they may also change gradually via a slow shift in their distribution. A precise documentation of the situation would have to take every location into account. Given the number of unique locations, e.g., in Germany, and the resources usually available to a linguist, this is completely impossible. The number of locations can only be determined with respect to the research goals. Someone investigating, say, varieties close to the standard can still achieve valid results with a less dense net of survey locations, since the geographic variation is on a broad scale. Someone wanting to record old base dialects, however, will require a rather closely meshed net of survey locations; linguistic forms can differ at locations just a few kilometers apart. The density of the net of survey locations is usually determined by the possibilities and resources at the researcher’s disposal, and by the research objectives. Mathematical procedures for determining the representativity of a net of survey locations are conceivable; they presuppose a good deal of knowledge about the number and distribution of the variants to be expected within an area, however. In principle, it would be necessary to devise a separate net of survey locations for each linguistic phenomenon, dependent on the number and distribution of the expected variants. The criteria determining the appropriate density for the net of survey locations for a dialectological survey can be summarized as in Figure 28.1. A narrow net of survey locations may help to minimize validity and reliability problems. In acquiring data, a higher probability of error is often accepted if it means the data can be collected with relatively little effort. This is true of the so-called indirect method for instance. Theoretically, someone who has measurement results from many locations could reduce the density of the net of survey locations by grouping neighboring locations together. When groups of, say, five locations are formed, an erroneous measurement at one of the locations is of less consequence than it was before the grouping. Within the model, there is now only one location but it is represented by five samples. Someone who combines five samples from five different informants from different but

less dense

net of survey locations

more dense

(1) few

(linguistic) variants (for a given content)

many

(2) constant

relative size of the “language areas”

irregular

(3) direct

survey method

indirect

(4) close to standard

form/level of speech

further from standard

(5) (rather) high

number of questions canvassed within the study

(rather) low

(6) large

area surveyed

small

왗

왘

Fig. 28.1: Criteria for determining the density of the survey location network

500

V. Data collection and corpus-building neighboring communities and then calculates a single value from these samples or weights the varying values according to frequency thus lowers one of the above quality criteria, namely (the density of the net of) locations, but also raises the value of another parameter, the (number of) speakers.

6.2. Speakers In the investigation of language in space, geographic space features as an independent variable, with language, varying from location to location, as a dependent one. Locations can only be operationalized via speakers and their social data. In terms of such social data, an ideal speaker would be someone who grew up in a particular place and still lives there, who is well integrated into social life there, and whose parents and ancestors also came from there. Such speakers are quite rare nowadays; they nevertheless remain the ideal, since all other sociopragmatic parameters need to be kept constant to enable study of the phenomenon “space” in isolation. Social science textbooks (cf., e.g., Haseloff and Hoffmann 1970: 146) cite a minimum sample size of around 30 for statistical significance because a normal distribution (bell curve) begins to emerge from this number of measurements. Bortz (2005: 103) talks of “smaller” samples where there are less than 30 tokens and “larger” ones where there are more than 30. Depending on the range (deviation) of the results, it may be possible to reduce the number of samples, but sometimes more may be necessary. This can be illustrated by the example of a number of small objects in a bag: if there are only two distinct, evenly distributed types of objects in the bag, only a few of them need to be pulled out to be able to make reliable statements about the types of objects and relatively reliable ones about their distribution. If the objects contained in the bag are of many types, shapes and colors ⫺ say 100 ⫺ then a sample of 30 is as good as useless. If the objects are well mixed, the size of the bag and the total number of objects it contains only play a marginal role here (cf. Szameitat and Koller 1958). Statistics gives us formulas with which to calculate for large populations the sample size required to capture phenomena of a given frequency with a defined level of probability (cf., e.g., Bortz 2005: 126⫺128; Kähler 2008: 319⫺354; König 1982). For example: in a sample of 30, objects which make up ten percent of a population are ninety-five percent likely to occur once; in a random sample of 300, even objects which make up just one percent will occur with the same probability (Sankoff 2005: 1000). If we relate these considerations back to the issue of informant selection, a small village would be comparable to the bag containing only a few variants, where there is a relatively high uniformity. Assuming that we want to explore this uniformity, it may suffice to work with a single, competent speaker in whose ability to perform as expected we are ⫺ thanks to outside data (origin, age, other social parameters) ⫺ confident. This is particularly the case when the community’s least standard speech level is the focus of attention. Nowadays this speech level normally differs considerably from the language usually spoken, so that it is no problem for a speaker to mentally isolate the old dialect and (re)produce it for the researcher. This entails that the number of sociopragmatically determined variants is relatively small and that the informant can be assumed to have a good command of the local forms of speech and generally even be able to make statements about their functionality in communicative events.

28. Methods and empirical standards

501

A larger locality can be equated with the bag holding substantially more distinct objects: the bigger a place, the more socially differentiated it usually is and hence the more variable its language is and the more factors there are which may affect results. Increasing the number of speakers may help to make the results more valid and representative, since it will reduce the sampling error. It is hard to raise the number of speakers (who should be as identical as possible with regard to social parameters) to a statistically significant level since their social data will then inevitably become less uniform. This can be a source of additional variation that is not geographically determined. Further, practical constraints (e.g., the sheer amount of analysis needed) often dictate that an increase in the number of speakers is only feasible when coupled with a reduction in the number of phenomena examined. Moreover, the less stable the linguistic phenomena under investigation are and the less conscious speakers are of them, the less controllable they are in an experimental set-up and the more likely it is that other factors will interfere. Examples of such interference factors might be contemporary “flexible” speakers’ (Macha 1991) role perceptions: these are hard to manipulate and subject both to ineradicable microshifts in the communicative context over the course of the test and to individual characteristics of the informants that can have an effect on speech production within an otherwise uniform experimental set-up. Here too, greater reliability and validity can be achieved by increasing the number of samples and hence informants, but extraneous factors limit any such increase. Figure 28.2 presents the parameters (names the criteria) that affect the number of speakers. We could think about how to derive concrete representative numbers of speakers from this scheme if we had sufficient knowledge about the dimensions listed in Figure 28.2 to be able to render them as numerical proportions. But research is far from this point. Practically too, there is little point making the effort to develop the relevant values, since consistent results can still be achieved without them. Therefore ⫺ and just as with the selection of an appropriate net of survey places ⫺ the number of informants is generally decided intuitively, dependent on past experience, the research objectives and the resources available. That it is possible to produce high quality results on this basis is due to the input of findings from earlier research and the fact that language is still ⫺ despite all its variability ⫺ a relatively uniform entity subject to social controls.

more

speakers per location

less

(1) variable

language at the location

relatively uniform

(2) large

number of variants for a given communicative function

small

(3) high

speakers’ awareness of the phenomenon

low

(4) large

size of the location

small

(5) strong

strength of the social network

weak

왗

왘

(6) indirect

survey method

direct

(7) small

number of phenomena examined

large

Fig. 28.2: Criteria for determining the number of informants per location

502

V. Data collection and corpus-building

6.3. Experimental situations: Register Anyone aiming to describe language in space needs to ensure that the values he/she wants to compare really are comparable. This requires limiting investigation to speech contexts, levels, styles, genres and registers that are as identical as possible. If the investigation is confined to narrowly defined manners of speaking that are kept steady, relatively little variation within a location need be anticipated. Where there is little variation, the example of a bag holding just a few objects is apt: one does not have to reach into it very often to gain an accurate picture of which objects are inside. The overwhelming majority of language-geographic surveys are based on purposebuilt corpora created within an experimental frame (see Wagener 1988 for general information). In these experimental or interview contexts, a narrowly sociologically defined type of speaker is stimulated to speak within a relatively fixed predetermined framework.

6.3.1. Direct method: Intended base dialect In the past, the construct “base dialect” has contributed to the creation, in both major and minor research projects (linguistic atlases, regional grammars), of language maps that bear high quality depictions of language variation in space. Although the variability of local languages has been known at least since Schuchardt (1885), the construct “(intended) base dialect” continues to be a successful heuristic concept in data collection for the production of language maps. Recent language change and variation processes in various regions have left this construct devoid of content, or retargeted at new regional languages in its stead. The concept will, however, retain its significance for research into the “lower” end of the register scale (also cf. Lenz 2004). The direct method was developed around 1900 by Romance scholars (Gillie´ron and Edmont 1902⫺1969) for the research of such “base dialects”. A trained fieldworker makes on-location recordings and works through a comprehensive questionnaire with the informant (cf., e.g., Hotzenköcherle 1962; Bellmann 1994; König 1997; for basic information on the method, see Werlen 1984). Normally, rather than speaking in his/her local variety with the usually previously unknown interviewer, the informant uses a form somewhat closer to the standard language. However, when an informant is told that the interview is about the old base dialects, he/she is generally capable of producing it, virtually as a series of quotations, for the fieldworker. The investigator and the informant are united in the attempt to unearth the oldest accessible form at a particular location (König 1997: 24); the informant scours the “antiquarian part” of his linguistic memory (Macha 1991: 86). This keeps the context constant and ensures that comparable language material appears on the maps. The informant is cast as an expert about the language situation at a particular location, demonstrating fluency and selecting and evaluating variants. This role also sufficiently motivates the informant to endure the usually lengthy interview. This experimental set-up is easy to establish and can furnish a language corpus which permits the production of maps as valid models of language change in geographic space. The observer’s paradox (Labov 1971: 135⫺136) does not play a role in it. Further, it is capable of generating a relatively homogeneous corpus in terms of register. It therefore permits work with a very low number of informants ⫺ in principle, one informant who

28. Methods and empirical standards meets the conditions described above is sufficient. Nonetheless, the results remain representative of what is termed the base dialect of a location. They can be generalized to situations where two or more speakers from a location talk about local topics. Normally, several people are interviewed in such surveys, each, for economic reasons, about different linguistic issues. A trained interviewer will quickly notice how competent an informant is and how reliably he/she can provide information about the location’s norms. Where uncertain, the interviewer may repeat a particular question or even go over whole parts of the questionnaire a second time with a second informant. This method also increases the number of experiments, which in turn increases the quality of the material obtained and thus the reliability, validity and representativity of the findings. In this way modern linguistic atlases provide not only a more-or-less complete phonology and morphology and comprehensive information about the lexicon, but also (in their latest generation) many syntactic maps (Eroms, Röder and Spannbauer-Pollmann 2006). They also, albeit not systematically, make available a wealth of information on recent linguistic developments. This happens whenever the fieldworker has also recorded speech forms that are inconsistent with the primary objective of the recording, the “base dialect”. Such forms are produced constantly and spontaneously by the fieldworker and the informant in the course of their common endeavor, since in most cases the whole interview takes place at a “higher” speech level, with the base dialect merely being quoted for conservation purposes. If “spontaneous forms” are plotted onto a map, impressive images of linguistic developments can be obtained (cf., e.g., SDS 1965: 16; SDS 1975: 256). With the direct method, the phonetic accuracy of the corpus is highly reliable, especially if the data is stored as a sound recording. Validity and representativity are also enhanced because the linguist is observing the communicative events, e.g., during spontaneous speech, he/she can compare the answers offered with expectations based on previous experience. Should contradictions arise, he/she has the opportunity to repeat the experiment by asking if there are other ways of referring to the topic at hand. The investigator may even make suggestions or propose particular forms. If different variants are produced, the fieldworker may ask questions about differences in usage. The replies need to be taken seriously since the informants usually have a competence that encompasses more than one variant. The modern, “flexible” speaker has a command of several “styles” and registers and makes use of them more or less intentionally. He/she is aware of which forms of speech conform to the old “local norm” and can provide information about the significance of “more recent” variants and the conditions under which they are appropriate. Needless to say, all this must be noted in the protocol and it must always be asked to what extent the information offered corresponds to actual usage, for this is not always the case (cf., e.g., König 2004; 1975: 352⫺353). There ought to be no objections to this procedure being employed as a dialect-geographical method. In most of the social sciences, interviews in which informants discuss social behavior are used, and this is much less subject to control than the information on linguistic behavior obtained in dialectological survey interviews, since language is in permanent use in these contexts and every spontaneous occurrence of an element as well as every comment on its use increases the number of samples and thus the validity and representativity of the material collected. Furthermore, the procedure is highly economical: all that is asked is where variants occur or where, on the grounds of previous knowledge or experience,

503

504

V. Data collection and corpus-building they are suspected to occur (e.g., in a neighboring location). Moreover, informants’ judgments may replace direct observations that can be very difficult to make. The situation described here also in general applies to telephone-based methods of language data collection like the one used for the Atlas of North American English (Labov, Ash and Boberg 2006).

6.3.2. Direct method: Intended standard In addition to the investigation of base dialects described above, there is a second context in which homogeneous material can be obtained. In it, the informant is asked to read out a written text. For example, he/she may be instructed to read the text as if in front of a school class. It is not hard for the informant to mentally evoke and maintain this imagined situation over a sustained period. What results is a form of speech that also occurs in “natural” contexts; the results can thus be generalized, in that they are representative of this level of speech. They are also reliable and valid for it, but they are confined to a very narrow linguistic aspect, namely, pronunciation. The fact that the text to be read is in written form makes it impossible to vary other descriptive fields. The results have no representativity in terms of the entire spoken standard (e.g., with regard to morphology or syntax; cf. König 1989; 1997: 27).

6.3.3. Direct method: Regional languages It is harder to set up experimental contexts and keep them stable over a longer period for speech forms that lie between the poles described above. This pertains to the surely most-spoken speech form today ⫺ everyday speech: vernaculars, regional languages, regional accents. This spectrum of variation between the poles of base dialect and standard is very unstable. The speech forms change according to the context and intended goal, depending on how the speaker understands the role he/she is playing and also on his/her attitude towards dialect or regional identity (Werlen 1984). The level of speech may also shift in line with the context or the topic. Speakers tend to be far less aware of their speech or of the norms governing it than with standard language or base dialect. The investigation of these speech forms is complicated by the fact that people are increasingly active in social groups that are not tied to a particular location, that seek (linguistic) role models in other areas and do not orient themselves to a local norm, however it may be defined. Thus, despite identical selection criteria (“ostensible” association with a particular location for at least two generations, a certain level of education), individual speakers’ modes of speech may characteristically deviate from those of the regionally orientated speakers under study. Nevertheless, there are successful projects which have managed to capture spatial structures in this intermediate stratum, e.g., the Mittelrheinischer Sprachatlas (Bellmann 1994; Bellmann, Schmidt and Herrgen 1994⫺2002; cf. Herrgen in this volume). In it, the base dialect of a younger, mobile group of speakers (commuters) was sought (data series 2) in addition to that of the older speakers (data series 1). In general, this method captured a speech form that can be referred to as regional dialect and which corresponds exactly with the intended “intermediate stratum” between base dialect and standard.

28. Methods and empirical standards Where the preconditions that applied for the Mittelrheinischer Sprachatlas are not met, it will be more difficult to survey regional languages empirically with regard to their geographic structuring, especially where there is also an interest in syntactic phenomena. For this, spontaneously spoken texts, preferably recorded under standardized conditions with informants from comparable social backgrounds, are necessary ⫺ i. e., speech recordings made under identical sociopragmatic conditions. As Labov (1966; cf. Kristiansen in this volume: section 2.2) demonstrated with his New York survey, creativity can be of assistance. A group conversation or the MapTask method developed in Edinburgh (Anderson et al. 1991) may also supply language material that is to some extent comparable across space. In the MapTask, stylized “maps” are handed to two dialogue partners. They contain a series of “landmarks”, or drawings representing notions. On one of the maps a path has been drawn from notion to notion, but not on the other. One of the informants, the “instruction giver” has to explain the path to the other informant, the “instruction follower”. Unbeknownst to the informants, the maps usually differ from one another, so that an animated conversation can develop. The “landmarks” are mentioned frequently and many spontaneously uttered tokens are obtained. The material produced in such quasi-natural situations can provide information on language use in analogous contexts at a particular location. With a corpus obtained in such a way, it is also possible to assign variants to particular sociopragmatic conditions. When describing geographic variation, particular contextual variants from undesired registers can be excluded, leading to higher quality results. Of course, this requires some knowledge about what is common at a location. Where this is not available, higher quality can be achieved by increasing the number of experiments, e.g., by increasing the number of informants per location, and then using quantitative analysis. For practical reasons, the number of informants cannot be increased indefinitely. However, the richness of variants in everyday language and the idiosyncratic inconsistency of today’s speakers seem to make this necessary. To obtain statistically sound figures, it would be necessary to know how many speakers with the same geographical background go their own individual linguistic way. Since there are no figures on this yet, any decision will need to be based on intuition and experience. This will ultimately be done on the basis of existing resources; but a compromise must be reached between two conflicting goals. The smaller the number of linguistic issues/variables to be dealt with, the higher the number of speakers may become and, vice versa, the higher the number of research topics, the longer the experiments will take and ⫺ if the expenditure and effort are held constant ⫺ the smaller the number of informants will need to be (see above). The number of experiments increases with the number of informants. But it is also possible to increase the former by conducting several experiments with one speaker, effectively increasing the amount of speech recorded ⫺ for instance, every realization of the negation particle not in a text can be regarded as an individual experiment. It is thus possible to reach conclusions about the use of the forms that occur, but not about variation between individuals: that can only be determined by increasing the number of informants.

6.3.4. Indirect method: Base dialects This method is above all associated with the Marburg School (Knoop, Putschke and Wiegand 1982; Wrede, Martin and Mitzka 1927⫺1956; Schmidt and Herrgen 2001⫺

505

506

V. Data collection and corpus-building 2009). Dialect data are collected with questionnaires and with the help of non-linguists. The school has achieved much and it continues to influence the scene, despite a striking reliability problem (Kleiner 2003). Orthographical transcription (into the standard alphabet) is useless for many problems; it is only adequate for recording phenomena the speaker is or can be made aware of. A distinction like that between Tiid and Zeit (‘time’) or between Appel and Apfel (‘apple’) can be captured well using the existing letters of the alphabet, and speakers are also aware of it. But they are not aware, for example, of distinctions in the aspiration of plosives, even though it could be rendered in writing, e.g., Phost could well be distinguished from Post (‘post’). The indirect method has provided very useful results for easily representable phenomena in phonology and morphology. Its particular strength is, however, the field of regional lexical variation. The speakers are aware of such differences (e.g., Fiedel vs. Geige ‘violin’); they are striking, and can be captured in normal written language. In this manner, Mitzka’s Deutscher Wortatlas (Mitzka [and Schmitt] 1956⫺1980) provides high quality results in the field of word geography, including in part phonetic/phonological distinctions. Even in the field of syntax, ingenious survey design can deliver findings of high reliability and validity, primarily by suggesting alternative forms and monitoring their level of acceptance in a region (Glaser 2000). In general it can be said that the more we know about the area under research, the more successful this method will be. It becomes increasingly possible to formulate questionaires in a language the informant understands and which makes him aware of alternatives. In case of aspiration, for example, one could ask, “In slow, carefully pronounced speech from where you live, does the p in Post sound more like a hard b or a p followed by h?”. Formulating the question in such a way ought to be enough to obtain a (not yet available) geographic distribution for this important phenomenon of regional German phonology, even via a written survey.

6.3.5. Indirect method: Regional languages The first atlas of German colloquial language is the Wortatlas der Deutschen Umgangssprachen by Jürgen Eichhoff (1977⫺2000). Two-thirds of his material (a total of 539 recordings) was collected indirectly by means of questionnaires, one-third via direct surveys in the field (Eichhoff 1977⫺2000, vol. 1, 15). Data from 402 survey locations is presented, the majority of which are towns with a medium to small population. In these towns, colloquial languages, i. e., regional accents, play the same role as the base dialects do in rural areas. Thus, geographically immobile informants with a “medium” level of education (prototypically low or middle-ranking public servants; Eichhoff 1977⫺2000: vol. 1, 14) are sufficiently aware of the “intermediate” speech form under investigation to be able to make competent statements about it. The controlled selection of informants and the restriction of the survey to lexical aspects contributed to the collection of a relatively uniform speech level and to the creation of maps of high validity and reliability. Today, the internet offers new opportunities for the indirect collection of materials. The Atlas der deutschen Alltagssprache (AdA, cf. Möller and Elspaß 2008; Elspaß and Möller 2006) is one of the first dialect-geographic projects to take avail of this method of obtaining a corpus. Here, the question of quality criteria takes a different form. The new method brings little freedom for the researcher; here ⫺ in contrast to the direct method ⫺ the question is not “What do I want to investigate?” followed by selection of

28. Methods and empirical standards the appropriate experimental design. Rather, it is “What can be investigated by this method; for which type of speaker, for which speech forms and linguistic issues does this method provide high quality results?” Informants are chosen more or less at random since their selection is not controllable: they are all computer users and internet surfers ⫺ predominantly 20 to 40 years old. As with the Deutscher Sprachatlas and the Deutscher Wortatlas, the sheer mass of informants provides valid results. The number of samples per location can be raised by reassigning the data from smaller regions to larger central locations. This often also increases the number of variants for a location or region and enhances quality (see section 6.1). Quantitative statements about usage can be made on the basis of the relative frequency of variants, but the conditions determining the use of a particular variant cannot be elicited. This method is especially suitable for the investigation of lexical variation. The problems of the indirect method with regard to other linguistic levels have already been mentioned (section 6.3.4). On the other hand, in taking stock of lexical variation or other phenomena of which the speaker is aware (or can be made more aware through sophisticated questioning), the speed and effectiveness of the method are unbeatable. Moreover, it is exceptionally inexpensive, in part because the materials are already available in machine-readable form. It is impossible to obtain as much social data via web-based surveys as in a direct on-location recording. However, one can learn the age of the informants (e.g., to within a decade), their gender, something about their education or occupation (classified into three to four broad groups), and about their level of geographic mobility (also in two or three categories). If this information is obtained, a high enough number of informants (five to ten thousand is realistically achievable in the German-speaking area) can ensure high validity and representativity. For example, 400 locations and 6000 replies yields an average of 15 results per location. These results are far more comprehensive than those obtainable with other recording methods. The variants obtained with this method always represent speakers’ judgments about which variants are common at a location. These speakers have, however, a much greater than average interest in language. Otherwise, they would not voluntarily complete the questionnaire. Moreover, they are of the most varied social status and tend to be better educated than the average. It is true that many of these speakers would not have been included in a direct method survey. However, they have lived in one place long enough to feel capable of judgments about the language common there. Thus, although there are problems with the language material obtained, it can yield high quality results when interpreted carefully. The high number of replies makes it possible to use filters to exclude particular groups or to concentrate exclusively upon them (e.g., mapping the utterances of over 60-yearolds, or people with a full secondary education, etc.). The replies of people who would otherwise have been excluded from a study with a purely areal focus provide further insights into the linguistic reality at a location or in a region. To date, the potential of this method of data collection is far from exhausted in linguistics. Coupled with sophisticated questioning, it is capable of significant future achievements ⫺ not just in the lexical field, but also in the areal study of idiom, morphology and syntax.

7. Existing corpora: Corpus linguistics Corpora of spoken and written language available on the internet are growing in significance, not least for linguistic geography. This is primarily due to the fact that they offer

507

508

V. Data collection and corpus-building new possibilities of analysis via electronic data processing. But such corpora, like those provided by the Institut für deutsche Sprache (‘Institute for the German Language’, IDS, Mannheim), the Bayerisches Archiv für Sprachsignale (‘Bavarian Archive for Speech Signals’, BAS), or found in the Digitales Wörterbuch der deutschen Sprache des 20. Jahrhunderts (‘Digital Dictionary of German 20th-Century Language’, DWDS) of the BerlinBrandenburg Academy of Sciences and Humanities for example, have not been systematically compiled with an eye to their representativity across space. Once major corpora of written and of transcribed spoken language that take systematic account of regional aspects (and are classified into text types) are available on the net, it will become possible to detect regionalisms via automated mass frequency analyses of hitherto unattainable dimensions ⫺ both qualitative and quantitative. The huge database that will then be at hand and the calculations of significance it will render possible will permit a comparison of areally based frequencies and the generation of results of exceptional quality. Unfortunately, the opportunities such corpora offer for areally oriented variation linguistics have yet to be explored in corpus linguistics. Within corpus linguists’ discussions about the question of representativity, issues of space and areality have regrettably played virtually no role to date (cf., e.g., Biber 1993; McEnery, Xiao and Tono 2006; McEnery and Wilson 2007).

8. Quality criteria and attitudinal research Here, the question is how people think about language and dialects, how they judge them, how much prestige they accord regionally colored speech forms and speakers, and how they assess the distribution of dialects (cf., e.g., Preston 2005). Only to a degree are these linguistic questions; they could just as well be posed by social scientists. The surveys made about the popularity of dialects demonstrate this. For this type of inquiry, the conditions described in section 6 are no longer valid; the quality criteria for surveys developed within the social sciences should be applied. Above all, the conditions that apply to the criterion of representativity are different. In order to make statements about the attitude of “the Germans” towards an interesting phenomenon, for example, a group which is “representative” in terms of the criteria of the social sciences needs to be polled. It needs to be kept in mind for surveys about the popularity of varieties in particular that sufficiently large groups from all of the regions in the survey area need to be surveyed, given that regional differences are especially significant for this research topic. Since the size of representative samples normally exceeds the means available to linguists, statements are generally only made about the group examined. A survey by Hundt (1992), which furnishes insights into the evaluation of regionally colored standard language by 175 informants, can arbitrarily be chosen as an example. Statistical tests which demonstrate the significance of some specific differences in the evaluation of regionally colored standard language create an impression of objectivity. In the absence of representativity, however, the impression is not warranted ⫺ for whom, aside from the 175 informants scattered all over Germany, can the result be regarded as valid? When social scientists, like those at the Institut für Demoskopie in Allensbach, conduct comparable surveys, they have no problems with representativity (1814 informants in the last survey

28. Methods and empirical standards in 2008, cf. Institut für Demoskopie Allensbach 2008), but very real ones with reliability and validity (e.g., with regard to how they label dialects). One would like to suggest that they seek the advice of dialectologists on future surveys.

9. Reerences Anderson, A. H., M. Bade, E. G. Bard, E. Boyle, G. Doherty, S. Garrod, S. Isard, J. Kowtko, J. McAllister, J. Miller, C. Sottillo, H. Thompson, R. Weinert 1991 The HCRC map task corpus. Language and Speech 34 (4): 351⫺366. Bellmann, Günter 1994 Einführung in den Mittelrheinischen Sprachatlas (MRhSA) Tübingen: Niemeyer. Bellmann, Günter, Jürgen Erich Schmidt and Joachim Herrgen (eds.) 1994⫺2002 Mittelrheinischer Sprachatlas. Tübingen: Niemeyer. Biber, Douglas 1993 Representativeness in corpus design. Literary and Linguistic Computing 8: 243⫺257. Bortz, Jürgen 2005 Statistik für Human- und Sozialwissenschaftler. Heidelberg: Springer. Eichhoff, Jürgen 1977⫺2000 Wortatlas der deutschen Umgangssprachen. 4 vols. Berne/Munich: Francke. Elspaß, Stephan and Robert Möller 2006 Internet-Exploration: Von den Chancen, die eine Online-Erhebung regional gefärbter Alltagssprache bietet. Osnabrücker Beiträge zur Sprachtheorie 68: 143⫺158. Eroms, Hans-Werner, Birgit Röder and Rosemarie Spannbauer-Pollmann 2006 Sprachatlas von Niederbayern, vol. 1: Einführungsband mit Syntaxauswertung. Heidelberg: Winter. Friedrichs, Jürgen 1990 Methoden empirischer Sozialforschung. 14th ed. Opladen: Westdeutscher Verlag. Glaser, Elvira 2000 Erhebungsmethoden dialektaler Syntax. In: Dieter Stellmacher (ed.), Dialektologie zwischen Tradition und Neuansätzen. Beiträge der internationalen Dialektologentagung, Göttingen, 19.⫺21. Oktober 1998, 258⫺276. Stuttgart: Steiner Gillie´ron, Jules and Edmond Edmont 1902⫺1910 Atlas Linguistique de la France (ALF). Paris: Champion. Haseloff, Otto Walter and Hans-Joachim Hoffmann 1970 Kleines Lehrbuch der Statistik. 4th ed. Berlin: de Gruyter. Hotzenköcherle, Rudolf 1962 Einführung in den Sprachatlas der deutschen Schweiz, Teil A. Berne: Francke. Hundt, Markus 1992 Einstellungen gegenüber dialektal gefärbter Standardsprache. Eine empirische Untersuchung zum Bairischen, Pfälzischen und Schwäbischen. (Zeitschrift für Dialektologie und Linguistik. Beihefte 78.) Stuttgart: Steiner. Institut für Demoskopie Allensbach 2008 Auch außerhalb Bayerns wird Bayerisch gern gehört. Allensbacher Berichte 4, Umfrage Nr. 10016. Kähler, Wolf-Michael 2008 Statistische Datenanalyse. 5th ed. Wiesbaden: Vieweg⫹Teubner. Kleiner, Stefan 2006 Geschriebener Dialekt in Bayerisch-Schwaben. Ein Vergleich indirekt erhobener dialektaler Laienschreibungen mit ihren lautschriftlichen Entsprechungen. (Phonai 48.) Tübingen: Niemeyer.

509

510

V. Data collection and corpus-building Knoop, Ulrich, Wolfgang Putschke and Herbert Ernst Wiegand 1982 Die Marburger Schule: Entstehung und frühe Entwicklung der Dialektogeographie. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke and Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, 38⫺92. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.1.) Berlin/New York: de Gruyter. König, Werner 1975 Überlegungen zu Aufnahmesituation und Informant bei sprachgeographischen Erhebungen. Mit einem Vorschlag zur Operationalisierung individueller Eigenschaften von Informanten. Deutsche Sprache 4/75: 346⫺364. König, Werner 1982 Probleme der Repräsentativität in der Dialektologie. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke and Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, 463⫺485. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.1.) Berlin/New York: de Gruyter. König, Werner 1988 Zum Problem der engen phonetischen Transkription. Zeitschrift für Dialektologie und Linguistik 55: 155⫺178. König, Werner 1989 Atlas zur Aussprache des Schriftdeutschen in der Bundesrepublik Deutschland. 2 vols. Ismaning: Hueber König, Werner 1997 Sprachatlas von Bayerisch-Schwaben, vol. 1: Einführung. Heidelberg: Winter König, Werner 2004 “Stimmen tut’s nicht, aber es ist was Besonderes!” Die Motivation/Intention von Informanten als Faktor für “hyperkorrekte” Antworten bei sprachgeographischen Erhebungen. In: Albrecht Greule, Rupert Hochholzer and Alfred Wildfeuer (eds.), Die bairische Sprache. Studien zu ihrer Geographie, Grammatik, Lexik und Pragmatik. Festschrift Ludwig Zehetner, 189⫺193. (Regensburger Dialektforum 5.) Regensburg: edition vulpes. Kromrey, Helmut 2006 Empirische Sozialforschung. Modelle und Methoden der standardisierten Datenerhebung und Datenauswertung. 11th ed. (UTB 1040.) Stuttgart: Lucius & Lucius/UTB. Labov, William 1966 The Social Stratification of English in New York City. Washington DC: Center for Applied Linguistics. Labov, William 1971 Das Studium der Sprache im sozialen Kontext. In: Wolfgang Klein and Dieter Wunderlich (eds.), Aspekte der Soziolinguistik, 111⫺194. Frankfurt: Athenäeum. Labov, William, Sharon Ash and Charles Boberg 2006 The Atlas of North American English: Phonetics, Phonology and Sound Change. Berlin/ New York: Mouton de Gruyter. Lamnek, Siegfried 2005 Qualitative Sozialforschung. 4th ed. Weinheim/Basel: Beltz. Lenz, Alexandra N. 2004 Zur Interpretation des intendierten Ortsdialekts. In: Alexandra N. Lenz, Edgar Radtke and Simone Zwickl (eds.), Variation im Raum, 113⫺131. (Variolingua 20.) Frankfurt: Peter Lang. Macha, Jürgen 1991 Der flexible Sprecher. Untersuchungen zu Sprache und Sprachbewusstsein rheinischer Handwerksmeister. Cologne: Böhlau. McEnery, Tony and Andrew Wilson 2007 Corpus Linguistics. An Introduction. Edinburgh: Edinburgh University Press.

28. Methods and empirical standards

511

McEnery, Tony, Richard Xiao and Yukio Tono 2006 Corpus-Based Language Studies. London/New York: Routledge. Mitzka, Walther [and Ludwig Erich Schmitt] 1956⫺1980 Deutscher Wortatlas. 22 vols. [From vol. 5 on; edited by Reiner Hildebrandt from vol. 18 on.] Gießen: Wilhelm Schmitz. Möller, Robert and Stephan Elspaß 2008 Erhebung dialektgeographischer Daten per Internet: Ein Atlasprojekt zur deutschen Alltagssprache. In: Stephan Elspaß and Werner König (eds.), Sprachgeographie digital ⫺ die neue Generation der Sprachatlanten, 115⫺132. (Germanistische Linguistik 190⫺191.) Hildesheim/Zürich/New York: Olms. Preston, Dennis R. 2005 Perceptual Dialectology. In: Ulrich Ammon, Norbert Dittmar, Klaus J. Mattheier and Peter Trudgill (eds.), Sociolinguistics: An International Handbook of the Science of Language and Society, vol. 2, 1683⫺1696. 2nd ed. (Handbücher zur Sprach- und Kommunikationswissenschaft 3.2.) Berlin/New York: de Gruyter. Sankoff, David 2005 Problems of Representativeness. In: Ulrich Ammon, Norbert Dittmar, Klaus J. Mattheier and Peter Trudgill (eds.), Sociolinguistics: An International Handbook of the Science of Language and Society, vol. 2, 998⫺1003. 2nd ed. (Handbücher zur Sprach- und Kommunikationswissenschaft 3.2.) Berlin/New York: de Gruyter. Schmidt, Jürgen Erich and Joachim Herrgen (eds.) 2001⫺2009 Digitaler Wenker-Atlas. Erste vollständige Ausgabe von Georg Wenkers “Sprachatlas des Deutschen Reichs”. Compiled by Alfred Lameli, Alexandra Lenz, Jost Nickel and Roland Kehrein, Karl-Heinz Müller, Stefan Rabanus. Marburg: Forschungszentrum Deutscher Sprachatlas. Schnell, Rainer, Paul B. Hill and Elke Esser 2005 Methoden der empirischen Sozialforschung. 7th ed. Munich/Vienna: Oldenbourg. Schuchardt, Hugo [1885]1928 Über die Lautgesetze. Gegen die Junggrammatiker. In: Leo Spitzer (ed.), Hugo Schuchardt-Brevier. Ein Vademecum der allgemeinen Sprachwissenschaft, 50⫺87. Halle: Niemeyer. Schulz, Thomas, Klaus- Peter Muthig and Karlfritz Koeppler 1981 Theorie, Experiment und Versuchsplanung in der Psychologie. Stuttgart: Kohlhammer. SDS 1965/1975 Sprachatlas der deutschen Schweiz, vol. II: Lautgeographie II: Vokalquantität; vol. III: Formengeographie. Compiled by Rudolf Hotzenköcherle and Rudolf Trüb. Berne: Francke. Szameitat, Klaus and Siegfried Koller 1958 Über den Umfang und die Genauigkeit von Stichproben. Wirtschaft und Statistik 10: 10⫺16. Wagener, Peter 1988 Untersuchungen zur Methodologie und Methodik der Dialektologie. (Deutsche Dialektgeographie 86). Marburg: Elwert. Werlen, Erika 1984 Studien zur Datenerhebung in der Dialektologie. (Zeitschrift für Dialektologie und Linguistik. Beihefte 46.) Wiesbaden: Steiner. Wrede, Ferdinand, Walther Mitzka and Bernhard Martin 1927⫺1956 Deutscher Sprachatlas. Auf Grund des von Georg Wenker begründeten Sprachatlas des Deutschen Reichs. Marburg: Elwert.

Werner König, Augsburg (Germany)

512

V. Data collection and corpus-building

29. Investigating language in space: Questionnaire and interview 1. 2. 3. 4. 5.

1.

Elicitation versus spontaneous discourse data Oral versus written elicitation Survey of oral elicitation techniques Survey of written techniques References

Elicitation versus spontaneous discourse data

1.1. Introductory remarks In spontaneous discourse, utterances are generated by a speaker in order to reach some communicative goal. What the speaker is saying, and the way s/he is saying it, is rooted in an interactive context. Elicited linguistic data are of a very different kind. They are utterances which the speaker is explicitly asked to produce by an investigator who is interested in the structure of the language itself. Elicited data are responses to linguistic tasks and thus arise from a situation that is more or less artificial, or experimental. Dialectological research has developed several data elicitation techniques. The question of which is the best elicitation method is perhaps misleading. Rather, the different techniques have specific profiles of strengths and weaknesses. The present article is intended to give the researcher guidelines for her/his choice of method. In practice, a combination of techniques often turns out to be useful. If several methods are applied in parallel, it is easier to estimate the extent to which a particular method impacts upon the results.

1.2. Comparability versus reliability o data - a contradiction? Investigating linguistic variation along the spatial dimension presents us with several methodological challenges regarding the collection of comparable and at the same time reliable data. Comparability with respect to the dialect-geographical structure of the landscape and reliability in relation to the language in actual use seem to contradict one another. On the one hand, relevant patterns of language variation across space can only be uncovered if comparable, i. e., more or less standardized, material is won or made available. On the other hand, all research into spatial linguistic variation deals with nonstandard varieties, often base dialects, accorded relatively low prestige in many cultures, so that it is a far from trivial question how the linguist can obtain “natural” data (i. e., data that are relatively free of standardization) in the course of an unnatural testing situation. There is no doubt that spontaneous discourse is among the most reliable, or authentic, sources of data available to linguists. However, discourse data are often not sufficient to achieve a dialect-geographically relevant picture (e.g., the precise geographi-

29. Questionnaire and interview cal distribution of variants). In this article we will discuss ways of resolving the dilemma of reliable versus geographically relevant data. It is important to note that each method is always a compromise between partially competing requirements.

1.3. Why elicit data? It is uncontroversial that records (and transcripts) of spontaneous discourse are the class of data least influenced by the investigator, especially discourse recorded in the investigator’s absence. This is especially true of dialects as nonstandard varieties. Large annotated corpora have become an indispensable tool for testing hypotheses of syntactic theory (cf., e.g., Bresnan et al. 2007). It is no surprise, therefore, that the methods of corpus linguistics have been adopted in dialectology, too. First to mention in this context is the pioneering work on dialects of England initiated by Bernd Kortmann and colleagues at the University of Freiburg (cf. Kortmann 2003 for a survey). Kortmann and his coauthors have compiled a large corpus of texts from all areas of England that were originally recorded in the context of oral history research (FRED, Freiburg English Dialect Corpus, cf. Kortmann and Wagner 2005). This line of research demonstrates that corpus analyses may reveal not only a great variety of nonstandard constructions and their conditions of use, but also, to a certain degree, remarkable dialect-geographical differences. It must be noted, however, that the corpus-linguistic approach is also subject to limitations (Glaser 2000; Cornips 2002). These limitations can be evaded only by means of controlled data elicitation. First, the geographical picture that can be revealed on the basis of corpora is still very rough. Although the geographical distribution of variables will show up when corpora from different localities are compared, the findings are still quite punctiform in comparison to the dense networks of measuring points used by dialect-geographical atlases. Time and financial constraints make it impossible for large corpora from as many localities as an atlas uses for investigation to be compiled. It is therefore very improbable that a corpus-based study would be able to identify the precise position of an isogloss (but this is exactly what dialectologists want to know). A dense network of maximally comparable data is mandatory for a linguistic atlas; this can only be achieved via standardization of the elicitation technique. Second, only for phenomena which occur very frequently in discourse can corpora guarantee the level of comparability between localities that is required in dialect geography. For less frequent phenomena some standardized elicitation technique is necessary to ensure comparability. Third, corpora do not contain negative evidence, i. e., the corpus does not tell us which variants do not occur in the area. For a precise geographical picture as well as for questions of more theoretical interest (e.g., questions on micro-parameters, the covariation of grammatical variables that are assumed to be connected at some level of structural organization, cf. Kayne 1996), it is sometimes necessary to know whether or not a particular pronunciation, construction, etc. with a particular intended meaning is possible. Moreover, for dialectological questions, it is sometimes very important to have access to the most archaic forms, e.g., in order to reconstruct the direction of a sound change not attested in written records. Sometimes such archaic forms are not used any more in everyday communication, but speakers still have a passive knowledge of the forms. Fourth, although the corpus-based approach seems to be especially well suited for syntax, it is less so for phonology, since the transcription of long texts (made in oral history

513

514

V. Data collection and corpus-building research, for instance) is usually not very phonetically precise, being too strongly oriented towards standard orthography. Fifth, and perhaps most importantly, large electronic corpora are simply not available for many areas. Elicitation is a relatively economical way of gaining a general orientation with respect to the combination of linguistic variables occurring in a certain area. In sum, corpus-based research on dialects is unrivaled as far as the reliability, or authenticity, of the data is concerned. It must be kept in mind that an elicitation procedure creates artifacts (task effects, repetition effects, order effects) arising from the unnatural situation the informant is exposed to. As Cornips (2002) points out in her study on the distribution of om and voor as infinitival complementizers in dialects of Dutch, the main differences between spontaneous discourse data and elicited data are found not so much in the purely linguistic distribution of the variants (voor is used more often in purposive contexts); rather, they are related to the social significance of the variants (speakers with a high level of education respond in test situations only with om, the prestige variant, whereas they use both om and voor in spontaneous discourse, Cornips 2002: 91). These results mean that standardized elicitation techniques may reveal the inventories and linguistic function of expressions at a given locality in a fairly reliable way, but not their social significance and stratification. These are certainly regrettable limitations, yet we think that they are tolerable, at least from the point of view of traditional dialect-geographical epistemological interest, which focuses on that layer of the language with maximal geographical differentiation. Of course, this sociolinguistically very limited (and perhaps over-idealized) picture can, and must, be supplemented by studies integrating the dynamic aspects of language variation on not only the spatial, but also the social and temporal dimensions, as realized (to a degree) by the two-dimensional Mittelrheinischer Sprachatlas in the German Rhine area (Bellmann, Schmidt and Herrgen 1994⫺2002), which worked with two generations of speakers (for further exploration of the dynamic approach cf. Schmidt in this volume and to appear). It is nevertheless important to note that for an appropriate interpretation of the findings of multidimensional variationist studies, knowledge of the linguistic layer with maximal geographical differentiation is mandatory (cf. e.g., Siebenhaar’s [2000] sociolinguistic study on the city dialect of Aarau, Switzerland, where his results are interpreted in comparison with those of the classical dialect atlas, Sprachatlas der deutschen Schweiz [Hotzenköcherle 1962⫺ 2003]).

2. Oral versus written elicitation 2.1. Brie historical background In oral elicitation, the informant and the investigator are present at the same time and place (sometimes, cf. Labov, Ash and Boberg 2006, oral elicitation is performed via telephone). The informant uses spoken language, which is recorded by the investigator. Methods based on oral elicitation are called direct methods. In indirect methods, the informant and the investigator are separated spatially and temporally. The informant fills out a written questionnaire, using the normal alphabet. In both methods a set questionnaire of linguistic tasks is used. In the following, I will make a terminological distinc-

29. Questionnaire and interview tion between the interview questionnaire used in the direct method, and the written questionnaire used in the indirect method. Historically, the contrast between direct and indirect methods coincides with, respectively, the Romance and the German traditions of dialect geography in the late nineteenth and early twentieth centuries. The ideal of the Romance tradition, most prominently represented by Jules Gillie´ron and his Atlas linguistique de la France, was to capture an instantaneous picture of the natural language use of the informant. This presupposes that only oral methods are to be applied, and that the informant is only minimally influenced by the elicitation technique. This degree of authenticity is reached at the expense of the density of survey locations for the linguistic atlas. The German tradition, represented by Georg Wenker and his Deutscher Sprachatlas (Wenker 1887⫺ 1923; Wrede, Martin and Mitzka 1927⫺1956), attempted to cover the area as densely as possible. It was only possible to achieve a rich network of survey locations by means of written questionnaires. Since 1876, 52,800 written questionnaires have been collected (König 2005: 139). The questionnaires contain 40 sentences in Standard German to be translated into the local dialects. (Insofar as the written questionnaires were sent to school teachers who were instructed to record how the sentences would be translated into the dialect of the local population ⫺ not necessarily that spoken by the teacher ⫺ this method perhaps includes a stronger direct component, with the teacher as investigator, than one would expect at first glance.) The Romance tradition has also had a significant influence on the methodology of newer atlases in the German-speaking world, e.g., atlases in the tradition of the Swiss Sprachatlas der deutschen Schweiz, which combine the “Romance” direct method with the “German” network density, whereby covering only a relatively small area (Hotzenköcherle 1962⫺2003).

2.2. Direct versus indirect methods: Advantages and disadvantages The main advantages of the indirect method lie in its relatively low costs in terms of time, finances and personnel (cf. Eichhoff 1982: 550). The indirect method is particularly well suited to an initial orientation to the distribution of variables found in an area. For example, the SAND project in the Netherlands and Belgium first undertook a pilot study with written questionnaires, the results of which served as the basis for later, more detailed oral interviews (Barbiers et al. 2004⫺: vol. 1, 8). The questionnaire material collected for the Deutscher Sprachatlas by Georg Wenker and successors could also be considered a pilot study insofar as much later work in German dialectology relies on the general picture provided by the Deutscher Sprachatlas material while at the same time refining it. The written data from the Syntaktischer Atlas der deutschen Schweiz (SADS) has served as a starting point for more detailed studies on particular phenomena (e.g., Seiler 2003; Bucheli Berger 2005). The physical absence of the investigator, although problematic in many respects, can be seen as an advantage of the indirect method, too. Some informants may feel disturbed by the presence of an unknown academic in their homes. Sending them written questionnaires is one way of dealing with the so-called observer’s paradox: “Our goal is to observe the way people use language when they are not observed” (Labov 1972: 61). Informants are less closely observed in filling out a questionnaire than they are in an interview situation.

515

516

V. Data collection and corpus-building In many respects, though, the direct method is superior. In the first place, the transcription into standard orthography by non-experts makes the precise phonetic interpretation of the results difficult. Written data is less problematic as far as vocabulary or syntax are concerned, cf. the Deutscher Wortatlas (Mitzka, Schmitt and Hildebrandt 1956⫺1980) as an example of a lexical atlas, the Atlas van de Nederlandse Dialectsyntaxis (Gerritsen 1991) and the SADS as examples of syntactic atlases, based on written material (cf. Kakhro 2005, where it is shown that material based on the 40 “Wenker sentences” originally collected for the Deutscher Sprachatlas yield good results for syntactic variation in German-speaking Switzerland; cf. also König 2005: 163, a syntactic map based on the original Deutscher Sprachatlas material). Note that Schmidt (2005) demonstrates that the original Deutscher Sprachatlas material has turned out to be fairly reliable, even for cases of phonological variation, since the results of the Deutscher Sprachatlas and the recent Mittelrheinischer Sprachatlas can easily be brought into relation with each other. However, interference from the standard variety is more probable in written tasks since most informants are not used to writing down their dialect. Furthermore, the informant is less subject to the investigator’s control when the indirect method is used. Not only does this mean that it is more difficult for the investigator to estimate how reliable, hesitant, concentrated, etc. the informant is, it also means that the investigator is not able to provide help when necessary, e.g., where a question is misunderstood (see section 4). In general, none of the informant’s reactions aside from the mere answers to the tasks (amusement, comments etc.), are recorded. It is impossible to compare the informant’s linguistic behavior in the interview with her/his ordinary speech. Written questionnaires must be relatively short in order to not exhaust the informant. In sum, the reliability of written material has perhaps been underestimated in dialectology. It seems that the indirect method yields reliable results as far as the geographical distribution of linguistic variables per se is concerned. However, its limitations are to be found in the exploration of the exact functional properties of the variants, such as context sensitivity or slight meaning differences.

3. Survey o oral elicitation techniques 3.1. Scene setting It seems to be very important that interviews are conducted in a place with which the informant is very familiar. Traditionally, the investigator visits informants in their homes (Hotzenköcherle 1962: 127). Bellmann (1994: 79) notes that some interviews for the Mittelrheinischer Sprachatlas were conducted in more public places, such as restaurant booths. A restaurant in the informant’s village turned out to be very suitable for the direct interviews conducted for the SADS. The place is still familiar to the informant, but there is no risk that the presence of the investigator could be perceived as invasive. Linguistic atlases in the tradition of the Sprachatlas der deutschen Schweiz work together with one informant at a time. The Mittelrheinischer Sprachatlas systematically conducted interviews with small groups of informants. Group interviews make it possible to abstract away from individual idiosyncrasies of one informant (competence-oriented approach, Bellmann 1994: 37⫺39). In oral interviews conducted for the SADS, very

29. Questionnaire and interview positive experiences have been reported with interviewing more than one person at a time, too. The informants often begin discussions with one another during the interview in a highly spontaneous manner. The benefits are particularly striking with mixed-sex couples: the informants are very familiar with each other and therefore relaxed, they are game enough to correct one another and their voices are easily distinguishable on tape (voice distinction is often a problem in group interviews). Furthermore, couple interviews guarantee a balanced ratio of male and female speakers. As for the investigators’ behavior, it is mandatory that they make clear to the informant that they are in the role of a learner, attempting to learn from the informant about the local language. One way of persuading the informant of her/his teacher-like position is to deliberately mispronounce and/or mistranscribe things, thus giving the informant an opportunity to correct and instruct the investigator. Another issue is the language to be used for communication between the investigator and the informant. This depends very much on the sociolinguistic environment of the region. During the interviews for the Sprachatlas der deutschen Schweiz, both investigator and informant used their native dialect, in accord with normal linguistic behavior in German-speaking Switzerland. The investigators for the Mittelrheinischer Sprachatlas in western Germany used the regional variety of the spoken standard language. Perhaps the most elaborate technique has been developed by the Syntactische atlas van de Nederlandse dialecten (SAND) in the Netherlands, where a great amount of interference from the standard language was expected during conversations with a non-local investigator. At each survey location, two informants were selected, one of whom became the target informant while the other was instructed to conduct the interview in the local dialect, with the investigator keeping out of the conversation as much as possible (Barbiers et al. 2004⫺: vol. 1, 8).

3.2. Loosely structured oral elicitation techniques Conversation dirige´e Although using an interview questionnaire, the standard methodology of dialect geography is oriented towards the ideal of recording dialect data that are only minimally influenced by the investigator. If possible, the word being asked for must not be mentioned by the investigator. Rather, it should be invoked by pointing at things, gestures, paraphrasing, etc. (Hotzenköcherle 1962: 127). Grammar in particular should be “hidden” in a conversation about concrete things or actions (Hotzenköcherle 1986: 24). The questions are grouped around complexes of objects, events, actions, in such a manner that the word(s) to be elicited ought to be mentioned spontaneously during the conversation. Ideally, the interview questionnaire consists predominantly of lists of words embedded in a functional context that takes account of aspects of external reality, whereby the sought after set of words includes all of the relevant linguistic variables. The method of conversation dirige´e (Hotzenköcherle 1986: 26) is therefore especially viable for phonological, lexical and, to some degree, morphological variables. However, all atlases use more structured methods for the investigation of syntactic variables.

Picture story Picture stories can be used for two purposes. First, they can serve as a starting point for the compilation of comparable discourse corpora where there is no specialized linguistic

517

518

V. Data collection and corpus-building variable to be determined. Zürrer (1999: 467) contains a picture story that has also been used during interviews for the SADS. It is easier to work with a picture story if more than one informant is present: one informant is given the task of telling the story to the other informants who cannot see the pictures. Second, for some very specific linguistic problems a picture story can serve as a suitable, and surprisingly rigid, means of exploration. Berthele (2006), investigating linguistic expressions for spatial relations, uses a longer picture story about a boy and his frog (Mayer 1969), originally developed for typological and psycholinguistic research. The story is made up in a fashion that encourages the informant to use spatial expressions of many different kinds. Another way of using pictures is illustrated by Fuchs (1993). Fuchs investigates the occurrence of morphological concord on predicative adjectives and participles in Valais German. With regard to the participles, concord morphemes appear only if the construction has resultant-state meaning, but not in the so-called analytic “perfect” tense (which, semantically, is simply periphrastic past). Fuchs presented the informants with drawings depicting situations that were clearly the results of earlier actions or events, including a pitcher that was being tipped over. Fuchs told the informants to describe the picture as precisely as possible in order for her to be able to draw the picture without seeing it. In doing so, informants said things like (1) Der Chrüe`ch isch umgikipptä the: sg:masc pitcher is tipped.over-sg:masc ‘The pitcher is (in the state of being) tipped over.’

(Fuchs 1993: 72)

Games Some phenomena can be investigated by means of a game to be played with the informants. For her study of Swiss German relativization, Rüegger (2003) developed a kind of memory game. Informants were presented a couple of play figures, each of which was related to a short story that the informant had to memorize. When asked about the identity of this or that play figure, the informant naturally answered with a relative clause: “That’s the guy who we always go swimming with”, etc.

3.3. Highly structured oral elicitation techniques Highly structured elicitation techniques within the direct method include translation and acceptability tasks.

Translation tasks In translation tasks, a form or a sentence is given in the standard language. The informant is asked to translate the model into her/his local dialect. Translation tasks formed a marginal part of the interview questionnaires in the tradition of the Sprachatlas der deutschen Schweiz; they are more prominent in more modern atlases which pay greater attention to syntactic constructions (SAND, Sprachatlas von Niederbayern [Eroms 2003⫺]). The issues with the translation method are well known (cf., e.g., Cornips and Poletto 2005). First, it is quite possible that the informant’s response is too pre-structured by the model, an issue which is particularly relevant in cultures where the social prestige

29. Questionnaire and interview of the local dialect is relatively low. Second, translating from one variety into another requires a high level of abstraction; if the informant does not abstract away from the literal form of the model, she/he will mechanically repeat the model. However, if he/she takes the explorer’s question, “How would you say …?”, too literally, he/she will perhaps say something which is too divergent from the model, i. e., where the intended identity of meaning is lost. Third, for many dialectal words, forms and constructions, no equivalent in the standard language exists. After all, it is a central advantage of the translation method that prior knowledge of the relevant linguistic variables is not as essential as it is for acceptability tasks. Translation tasks represent a relatively easy way of detecting the relevant phenomena at all, or of collecting expressions from previously undescribed dialects. The results of an investigation based on translation tasks may provide a useful basis for later, more narrowly focused studies.

Acceptability tasks Some of the issues raised by the translation method are avoided in another elicitation technique ⫺ acceptability tasks. The informant is presented with a series of variants and asked to judge their acceptability. The informants for the Syntactische atlas van de Nederlandse dialecten were presented with the variants by a local assistant investigator using the local pronunciation (Barbiers et al. 2004⫺: vol. 1: 7). Note that “acceptability” may mean at least two different things which do not necessarily coincide: first, whether or not the informant would use the variant, and second, whether the variant is more generally common in the dialect, i. e., passively known to the informant. Further, if an informant is asked, “Would you say this?”, she or he might answer in the negative for nonlinguistic reasons, i. e., because the intended meaning is uncommon in the informant’s social context. For example, in the northeast of German-speaking Switzerland, a special agreement marker for depictive predicatives is used: (2) si säät, eren Vatte hei de⫽Kafi all schwaz-e trunke she says her father has the:sg:masc-coffee always black-sg:masc drunk ‘She says that her father always drank his coffee black.’ (Appenzell, Switzerland; Bucheli Berger 2005: 162) In a social context in which people simply do not drink black coffee (as in Appenzell, where coffee usually is drunk with milk), the sentence could be rejected by the speakers on semantic grounds even if the morphosyntactic structure were fine (cf. Cornips and Poletto 2005 for further discussion of the notion of acceptability). Finally, it is important to note that the success of any acceptability task relies on previous knowledge of the relevant variants. Therefore, acceptability tasks are particularly useful in situation where the inventory of variants is already known but not their exact geographical distribution. For these purposes, as well as for theoretically oriented research into syntax, it is mandatory to know which variants are not acceptable in a particular dialect. Moreover, only a procedure involving acceptability judgments gives access to the full range of variation within a speaker’s competence.

Combination o translation and acceptability tasks The syntactic part of the Sprachatlas von Niederbayern (Eroms 2003⫺) uses a combination of translation and acceptability tasks. The informants were first confronted with a

519

520

V. Data collection and corpus-building model sentence in Standard German, which they were asked to translate into their local dialect. That variant was recorded as the one that was spontaneously mentioned. Then alternative variants were suggested, the acceptability of which was also noted. An example is (3), where the occurrence of complementizer doubling was being investigated: (3) a. Model sentence in Standard German: ich muss warten, bis der Vater heimkommt I must wait until the father home.comes ‘I have to wait until Father comes home.’

(Eroms 2003⫺: vol. 1, 294)

b. Suggested alternative variants: i. … bis da Vater … until the father ii. … bis dass da Vater … until that the father This method combines the advantages of both translation and acceptability tasks. It provides evidence about the informants’ first, most spontaneous reaction to the Standard German stimulus as well as the full range of possible and rejected variants. The method makes possible a more finely graded cartographic depiction of the geographical variant distribution. The symbols on Sprachatlas von Niederbayern maps (e.g., Eroms 2003⫺: vol. 1, 295) distinguish between “spontaneous”, “accepted” and “rejected” grades for a given variant. In interviews for the SADS, the model sentences for translation and acceptability tasks were always embedded in a small context story in order to make sure that the informant captured the desired semantic/pragmatic interpretation of the relevant expression. One desirable side effect of this procedure is that informants concentrate on something other than the purely linguistic problem. Often, the answers are therefore given in a more spontaneous way. It is important to note that a combination of different interview techniques can or must be applied, especially when multiple dimensions of variation are the focus of interest. In the Atlas Lingüı´stico Diato´pico y Diastra´tico del Uruguay (‘Diatopic and Diastratic Linguistic Atlas of Uruguay’; Thun and Elizaincı´n 2000⫺), reading tasks, an interview questionnaire and free conversation are combined and complement each other (Schmidt, to appear).

4. Survey o written techniques Length o the questionnaire As a word of caution, a written questionnaire should not be overloaded with too many tasks. Filling in a long questionnaire is not only tiring for the informants, it also creates undesired repetition effects (after a couple of questions on the same phenomenon many informants give answers in an increasingly mechanical way, without paying enough attention to changes in context). A questionnaire should be able to be completed within an hour. The questionnaire for the SADS was split into four series which were sent to the informants over a period of several months (Bucheli and Glaser 2002: 52).

29. Questionnaire and interview

Writing in dialect Investigations based on written questionnaires presuppose that the informants have some skills in reading and writing down answers in dialect. In an area such as Germanspeaking Switzerland, the prestige of the dialects is relatively high and most people are not totally unfamiliar with reading words and sentences in dialect (since this also occurs in advertising for instance), even though the standard variety is normally used for writing. Informants for the SADS were instructed to read aloud the model sentences in dialect and to write down their own variants just as they are usually spoken, without regard for orthography; this instruction proved adequate for virtually all informants (only a few, who were later removed from the list of informants, gave answers in the standard language). It is obvious that such vague instructions are not suitable if phonetic detail is the focus of interest (as is not the case with the SADS, it being a syntactic atlas).

Giving a context Especially (but not exclusively) for syntactic phenomena, it is advisable to embed the model sentences in a little contextualizing story, in order to guarantee the right semantic/ pragmatic interpretation of the sentence. For word-order phenomena, information structure can be crucial. It is often only possible to judge the grammaticality of a given word order if the intended topic⫺comment or focus⫺background structure is made plausible. Another example is the “resultant state versus perfect” distinction in some dialects of Swiss German. In German, the combination of have/be ⫹ past participle has (at least) two meanings: (i) the description of a state which is the result of an action performed some time in the past (without determining by whom it has been performed); (ii) socalled perfect tense (which, semantically, is just periphrastic past for most dialects), i. e., the expression of an action performed in the past by the subject of the sentence. In some archaic Swiss German dialects (e.g., Freiburg, Valais) the participle form agrees in gender and number with the noun phrase it refers to, but only if the construction has the resultant-state meaning. In the SADS questionnaire, the intended resultant-state meaning has been achieved by constructing a context that makes this meaning plausible and adding the adverbial immer no ‘still’: (4) Context: Your brother has hurt his hand during gardening. Ms Terchert asks you how he is now. You answer: a. Er hät d Hand immer no iibunde (yes/no) he has the:sg:fem hand always still bound b. Er hät d Hand immer no iibundni (yes/no) he has the:sg:fem hand always still bound-sg:fem ‘He still has his hand bandaged.’ Moreover, the context stories serve another purpose. They make the questionnaire more entertaining for informants, and they distract them from the abstract linguistic problem. But misinterpretations of a question do still occur. Again, an example is found in the SADS material. In some archaic Swiss German dialects, a pronominal partitive form, re (ra, ru), a relict genitive case marker (here: feminine singular), is used when referring to an indeterminate amount of a substance or things, whereas in the majority of Swiss dialects this relation remains unexpressed. In the context story, one person (female) says

521

522

V. Data collection and corpus-building there is no milk left. The other person (male) asks whether he should go and buy some. Two variants of this response are suggested: (5) a. Söll i go chaufe? should I go buy b. Söll i re go chaufe? should I partitive go buy ‘Should I go and buy some (i. e., milk)?’

(yes/no) (yes/no)

One person from an area where the partitive pronoun is completely unexpected rejected the first (expected) variant while accepting the second. Moreover, the person noted the following as his own variant: (6) söll der go chaufä? should you:dative go buy ‘Should I go and buy you some (i. e., milk)?’ Obviously, the informant wanted to insert a (benefactive) dative object to denote the female from the context. The second variant in (5) comes closest to this, since the form re can also be understood as the dative singular feminine, although in the third person. This possible source of confusion had not been taken into account when the SADS questionnaire was constructed. The confusion could have been easily avoided if the actors’ genders had been reversed in the context story (so that the female asks the male the question). Confusions of this type are symptomatic of the indirect method. They could easily be diagnosed and compensated for in an oral interview. See Bucheli and Glaser (2002: 54) and Cornips (2002: 79⫺82) for further sources of confusion in written questionnaires.

Translation tasks The advantages of and issues surrounding translation tasks in written questionnaires resemble those for direct interviews. To mention one additional aspect, optional dialect structures (with no equivalent structure in the model sentence) tend to appear less frequently in translation tasks than in acceptability tasks. For example, in many Swiss German dialects, a dative object can be introduced by a prepositional marker, e.g., in wem ‘dative marker ⫹ who:dative’ rather than Standard German wem ‘who:dative’ (cf. Seiler 2003). For most speakers, this kind of prepositional marking is optional. In a translation task in the SADS questionnaire 336 informants spontaneously produced the prepositional dative maker. But 1150 informants accepted prepositional marking in an acceptability task, with 280 informants even judging it to be obligatory. That is, while prepositional marking is optional for 870 (i. e., 1150⫺280) informants according to the acceptability task results, only a few ⫺ 56 (i. e., 336⫺280) ⫺ of those informants made use of it in the translation task.

Acceptability tasks When acceptability judgments are elicited in written questionnaires, the model sentences must be presented in an orthographical and lexical form that is acceptable to the infor-

29. Questionnaire and interview mants. Thus, whereas the Atlas van de Nederlandse Dialectsyntaxis used spelling conventions close to those of the Dutch standard language, the SADS suggested variants in a dialectal form because the influence of the standard language was supposed to be minimized (note that Switzerland is diglossic, with a sharp discontinuity between Standard German and the dialect. If the spelling of the model sentences resembled Standard German, it would have been extremely difficult to judge their acceptability from a dialectal perspective). But which dialect ought to be used? The SADS worked with three variants: one for Berne, one for Valais, and one for the rest (in Zurich German phonetics that avoided too local-sounding forms). Although most speakers are used to listening to dialects other than their own, for extralinguistic reasons some Bernese informants were unable to cooperate if the model sentences sounded too Zurich-like. Speakers from Valais use a variety that is very different from their local dialect when communicating with speakers from other parts of the country. In order to not elicit the outsider variety, the questionnaire was translated into Valais German. Most questionnaires use a simple yes/ no-distinction to record the acceptability of the suggested variants. The written pilot study for the Syntactische atlas van de Nederlandse dialecten used a five-grade scale (Cornips and Poletto 2005; see also Simon 2008 on graded acceptability elicitation). The Atlas van de Nederlandse Dialectsyntaxis and the SADS questionnaires distinguish between acceptable and preferred variants. In the SADS, informants were first asked to indicate which of a choice of variants were acceptable (“Which ones can be used in your dialect?”), then to decide which is the preferred (“most natural”) one, and then to write down an additional variant if the informant would not use any of those suggested. This design makes it possible to distinguish between unacceptable, acceptable, preferred and obligatory variants in the cartographic analysis of the results. Interestingly, geographical transitions from one variant to another are often gradual; for example, a variant is obligatory in the west, preferred but in competition with others in more central areas, acceptable but dispreferred farther eastwards, and unacceptable in the easternmost areas (see Seiler 2004 for the gradual transition from western ascending to eastern descending orderings in verb clusters and Seiler 2005 for a similar picture concerning purposive infinitives). It is of crucial importance that the informants also have the opportunity to write down their own variant in acceptability tasks (“Would you say the same thing in a different way? How?”). Very often, informants reject all of the suggested variants for lexical or phonetic reasons, but their own variant is identical to one of the suggested model sentences as far as the syntactic structure is concerned. Moreover, in many cases, new (i. e., previously unknown to the investigator) variants or insights into relevant but as yet unknown semantic properties of the suggested variants appear here.

Completion tasks Completion tasks consist of a sentence fragment, together with an instruction concerning the intended meaning of the completed sentence as a whole. Completion tasks have sporadically been applied in collecting data for the SADS (see Bucheli and Glaser 2002: 61⫺62 for details).

Assessing the results o written questionnaires As mentioned in section 2.2, the reliability of results gained by means of written questionnaires is more difficult to control, mainly because the behavior of the informant is not immediately observable. However, as more recent research suggests, the results of

523

524

V. Data collection and corpus-building oral and written elicitation match remarkably well (cf. Bucheli, Berger and Glaser 2004 and Seiler 2003 for comparisons of the orally elicited Sprachatlas der deutschen Schweiz data and the SADS questionnaire study; Schmidt 2005 on the comparability of the written Deutscher Sprachatlas and oral Mittelrheinischer Sprachatlas data). The reliability of written questionnaire material can be estimated if the following points are kept in mind. ⫺ Mass comparison: To some degree, the lack of direct control can be compensated for by the great amount of collected data. Problematic answers ⫺ which are certain to be found in any questionnaire study ⫺ become marginalized. This idea already underlay the data collection for the Deutscher Sprachatlas: “It is precisely the number of localities represented which protects us from incorrect data from a single one” (Wrede 1895: 35, quoted in Niebaum and Macha 2006: 16; my translation). For the same reason, the SADS works together with several informants at each survey location (seven on average). Taking as an example the use of an impersonal passive expletive in the middle field position in Swiss German (do wird’s gwärchet, lit. ‘here becomes-it worked’, Kulm/Aargau, Frey 1906: 28), the construction has been very sporadically judged as acceptable in various areas of German-speaking Switzerland. However, it is only in a central area (Aargau, Lucerne) that the construction is accepted significantly more often, i. e., by more than one informant per locality. ⫺ Repetition: The same phenomenon can be elicited several times with superficial changes, especially when a questionnaire is split into parts. ⫺ Analysis of comments: Some informants make extensive use of opportunities to comment on the suggested variants, the way the questionnaire is structured, etc. A careful analysis of such comments helps to assess whether the task is understood in the manner intended, whether the task is confusingly formulated, whether all relevant variants are suggested in acceptability tasks, etc. ⫺ Oral interviews: In addition to written questionnaires, it is recommended that investigators collect a control sample of data via oral interviews, where possible with different informants, in order to avoid memory effects. It might be the case that the results of oral versus written elicitations deviate from one another, but the control sample can tell us in which direction and to what extent. ⫺ Deviations from the standard language: A notorious problem with translation tasks is the influence of the model sentence written in the standard language. While there is no way of avoiding that influence, it has a clear advantage, too. When a translation closely corresponds to the standard language model, the interpretation remains unclear: is the variant given by the informant really what she/he uses in dialect, or is it merely an imitation of the model sentence? However, cases where the informant’s translation deviates from the standard language model can be taken as clear evidence that the variant is rooted in the dialect (Glaser 2000: 263). ⫺ Dialect-geographical differences: It is clear that all elicitation procedures, oral or written, represent actual use only in a distorted way. But whilst it is difficult to measure the amount of distortion vis-a`-vis actual use (see section 1.2), it needs to be kept in mind that, as a desired result of standardization of elicitation procedures, the distortion, however great it might be, is identical for all informants. If a standardized experiment yields systematic geographical differences in the results, it is highly probable that there is also a difference in actual use in the areas. In other words, if no dialect-geographical differences are revealed on the basis of elicited data, the interpretation remains unclear: either there is simply no dialect difference with respect to

29. Questionnaire and interview the elicited variable, or the experiment has failed, e.g., because the task has been formulated in a misleading fashion. By contrast, if an isogloss shows up as a result of the analysis of elicited data, there must be some concomitant difference in actual use.

5. Reerences Barbiers, Sjef, Leonie Cornips and Susanne van der Kleij (eds.) 2002 Syntactic Microvariation. Amsterdam: Meertens Instituut. Barbiers, Sjef, Hans Bennis, Gunther De Vogelaer, Magda Devos, Margreet van der Ham, Irene Haslinger, Marjo van Koppen, Jeroen Van Craenenbroeck and Vicky Van den Heede (eds.) 2004⫺ Syntactische atlas van de Nederlandse dialecten [Syntactic Atlas of the Dutch Dialects]. Amsterdam: Amsterdam University Press. Bellmann, Günter 1994 Einführung in den Mittelrheinischen Sprachatlas (MRhSA). Tübingen: Niemeyer. Bellmann, Günter, Jürgen Erich Schmidt and Joachim Herrgen (eds.) 1994⫺2002 Mittelrheinischer Sprachatlas. Tübingen: Niemeyer. Berthlele, Raphael 2006 Ort und Weg. Die sprachlichen Raumreferenz in Varietäten des Deutschen, Rätoromanischen und Französischen. Berlin/New York: de Gruyter. Besch, Werner, Ulrich Knoop, Wolfgang Putschke and Herbert Ernst Wiegand (eds.) 1982 Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung. 2 vols. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.2.) Berlin/New York: de Gruyter. Bresnan, Joan, Anna Cueni, Tanya Nikitina and Harald Baayen 2007 Predicting the dative alternation. In: Gerlof Boume, Irene Kraemer and Joost Zwarts (eds.), Cognitive Foundations of Interpretation, 69⫺94. Chicago: University of Chicago Press. Bucheli Berger, Claudia 2005 Depictive agreement and the development of a depictive marker in Swiss German dialects. In: Nikolaus P. Himmelmann and Eva Schultze-Berndt (eds.), Secondary Predication and Adverbial Modification: Crosslinguistic Explorations in the Syntax and Semantics of Depictives, 141⫺171. Oxford: Oxford University Press. Bucheli, Claudia and Elvira Glaser 2002 The Syntactic Atlas of Swiss German Dialects: Empirical and methodological problems. In: Barbiers, Cornips and van der Kleij (eds.), 41⫺74. Bucheli Berger, Claudia and Elvira Glaser 2004 Zur Morphologie des (ko)prädikativen Adjektivs und Partizips II im Alemannischen und Bairischen. In: Franz Patocka and Peter Wiesinger (eds.), Morphologie und Syntax deutscher Dialekte und Historische Dialektologie des Deutschen. Beiträge zum 1. Kongress der Internationalen Gesellschaft für Dialektologie des Deutschen, Marburg/Lahn, 5.⫺8. März 2003, 189⫺226. Vienna: Praesens. Cornips, Leonie 2002 Variation between the infinitival complementizers om/voor in spontaneous speech data compared to elicitation data. In: Barbiers, Cornips and van der Kleij (eds.), 75⫺96. Cornips, Leonie and Cecilia Poletto 2005 On standardising syntactic elicitation techniques. Part I. Lingua 115(7): 939⫺957. Eggers, Eckhard, Jürgen Erich Schmidt and Dieter Stellmacher (eds.) 2005 Moderne Dialekte ⫺ Neue Dialektologie. Akten des 1. Kongresses der Internationalen Gesellschaft für Dialektologie des Deutschen (IGDD). Stuttgart: Steiner.

525

526

V. Data collection and corpus-building Eichhoff, Jürgen 1982 Erhebung von Sprachdaten durch schriftliche Befragung. In: Besch et al. (eds.), vol. 1: 549⫺554. Eroms, Hans-Werner (ed.) 2003⫺ Sprachatlas von Niederbayern. Heidelberg: Winter. Frey, Arthur 1906 Beiträge zur Syntax des Schweizerischen. In: Anton Glock and Expeditus Schmidt (eds.), Analecta Germanica. Hermann Paul zum 7. August 1906, 19⫺42. Amberg: Böes. Fuchs, Gabriela 1993 Das prädikative Adjektiv im Walliserdeutschen. In: Helen Christen (ed.), Variationslinguistik und Dialektologie. Ergebnisse aus studienabschliessenden Arbeiten an der Universität Freiburg/Schweiz, 65⫺79. Freiburg (Switzerland): Editions universitaires. Gerritsen, Marinel 1991 Atlas van de Nederlandse Dialectsyntaxis. Amsterdam: Meertens Instituut. Gillie´ron, Jules and Edmond Edmont 1902⫺1910 Atlas linguistique de la France. Paris: Champion. Glaser, Elvira 2000 Erhebungsmethoden dialektaler Syntax. In: Dieter Stellmacher (ed.), Dialektologie zwischen Tradition und Neuansätzen. Beiträge der internationalen Dialektologentagung, Göttingen, 19.⫺21. Oktober 1998, 258⫺276. Stuttgart: Steiner. Hotzenköcherle, Rudolf 1962 Einführung in den Sprachatlas der deutschen Schweiz, Part A. Berne: Francke. Hotzenköcherle, Rudolf 1986 Zur Methodologie der Kleinraumatlanten. In: Rudolf Hotzenköcherle, Robert Schläpfer and Rudolf Trüb (eds.), Dialektstrukturen im Wandel, 13⫺29. Aarau: Sauerländer. Hotzenköcherle, Rudolf (ed.) 1962⫺2003 Sprachatlas der deutschen Schweiz. Founded by Heinrich Baumgartner and Rudolf Hotzenköcherle; in cooperation with Konrad Lobeck, Robert Schläpfer, Rudolf Trüb and with the assistance of Paul Zinsli. Berne: Francke. Kakhro, Nadja 2005 Die Schweizer Wenkersätze. In: Linguistik Online 24: 155⫺169. König, Werner 2005 dtv-Atlas deutsche Sprache. 15th ed. Munich: Deutscher Taschenbuch Verlag. Kortmann, Bernd 2003 Comparative English dialect grammar: A typological approach. In: Ignacio M. Palacios, Marı´a Jose´ Lo´pez Couso, Patricia Fra and Elena Seoane (eds.), Fifty Years of English Studies in Spain (1952⫺2002). A Commemorative Volume, 63⫺81. Santiago de Compostela: University of Santiago. Kortmann, Bernd and Susanne Wagner 2005 The Freiburg English Dialect Project and Corpus. In: Bernd Kortmann, Tanja Herrmann, Lukas Pietsch and Susanne Wagner (eds), A Comparative Grammar of British English Dialects: Agreement, Gender, Relative Clauses, 1⫺20. Berlin/New York: de Gruyter. Kayne, Richard S. 1996 Microparametric syntax: Some introductory remarks. In: James R. Black and Virginia Motapanyane (eds.), Microparametric Syntax and Dialect Variation, ix⫺xxviii. Amsterdam: Benjamins. Labov, William 1972 Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Labov, William, Sharon Ash and Charles Boberg (eds.) 2006 The Atlas of North American English: Phonetics, Phonology and Sound Change. Berlin/ New York: de Gruyter. Mayer, Mercer 1969 Frog Where Are You? New York: Dial.

29. Questionnaire and interview

527

Menge, Heinz H. 1982 Erhebung von Sprachdaten in “künstlicher” Sprechsituation (Experiment und Test). In: Besch et al. (eds.), 544⫺549. Mitzka, Walther, Ludwig Erich Schmitt and Rainer Hildebrandt (eds.) 1956⫺1980 Deutscher Wortatlas. Giessen: Schmitz. Niebaum, Hermann and Jürgen Macha 2006 Einführung in die Dialektologie des Deutschen. 2nd ed. Tübingen: Niemeyer. Rüegger, Tamara 2003 Der Dativ im Relativsatz des Thurgauer Dialekts. Unpublished proseminar paper, University of Zurich. SADS: Syntaktischer Atlas der deutschen Schweiz [Syntactic Atlas of German-Speaking Switzerland]. Ongoing research project at the University of Zurich, German Department, funded by the Swiss National Science Foundation, directed by Elvira Glaser. Schmidt, Jürgen Erich 2005 Sprachdynamik. In: Eggers, Schmidt and Stellmacher (eds.), 15⫺44. Schmidt, Jürgen Erich to appear Pluridimensionale Sprachgeographie und Sprachdynamik. In: Fred Boller (ed.), Die pluridimensionale Sprachgeographie in der Diskussion. Festschrift für Harald Thun. Kiel: Westensee. Seiler, Guido 2003 Präpositionale Dativmarkierung im Oberdeutschen. Stuttgart: Steiner. Seiler, Guido 2004 On three types of dialect variation, and their implications for linguistic theory. Evidence from verb clusters in Swiss German dialects. In: Bernd Kortmann (ed.), Dialectology meets Typology. Dialect Grammar from a Cross-Linguistic Perspective, 367⫺399. Berlin/ New York: de Gruyter. Seiler, Guido 2005 Wie verlaufen syntaktische Isoglossen, und welche Konsequenzen sind daraus zu ziehen? In: Eggers, Schmidt and Stellmacher (eds.), 313⫺341. Siebenhaar, Beat 2000 Sprachvariation, Sprachwandel und Einstellung. Der Dialekt der Stadt Aarau in der Labilitätszone zwischen Zürcher und Berner Mundartraum. Stuttgart: Steiner. Simon, Horst 2008 Methodische Grundfragen zu einer vergleichenden Syntax deutscher Dialekte. In: Karin Donhauser, Elvira Glaser and Marcel Vuillaume (eds.), Empirische Grundlagen moderner Grammatikforschung, 59⫺70. Berne: Lang. Thun, Harald and Adolfo Elizaincı´n 2000⫺ Atlas Lingüı´stico Diato´pico y Diastra´tico del Uruguay [Diatopic and Diastratic Linguistic Atlas of Uruguay]. Kiel: Westensee. Wenker, Georg 1887⫺1923 Sprachatlas des Deutschen Reichs. Laut- und Formenatlas. 1647 hand-drafted multicolor maps held in the archives of the Forschungszentrum Deutscher Sprachatlas in Marburg and at the Staatsbibliothek Stiftung Preußischer Kulturbesitz in Berlin. Wrede, Ferdinand 1895 Über die richtige Interpretation der Sprachatlaskarten. In: Der Sprachatlas des deutschen Reiches. Dichtung und Wahrheit, 31⫺52. Marburg: Elwert. Wrede, Ferdinand, Bernhard Martin and Walther Mitzka (eds.) 1927⫺1956 Deutscher Sprachatlas. Marburg: Elwert. Zürrer, Peter 1999 Sprachinseldialekte. Aarau: Sauerländer.

Guido Seiler, Manchester (Great Britain)

528

V. Data collection and corpus-building

30. Investigating language in space: Experimental techniques 1. 2. 3. 4.

Theoretical and definitional considerations Experimental construction of language-use data Experimental construction of language-attitudes data References

1. Theoretical and deinitional considerations The appropriateness of the methods we use to collect data is determined by our theories about the nature of the object under study and the questions we ask about it. The nature of language (in space, as well as in general) is no uncontroversial issue, as humans are fundamentally both natural and social beings, with an ability to communicate that also has both natural and social aspects to it. Both natural and human scientists use experimentation as a method of data collection, but the experimental approach is commonly held to be less appropriate within the human sciences than within the natural sciences. So, what are we to understand by experimentation in the context of language in space? As there does not seem to be any simple and straightforward answer to this question, it will reappear throughout this article. To begin with, this first section introduces some conceptual distinctions that may help us explicate the notion of “experimentation with language in space”. These introductory remarks shall be very brief with regard to the object (language in space; section 1.1), and a little more elaborated with regard to the method (experimentation; section 1.2).

1.1.

The object: Language in space

1.1.1. Space: Cultural entities Discourse about how and why language varies and changes across the spatial dimension often reflects the “natural” and “social” foundations of human beings and their languages in a direct and “naı¨ve” way. While language differences in the social landscape are understood in terms of “social barriers” (between classes, genders, ethnic groups, etc.), language differences in the geographic landscape are understood in terms of “natural barriers” (such as mountains, woods, bodies of water, etc.). In this article we shall take it as a given that both geographic and social spaces ⫺ beyond having a material existence and being “hard facts” ⫺ are also “imagined” (in the sense of Anderson 1983), and that the issue of how and why language varies and changes is mainly to do with space as an array of ideologically constructed, cultural spaces.

30. Experimental techniques

1.1.2. Language: Use and attitudes, production and perception, conscious and subconscious In order to understand language variation and change across cultural spaces, we need to study not only the linguistic facts as such but also their social meaning. We need to collect data at the level of language use, and we need to collect data at the level of social evaluation (or attitudes). This distinction is at the root of the main division of this article into a section on the construction of language-use data (section 2) and another on the construction of language-attitudes data (section 3). Both language use and language attitudes do of course have productive as well as perceptive (or receptive) aspects to them, but in our context here the distinction between production and perception will often refer to very much the same distinction as that between use and attitudes. Likewise, it goes for both use/production and perception/evaluation that these processes may be either conscious or subconscious, and this distinction is by no means the least pertinent in connection with experimental research.

1.2. The method: Experimentation As used in practice, there is no clear-cut dividing line between data-collection methods referred to as “experiments” on the one hand (treated in this article) and “questionnaires and interviews” on the other (cf. Seiler in this volume). As we shall see, an interview may not only contain an experiment, it may in itself be an experiment, or, put the other way round, an experiment may take the form of an interview. It therefore seems appropriate to begin by determining in what fundamental way experimentation is something other than questioning and interviewing. In so doing, we shall distinguish between experimental and observational approaches to research (leaving aside the obvious fact that “making observations” is a central aspect of research in general) and try to point out what differentiates the former from the latter.

1.2.1. Deining experiment The Danish National Encyclopedia (giving the floor, not surprisingly, to a nuclear physicist) defines an experiment as a “set up or arrangement with the purpose of discovering new connections or verifying a scientific hypothesis or theory” (SDE 1996: 447; my translation). The definition points to the deliberately staged nature of an experiment: it is “an arrangement with a purpose”. As to the substance of the purpose or aim of experimentation, the definition points at what is commonly seen as two levels of ambition. The less ambitious aim is to describe (“discover new connections”), the more ambitious aim is to explain (“verify a hypothesis or theory”). There is no one-to-one relationship between such ambitions and research approaches: neither the observational nor the experimental approach necessarily aims beyond the descriptive establishment of connections (correlations), but both of them may, and often will, aim at answering the why

529

530

V. Data collection and corpus-building question of explanation as well as the what/who/when/how questions of description. Nevertheless, the quest for explanation is often seen as particularly characteristic of experimental research.

1.2.2. Techniques or securing control Explanation in experimental research is based on the comparison of measures obtained under different conditions. Valid and reliable comparison is only possible with full control of the different factors or phenomena (entities and processes) that make up the experimental arrangement. Thus, the basic techniques of experimentation are techniques for securing control. Control is searched for through operationalization of the experimental factors into rules (scales) on which they can be measured. Several rules may be candidates for the measurement of the phenomenon we want to study, and the chosen rule may be constructed in several ways. By choosing a rule that measures what we want to measure, we solve what is known as the validity problem in data collection. By constructing the rule so that it measures in the same way each time it is used, we solve what is known as the reliability problem in data collection. Only by securing the validity and reliability of our data do we make them replicable, and hence comparable with other data (cf. König in this volume). The phenomenon we want to study is operationalized as the dependent variable. Factors we suspect of having an impact on the phenomenon are operationalized as independent variables. Both dependent and independent variables need to be divided into clear divisions (units, steps, entities). The dependent variable “collects” data that are said to attain either a nominal, ordinal, interval or ratio level of measurement. On a nominal scale the relationship between its units is a matter of “either/or” (X or Y), on the other scales it is matter of “more or less” of the same on a continuum (X1 or X2 or X3 …). The levels of measurement are of great importance for the statistical analyses of the data.

1.2.3. Obtrusiveness and awareness In order to secure full control of all variables, the experimental arrangement will often be set up as a kind of “laboratory”. It follows that the search for control tends to destroy “naturalness”. Normally, however, experimenters (as well as observers) are interested in collecting “natural” data ⫺ data as they are, independent of being collected or made data, so to speak. This is known as the experimenter’s (as well as the observer’s) paradox. As laboratories are not part of people’s everyday lives, data obtained in laboratories or laboratory-like settings are by their nature non-natural. Attempts at solving the nonnaturalness problem involve moving the data collection to everyday situations which can be “used” as experimental arrangements. Such experimental exploitations of real-life situations are known as field experiments. While laboratory experiments are by nature obtrusive, field experiments aim to be unobtrusive. The reason why obtrusive data collection may be problematic has to do with human consciousness and awareness. When some aspect of human mind or behavior is the object of study, awareness is often thought to impede naturalness, in the sense that the

30. Experimental techniques informants’ answers or reactions will be influenced by their ideas about what the researcher expects from them, or about social acceptability in general and how they present themselves. Thus, attempts at solving the paradox of data collection typically seek to reduce or avoid the participants’ awareness, so that their answers or reactions are subconsciously offered. In field experiments, it may be possible to attain this by simply avoiding informing the participants that data collection is taking place. This will never be the case in laboratory-like settings, where the awareness problem may instead be addressed by asking subjects to accept that they will not be informed about the purpose of the experiment until after its completion, or by using a deception technique (e.g., a cover story) to give subjects the wrong impression of what the data collection is about.

2. Experimental construction o language-use data Being based on self-reported data obtained from respondents in interviews or questionnaires, traditional dialectology can be considered observational in its approach to data collection. Whether they themselves traveled for years as interviewers from village to village (like Aasen in Norway or Edmont for Gillie´ron in France) or covered the geographical space by means of mail-distributed questionnaires (like Wenker in Germany), the aim of dialectology’s first great data collectors, and their followers, was to describe the geographically distributed variation of language. It is not that they were necessarily uninterested in explanations, rather that their theories constructed the facts of geographical distribution as the necessary and sufficient empirical evidence for how variation and change in language should be understood, either in terms of sound laws, or, alternatively, in terms of “to every word its own history” (cf. Murray, Schmidt in this volume). Experimentation enters the study of language in space with the development of variationist sociolinguistics and its new focus on explanations in terms of social motivation and social mechanisms. So let us start by looking at how our theoretical concepts are rendered in two of sociolinguistics’ classic studies, namely Coupland’s (1984) travel agency study (section 2.1) and Labov’s (1966) department store study (section 2.2).

2.1. No deliberate stimulus-variable manipulation: The Cardi travel agency study The aim of this study was to test the claims of speech accommodation theory (Giles and Powesland 1975) that interlocutors accommodate linguistically, either by converging or diverging, and to examine the nature of these processes as either “direct linguistic matching” or “interpretively based persona construction”. The data collection was conducted in a travel agency in Cardiff and yielded tape-recorded interactions between one assistant and her clients. The “paradox” of data collection was solved by not making the clients aware that they were being recorded until they left the agency, at which point they were debriefed and asked for permission to use the recorded material; the assistant knew about the recording, but thought that the object of study was the clients’ speech alone. In fact, the analysis focuses on whether and how the assistant accommodates to her clients, as representatives of categories in social space. The measurement of accommoda-

531

532

V. Data collection and corpus-building tion was done on four phonological variables (dependent variables), all of which were divided into two variants. Social space was operationalized as an occupational class variable with six units (independent background variable), on which the clients were distributed on the basis of information they gave about themselves during the debriefing. This clearly is an example of a deliberate set up which aims to examine a scientific theory and explain language use in social space, and which comes far in securing the unobtrusive and yet highly controlled character of the whole arrangement. In other words, the study is an experiment according to our definition. It is also an experiment according to A Dictionary of Sociolinguistics, which defines experiment as “the study of any phenomenon under controlled conditions designed to assess the effect of one variable on another” (Swann et al. 2004: 105). However, Coupland himself explicitly talks of an “observational approach” in a “non-experimental setting” (1984: 52). Coupland offers the travel agency study as a contribution to the development of social-psychological speech accommodation theory. So his labels should first and foremost be understood in contrast to the “experimental approach” in “laboratory settings” that is typical of the relevant social-psychological research. Nevertheless, it seems clear that our definition renders much research which researchers themselves consider “observational” as “experimental”. In fact, if control alone defines experimentation, variationist studies in general are experiments. This may be seen as a reason for narrowing down the definition. We can achieve that by distinguishing between two kinds of independent variables: stimulus variables and background variables. Background variables are “naturally present” in the material and are exploited by the researcher in the set up of the experiment. “Naturally present” differences may also be exploited to construct the stimulus variable, but more often the stimulus variable is introduced into the setting by the researcher. In the travel agency arrangement, what was measured was the assistant’s language use in response to the clients’ language use. In other words, the clients’ linguistic behavior was the stimulus variable. The differences in this behavior were not a result of any manipulation by the researcher. Coupland did not introduce the differences into the setting; he just exploited “naturally present” differences in the construction of the stimulus variable.

2.2. Deliberate stimulus-variable manipulation: The New York department stores study Labov’s (1966) famous department store study is well suited for illustrating an experimental design that includes a deliberately manipulated stimulus variable, and also for illustrating the characteristics of an experiment mentioned in section 1. It should be mentioned, though, that Labov himself usually refers to this study as a “rapid and anonymous survey” based on interviews (a problematic label, one might argue, as we are talking about encounters yielding sequences of four short, more or less identical, utterances, described in section 2.2.1). In the programmatic article on the field methods used in the Philadelphia project, a list of eight different “field experiments” is discussed in the “Neighborhood Studies” section (Labov 1984: 42⫺45) while “the method of rapid and anonymous surveys” (1984: 49) is accorded a section of its own. But, Labov has also used the term experiment in reference to the department store study. Thus, the relevant chapter in the seminal book Sociolinguistic Patterns ends with the statement that

30. Experimental techniques We see rapid and anonymous observations as the most important experimental method in a linguistic program which takes as its primary object the language used by ordinary people in their everyday affairs (Labov 1972: 69, italics added).

Being set up to do exactly that ⫺ i. e., study “language used by ordinary people in their everyday affairs” ⫺ the department store study can clearly be classified, just like Coupland’s travel agency study, as an unobtrusive field experiment and a study of language in social space. In the following paragraphs, we shall show in some detail, under (a), how the linguistic and socio-spatial variation was operationalized into variables. Under (b), we shall exemplify some of the problems posed by the experimental approach for the study of language in space, drawing heavily on Labov’s own instructive discussions of operationalization problems with respect to both dependent and independent variables.

2.2.1. The stimulus variable (a) The data collection procedure ⫺ which secures the unobtrusiveness of the event and the informants’ ignorance of their status as informants ⫺ is described as follows: The interviewer approached the informant in the role of a customer asking directions to a particular department. The department was one which was located on the fourth floor. When the interviewer asked, “Excuse me, where are the women’s shoes?” the answer would normally be, “Fourth floor.” The interviewer then leaned forward and said, “Excuse me?” He would usually then obtain another utterance, “Fourth floor,” spoken in careful style under emphatic stress. The interviewer would then move along the aisle of the store to a point immediately beyond the informant’s view, and make a written note of the data. (Labov 1972: 49)

In the way he asks for the information twice, Labov deliberately manipulates the contextual condition under which “fourth floor” is uttered by the informants. (b) There were problems with this operationalization, however. Because it turned out that the simple request for repetition had only a limited effect on the dependent variable, Labov suggests that the stylistic range might be enlarged “by emphasizing the difficulty in hearing by one technique or another” (Labov 1972: 66).

2.2.2. The dependent variable (a) The dependent variable is the manifestation of postvocalic /r/ in the words fourth and floor. The variable is divided into two units (variants) ⫺ “presence” or “absence” ⫺ and yields data at a nominal level of measurement. In other words, a naturally continuous variable is operationalized into discrete units: For each plainly constricted value of the variable, (r-1) was entered; for unconstricted schwa, lengthened vowel, or no representation, (r-0) was entered. Doubtful cases or partial constriction were symbolized d and were not used in the final tabulation. (Labov 1972: 50)

533

534

V. Data collection and corpus-building (b) The transformation of something continuous into something discrete will always raise problems. In this connection, Labov points to the risk of unconscious experimenter bias, because “the transcriber, myself, knew what the object of the test was” (Labov 1972: 66). In general, and independently of when the transcription takes place (during data collection as in this case, or during data preparation in the case of tape-recorded speech), any transcriber should treat doubtful cases in accordance with strict guidelines.

2.2.3. The background variables (a) Social space is operationalized by means of several background variables. The main one of these is the department store variable, which is divided into three units: Sacks, Macy’s, and S. Klein. These units are characterized and ranked as to social status, on the basis of their advertising and price policies, the physical plant of the stores, working conditions and general prestige. Other “naturally given” distinctions exploited by Labov for further operationalization of the social space included the informants’ sex (men, women), age (estimated in blocks of five years), occupation (floorwalker, sales, cashier, stockboy), race (black, white), and accent (foreign, regional, “none”). In addition, informants’ working place in terms of floor and department was also registered. (b) How to operationalize naturally continuous variables is a problem with background variables too. Age was operationalized as the researcher’s estimate, during data collection, of the informants’ age within five-year intervals. But as “these figures cannot be considered reliable for any but the simplest kind of comparison” (Labov 1972: 57), the informants were regrouped on an age variable with only three units (15⫺30, 35⫺50, 55⫺70) for the purpose of more reliable analysis. Another sampling problem was the result of selecting the most available subject in a given area. Labov certainly subscribes to the view that good sampling in survey studies aim at representativity, but points out that an effort to avoid sampling bias should not be allowed to interfere with the unobtrusive character of the speech event (Labov 1972: 66). A general problem with background variables as exploitations of naturally present distinctions is that they may show up too rarely in the collected data to be made the object of analysis. In the New York department stores study, this turned out to be the case with the department variable: “the numbers for individual departments are not large enough to allow for comparison” (Labov 1972: 49, footnote 7).

2.2.4. The quest or explanation The operationalization of the interactional context (the stimulus variable) is grounded in certain theoretical assumptions about the nature of intra-individual variation, namely that it is motivated by different degrees of “attention to speech”, and more specifically in such a way that more attention yields more careful speech. The theoretical framework furthermore predicts that careful speech will contain more “present” /r/s than casual speech, because careful style is assumed to favor prestigious speech variants, and pronounced postvocalic /r/ is assumed to be a prestige feature. The notion of prestige, in turn, is tied up with theoretical conceptualizations of social space. In the final instance,

30. Experimental techniques the data collection in the department stores is based on a complex theory, both psychological and social in nature, the ultimate aim of which is to explain (not just describe) the distributional use of language across social space. The whole “arrangement” is set up to test a hypothesis and validate a theory.

2.2.5. Consequences o narrowing down the deinition o experiment One might object, rightly I think, that a requirement that stimulus-variable manipulation be deliberate puts the focus on experimentation’s quest for explanation ⫺ to the detriment of the more modest aim of description, which was included as a possibly solitary goal in our encyclopedic definition (cf. section 1.2.1). However, if experimentation is commonly felt to be particularly explanatory in nature (as suggested in section 1.2.1), it is probably because the common notion of experimentation includes this “deliberate manipulation of a stimulus variable”. We may note that this narrower definition is also found in the literature, e.g., in the glossary section of Introduction to Social Psychology, where experiment is spelled out as “a method in which the researcher deliberately introduces some change into a setting to examine the consequences of that change” (Hewstone et al. 1988: 449). This narrower definition omits studies that, notwithstanding the general use of experimental techniques to secure control (operationalization of variables), are commonly seen as “observational”, like the Cardiff travel agency study and many, many others. The narrower definition also shuts out contrastive sociolinguistic studies that “stage” existing data sets as “laboratories” in which “experiments” may be carried out (e.g., Sandøy 2004). In any case, as background-variable manipulation ⫺ i. e., exploitation of naturally present distinctions in terms of “social group belongingness” ⫺ will never be interestingly (fundamentally) different from what we have already touched upon (section 2.2.3), we shall focus the following discussion of experimental techniques on stimulus variables and their manipulation.

2.3. The sociolinguistic interview as experiment While it is true that interviewing is normally done with aims that do not require the kind of control striven for in experimental approaches to data collection, it may also be the case that an interview takes the form, wholly or partly, of an experiment. In fact, the most common data collection method within variationist sociolinguistics, the sociolinguistic interview, often has experimental aspects to it ⫺ not only in the broader sense of carefully controlled variables, but also in the narrower sense of a deliberately manipulated stimulus variable. Indeed, it is an important aspect of the sociolinguistic interview as developed and used by Labov that it is a stimulus variable in the same way that the two-question “interview” is the stimulus variable in the department stores experiment (section 2.2.1). In either case, the stimulus variable operationalizes the “context of use” and is deliberately manipulated by the researcher ⫺ in order to measure the effect on “language use” (operationalized as dependent variables) in “space” (operationalized as background variables).

535

536

V. Data collection and corpus-building Furthermore, the operationalization of the “context of use” builds, of course, on the same theoretical assumptions about the nature of intra-individual variation (outlined in section 2.2.4). As a sociolinguistic interview is an obtrusive arrangement (the informants know that they are taking part in something that is not an everyday affair), the default interview context is assumed to yield “high awareness” and “careful speech”. Consequently, elicitation of more casual speech is a priority in sociolinguistic interviewing ⫺ operationalized as speech produced with a minimum of monitoring awareness. In order to manipulate the context of use in the sociolinguistic interview (or to manipulate the sociolinguistic interview as a context of use), Labov and his many followers in variationist sociolinguistics have therefore used techniques aimed first and foremost at reducing interviewees’ attention to their speech, but techniques aimed at increasing such attention have also been applied (section 2.3.1). Or, to put it another way, techniques have been used to construct “formal” versus “informal” contexts of use (section 2.3.2).

2.3.1. Constructing casual and careul speech data by manipulating attention In order to reduce attention ⫺ by influencing the psychological “climate” of the encounter in such a way that the interviewee, ideally, forgets about being in an interview, and, in particular, forgets all about its focus on language ⫺ two techniques in particular have been used: (i) on the assumption that interviewee awareness will decrease as duration increases, interviewers make the encounter last one or two hours, or even more; (ii) on the assumption that interviewee awareness will decrease as emotionality increases, interviewers introduce topics and ask questions that are likely to arouse emotions. We should not expect any automatic awareness reduction effect, however, neither from longer duration nor from stronger emotions. Other factors influence the psychological climate as it develops during an interview, including first and foremost the individual personalities of both interviewee and interviewer, and the interpersonal “chemistry” these give rise to. Hence, what should be stressed here is, first of all, the need for ingenuity and flexibility in the design and use of awareness reduction techniques. The “danger of death” question offers a classic illustration of the need to adapt these techniques to the local context under study. In order to arouse emotion and thus reduce attention to speech in the narration about the event, Labov successfully asked his New York informants, Have you ever been in danger of death? Researchers in other and very different communities (including myself in Denmark) have subsequently asked their informants the same question ⫺ with no success at all. Thus, if most people in your community have never been in danger of death, the challenge, of course, is to figure out what other topics might elicit emotional narratives. We may add that it seems more problematic, and perhaps even contradictory, to think along these lines within theoretical frameworks that conceive of interactions as continuous negotiations and renegotiations of relationships and identities. Now, let us move on to techniques that are intended to influence informants’ awareness in the opposite direction, so to speak, and cause increased attention to language. We saw in section 2.2.1 how Labov used the hard-of-hearing technique in the department store study. Such techniques will largely be the ones developed and used to establish linguistic structure in the first place. In their search for sameness and difference amidst the chaos of linguistic variability, linguists will often need to adopt a hard-of-hearing

30. Experimental techniques persona in order to make their informants repeat. Thus, the Danish linguist Jørgen Rischel reported that he was called “The Deaf Ear” by the Mlabri people in Northern Thailand during his many years of fieldwork among them. Moreover, the obvious way of putting focus on language is of course to make some aspect of language the topic of the interview. Finally, as reading can be assumed to require more attention to speech than talking, informants in sociolinguistic interviews may be asked to read aloud, either a text or a list of words. Having informants read “minimal pairs” is used as the ultimate way of directing attention to speech in the Labovian tradition.

2.3.2. Constructing casual and careul speech data by manipulating ormality In addition to awareness versus non-awareness, casual and careful speech is often theorized in variationist research in terms of formality: casual speech is informal, careful speech is formal. We have several options at our disposal if we want to manipulate the formality of the recording situation. i.

We may change our own appearance and behavior as interviewers. A pair of jeans and a T-shirt easily construe another situation than a suit and a tie; serving a beer may change a situation, and so on and so forth. ii. We may change the location of the interview. A first part may be conducted in a private context, e.g., the interviewee’s own home, a second part in a public context, e.g., the interviewee’s working place, etc. iii. We may change the numbers and kinds of participants in the recording situation. A husband or wife, or a friend, may be included, either as silent bystanders or as participants in their own right. With two or more persons present, the interaction can be changed into a conversation without the interviewer if the latter leaves the scene for a time; or the encounter can be set up as a group conversation from the very beginning. Again, as with awareness, we should not expect any automatic effects from manipulating these contextual factors. There is no direct relationship between, say, numbers of participants and degrees of formality. Nor should we expect any simple relationship between degrees of awareness and degrees of formality. While the presence of a friend during an interview, or of several friends during a group conversation, may well reduce the formality of the situation, the awareness issue is, in the first instance, likely to become more complex the more participants there are in an event. So, the bottom line is that, more than anything else, we need to bring reflexivity to our attempts to manipulate the interview context.

3. Experimental construction o language-attitudes data If we assume that people may hold, or express, quite different attitudes towards language use differences ⫺ dependent upon whether or not they are aware of displaying language attitudes ⫺ we may, just as for language use, collect data under “awareness” versus “non-awareness” conditions and register how this manipulation affects the expression of attitudes.

537

538

V. Data collection and corpus-building

3.1. Experimental use o interviews and questionnaires Interviews and questionnaires (cf. Seiler in this volume) belong here to the extent that these methods take on an experimental character when construed as tests of the effects of “question manipulations”. Such manipulations have to do with the type and phrasing of questions (section 3.1.1), the scope and ordering of questions (section 3.1.2) and the further contextualization of questions (section 3.1.3). One should of course try to avoid manipulating in such a way that respondents feel manipulated.

3.1.1. Manipulations o question types and phrasings The type issue is best illustrated with reference to the distinction between open-endedness and closedness in questioning research. For instance, people’s expressed feelings about living in their area or town will be of interest in a study of language variation in space, and the potential difference between answers to open-ended and closed questions may well in itself shed valuable light on the issue. Thus, in our own studies of language attitudes in the Danish town of Næstved, close to Copenhagen, we raised the topic of “local belongingness”, and other topics judged to be of relevance to the community’s language ideology, in both an open-ended way (What is it like to live in the Næstved area?) and in a closed format (How do you feel about living in the Næstved area? plus a choice of predetermined answers), and have sought to integrate data from both types of questioning into the analyses (Kristiansen 2004). In closed (structured) questioning, raising the same issue in different phrasings may well result in quite different answering patterns, and thereby yield possibly valuable data to interpret.

3.1.2. Manipulations o question scope and ordering In questions that involve some kind of evaluative comparison, the range of “objects” included may be altered in order to test the effect of any change in comparative scope. Thus, in the Næstved studies (Kristiansen 1991), one group of adolescents was asked in a questionnaire study to rank sjællandsk, københavnsk and rigsdansk, i. e., the names of the three ways of speaking Danish that are important for social identification in the Næstved area (the “local” variety, the “Copenhagen” variety and the “standard” variety, respectively), while another group was asked to rank a list which included another four such names and thus enlarged the perspective to comprise the whole of Denmark. (In the Næstved case, this widening of the comparative scope did not influence the relative ranking of the three varieties under study). So-called funneling is a technique used to progressively narrow down the focus of questioning. It is an indirect way of approaching the heart of the matter, and may be used as a way of manipulating the respondents’ degree of awareness of the object of inquiry. For instance, if the object of study is the role of language use and attitudes in feelings of local affiliation and identity, one might want to start by addressing the issue with more general area-related questions like What is it like to live in the Næstved area?, Do you often go to Copenhagen?, and end with more specific language-related questions like Have you ever been teased because of your language?, For instance in Copenhagen?. The effect of such focus manipulations may be

30. Experimental techniques studied in answers from individual respondents, but can not really be controlled for unless we administer different sequences of reaction-differentiating questions to two or more groups of respondents that can thus be compared.

3.1.3. Manipulations o question contextualization In fact, the question manipulations we have discussed in the two preceding paragraphs can all be grouped under the heading of changes in the context of questioning. However, another kind of question contextualization technique should be added, which, unlike those already mentioned, does not consist in a manipulation of the context in terms of the questions themselves but in terms of other aspects of the inquiry setting. One way of doing this is to ask the questions first in a questionnaire and then again in a subsequent interview, or the other way round: first in an interview and then followed up afterwards in a questionnaire. Perhaps a more common way of manipulating the contextual condition is to raise issues with the respondents first in individual interviews and then in group conversations.

3.2. Experimental approaches to recognition and representation Social psychologists tell us that evaluation presupposes categorization and recognition, and that categorization in turn involves comparison and evaluation: we can not evaluate something without recognizing it, and we can not recognize something without evaluating it (Hogg and Abrams 1988). In other words, the cognitive and affective aspects of our relationship with the world are two faces of the same coin. In studies of language attitudes, we may therefore feel a need to use some kind of recognition test in order to check on what people (think they) evaluate: what exactly is it that they recognize?

3.2.1. Mental maps Following up on techniques used within cultural geography, and also previously employed in Japanese and Dutch dialectology, Dennis Preston has headed the development of perceptual dialectology (cf. Preston in this volume). Whereas dialectologists are traditionally engaged in the localization of boundaries and areas as they exist by virtue of similarities and difference in people’s use of language, perceptual dialectology’s interest lies in establishing the language-in-space categories people operate with: what do their mental maps of regional speech look like? Where do people think linguistically distinct places lie, and what do they think about them? Answers to these questions are obtained by giving people a blank map, of the United States for instance, and asking them to first encircle what they believe to be distinct dialect areas and then attach some characterizing label to these areas. Computer-assisted analyses of large numbers of such hand-drawn maps allow for generalizations about subjectively extant language-in-space categories. Thus, a map of the USA divided into 14 such categories emerges as the typical southeastern Michigan mental map, whereas the typical southern Indiana mental map consists of just 10 speech regions (Preston

539

540

V. Data collection and corpus-building 1989: 343⫺344). (For illustrations of the mental-map technique and its results, cf. Preston 1989b, 1999; Niedzielski and Preston 2000; Long and Preston 2002.) In a study of perceived English dialect areas in Wales, Garrett, Coupland and Williams (2003: 92) adopted a somewhat different mental-map approach in that they asked their respondents to outline a maximum of eight English dialect areas, followed by the usual invitations to label and characterize the areas identified. The choice of eight was based on piloting work that pointed to this as an optimal number. Respondents generally identified the maximum number requested (2003: 112). In the research referred to above, the mental-map technique is not used experimentally, in the sense of a stimulus variable being manipulated. However, it is easy to see how this could be done, and our examples do indeed suggest that it might be interesting to present the map task in different ways to comparable groups in order to see whether and how such manipulations influence the numbers and types of language-in-space categories that are recognized.

3.2.2. Locating speech samples in space The basic technique for testing people’s ability to “localize” speech differences consists in having them listen to and assess the geographical (or social) “belongingness” of recorded speakers. The concrete design may vary considerably, reflecting the fact that the test will often respond to the divergent research interests of different research programs. In nationwide studies of language attitudes among young Danes (the LANCHART studies; see Kristiansen to appear), the recognition test formed part of a battery of tests which was administered to subjects in a strict procedural order, and the main purpose of which was to manipulate “the awareness condition” under which evaluations took place. The recognition test was presented just after the subjects had finished reacting to audio recordings of twelve young speakers (two boys and two girls for each of the three locally relevant accents, i. e., the local, the modern, and the conservative accents of the standard language) under conditions in which they did not realize they were evaluating differently accented speech. This was now revealed to them, and they were asked to judge, while listening to the tape again, whether they thought the speakers were from Copenhagen or from their own local town. The picture that emerged from this experiment shows high “recognition” rates for young Danes in general: speakers with the modern accent were mostly (i. e., by two out of three subjects) judged to be Copenhageners, speakers with the local accent were predominantly (three times out of four) recognized as locals, whereas the distribution for speakers with the conservative accent was a 50/50 split, thus sustaining a generally held view that what characterizes this latter accent is its non-localizability. In the recognition task included in the Welsh studies reported in Garrett, Coupland and Williams (2003), the informants were asked to choose between seven given geographical names (six Welsh names plus “England”) as they listened to fourteen young speakers (two boys from each of the seven places). Even though the linguistic differences were no doubt greater and should thus have made the task easier, the greater complexity of having to choose between seven places, instead of two as in the Danish test, may well have been the main reason why Welsh adolescents showed much lower percentages of

30. Experimental techniques correct recognitions, in the order of 15 to 40 percent, with a mean of “just below 45 percent” for recognition of speakers of their own local variety (2003: 200⫺201). In his studies of folk perceptions of the north⫺south dimension of US English, Preston (1989a) used speech samples drawn from a geographical continuum of nine sites representing this dimension from Saginaw, Michigan in the north to Dorthan, Alabama in the south. Two groups of judges, representing the northern (Michigan) and middle (Indiana) parts of the continuum, assigned site numbers from 1 (northernmost) to 9 (southernmost) to the voices they heard. The mean scores for both groups indicated that the north⫺south dimension consisted of only two categories of linguistic distinctiveness, with the six northernmost sites belonging to one area, and the three southernmost sites belonging to another area. In a recent web-based experiment, Plichta and Preston (2005) used an analysis/resynthesis technique to prepare a stimulus variable ⫺ a seven-step continuum of monophthongization of the phoneme /ay/ in the word guide ⫺ which respondents reacted to by assigning site numbers from 1 through 9 to the 42 randomized tokens of guide they listened to (three tokens for each of the seven steps for two voices, one male and one female). The results here did not group the stimulus speech samples in terms of a northern and a southern area, but instead showed regularly increasing mean scores for each step along the monophthongization continuum. The perfect and robust discrimination of steps appears as a subconsciously produced pattern at the group level, as respondents generally commented that they did not feel capable of discriminating the varying degrees of monophthongization presented on the tape. In addition, for the same synthesized /ay/ qualities, the subjects as a group ranked the male voice as “more southern” than the female voice, again in a perfect and statistically robust way across all seven steps, thus producing a nice illustration of the intertwined nature of cognition and affection. The authors conclude by expressing a strong belief that “… finely-grained acoustic differences […] may be manipulated in experimental settings to confirm perceptual salience in such tasks as social category identification” (Plichta and Preston 2005: 128). We notice that both qualitative and quantitative aspects of the stimulus voices can easily be manipulated in many ways, and add that, just as for mental-map drawing (section 3.2.1), manipulations of the answering format might also prove revealing, as suggested by Clopper and Pisoni (2007) who report on dialect classification experiments using the same stimulus voices in both a six-alternative forced-choice categorization task and a free classification task.

3.3. Experimental approaches to evaluation Although (re)cognition and affection are two sides of the same coin, it is not only possible but also analytically necessary to focus on one side at the time. Without forgetting that categorization is “thinking” (representations, beliefs), we now change focus to consider categorization in terms of “assessing” (values, feelings). The stimulus material may be presented as either names (labels) for, or examples of, different “ways of speaking” or “speech styles”. We shall distinguish between the evaluation of language in space based on conceptual labels (section 3.3.1) and evaluation that involves speech samples (section 3.3.2).

541

542

V. Data collection and corpus-building

3.3.1. Evaluation based on conceptual labels Here, a list of names or labels for different “ways of speaking” is presented to subjects in questionnaire form together with some format for assessing the conceptualized speech styles. A small number of labels makes it possible to simply ask subjects to rank them according to some criterion. This was the approach adopted in the Næstved studies, where the two label-ranking tasks included either just three or seven labels (mentioned in section 3.1.2), and it was also the approach chosen in the LANCHART studies (mentioned in section 3.2.2), in which no label-ranking task included more than ten labels. Ranking has also been used by Preston (1989b) in his studies of how the accents of the 50 United States (plus Washington DC and New York City) are evaluated in terms of “correctness” and “pleasantness”. With a large number of labels, ranking becomes a rather absurd task, and rating scales are used instead. Thus, in another study by Preston (1989a), subjects rated the many state names for “correctness” and “pleasantness” on ten-point scales. In a classic study within this approach, Giles (1970) operated with sixteen different labels for English accents (regional, social, and foreign) which were rated on seven-point scales for “status content”, “aesthetic content”, and “communicative content” (a measure of perceived ease of interaction with the speaker). In a recent online survey, more than 5000 informants from across the UK rated 34 such labels on two seven-point scales in terms of “prestige” and “pleasantness” (Coupland and Bishop 2007). The experimental nature of these studies made it possible to treat a subset of the 2007 study as a near-replication of the 1970 study, thus accomplishing a real-time comparison of language attitudes (Bishop, Coupland and Garrett 2005).

3.3.2. Evaluation involving speech samples Categorizations of “others” and of “self” are no less intertwined than those of cognition and/or affection (Turner et al. 1987) ⫺ but, again, we may analytically focus on one at a time: first, self-evaluations (section 3.3.2.1), and second, evaluations of others (section 3.3.3.2).

3.3.2.1. Report or evaluation o ones own speech The point of self-report tests is that people “place” their own speech in relation to some other speech. Accomplishing the placement task clearly involves processes of both recognition and emotion. This double-sidedness of the placement task has been exploited in a series of self-evaluation experiments designed to highlight crucial aspects of the “language use/attitude constraints” on social identification in Danish communities (Kristiansen 2004 reports on such an experiment in the Næstved area). These constraints are conceived of as a “normative field” defined by three “norm ideals for speech”. In the experiment, this three-poled normative field is presented to informants (i) on paper as a large triangle with a small circle at each of the three angles, and (ii) on audiotape as three speakers representing each of the small circles at the angles of the triangle. Having listened to the speakers, the informants’ task is to place their own speech wherever

30. Experimental techniques they think appropriate, either within one of the circles (indicating similarity with the corresponding voice) or somewhere within the triangle indicating degrees of judged similarity and difference in relation to the three field-defining voices ⫺ or even outside the triangle if that is judged appropriate. If the resulting data shows significant correlations between self-evaluations and social groups, such patterns can be taken partly as testimony to (the informants’ recognition of) the existence of norm ideals in the speech community and partly as testimony to (the informants’ assessments of) the social meanings that are involved. On the assumption that social identifications (categorizations/ evaluations of the self and others) involve accentuation of resemblances and differences in relation to ingroups and outgroups, respectively, the experimental situation is assumed to provoke stereotyping thought processes such as: “I’m young, so my language is like … and not like …”, or “I’m from town, so my language is like … and not like …”, etc. In other words, the interesting question is not so much whether people assess their own speech correctly from a purely linguistic point of view, but rather which patterns of social meaning emerge. More commonly, self-report tests are used to investigate the relationship between, on the one hand, what people say that they say, and, on the other hand, what they actually do say. For instance, subjects may be presented with two different ways of pronouncing a word and asked to indicate which pronunciation they themselves use. The classic examples of the use of self-report tests are from Labov (1966), who found a general tendency for New Yorkers to “over-report” (i. e., people reported themselves to speak more standardly than they actually did), and from Trudgill (1972), who found an opposite tendency among Norwich men to “under-report” (i. e., they reported themselves to speak more nonstandardly than they actually did). Again, the research interest lies not so much in whether the self-report is linguistically correct or not, but rather in the interpretive implications when it is wrong: what does the phenomenon of over or under-reporting tell us about the system of social values that imbues and surrounds variation in use? In linguistic insecurity tests, another of Labov’s “field experiments” (Labov 1984: 45), the research interest lies in the differences that emerge when people are first asked to indicate which one of two word pronunciations they think is correct, and then to do the same with regard to their own pronunciation. The difference is taken as a measure of “linguistic insecurity” (Labov 1966: chapter 12).

3.3.2.2. Evaluation o the speech o others Data collection methods that somehow ask people directly about the object of study, in our case language variation, are known as direct methods. This has been a characteristic of the approaches we have looked at up until now (except that “funneling” was mentioned as an indirect way of approaching the object of inquiry, in section 3.1.2). In contrast, methods used to elicit assessments of the speech of others are typically construed as indirect ⫺ the reason being that respondents are asked questions about the speakers and not their speech. In itself, this distinction is unproblematic (but readers should note that these terms are also used to label a separate distinction, cf. Seiler, König, in this volume). However, since the reason typically given for indirect questioning is to avoid informants becoming aware of what the inquiry is really about, the distinction between direct

543

544

V. Data collection and corpus-building and indirect approaches to data collection is also associated with the distinction between conscious and subconscious reactions. This association is far from unproblematic. In particular, we should take care not to assume that a speaker evaluation experiment (SEE) secures a subconscious reaction in and of itself. With attitudes towards accent differences as our object of study, we may well claim to have collected indirect data if our respondents reacted to speakers, but we can not claim to have collected subconscious data if the respondents realized that what was at stake were their attitudes towards accent differences. In what follows, we shall review and discuss speaker evaluation experiments in terms of (i) the stimulus speakers, (ii) the measurement instrument, (iii) the procedure for collecting subconscious evaluations, and finally (iv) the scope of evaluation: variety or variant? (i)

The stimulus speakers. The central data collection method in speaker evaluation experiments has been the matched guise technique (MGT), created by the Canadian social psychologist Wallace Lambert and his colleagues in the late 1950s and 1960s (Lambert 1967) and subsequently further developed and applied in a plethora of studies within the social psychology of language field (Ryan and Giles 1982; Giles and Coupland 1991, chapter 2). The same individual appears twice (or even more often) on the stimulus tape, each time exhibiting a new speech style, a new guise. Subjects mostly remain unaware of this because the matched guise voice appears at some distance on the tape with filler voices in between. The idea behind the matched guise technique is to manipulate speech style while controlling for all other factors that might have an impact on the assessment. Extralinguistic features like voice and delivery quality are controlled for by virtue of the matched guise voice being one and the same person; “content” is controlled for by having the matched guise voice (and the filler voices) read one and the same text. The recorded voices are kept equal in terms of duration; most excerpts used in speaker evaluation experiments are some 20 to 50 seconds long. The linguistic contrasts of the matched guise voice may be manipulated in many ways. Some are of more interest for our language-in-space perspective than others. We should mention, firstly, manipulations in terms of national languages, and recall Lambert’s studies in Montreal in which the matched guise voice spoke English Canadian in one guise and French Canadian in the other. Secondly, and obviously pertinent to the study of language in space, matched guise voice manipulations in terms of dialects and accents are rich in possibilities. The Giles (1970) study included, in addition to the label-ranking task mentioned in section 3.1.2, a speaker evaluation experiment with a matched guise voice which performed thirteen regional, social, and foreign accents of English. Thirdly, matched guise voice manipulations in terms of more or less “anglicized” speech may be a promising approach to language-in-space issues in the era of globalization. Thus, we used “Englishcolored” guises versus “pure national” guises in MGT-based speaker evaluation experiments carried out in order to study and compare attitudes towards the influence of English in all Nordic countries (Kristiansen 2006). It is often felt that the matched guise technique, because of its experimental preoccupation with control, produce data that suffers heavily from low ecological validity in many respects. The matched guise voice is not the least of the problems. It may prove simply impossible to find a person who is able to speak “naturally”

30. Experimental techniques in the accents we want for our stimulus tape. The only solution is to use different speakers ⫺ thus achieving a speaker evaluation experiment design which is no longer really based on matched guises, but on what have been called “verbal” guises (initially suggested by Cooper 1975: 5). We may also want to relax our control over content and use voices that, instead of reading the same text aloud, speak “naturally”, either about one and the same topic, or about any “neutral” topic. Control can still be achieved by using two or more voices for each of the varieties to be evaluated. For us to argue that variety differences have a decisive role in such evaluations, the result pattern must indicate that voices representing the same variety have by and large been treated the same. It goes without saying that such relaxations of control yield data that need to be examined and analyzed with a sharpened sensitivity to how multiple factors interact when speakers and their speech are socially evaluated. (ii) The measurement instrument. The usual measuring instruments in speaker evaluation experiments are some kind of scale: either bipolar adjective scales, known as semantic differentials (How do you find this speaker? fascinating L boring) or unidimensional scales (How fascinating do you find this speaker? not at all L very much). Typically, the number of steps on these scales is either five or seven, leaving open the possibility for subjects to choose the “neutral” midpoint. If several scales are used, the scale ratings may be factor-analyzed in order to discover underlying evaluative dimensions. It is considered a well-documented finding of language attitudes research that all such scales load into one of only a couple of dimensions: status and solidarity (Ryan, Giles and Sebastian 1982), or superiority, attractiveness and dynamism (Zahn and Hopper 1985). In concrete research projects, scales are often chosen to represent these dimensions because of their established position in previous research. An alternative to choosing scale items along well-known lines is the “keywords” approach. It simply refers to an open-ended questioning format which invites subjects to write down all their immediate impressions, or maybe better their first three impressions, as they listen to the speakers on the tape. The resulting material may not only be analyzed in its own right, but also used to find the “right” labels for evaluative scales (Garrett, Coupland and Williams 2004; Maegaard 2005). (iii) Collecting subconscious evaluations. In fact, speaker evaluation experiments are rarely designed and carried out in ways that allow for the elicitation of subconscious language attitudes. The first prerequisite for this to be possible is, of course, not to inform the subjects about the nature of the experiment until after its completion. Next, we must take care that the nature of the experiment is not revealed by the stimulus material itself. If you are asked to evaluate voices speaking in different languages, or clearly different dialects or accents, it would be strange not to think that “this is about my language attitudes”. However, as long as the stimulus voices are kept within the “natural” range of variation for the speech communities under study, our experience reveals that subjects do not become aware that they are evaluating accent differences. As to the measurement instrument, questions addressing the speakers’ language in any way, for instance in terms of “correctness” or “dialect color”, should be avoided. Such questions bring us back into direct questioning, of course, and take us far from the realm of subconscious language attitudes. Even if we succeed in maintaining a lack of awareness with regard to the object of study, our subjects will always know that they are taking part in an experiment.

545

546

V. Data collection and corpus-building The data collection is an “obtrusive” event in, say, normal school-day activities. If we want to pursue unobtrusively collected data, we need to find an everyday setting where the manipulation of stimulus voices can go unnoticed while reactions can be reliably registered and background variables controlled at the same time. There are few examples of such studies, no doubt because they are rather difficult both to design and to execute. However, the basic idea of the short “planted encounter” has been exploited in studies using field researchers who, instead of feigning hardness of hearing like Labov in the New York department stores (cf. section 2.2.1), shift between two linguistic codes as they address people on the street and ask for some information (Rosenbaum et al. 1977; Bourhis 1984). If the reactions to the differently spoken addresses are measured in terms of, say, “readiness to help”, the results may lend themselves to interpretation in terms of “language attitudes”. If we want to use tape-recorded stimulus voices in unobtrusive speaker evaluation experiments, we will have to operate in social situations where public address systems are “natural”. The loudspeaker system in a Cardiff film theatre was used by Bourhis and Giles (1976) to ask audiences to help plan future programs by filling out a questionnaire form. The announcement was presented in four matched guises, one per evening: RP, mildly and broadly South Welsh-accented English, and Welsh. The same method was adopted in a study in Næstved, but with a verbal guise design in that the announcement was read out by four different speakers: in modern and conservative standard, and in mildly and broadly locally accented speech (Kristiansen 1997). In both studies, the degree of cooperation was measured as the ratio of questionnaires completed to tickets sold. (iv) The scope of evaluation: Variety or variant? Finally, it should be noted that speaker evaluation experiments may be designed to gather assessments either of “whole” varieties or of variants of a particular variable. The “holistic” approach effectively imposes itself as soon as a stimulus voice speaks for more than a few seconds. This means, however, that we cannot, no matter what detailed linguistic analyses the stimulus voices are subjected to, infer anything certain about the contribution of isolated variables to the assessment of a voice. If we want to focus on particular variables, the stimulus tape has to be created by “cutting and pasting” different values of the variable under study into short versions of the same sentences. Originally, this was done by Labov, who used the cut-and-paste method to study the social evaluation of postvocalic /r/ and other phonological variables in New York (Labov 1966: chapter 11) and subsequently also in Philadelphia (Labov 2001: 206⫺ 222). Campell-Kibler (2007) seeks to combine the holistic and particularistic approaches in her recent studies of the (ING) variable and its social meanings in complex interplay with Southern and gay accents in the United States.

4. Reerences Anderson, Benedict 1983 Imagined Communities. London: Verso Bishop, Hywel, Nikolas Coupland and Peter Garrett 2005 Conceptual accent evaluation: Thirty years of accent prejudice in the UK. Acta Linguistica Hafniensia 37: 131⫺154.

30. Experimental techniques Bourhis, Richard Y. 1984 Cross-cultural communication in Montreal: Two field studies since Bill 101. International Journal of the Sociology of Language 46: 33⫺48. Bourhis, Richard Y. and Howard Giles 1976 The language of cooperation in Wales: A field study. Language Sciences 42: 13⫺16. Campbell-Kipler, Kathryn 2007 Accent, (ING), and the social logic of listener perception. American Speech 82(1): 32⫺64. Clopper, Cynthia G. and David B. Pisoni 2007 Free classification of regional dialects of American English. Journal of Phonetics 35: 421⫺438. Cooper, Robert L. 1975 Introduction to language attitudes II. International Journal of the Sociology of Language 6: 5⫺9. Coupland, Nikolas 1984 Accommodation at work: Some phonological data and their implications. International Journal of the Sociology of Language 46: 49⫺70. Coupland, Nikolas and Hywel Bishop 2007 Ideologised values for British accents. Journal of Sociolinguistics 11(1): 74⫺93. Garrett, Peter, Nikolas Coupland and Angie Williams 2003 Investigating Language Attitudes. Social Meanings of Dialect, Ethnicity and Performance. Cardiff: University of Wales Press Garrett, Peter, Nikolas Coupland and Angie Williams 2004 Adolescents’ lexical repertoires of peer evaluation: Boring prats and English snobs. In: Jaworski, Coupland and Galasin´ski (eds.), 193⫺225. Giles, Howard 1970 Evaluative reactions to accents. Educational Review 22: 211⫺227. Giles, Howard and Peter F. Powesland 1975 Speech Style and Social Evaluation. London: Academic. Giles, Howard and Nikolas Coupland 1991 Language: Contexts and Consequences. Milton Keynes: Wadsworth. Hewstone, Miles, Wolfgang Stroebe, Jean-Paul Codol and Geoffrey M. Stephenson (eds.) 1988 Introduction to Social Psychology: A European Perspective. Oxford: Blackwell. Hogg, Michael A. and Dominic Abrams 1988 Social Identifications. A Social Psychology of Intergroup Relations and Group Processes. London: Routledge. Jaworski, Adam, Nikolas Coupland and Dariusz Galasin´ski (eds.) 2004 Metalanguage. Social and Ideological Perspectives. Berlin: Mouton de Gruyter. Kristiansen, Tore 1991 Sproglige normidealer pa˚ Næstvedegnen. Kvantitative sprogholdningsstudier [Norm ideals for speech in the Næstved area. Quantitative language attitude studies]. Unpublished Ph.D. dissertation, Copenhagen University. Kristiansen, Tore 1997 Language attitudes in a Danish cinema. In: Nikolas Coupland and Adam Jaworski (eds.), Sociolinguistics. A Reader and Coursebook, 291⫺305. London: Macmillan. Kristiansen, Tore 2004 Social meaning and norm-ideals for speech in a Danish community. In: Jaworski, Coupland and Galasin´ski (eds.), 167⫺192. Kristiansen, Tore to appear The macro-level social meanings of late-modern Danish accents. Acta Linguistica Hafniensia 41. Kristiansen, Tore (ed.) 2006 Nordiske sprogholdninger. En masketest [Nordic language attitudes. A matched-guise study] Oslo: Novus.

547

548

V. Data collection and corpus-building Labov, William 1966 The Social Stratification of English in New York City. Washington, DC: Center for Applied Linguistics. Labov,William 1972 Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press Labov, William 1984 Field methods of the Project on Linguistic Change and Variation. In: John Baugh and Joel Scherzer (eds.), Language in Use: Readings in Sociolinguistics, 28⫺53. Englewood Cliffs: Prentice-Hall. Labov,William 2001 Principles of Linguistic Change, vol. 2, Social Factors. Malden: Blackwell. Lambert, Wallace 1967 The social psychology of bilingualism. Journal of Social Issues 23: 91⫺109. Long, Daniel and Dennis R. Preston (eds.) 2002 Handbook of Perceptual Dialectology, vol. 2. Amsterdam: John Benjamins. Maegaard, Marie 2005 Language attitudes, norm and gender. A presentation of the method and results from a language attitudes study. Acta Linguistica Hafniensia 37: 55⫺80. Niedzielski, Nancy A. and Dennis R. Preston 2000 Folk Linguistics. Berlin: Mouton de Gruyter. Plichta, Bartłomiej and Dennis R. Preston 2005 The /ay/s have it. The perception of /ay/ as a north-south stereotype in United States English. Acta Linguistica Hafniensia 37: 107⫺130. Preston, Dennis R. 1989a Standard English spoken here: The geographical loci of linguistic norms. In: Ulrich Ammon (ed.), Status and Function of Languages and Language Varieties, 324⫺354. Berlin: de Gruyter. Preston, Dennis R. 1989b Perceptual Dialectology. Nonlinguists’ Views of Areal Linguistics. Dordrecht: Foris. Preston, Dennis R. (ed.) 1999 Handbook of Perceptual Dialectology, vol. 1. Amsterdam: Benjamins. Rosenbaum, Yehudit, Elizabeth Nadel, Robert L. Cooper and Joshua A. Fishman 1977 English on Keren Kayemet Street. In: Joshua A. Fishman, Robert L. Cooper and Andrew W. Conrad (eds.), The Spread of English: The Sociology of English as an Additional Language, 179⫺196. Rowley, MA: Newbury House. Ryan, Ellen B. and Howard Giles (eds.) 1982 Attitudes towards Language Variation. Social and Applied Contexts. London: Arnold. Ryan, Ellen B., Howard Giles and Richard J. Sebastian 1982 An integrative perspective for the study of attitudes toward language variation. In: Ellen B. Ryan and Howard Giles (eds.), 1⫺19. Sandøy, Helge 2004 Types of society and language change in the Nordic countries. In: Britt-Louise Gunnarsson, Lena Bergström, Gerd Eklund, Staffan Fridell, Lise H. Hansen, Angela Karstadt, Bengt Nordberg, Eva Sundgren and Mats Thelander (eds.). Language Variation in Europe: Papers from the Second International Conference on Language Variation in Europe, ICLaVE 2 Uppsala University, Sweden, June 12⫺14, 2003, 53⫺76. Uppsala: Department of Scandinavian Languages, Uppsala University. SDE ⫽ Danmarks Nationalleksikon 1996 Den Store Danske Encyclopædi, vol. 5. Copenhagen: Danmarks Nationalleksikon. Swann, Joan, Andrea Deumert, Theresa M. Lillis and Rajend Mesthrie 2004 A Dictionary of Sociolinguistics. Edinburgh: Edinburgh University Press.

30. Experimental techniques

549

Trudgill, Peter 1972 Sex, covert prestige and linguistic change in urban British English. Language in Society 1: 179⫺195. Turner, John C., Michael A. Hogg, Penelope J. Oakes, Stephen D. Reicher, Margaret S. Wetherell 1987 Rediscovering the Social Group: A Self-Categorization Theory. Oxford: Blackwell. Zahn, Christopher J. and Robert Hopper 1985 Measuring language attitudes: The speech evaluation instrument. Journal of Language and Social Psychology 4: 113⫺123.

Tore Kristiansen, Copenhagen (Denmark)

VI. Data analysis and the presentation o results 31. Measuring dialect dierences 1. 2. 3. 4. 5. 6. 7.

Why measure? Categorical data String distance measure An example Evaluation Emerging issues References

1. Why measure? Dialectology thankfully inherits large data reserves from earlier practitioners, especially the compilers of dialect atlases designed to display variation in comparable linguistic items. These compilers collected linguistic variants such as the lexical realization of the word for ‘house fly’, the common order of pronominal objects, or the pronunciation of (the first vowel in) the word marry throughout a large selection of sites in a language area. We refer to such items of variation as features or variables. Large collections may include several hundred sites at which hundreds of features are documented. One of the primary tasks of dialectology is to characterize this variety of speech forms. Naturally there are many other tasks, e.g., determining the extralinguistic correlates of variation, modeling the cognitive processes needed to deal with variation in form, and understanding the tension between variation and effective comprehension, but many of these require good primary characterizations of the linguistic variation. The data reserves are so large and complex that simple attempts to characterize the differences always encounter problems of different sorts, many of which are well documented. Bloomfield (1933: chap. 19) discusses how linguistic variants often fail to map neatly to geography, noting exceptions of two types ⫺ linguistic exceptions, in which linguistically similar material does not project to geography in the same way (the same etymological vowel is realized differently in two different words) and geographic exceptions, where linguistic variants project fairly simply but where pockets of exception remain. A larger problem is that even very thorough studies (Kurath and McDavid 1961; König 1994) inevitably examine only a very small fraction of the features documented in substantial atlases. This leaves dialectology open to the charge statisticians have dubbed “cherry picking”, i. e., picking variables that confirm the analysis one wishes to settle on. Given the tens of thousands of features which may differ between languages (in lexical items alone), we would do well to take measures to avoid the danger of fortuitous selection. Finally, we are motivated to develop measurement techniques in order develop hypotheses about variation that require abstract characterizations. These indeed seem within reach given the right level of abstraction.

31. Measuring dialect differences We preview briefly how the use of measurement schemes attacks these issues. By measuring differences we map them to numbers. This numerical characterization enables us to integrate information from large samples of linguistic variables ⫺ just by adding the differences. We postulate that the pronunciation difference between night as pronounced diphthongally in standard American, [na=t], and the same word pronounced monophthongally in the American south, [nat], is a number that may be added to other measures of difference from a large sample. Given this, we may collect many measurements involving many pairs of pronunciations, enabling an aggregate and therefore more abstract characterization of the data. Once we have an aggregate, then we need not be distracted by linguistic exceptions, which simply contribute differently ⫺ perhaps more and perhaps less ⫺ to the sample. Geographic exceptions may of course remain, but only if the weight of a great deal of linguistic evidence bears this out. The single exceptional feature no longer spoils a characterization, which is based on the tendencies of many features. We avoid the dangers of relying on fortuitously chosen data by including a great deal of data, and abstract characterizations inherit the reliability of the mass of data on which they are based.

2. Categorical data The strategy of characterizing dialects on the basis of large aggregates of data samples was pioneered in dialectometry by Jean Se´guy and Hans Goebl (Se´guy 1973; Goebl 1984), who analyze large samples at a nominal, or categorical level. Categorical data analysis views data as belonging either to the same or to different categories. One may then, for instance, measure the similarity between the lexicalizations for the concept dragonfly, i. e., as zero (0) where the variants, such as darning needle and dragonfly, are not the same and one (1) if they are. Alternatively, one may measure the dissimilarity or distance between the two variants, in which case the values switch. Henceforth we focus on distances rather than similarities, but we maintain that the two perspectives are interchangeable. It is also possible to weight some items as more important than others. Goebl (1984), for example, advocates the use of a similarity measure weighted to favor infrequent coincidence, which Nerbonne and Kleiweg (2007: 160⫺163) evaluate positively. We shall ignore such weightings in the remainder of this article, but we note that it is straightforward to apply weightings such as Goebl’s to the all the aggregating procedures discussed below. We also need to deal with competing forms. In the simple case where a single form at one site is compared with two at another, the mean of the two distances is used. Nerbonne and Kleiweg (2003) generalize this idea. Both Se´guy and Goebl include pronunciation differences among the variables they quantify. If the (first vowel of the) word marry is among the pronunciation variables, then one may review the variants to determine whether the pronunciation is [æ], [e], [e], or even [æw], [æß ], [eœ] or [e˜]. We then check straightforwardly whether two pronunciations are the same or not. A large number of such comparisons provides a reliable basis on which to measure varietal similarity. Exact identity may be too demanding a criterion, in which case one needs a procedure testing whether two sounds belong to the same

551

552

VI. Data analysis and the presentation of results class or not, effectively dealing with the problem of individual or irrelevant variation. In the running example, one might distinguish [æ] and its variants (including [æw ] and [æß ]) from raised variants such as [e] and [e] (including [eœ] and [e˜]). Naturally, it is important to defend the classification chosen. But if the classification is sound, then it is good methodology to include this 0/1 measurement as one of a large number on which an aggregate relation is assayed. Even though their methodology is sound, we have sought to go beyond the Se´guy/ Goebl categorical level of analysis for pronunciation differences, first, because it requires manual intervention in isolating specific aspects of pronunciation, where an automated process is also possible. Second, we wished to make fuller use of the rich dialect atlas transcripts, rather than being limited to selected variables.

3. String distance measure Fortunately, there are effective algorithms available for comparing strings, i. e., sequences of symbols. Gusfield (1999: chap. 11) provides an excellent summary of the current state of the art. Rather than try to summarize all the techniques, we present edit distance here, also known as Levenshtein distance and string (edit) distance. We can approach edit distance from two perspectives, and it will be instructive to use both in this overview. One the one hand, one may ask how many operations of a simple sort are required to transform one string into another. We illustrate how one dialectal pronunciation of German Durst ‘thirst’, namely [twes] (Aachen) is transformed into another, [tcst] (Vielbrunn). By writing each derivation step, we can see the operations at work (Figure 31.1). t t t t

w c c c

e e

s s s s

substitute [c] for [w] delete [e] insert [t] t

Fig. 31.1: A set of transformations to map [twes] to [tcst]. Edit distance associates costs with each step and is then defined as the total cost of the transformation (the sum of the costs for each step)

If each operation is associated with a cost, e.g., one, then the Levenshtein distance is the sum of the least costly set of operations mapping one string to another. Since the three operations above are indeed minimal, the distance between the two strings is the sum of the cost of the operations, three. This is naturally rough; we examine more sensitive costs in section 3.3. Gusfield (1999: 11.1⫺11.3) presents a dynamic programming algorithm for calculating Levenshtein distance efficiently. The alternative perspective is that of alignment. Proceeding from the least costly derivation, we align identical segments that remain constant during the derivation and all pairs of segments where one was substituted for the other. Finally we align symbols with an empty segment in case they are involved in deletions or insertions. The result in this case is shown in Figure 31.2.

31. Measuring dialect differences t t Costs

553

w c

e

1

1

s s

t 1

Fig. 31.2: The calculation of the edit distance between two strings implicitly aligns the component elements. This is the alignment of the transcriptions based on the operations in Figure 31.1. Note that insertions and deletions create correspondences where only one segment is specified

Given a derivation, the alignment may be recovered. Alignments are an important check on the quality of measurements (see section 5), and we may search the aligned segment pairs ([w/c], [e/ ] and [ /t] in Figure 31.2) for regular correspondences (see Kondrak 2002; Prokic´ 2007). The procedure is normally applied to the entire dialect material available. We measure the differences not only between all pairs of transcriptions for Durst ‘thirst’, but also those of dozens to hundreds more. It is sensible to use material which represents true pronunciation differences as purely as possible, and which therefore contains as few differences as possible attributable to inflectional morphology, sandhi, and intonation. After determining all of the pronunciation distances for all of the words, the pronunciation distance between sites is simply the mean pronunciation distance of all the words in the atlas’s sample. These site ⫻ site differences may be further analyzed using hierarchical clustering or multidimensional scaling (see Nerbonne, Heeringa and Kleiweg 1999; Heeringa 2004; or below for examples). Peter Kleiweg maintains an interactive demo at which includes various operation weights as well as the opportunity to use segment distances derived from features. We turn now to the linguistic interpretation of this sort of measurement. We note first that the measurement may be automated so as to require no manual intervention (see RuG/L04, ), and second, that it is sensitive to a large number of pronunciation differences, not merely to a limited set chosen by the investigator. Third, however tempting it might be to interpret the insertions, deletions and substitutions as historical sound changes, we warn that the operations work strictly on the surface. Alignments provide documentation of sound correspondences, but the procedure is not designed to discover or simulate historical changes. Fourth, even though there is a unique least cost of operations mapping from one string to another, there may be different sets of operations that correspond to different alignments. This is one reason we are motivated to explore more sensitive measures of phonetic overlap (see section 3.3). Fifth and finally, there are many potential refinements of the procedure which might be expected to yield more precise measurements. In section 3.2 we discuss how one might discount the effects of fast speech rules, and in section 3.3 how to build more phonetic sensitivity into the measure, e.g., by weighting the substitutions phonetically. In section 3.5 we consider enforcing a syllabicity constraint, making the calculation sensitive to context, and including metathesis as a basic operation.

554

VI. Data analysis and the presentation of results

3.1. Formal properties We note some formal properties of edit distance, reviewing material in Kruskal (1999) and Gusfield (1999). First, edit distance is a generalization of Hamming distance, which makes no provision for insertions and deletions. We illustrate Hamming distance in Figure 31.3. t t Costs

w c

e

1

1

s s

t

t t

1

w c

e s

s t

1

1

1

Fig. 31.3: Two alignments for the string pair [twes / tcst]. Even though the alignments do not differ in costs, each with a total cost of three, the left-hand alignment is linguistically preferable because it never aligns vowels with consonants

Even though the distance happens to be the same in this case, it is clear that the Hamming procedure misses the coincidence of [s] in the two strings, so that it is less suitable for dialectological application. The psycholinguistics of spoken word recognition is extremely sophisticated in some aspects, but it uses Hamming distance ⫺ on carefully controlled material in which insertions and deletions do not occur ⫺ as a measure of word similarity (Luce and Pisoni 1998; see also section 6). There are even less sophisticated measures, such as Jaccard distance and Dice, which ignore the order of segments (Manning, Prabhakar and Schütze 2008). Second, we note that Levenshtein measures are normally distances in the mathematical sense, i. e., numbers greater than or equal to zero; zero just in the case where two strings are the same; symmetric, so that the distance from string s1 to s2 is always the same as the distance from s2 to s1; and in conformance with the so-called triangle inequality: the distance from s1 to s2 is less than or equal to the sum of the distances from s1 to s3 and s3 to s2, for any string s3. The fact that the distances are symmetric makes them unsuitable for modeling some problems, for example, the confusability matrices phoneticians compile from how frequently one segment may be mistaken for another (Johnson 2004: chap. 4). If one modeled confusability with similarity, implemented by edit distance on samples of the segments’ spectrograms, the attempt would be limited by the inherent symmetry of edit distances.

3.2. Herrgen and Schmidts dialecticality measures Independent of the work on Levenshtein distance, Herrgen and Schmidt (1989) sketched a procedure for comparing pronunciations of entire words which has been applied in several projects and publications. Our presentation follows Lameli (2004: chap. 5). The Herrgen and Schmidt procedure shares the goal of being able to evaluate a large sample of comparable pronunciations, and to include all the material from every word transcription rather than just selected differences. The procedure assumes feature-based descriptions of each segment are available; these are interpreted numerically. For exam-

31. Measuring dialect differences ple if there are four vowel heights, corresponding to [i], [e], [e] and [a], then the height difference between [i] and [e] is one, and that between [i] and [a] three. It is challenging to specify a segment distance table exhaustively (see section 3.3), but Lameli completes the table successfully. The segment distance table, thus derived, is then used as the basis for calculating word differences, which are simply the sum of differences in aligned segments. There are several unique aspects to this research line. First, the work focuses less on measuring dialect differences, more on measuring differences between a dialect on one hand and the standard language on the other. This reflects the interests of the researchers, but it also allows them to incorporate a second unique feature, a normalization with respect to fast-speech rules. The Herrgen and Schmidt procedure does not regard differences as genuine if they might have arisen through the application of a fast-speech rule. For example German Lippen ‘lips’ may have a canonical standard pronunciation as [1l=pen], but it is pronounced in fast speech as [1l=pm]. Dialect pronunciations which 1 elide the schwa and assimilate the final [n] should not be measured as differing from Standard German. In unpublished work at Groningen we have experimented with implementing several fast-speech rules in Standard German, measuring the distance from a dialectal form to each “allegro” variant, and then using the least value as the distance. Third, Herrgen and Schmidt’s rules are sensitive to small differences, but they also set maximal segment difference values so that word measurements are not overwhelmed by single-segment differences. Fourth, there is still no automatic procedure for applying Herrgen and Schmidt’s difference metric (but Björn Lüders of Marburg is currently engaged in an implementation). This probably derives from detailed and complex rules for handling some special cases such as the German /a/ and foreign borrowings. This overview is too brief to review all of the details.

3.3. Segment distances It is linguistically natural to wish to incorporate more phonetic sensitivity into the string distance measure. After all, an English dialect with path as [paθ] is closer to one with [pæθ] than to a third with [peθ]. There is also a technical reason for preferring more sensitive measures of segment difference, namely the wish to avoid multiple alignments. The more sensitive the phonetic measure, the less likely these are. Accordingly, not only Herrgen and Schmidt (section 3.2), but also Kessler (1995), Nerbonne and Heeringa (1997), Kondrak (2002: chap. 4.5), Heeringa (2004: chap. 3⫺ 4), McMahon, Heggarty, McMahon and Maguire (2007) and others have proceeded from segment distance tables, and, for the most part, have used the Levenshtein procedure sketched above. Heeringa (2004) devotes ninety pages to a comparison of segment distances derived from the feature system in The Sound Pattern of English (Chomsky and Halle 1968), two different systems developed to score transcription quality, and a fourth system Heeringa develops based on curve distance in canonical spectrograms. All of the systems allow for the representation of diphthongs, affricates, stress, length, syllabicity, and a wide range of secondary articulations. In the same spirit as Herrgen and Schmidt’s maximal values for segment distances, Heeringa uses a logarithmic correction to limit the impact of differing segments, following Stevens’ (1975) idea that psychophysical reactions scale logarithmically.

555

556

VI. Data analysis and the presentation of results Heeringa achieved modestly superior analyses using spectrogram-based segment distance tables. (See also sections 5 and 6 on comparing putative improvements.) We note further that, given a table of segment distances, however derived (called alphabetic weights in Gusfield 1999), the dynamic programming algorithm computing Levenshtein distance always returns the optimal alignment with respect to that table (Kruskal 1999: section 5; Gusfield 1999: 11.5), i. e., the alignment that minimizes the sum of the aligned segments’ differences. We therefore say that the Levenshtein procedure “lifts” the segment distance table to a sequence distance measure. McMahon and coauthors (2007) aim to align words with respect to etymological frames, which would seem to require moving from two-string alignments to the simultaneous alignment of three strings, but their procedure is not explained in detail (see Gusfield 1999: chap. 14 on the difficult problem of multiple string alignment).

3.4 Other related work Vierrege, Rietveld and Jansen (1984) developed a system for evaluating student transcribers that assigned varying penalties to mistranscriptions. Kessler (1995) first applied edit distance to dialect material. Like Herrgen and Schmidt, he proceeded from a segment distance table, which he also compared to a binary scheme in which segments were either alike or different, concluding that the latter was superior. Kessler applied clustering to check whether the edit distances could delineate dialect areas. Nerbonne, Heeringa and Kleiweg (1999) followed with an analysis of Dutch and introduce multidimensional scaling (MDS) as a means of further analyzing the average pronunciation differences. Both clustering and MDS are illustrated in section 4. Heeringa (2004) analyzed Norwegian and Dutch, comparing many options for phonetic representation (see section 3.3), clustering and MDS. This is the most thorough treatment of the subject to date. There are also analyses of German, Bulgarian, American English, Italian, Sardinian, Bantu, and some Indo-Iranian and Turkic varieties. Kondrak (2002) modified edit distance to discount mismatches near the beginnings and ends of words and applied his algorithm to diachronic phonology, experimenting on Indo-European and Algonquian languages. By keeping track of frequent operations, he could detect regular sound correspondences and cognates. McMahon, Heggarty, McMahon and Maguire (2007) apply an alignment algorithm to English dialects and analyze results using phylogenetic algorithms designed to infer genealogical trees (and/ or networks).

3.5. Other reinements The success of the techniques raises questions about its linguistic underpinnings, some of which we explore here. Early on, inspection of the alignments induced by the dynamic programming algorithm revealed alignments such as the following two: t t

w c

e s

s t

t t

w c

e

s s

t

31. Measuring dialect differences Each of these requires three operations, yielding distances of three, but intuitively the first violates a syllabicity constraint by allowing the consonant [s] to replace the vowel [e]. It is a straightforward matter to enforce this constraint, and Heeringa, Kleiweg, Gooskens and Nerbonne (2006) claim that this yields superior analyses. While almost all applications have ignored the context of sound correspondences, Heeringa and coauthors (2006) operationalize the effects of context by applying the algorithm not to single segments, but rather to bigrams, a standard technique for incorporating context in computational systems. Results were slightly, but consistently better. There are also extensions of the Levenshtein procedure allowing metatheses (also known as “swaps” or “transpositions”); see Kruskal (1999: 9⫺10). Metathesis is rare in languages analyzed to date, with Bulgarian a notable exception. Adding swaps to the dynamic programming algorithm can be quite difficult when segment distance tables are used (Wagner 1999). Stress can be straightforwardly interpreted as a vowel feature. Under this scheme the verb contract [ken.1trækt] and the noun [1kcn.trækt] differ at two positions, the vowels. Without special treatment, the words would have differed in the first vowel, but not in the second. Because tone is sometimes not realized on single segments, but rather in complicated ways dependent on the sequence of syllables, incorporating it into the distance measure in a general way is less straightforward. Gooskens and Heeringa (2006) contrast the three tone patterns of Norwegian and compare the degree to which perceptual distances can be predicted by prosodic differences with the degree to which they can be predicted by segmental phonological differences measured using edit distance. Finally, we would like to add that, although it is good scientific practice to prepare one’s data carefully to eliminate potential confounding factors and “noise”, it is not always practical. Fortunately, the procedures we sketch are robust enough to function well even on noisy data. The Dutch dialect source used by Heeringa (2004) was analyzed in two ways: once restricted to cognate words (allowing morphological differences, cf. dog and doggy), and once comparing all semantically similar words, including unrelated words (cf. friend and buddy). Given a 125-word sample, results correlate nearly perfectly (r ⫽ 0.98). The replication on a Norwegian sample was slightly lower (r ⫽ 0.95, n ⫽ 58). We consider other potential refinements in section 6: “Emerging issues”.

4. An example The Linguistic Atlas of the Middle and South Atlantic States (LAMSAS) comprises material collected on the eastern seaboard of the US between 1933 and 1974. The area extends from northern Florida through to New York State and includes all the intermediate states on the Atlantic coast, plus West Virginia. Our focus here is on the data collected by Guy Lowman in 1933⫺1936, which makes up roughly seventy percent of the total (see Map 31.1). We focus on just Lowman’s work to avoid the confounding influence of differing fieldwork techniques and/or transcriptions (Nerbonne and Kleiweg 2003). The data were obtained using a questionnaire in which respondents were asked how they expressed everyday things and events, e.g., “If the sun comes out after a rain, you say the weather is doing what?” (used to elicit clearing up, fairing off and forty other dialectal variants. The LAMSAS material is publicly accessible for reanalysis (see

557

558

VI. Data analysis and the presentation of results ; Kretzschmar 1994) and contains the responses of 1162 informants interviewed in 483 communities. The responses to 151 items are included in the web distribution, the basis for the work here. We analyze Lowman’s data in what follows. In Lowman’s section of the database there are 92,537 transcriptions involving 1.3 million phonetic tokens collected from 363 locations, and 797 informants. There are, on average, 14.0 characters per string, which are parsed into 7.9 sound tokens (IPA segments, often with diacritic) per string. There are 1,677 unique sounds (combinations of base segments and various diacritics) and 1,132 unique vowel sounds alone. LAMSAS usefully contains both phonetic and orthographic transcriptions, allowing us to focus measurements on comparable material, ignoring lexical and morphological differences. We implemented the LAMSAS feature system as a segmental basis. We present the features for vowel system in Table 31.1, and suppress discussion of the consonantal features in the interest of saving space. If one chooses to adopt an existing feature system, as we did, the question of how to numerically interpret the feature values arises. Thus Kurath and Lowman distinguished nine different vowel advancement positions, and fifteen vowel heights (Kretzschmar et al. 1994: 114⫺120), but we need to assign numbers to the positions if we are to use them to measure segment differences. Table 31.1 shows how we did this, allowing six as a maximum advancement difference and only three-and-a-half as a maximum height difference. Wherever the values of one feature may differ more than those of another, the scheme in Table 31.1 effectively weights the first feature more heavily. As just noted, advancement may differ by as much as six, while rounding can not differ by more than one, so differences in advancement can count much more. Diacritics representing stress, rhotism, pharyngealization and devoicing were each capable of adding maximally one unit of difference, and intermediate differences, including those indicated by diacritics, were interpolated. The differential weightings of the features are given implicitly by the difference in a feature’s extreme values. The decision to make advancement count more heavily than height is based on the idea that a change in vowel advancement marks a dialect speaker more saliently than a difference in vowel height, but we concede that this decision is moot. One vowel feature, “direction”, was not instantiated anywhere in the LAMSAS database, so it is omitted from Table 31.1, and Lowman and Kurath allow for six degrees of rounding (Kretzschmar et al. 1994: 119), which we simplified to five when we found only five levels distinguished in the data. The feature names reflect their normal phonetic (articulatory) interpretation. The stress which is marked on a syllable is interpreted as a property of the vowel, which is why it appears in Table 31.1. Vowels receive either stress, secondary stress, or no stress. Vowels were interpreted as voiced except when explicitly marked as voiceless, in which case they bore the feature [⫺ voice]. On rare occasions, Lowman added a diacritic indicating the pharyngealization of a vowel; where it occurs this is interpreted with the [v-pharyng.] feature. Vowels transcribed as superscripts (e.g., the latter halves of laxing diphthongs) are interpreted not by a [( super] feature, but rather through a weighting ⫺ the cost of comparisons involving superscripted vowels is only fifty percent of what it would have been had the segments not been superscripted. The idea, of course, is that such minor articulations contribute less to pronunciation difference. We calculate the distance between two vowels first by simply summing the differences of all the feature values, Σ f | fv ⫺ fv1 |. In order to emphasize the importance of slight

31. Measuring dialect differences Tab. 31.1: Vowel features used in LAMSAS together with the numerical interpretation needed to interpret feature differences Vowel feature

Possible values

v-advanced

⫺3, ⫺2, ⫺1, 0, 0.4, 1, 1.4, 2, 2.4, 3

v-high

⫺1.75, ⫺1.5, ⫺1.25, ⫺1, ⫺0.75, ⫺0.5, ⫺0.25, 0, 0.25, 0.5, 0.75, 1, 1.25, 1.5, 1.75

v-rounded

⫺1, ⫺0.5, 0, 0.5, 1

v-long

⫺0.5, 0, 0.5, 1

v-stress

0, 0.35, 0.7

v-nasal

0, 1

v-rhotic

0, 1

v-pharyng.

0, 1

v-voice

0, 1

differences as opposed to larger ones, we work with the logarithm of that sum, as introduced above. Finally, we wish to work with a scale with a genuine zero: d (v,v1) ⫽ log (1 ⫹ Σ f | fv ⫺ fv1 |) We applied the unigram Levenshtein model using the segment distances just sketched, without any word-length normalization. We compared the pronunciations of each pair of sites in Lowman’s data, using all of the words common to each pair. The result is a large pronunciation distance table, or matrix, providing a mean pronunciation distance for each pair of sites. As distances are symmetric, we ignore half of the cells in the table. We turn now to the results, beginning with the dialect areas which emerge from the pronunciation comparison. They are detected via hierarchical clustering (Nerbonne et al. 2008), a technique for recognizing groups in data, and are shown in Map 31.1. Map 31.1 shows the entire LAMSAS area, with the regions covered by fieldworkers other than Lowman shaded gray. Note that even though the clustering procedure is based solely on a matrix of pronunciation distances, it nonetheless detects geographically coherent areas, a first indication that the analysis is working. The dendrogram on the right shows that the major division in the data separates the two most southern subareas from the four northern ones, with the split running to the south of Virginia’s northern border; in the northern cluster, there is also a split running east⫺west through north Pennsylvania. This is a defensible subdivision of the LAMSAS speech area, even if it differs from Kurath’s (1949) assessment in failing to confirm his “midland” area, extending from western Pennsylvania southward along the Appalachian Mountains. Because the pronunciation distance is numerical, we can also apply multidimensional scaling to our results and attempt to view it in a more simplified form. The result is shown in Map 31.2. Heeringa (2004: chap. 6.2) explains MDS in more detail. We note only that MDS tries to place each site in a coordinate space of few dimensions. We use a three-dimensional solution which accounts for over ninety percent of the variation in the original distance table. If we map each of the dimensions in the three-dimensional solution to intensities of red, green, and blue, we obtain Map 31.2.

559

560

VI. Data analysis and the presentation of results

Map 31.1: Left, the dialect areas emerging from the aggregated pronunciation difference measurements and right, the dendrogram, providing a key (see text for discussion)

The MDS analysis is striking for the prominence it grants to southeast Pennsylvania. The explanation for the area’s unusual status is suggested by pronunciations such as [1tscr.tse] ‘Georgia’, where the devoiced syllable onsets (expected [D]) betray the influence of German on the speakers interviewed (speaker PA7C in Lancaster in fact spoke German as his first language). Nerbonne (2006) traces which linguistic features contribute the most to the aggregate dialectal differences shown in Map 31.2, but concentrates on the four southernmost areas shown in Map 31.1. We focus on the most important distinctions here. Using factor analysis it could be shown that the different reductions of unstressed vowels ([e] vs. [ = ], the latter including [I]) in the unstressed closed syllables of words such as closet, kitchen, and Baltimore (second syllable) were most strongly coherent across the collection sites. Closely aligned with this shift, we noted that the varieties that used the higher version of the reduced vowel ([ = ]) also fronted the [u] to [⁄] in words such as St. Louis and Tuesday, likewise fronting the onset of the diphthong in the second syllable in Missouri. Finally, the [c/a] distinction in hog(pen) and Florida aligned well with [e/= ] distinction. Note that only the [c/a] distinction is phonemic in some varieties of American English, while the other two are subphonemic. Nerbonne (2006) also examines several less prominent alternations, including (i) rhotic vs. non-rhotic pronunciations, where the latter varieties likewise demonstrate a lowering of [c] in words such as forty or storm; (ii) a contrast between raised and unraised [I] as the last syllable in Tuesday, foggy or thirty; (iii) another [e/= ] distinction, but this time in open syllables such as the final

31. Measuring dialect differences

Map 31.2: The three-dimensional MDS analysis of the pronunciation distances, better suited to portraying dialect continua (see text for discussion)

syllable of sofa, Georgia or Russia; (iv) the raising of [e], most extremely to [= ], e.g., in the first syllable of Tennessee (in general before [n]); and (v) fronted vs. non-fronted versions of the lax [w] in words such as wood and good. It is striking that the [a=/a] shibboleth was not among the most prominent distinctions. Kurath and McDavid (1961) also discuss each of the features. The aggregate, dialectometric perspective adds the opportunity to quantify the importance of features which Kurath and McDavid approached cartographically.

5. Evaluation In the foregoing, we have criticized the traditional method for having little to say about which features and which distributions are important. This deficit arises in traditional methodology because there are simply too many features and hence too many distribu-

561

562

VI. Data analysis and the presentation of results tions and isoglosses to choose between, leaving the method quite underdetermined. This problem also arises in the evaluation of older methodology. But it would be naı¨ve to think that we render ourselves immune to methodological woes simply by aggregating. We do indeed avoid the problem of which material to choose by obtaining a large, representative sample, preferably via (standard) random selection (Bolognesi and Heeringa 2005). But many options for measuring pronunciation differences remain (see sections 3.3 and 3.5). Defining segment distances alone involves distinguishing about twenty features and five or so values per feature. In addition we may process unigrams or bigrams, treat diphthongs and/or affricates as one segment or two, normalize based on word length or on segment or word frequency, aggregate feature differences via addition or via Euclidean distance, etc. The list is substantial, leading to the question: If we experiment extensively, and finally hit upon a pleasing analysis, have we attained substantial insights or have we struck lucky thanks to the plethora of analytical options? This is our evaluation problem. As we suggest by introducing the issue this way, we believe that the evaluation problem has always been present in dialectology, but it is more obviously present in complex systems involving measurements where many options are available to analysts. In the present discussion we only consider the problem of evaluating pronunciation measurements, and ignore issues concerning further analysis, in particular those involving clustering. Psychometrics divides these problems into two sub-issues, consistency and validity. Consistent measurements tend to provide the same information. We are encouraged when measurements continue to function well when used on new data, a simple sort of consistency check. Cronbach’s alpha is a more formal check on consistency (Nunnally 1978). We first measure the inter-item correlation, e.g., between the word dog and the word cat, by first obtaining two site ⫻ site tables of pronunciation distances, one for dog and another for cat. We then calculate the correlation between the two tables, measuring roughly the degree to which the two words provide the same indication of linguistic distance. We repeat this process for every pair of words in the sample (a time-consuming task), thereby obtaining the mean inter-item correlation. Figure 31.4 graphs Cronbach’s alpha as a function of r, the mean inter-item correlation coefficient, and n, the number of items. Given a mean inter-item correlation, Cronbach’s alpha indicates whether enough material is present in the sample. Nunnally (1978) suggests that 0.7 or 0.8 is a satisfactory level, and Cronbach’s alpha ⫽ 0.97 for the LAMSAS sample, indicating that the signal is very consistent, given the amount of material. We emphasize that Cronbach’s alpha depends on the data set being analyzed. In our experience thirty-word sets normally show levels of above 0.8, but new data, especially with low inter-item correlation, could differ. We say that a measure is valid if it measures what it purports to, and this leads immediately to more reflective questions about the goals of dialectology. We can claim to measure similarity in pronunciation, but similarity with respect to what? Our thinking on this point has evolved. Taking expert opinion to be well founded, we once calculated the degree to which our analyses coincided (Heeringa, Nerbonne and Kleiweg 2002). Although this is worthwhile, we would still, if we are ambitious, like to improve upon earlier expert opinion where possible. We now prefer to begin from the premise that one of the goals of dialectology is to characterize the signals of provenance normally present in speech. Naturally this process begins with the detection of those signals and proceeds

31. Measuring dialect differences

Fig. 31.4: Cronbach’s alpha is a function of the inter-item correlation r (left axis), which measures whether different items are giving the same indication, and n (right axis), the sample size. It shows whether samples are large enough to provide consistent signals

to an investigation of their structure. We speak of signals in order to emphasize that people should be able to receive them, and this leads immediately to the idea that we ought to validate our measurements by comparing them to dialect speakers’ judgments. Gooskens and Heeringa (2004) have developed this idea and rendered it operational. They analyzed Norwegian pronunciations along the lines sketched in section 3: and played recordings of the pronunciations to dialect speakers, who judged how similar the pronunciations were to their own. Gooskens and Heeringa then measured the correlation between the perceptual judgments and pronunciation distances (r ⫽ 0.7). We have also applied this method when comparing different versions of the pronunciation difference measure (Heeringa et al. 2006), and it seems the best way of independently validating the work. We should add that most analyses indicate that the various versions of the pronunciation measure do not differ significantly from one another when evaluated strictly. It is surprisingly difficult to demonstrate the superiority of the more discriminating systems.

6. Emerging issues Once we are satisfied that pronunciation distance measures are a valuable contribution to the dialectologist’s toolbox, several opportunities arise and several additional undertakings suggest themselves. Perhaps the most exciting opportunity is the chance to attempt a characterization of the general relation between geographic and linguistic distance. Se´guy (1971) has shown that dialectometric distances derived from French data grow sublinearly as a function of geography, a result which most studies to date have confirmed. Nerbonne and Heeringa (2007) replicate Se´guy’s finding using Dutch data and argue that this contradicts Trudgill’s (1974) well-known “gravity model”. The pres-

563

564

VI. Data analysis and the presentation of results ent overview is too brief to go into further detail, but our argument makes essential use of the measurement techniques presented here. The success of the technique suggests that it should be possible to detect recent borrowings, since such words show an unexpectedly low distance given overall varietal distances (van der Ark et al. 2007), and it is exciting to consider what other cultural markers might be studied quantitatively to explore the degree to which linguistic variation behaves like other cultural traits (Manni et al. 2008). Finally, we wish to discover how the different linguistic levels ⫺ lexical, phonological and syntactic ⫺ correlate as signals of provenance. There are also several points at which we feel the techniques presented here could be improved. It should be clear that we are not satisfied with the minor benefits that have been adduced for systems which take account of segment distances. Linguistically, we are quite certain that finer distinctions may be reliably drawn, and it is a puzzle as to why their benefit should be so hard to demonstrate. Our current hypothesis is that the high level of aggregation may be hiding the benefits. It would also be prudent to explore the relation between dialect perception and other aspects of phonetics, asking whether they share traits and, if so, which (Wieling and Nerbonne 2007). Finally, we look forward to the development of better techniques for identifying the major linguistic factors in aggregate comparison.

Reerences Bloomfield, Leonard 1933 Language. New York: Holt, Rinehart and Winston. Bolognesi, Roberto and Wilbert Heeringa 2005 Sardegna fra tante linge. Il contatto linguistico in Sardgna dal Medioevo a oggi. Cagliari: Condaghes. Chomsky, Noam and Morris Halle 1968 The Sound Pattern of English. New York: Harper and Row. Goebl, Hans 1984 Dialektometrische Studien: Anhand italoromanischer, rätoromanischer und galloromanischer Sprachmaterialien aus AIS und ALF. 3 vols. Tübingen: Niemeyer. Gooskens, Charlotte and Wilbert Heeringa 2004 Perceptive evaluation of Levenshtein dialect distance measurements using Norwegian dialect data. Language Variation and Change 16(3): 189⫺207. Gooskens, Charlotte and Wilbert Heeringa 2006 The relative contribution of pronunciation, lexical and prosodic differences to the perceived distances between Norwegian dialects. Literary and Linguistic Computing 21(4): 477⫺492. Gusfield, Dan 1999 Algorithms on strings, trees and sequences. Computational Science and Computational Biology. 2nd ed. Cambridge: Cambridge University Press. Heeringa, Wilbert 2004 Measuring dialect pronunciation difference using Levenshtein distance. Groningen: University Library Groningen. [PhD dissertation, University of Groningen. Available via printing on demand from .] Heeringa, Wilbert, John Nerbonne and Peter Kleiweg 2002 Validating dialect comparison methods. In: Wolfgang Gaul and Gerd Ritter (eds.), Classification, Automation, and New Media. Proceedings of the 24th Annual Conference of the Gesellschaft für Klassifikation, 445⫺452. Heidelberg: Springer.

31. Measuring dialect differences Heeringa, Wilbert, Peter Kleiweg, Charlotte Gooskens and John Nerbonne 2006 Evaluation of string distance algorithms for dialectology. In: John Nerbonne and Erhard Hinrichs (eds.), Linguistic Distances [Proceedings of the workshop held at ACL/COLING, Sydney 2006], 51⫺62. Shroudsburg, PA: Association for Computer Linguistics. Also available from . Herrgen, Joachim and Jürgen Erich Schmidt 1989 Dialektalitätsareale und Dialektabbau. In: Wolfgang Putschke, Werner Veith and Peter Wiesinger (eds.), Dialektgeographie und Dialektologie. Günter Bellman zum 60en Geburtstag von seinen Schülern und Freunden, 304⫺346. (Deutsche Dialektgeographie 90.) Marburg: Elwert. Johnson, Keith 2004 [1997] Acoustic and Auditory Phonetics. London: Blackwell. Kessler, Brett 1995 Computational dialectology in Irish Gaelic. In: European Chapter of the Association of Computational Linguistics, Seventh Conference of the European Chapter of the Association for Computational Linguistics, 60⫺67. Dublin: Morgan Kaufmann. Kondrak, Grzegorz 2002 Algorithms for language reconstruction. Unpublished PhD dissertation, University of Toronto. König, Werner 1994 [1978] dtv-Atlas zur deutschen Sprache. München: Deutscher Taschenbuchverlag. Kretzschmar, William Jr., Virginia G. McDavid, Theodore K. Lerud and Ellen Johnson 1994 Handbook of the Linguistic Atlas of the Middle and South Atlantic States. Chicago: University of Chicago Press. Kruskal, Joseph 1999 An overview of sequence comparison. In: Sankoff and Kruskal (eds.), 1⫺44. Kurath, Hans 1949 A Word Geography of the Eastern United States. Ann Arbor: University of Michigan Press. Kurath, Hans and Raven McDavid 1961 The Pronunciation of English in the Atlantic States: Based upon the Collections of the Linguistic Atlas of the Eastern United States. Ann Arbor: University of Michigan Press. Lameli, Alfred 2004 Standard und Substandard. Regionalismen im diachronen Längschnitt. (Zeitschrift für Dialektologie und Linguistik. Beiheft 128.) Wiesbaden: Steiner. Luce, Paul A. and David B. Pisoni 1998 Recognizing spoken words: The Neighborhood Activation Model. Ear and Hearing 19(1): 1⫺36. Manni, Franz, Wilbert Heeringa, Bruno Toupance and John Nerbonne 2008 Do surname differences mirror dialect variation? Human Biology 80(1): 41⫺64. Manning, Christopher D., Prabhakar Raghavan and Hinrich Schütze 2008 Introduction to Information Retrieval. Cambridge: Cambridge University Press. McMahon, April, Paul Heggarty, Robert McMahon and Warren Maguire 2007 The sound patterns of Englishes: Representing phonetic similarity. English Language and Linguistics 11(1): 113⫺142. Nerbonne, John 2006 Identifying linguistic structure in aggregate comparison. Literary and Linguistic Computing 21(4): 463⫺476. Nerbonne, John and Wilbert Heeringa 1997 Measuring dialect distance phonetically In: John Coleman (ed.), Workshop on Computational Phonology, 11⫺18. Madrid: Special Interest Group of the Association for Computational Linguistics.

565

566

VI. Data analysis and the presentation of results Nerbonne, John and Wilbert Heeringa 2007 Geographic distributions of linguistic variation reflect dynamics of differentiation. In: Sam Featherston and Wolfgang Sternefeld (eds.), Roots: Linguistics in Search of its Evidential Base, 267⫺297. Berlin: Mouton de Gruyter. Nerbonne, John, Wilbert Heeringa and Peter Kleiweg 1999 Edit distance and dialect proximity. In: Sankoff and Kruskal (eds.), v⫺xv. Nerbonne, John and Peter Kleiweg 2003 Lexical distance in LAMSAS. Computers and the Humanities 37(3): 339⫺357. Nerbonne, John and Peter Kleiweg 2007 Toward a dialectological yardstick. Journal of Quantitative Linguistics 14(2⫺3): 148⫺166. Nerbonne, John, Peter Kleiweg, Wilbert Heeringa and Franz Manni 2008 Projecting dialect differences to geography: Bootstrap clustering vs. noisy clustering. In: Christine Preisach, Lars Schmidt-Thieme, Hans Burkhardt and Reinhold Decker (eds.), Data Analysis, Machine Learning, and Applications. Proceedings of the 31st Annual Meeting of the German Classification Society, 647⫺654. (Studies in Classification, Data Analysis, and Knowledge Organization). Berlin: Springer. Nunnally, Jum C. 1978 Psychometric Theory. New York: McGraw-Hill. Osenova, Petya, Erhard Hinrichs and John Nerbonne (eds.) 2007 Proceedings of the RANLP [Recent Advances in Natural Language Processing] Workshop on Computational Phonology. Borovetz: RANLP. Prokic´, Jelena 2007 Identifying linguistic structure in a quantitative analysis of dialect pronunciation. Proceedings of the ACL 2007 Student Research Workshop, Prague, 61⫺66. Prague: The Association for Computer Linguistics. Available from: . Sankoff, David and Joseph Kruskal (eds.) 1999[1983] Time Warps, String Edits and Macromolecules: The Theory and Practice of Sequence Comparison. Stanford: CSLI. Se´guy, Jean 1971 La relation entre la distance spatiale et la distance lexicale. Revue de Linguistique Romane 35: 335⫺357. Se´guy, Jean 1973 La dialectome´trie dans l’Atlas linguistique de Gascogne. Revue de Linguistique Romane 37: 1⫺24. Stevens, S. Smith 1975 Psychophysics: Introduction to its Perceptual, Neural and Social Prospects. New York: John Wiley. Trudgill, Peter 1974 Linguistic change and diffusion: Description and explanation in sociolinguistic dialect geography. Language in Society 2: 215⫺246. van der Ark, Rene´, Philippe Mennecier, John Nerbonne, Lolke van der Veen and Franz Manni 2007 Preliminary identification of language groups and loan words in Central Asia. In: Osenova, Hinrichs and Nerbonne (eds.), 13⫺20. Vieregge, Wilhelm H., Antonius C. M. Rietveld, and C. I. E. Jansen 1984 A distinctive feature based system for the evaluation of segmental transcription in Dutch. In Van den Broecke, M. P. R. and Cohen, A., editors, Proceedings of the 10th International Congress of Phonetic Sciences, pages 654⫺659, Dordrecht: Foris. Wagner, Robert A. 1999 On the complexity of the string-to-string correction problem. In: Sankoff and Kruskal (eds.), 215⫺235.

32. Linguistic atlases

567

Wieling, Martijn and John Nerbonne 2007 Dialect pronunciation comparison and spoken word recognition In: Osenova, Hinrichs and Nerbonne (eds.), 71⫺78.

John Nerbonne, Groningen (The Netherlands) Wilbert Heeringa, Amsterdam (The Netherlands)

32. Linguistic atlases - traditional and modern 1. 2. 3. 4. 5.

Introduction Preliminary events Methodological grounding Further development References

1. Introduction Without doubt, our understanding of human language and communication benefits from the geographical approach, especially the knowledge organized via maps and atlases. Given that institutes dedicated to the production of cartographic compendia have in the past been seen as linguistic laboratories (Stanforth 1971: 12), it should be clear that, in addition to recent developments, the traditions in linguistic cartography are well worth investigating. A resurgence in international atlas projects over recent decades has led to a strong (and continuing) interest in cartographic methodology. Above all, the computational handling of maps and atlases needs to be seen as a current focus of attention. But, due to the reassessments of historical data necessary for the analysis of language change, there is also a strong interest in the history of language cartography. While the first focal point is widely described in both papers and manuals, to date there has been no comprehensive cross-philological overview of the traditions in language mapping, especially one dedicated to the development of linguistic atlases (approaches are presented in, e.g., Thun 2000; Kirk 2001; Veith 2006; comprehensive but out of date is Pop 1950). This is unfortunate, considering that every cartographic work has a unique subjective dimension that is bound up with, first, the decisions of the individual authors and, second, the particular context of origin. Understanding how maps and atlases are organized is thus a precondition for their correct interpretation. The production of linguistic maps became a relevant linguistic field in the nineteenth century, at which time we find an abrupt rise in the amount of cartographic work, making it possible to speak of the origin of the cartographic paradigm in linguistics. Right from the outset, atlases, not just maps, were seen as a productive tool in the pursuit of geolinguistic interests. The extent to which this was a part of a more general zeitgeist becomes clear when one reflects that geographers were also publishing the first thematic atlases at this time. Nevertheless ⫺ as will be shown ⫺ an atlas is not necessarily the

568

VI. Data analysis and the presentation of results result of cartographic impulses. In these early times we find so-called “atlases” that were based on spatial but not cartographic information; at first approximation an atlas should thus be understood as a phenomena-based compendium of spatial data of (at best) empirical provenience. This is unusual, inasmuch as an atlas is currently defined as a systematic collection of maps (cf., e.g., Bollmann and Koch 2001: 39). But we also find similar approaches today, as with the modern linguistic atlases that are occasionally organized without any maps. This is related to the existence of different types of atlas organization and different historical stages of atlas production. But it also shows that there is to date still a problem in defining a linguistic atlas. This article meets the challenge by arguing for a more comprehensive understanding of the initial phase of atlas cartography than has previously been adopted. Taking methodological progress into account, it becomes obvious ⫺ at least in the cases discussed ⫺ that there are stronger similarities of intent between the different approaches followed within map and atlas-based projects than has previously been conceded. Accordingly, section 2 highlights both the growing interest in atlases and the very earliest approaches which led to the creation of phenomenal maps. In addition, section 3 describes the development of the first linguistic atlases in Germany, Switzerland and France in the late nineteenth and early twentieth centuries. Taking these together with the work of their immediate successors, it becomes apparent that the most relevant issues in atlas cartography had already been worked through by the first half of the twentieth century, be they thematic structuring, onomasiological issues, the principles of map and data interpretation, the integration of extralinguistic factors, etc. Alongside this, section 4 shows that current atlases have set a new focus, directed not only at individual varieties but also at the communicative practices of their speakers. In this regard, the pluridimensional atlases are a promising tool for a more comprehensive interpretation of language in space. An expanded data basis is also emphasized in the design of current information systems, which extend well beyond traditional atlas design. The computational aspects involved in such systems are also discussed in section 4. All together, it will be shown that there are chains of strong (mostly) personal connections which constitute lineages of tradition of which linguists are not always fully aware. But, more importantly, the main objective of this contribution is to demonstrate the progress, both ideological and methodological, which connects individual projects more closely than has previously been realized. Of course, it is not possible to offer a satisfying systematization of both chronology and methodology within the confines of a short essay, especially given that the current state of discussion is relatively meager. This contribution therefore highlights important aspects insofar as they are necessary to understand the central theme. For a more detailed treatment of the whole topic, readers are referred to the Language Mapping handbook in this series (Lameli, Kehrein and Rabanus, to appear). All of the maps and atlases mentioned in the explanations which follow have been included there (in the second volume) as examples.

2. Preliminary events 2.1. Linguistic maps Cartography as a method for the description and explanation of spatial relations has a tradition that can be traced back to ancient times. Its subject is the illustration of a ⫺

32. Linguistic atlases however it may be defined ⫺ geographic reality on a two-dimensional plane. Important epochs can be defined in terms of both the different stages in human conceptualization of the planet and the development of geometric techniques. Rather late in its history, a correlative approach that visually combines the terrestrial dimension with specific natural or social phenomena becomes apparent. Aside from ancient cartographic descriptions of space as affected by man (e.g., the Tabula Peutingeriana [c. 350 b.c.], or the so-called mappae mundi that represent the Christian view of the world during the Middle Ages, e.g., Erhard Etzlaub’s Romweg Map of 1500), an explicitly thematic cartography does not clearly evolve until the eighteenth century. Then, however, we find a diversity of analytical approaches, in many cases based on quantitative measurements, as represented by Edmont Halley’s isogonic Map (1701), which is commonly seen as one of the first truly thematic maps, if not the first. It is no coincidence that this map was intended for practical, i. e., naval purposes, namely calculation of the variation in the magnetic field around the Americas: maps confer practical benefits and are therefore used as instruments. But human spatial phenomena soon became more relevant for cartographers, such as, for instance, demographic questions arising from emergent statistics, and maps thus began to help interpret the conditions of human life. This also includes the spatial organization of language. However, it was not linguists but geographers who were the first scholars to demonstrate the regional distribution of languages or language phenomena (e.g., ten Kate 1723). It was Gottfried Hensel (1741) who produced early maps of linguistic variation that show, for instance, the different realizations of the Lord’s Prayer in Europe or other details of the world’s languages. Although the focus so far has been on thematic cartography, it should not be forgotten that evidence of linguistic particularities is already found in the earliest geographical compendia. Thus, for example, in the commentary accompanying the map collection which forms part of his Cosmographiae Universalis, the German geographer Sebastian Münster mentions differences within the region of Silesia that make his essentially topographic map linguistically productive. He identifies the Oder River as a language border: Utitur pro maiori parte lingua germanica. Ultra Oderam vero loquuntur Sclavonice, quae lingua communis est Bohemis, Polonis, Sclavis, id est, Illyrijs, Moscovitis, Lithvanis et multis alijs populis. ‘In most parts the German language is used. On the other side of the Oder, however, Slavonic is spoken, which is the common language of Bohemians, Poles, Slavs, that is, Illyrians, Muscovites, Lithuanians and many other peoples.’ (Münster 1572 [1544]: Liber I, map 14 “slesiae descriptio”; my translation)

Nevertheless, an enduring desire for knowledge, including the investigation of language, first evolved during the Enlightenment in Europe. The Enlightenment focus on the human individual made both the subjectivity of the individual and the subjectivity of perception topics of scholarly discourse. This interest in the res humana led to a particular tenor of observation that also implicated language. Important consequences were the first empirical scientific attempts at collecting linguistic data within geography, thus helping to negotiate the theoretical-cum-reflective attitude that can be traced back to the early modern period. An empirical interest in the spatial dimension of language first appeared in nations with a marked linguistic heterogeneity, such as Germany, Italy, Russia or Switzerland, even if the actual aim of the early geographically based studies was not the production of maps, but rather the publication of the established linguistic

569

570

VI. Data analysis and the presentation of results instrument of a lexicon (cf. Freudenberg 1965). Undoubtedly, the indirect global exploration of mass linguistic data commissioned by Catharine II of Russia is one of the most impressive examples. It led to the Linguarum Totius Orbis Vocabularia comparativa, realized by Simon Pallas (1977⫺1978 [1786⫺1789]). It did not take long for the first linguistic atlases to emerge from such undertakings, especially taking into account that the tradition of thematic atlases began in the same period, with Carl Ritter’s Sechs Karten von Europa (1806). According to Pop (1950: 18), a French sound atlas was being planned within the Societe´ Royale des Antiquaires as early as 1823, although it was never completed. But in the same year, 1823, the first explicitly labeled (in the subtitle) Sprachatlas was published by Julius Klaproth. Its actual name is Asia Polyglotta, in reference to the multilingual relations in Asia which attracted scientific interest. From a current perspective, this work is anything but typical. In particular, it features lists of words ordered into tables according to certain language types, which, where necessary, are subdivided into individual roots (Stämme). At the end of the volume the reader finds a map where the linguistic information is put into geographical relation. However, on his map, Klaproth explicitly refers not to the listed languages but to their speakers (“Georgier, Kaukasier, Samojeden, Jeniseier, Ost-Finnen”, etc.). In so doing, Klaproth employs a relevant ethnological argument, which may well have been of only minor importance to him. The argument reifies apparent parallels between languages and ethnic groups, e.g., the depicted linguistic proximity of peoples in “Hindustan” and “Persien”. In this regard, the map contains additional information and also acknowledges the fact that language is dependent upon human carriers and does not exist autonomously. Further, it includes topographical details. Hence, the reader sees that the border between the “Indo-Germanen” and the “Tübeter” coincides with the “Himalaja”, for instance. Thanks to the visual impressions it generates, which simultaneously abbreviate and broaden the linguistic information on offer, the map also serves as a potential starting point for linguistic deliberations. In fact, this proves helpful, since some groups of speakers, such as the Koreans or the Japanese, are not described in the preceding tables. But such rare cases aside, the map is also to be understood as a “register” which refers to the contents of the table and helps to interpret the regional picture in more depth. This methodology is already identifiable in the early work of ten Kate (1723), which features a classification of European languages, together with textual explanations that also refer to the variation of specific phenomena. But above all, Klaproth’s methodology is somewhat reminiscent of that of Johann Andreas Schmeller (1821), even if Schmeller’s single map is not as elaborately illustrated, nor his work explicitly designated an atlas. Nonetheless, the organizational parallels to Klaproth alone mandate mentioning it in this context. Schmeller’s objective was an exhaustive description of the dialects of the Kingdom of Bavaria via a comprehensive set of directly collected data. In order to achieve a geographical overview, he designed a coded map that enables the identification of individual language phenomena listed in a particular chapter of his monograph (1821: 427⫺432). Each code represents a dialect grouping, and the map is divided by lines into six individual groups at the first hierarchical level. The starting points for the areal structuration are primarily rivers, after which the dialect groups are named. Once again, based on the chapter about these dialect groupings, further information can be found in other parts of the book via an identifying number (1821: 31⫺426), where the reader finds a phonetic description of individual dialect phenomena. In such a way, Schmeller

32. Linguistic atlases develops a distinct dialect typology. Here too, the map functions like a register that gives access to linguistic information while at the same time also providing a spatial impression of the variability of language. Against this background, the beginnings of language cartography can be seen to be characterized by rather similar approaches, with no single clear criterion for the use of the atlas label ⫺ which should not really be surprising, given that a tradition in thematic cartography had not yet been established. In contrast, it becomes quite clear that the terminology of the time is imprecise when we consider Adrien (or Adriano) Balbi’s Atlas ethnographique du globe (1826), which, despite being called “atlas”, contains no cartographic material. What makes Balbi (who was a cartographer with an interest in ethnology, not a linguist like Klaproth) important from a linguistic point of view is the fact that he aimed to define human space, i. e., relations between peoples, in terms of the differences between the languages spoken by them (cf. the subtitle of his atlas: ou Classification des peuples anciens et modernes, d’apre`s leurs langues, pre´ce´de´ d’un discours sur l’utilite´ et l’importance de l’e´tude des langues applique´es a` plusieurs branches des connaissances humaines). As a consequence, his work is in effect a classification of languages, or rather a classification of groups of speakers. In terms of the notion of maps and atlases within the linguistic context, it is important to note that Balbi included in his work a so-called “mappemonde” that is nothing more than a “tableau ethnographique” and thus not a cartographic work. At the end of his volume, the reader finds translations of individual words into the languages of the world as classified therein. It should be pointed out that the entire work is explicitly located within the context of Catherine II’s project mentioned above; such early authors were obviously quite close to the lexicon projects of the time. For authors like Klaproth and Balbi, the most important defining criterion for an atlas was a reference to variability in space, not the presence of cartographic material. Consequently, their work occupies the middle ground between lexica and more modern atlases. One important difference between phenomena-based projects (like Klaproth’s atlas) and cross-linguistic comparative lexica (like Catherine’s project) lies in their macro and micro-structural organization. But possibly the most important difference is that maps and atlases aim to achieve a particular spatial structuring of language, while this is usually a precondition for lexica, not their objective.

2.2. Collections o maps If, however, one defines an atlas from a contemporary perspective as a “systematic collection of maps in book format or as loose sheets for a specific goal and purpose” (Bollmann and Koch 2001: 39; my translation), then the initial phase of actual linguistic atlas cartography must be dated a little more recently. The most important starting point from this perspective is Bernardino Biondelli’s Atlante Linguistico d’Europa of 1841. Here too, the focus is on the geographical relations between languages, or rather language families. Specific linguistic phenomena are also excluded from this work. The visual means of argumentation are color and lines, techniques it shares with virtually every thematic map from earlier periods. Above all, the use of (colored) lines proved to be a useful structuring method for evoking a spatial impression. When one considers, for instance, Bartolomeo Scultetus’ Lusatia Superior map (1593) which displays the regional

571

572

VI. Data analysis and the presentation of results difference between “deutsch” and “wendisch” (‘German’ vs. ‘Sorbian’) using colored lines, it can be claimed that this method has also been used for linguistic purposes from the very outset of modern cartography (cf. Thun 2000: 70⫺71). In Biondelli’s case, the integration of further information via specific explanation is also important. In contrast to Klaproth, Biondelli’s individual maps are supplemented by additional information in the form of texts, not tables, albeit not as voluminous as Schmeller’s comments and lacking Klaproth’s detail about concrete realizations. It is nevertheless noteworthy that they convey, for instance, regional subgroupings or historical information not deducible from the map. It thus becomes obvious that, from the authors’ perspective, the map as such is not self-explanatory. In contrast, these authors feel an unquestionable need for commentary. This too is an element familiar from the first non-linguistic map collections, such as Sebastian Münster and his explanation of the Silesian map quoted above. Thus, bearing in mind that the initial nineteenth-century approaches shared an emphasis on the textual dimension and that the cartographic aspects gradually became more important, we can identify a line of development in the constitution of linguistic atlases, which ⫺ from a latter-day perspective ⫺ can be captured by at least two definitions distinguished from one another by a methodological advance. The first definition, which characterizes the first period up to and including the work of Balbi, conceives a linguistic atlas to be (1) a (non-lexicographic) collection of language data based on regional variation. However, considering that Schmeller, whose work fits within this first definition, did not make use of the term atlas, it becomes obvious that the actual nature of language atlases was in practice rather unclear and dependent on individual aspects. The second definition ⫺ which characterizes the later period starting more or less with Biondelli (Freudenberg [1965: 172] discusses an earlier unpublished project) ⫺ conceives of a linguistic atlas as (2) a collection of thematic maps showing the regional distribution of language. What holds for both definitions/periods is, firstly, that the mapping techniques are linked to traditions in cartographic visualization well known to the cartographically adept authors. Secondly, with the inclusion of textual elements we find the use of techniques that transcend the illustration as such. At the same time, it is also obvious that linguistic atlases are from the outset explicitly bound to the concrete phenomena of spoken language (i. e., words), be it via maps or texts. If the comments are seen as an extension of the actual map information ⫺ and contrary to usual practice they should be ⫺ then there can be no doubt that the mapped registers included in the work of Klaproth or Schmeller ⫺ alongside earlier maps like Hensel’s, that remained unknown to nineteenthcentury linguists ⫺ are the direct forerunners of later phenomena-based maps. That is, the distinction between maps of a language (Sprachenkarten), which detail the distribution of a language, and maps of linguistic phenomena (Sprachkarten), which describe concrete features (cf. Thun 2000: 71⫺73), is a theoretical simplification, although it is of practical benefit. For most of the authors mentioned, even if they produced maps of languages, the inherent information was phenomenal in nature. Of Schmeller at least, we know that he revised his early map to create a cartographic overview of linguistic

32. Linguistic atlases variants in the course of a later, commissioned work (Schmeller 1956 [1846]: 419). This interest in phenomena will be explored further in the following section, where the notion of “atlas” is that of the second definition.

2.3. The problem o phenomena As shown, at the beginnings of linguistic cartography the kind of information that should properly form part of an atlas or a map seems to have been rather unclear. The simplest approach to the mapping of language is ⫺ at first glance ⫺ to show the regional distribution of a language as such. But given that ethnological constraints do not apply, this is only possible if comparable data that help to define a particular geographical distribution are available. This makes the linguistic-mapping approach exceptionally complicated. Even if scholars set out to create a phenomenal overview, the data ⫺ if available ⫺ is so voluminous that (the problem of transcription aside) it is not possible to present it all on one page. Scholars therefore attempt to classify their data. The register maps demonstrate that cartography was at first blush seen as a means of presenting such a linguistic classification. This makes it clear that, even if the dependent variable of the geographical approach was “linguistic item (sound, word, etc.)”, the dependent variable of the cartographic approach was, for methodological reasons, “language”. Nevertheless, the linguistic items were connected to the maps, via the textual explanations. The general problem can best be understood by looking at other topics of thematic cartography. The aim of linguistic cartography is not substantially different from the objective of mapping “weather”, for instance. However, even the early cartographers (cf., e.g., Halley’s weather map from 1686) define weather via particular attributes such as barometric measures, barometric changes, wind, temperature, and so on. These attributes can then be individually presented on a single map. In linguistics, however, the definiens is often much more extensive and, moreover, there are fewer or no analytical instruments available that provide an intersubjective definition and classification of language “attributes”. It took some time before scholars like Schmeller managed to abandon their comprehensive approach and identified the possibility of using maps for the documentation of individual phenomena. When they did, the dependent variable was finally defined as “linguistic item” in linguistic cartography as well, and it made sense to assemble such individual maps into an atlas. As this orientation took hold in linguistics, the full potential of linguistic cartography was finally tapped.

2.4. A look at the zeitgeist Finally, the scholarly resonance of this early stage needs to be addressed. Some of the projects mentioned are not well known to contemporary scholars; similarly, the prescientific maps of Hensel, ten Kate or Scultetus were also probably not known to the geolinguistic pioneers of the nineteenth century. But from 1800 on, most contemporary works in emergent linguistics rapidly became widely known after publication, at least within their individual philological traditions (although some occasionally sank rapidly into philological oblivion). The following anecdote gives a good impression of the situa-

573

574

VI. Data analysis and the presentation of results tion at the time. In the Bavarian Academy of Science’s journal, the Gelehrte Anzeigen, an anonymous review of Biondelli’s atlas was published in 1842. With a vague reference to Balbi and following a detailed description of Biondelli’s work, the enthusiastic writer makes it clear that linguistic cartography is only in the preliminary stages of development and that he therefore hopes Biondelli’s atlas will act as an “effective admonition” (“wirksame Mahnung”; Anonymous 1842: 80) to go further. His enthusiasm is inspired not least by Biondelli’s announcement of future volumes dedicated to the European languages and dialects. Biondelli’s atlas is thus seen as an “excellent first attempt” (“erster großartiger Versuch”; Anonymous 1842: 80) on the way to an exhaustive description of geographical variation in language. Two years later, in 1844, the same anonymous reviewer compliments Sˇafarˇik and Bernhardi on their individual non-phenomenal language maps of Slavia (1842) and Germany (1843), respectively. But at the same time the reviewer points out that, in future, the main effort of linguistics should be pitched at the level of phenomena, not that of individual languages, which are heterogeneous sets of varieties and variants (Anonymous 1844: 573). In the reviewer’s opinion, these three authors ⫺ Biondelli, Sˇafarˇik and Bernhardi ⫺ are nevertheless the founders of an explicitly linguistic cartography. This is remarkable in light of the identity of the reviewer in question: it is none other than the pioneer of empirically based (Bavarian) dialectology named above, Johann Andreas Schmeller, as becomes apparent from his voluminous diaries (Schmeller 1956 [1842]: 328). There is no mention of Klaproth in these, so it is possible (albeit unlikely) that Schmeller had failed to notice the Asian work. But most striking of all is the fact that Schmeller ⫺ in contrast to virtually every linguist since ⫺ obviously does not regard his own map from 1821 as one of the first dialect maps, a fact which serves to underline its abovementioned “register” function.

3. Methodological grounding Linguistic atlases that are strictly confined to the documentation of linguistic phenomena have so far been portrayed as the result of developments in the history of language geography, and it is the German Georg Wenker and the Swiss Jules Gillie´ron who independently realized such works in their earliest attempts in 1878 and 1880. Their projects are commonly seen as linguistic milestones and hence a large volume of literature on the basic principles guiding their work has been generated, so that little needs to be said here (for further information cf. Pop 1950; Werlen 1996: 436⫺444; Chambers and Trudgill 1998: 13⫺31). What makes their atlases interesting in the given context is the fact that they ⫺ not just metaphorically ⫺ enter uncharted territory and thereby set a new focus in geolinguistics. But conceptually the authors are not that far apart either; recall that both started out with the documentation of relatively small areas and ended up working with nations. Naturally, there were other highly creative scholars who addressed the phenomenal dimension of linguistic atlases, e.g., Fischer (1895). But together with their immediate successors, Wenker and Gillie´ron developed the basic principles of language cartography that remain indispensable to this day. This is due not least to the sheer volume of their work and the central role it has played within the linguistic discourse of their time and since. This section therefore sets out to explore the underlying lines of their work which place them directly within the above-mentioned tradition.

32. Linguistic atlases

3.1. Georg Wenker Georg Wenker is renowned for his exploration of German dialects using a questionnaire of 40 standardized sentences that he distributed to German schools by mail. There, the questionnaire was translated into the individual dialects by pupils or teachers. Proceeding thus, Wenker collected information from over 45,000 locations. In a second step, this data was plotted by hand onto over 1,500 maps. His mapping procedure was refined in several stages within the project, in line with overall progress in linguistic cartography (cf. Knoop, Putschke and Wiegand 1982; Herrgen 2001). In light of the above considerations (section 2.3), it is not surprising that Wenker’s first cartographic work was a classification of the dialects around his homeland (Wenker 1877). In separating regions via borders ⫺ e.g., the “Grenze von Benrath” ‘border of Benrath’ ⫺ which he refers to in the text as lines ⫺ e.g., the “Benrath line” (Benrather Linie) ⫺ he employs explicitly cartographic arguments, thereby exhibiting a stronger emphasis on cartography in linguistics than did his forerunners. This also becomes apparent when we consider that Wenker used a geographically detailed and coherent base map. It is noteworthy that with this map Wenker also aimed to develop a classification dependent on various dialect group hierarchies. As with Schmeller (whose work was known to him), we thus find a cartographic structuring of language in space. But, in contrast to Schmeller, Wenker’s map does not directly refer to textual elements, so that to a certain degree the map stands alone and can only be interpreted by reading the text before looking at the map. It should be kept in mind that Wenker’s small volume is explicitly written for non-linguists (i. e., his informants) and thus addresses a different audience to Schmeller’s monograph. Things changed with Wenker’s next undertaking. His Sprachatlas der Rheinprovinz nördlich der Mosel und des Kreises Siegen (1878) contains 25 printed base maps with colors that show the distribution of dialect sounds and forms added by hand. Developed as a demonstration of the validity of the methodology, it never reached the international book stores (only two copies existed). Nevertheless it has official status, since, as becomes clear from Wenker’s letters, it was prepared for and considered by the German Akademie der Wissenschaften in the context of a proposal for further project funding. Indeed, this is the first real language atlas in the sense of a collection of phenomenabased maps. His methodological approach was to combine on one map several items that ⫺ from Wenker’s point of view ⫺ stood in a certain linguistic relation to one another or showed similar regional distributions. Larger texts were obsolete and further information could be integrated into a corner of the map as a comment. An onomastic map and a topographic map complete the volume and offer further information to assist interpretation. But, more importantly, areas in which a specific word-bound element (sound, syllable, etc.) is realized in similar ways are indicated using contour lines. It is possible to speak here of isoglosses (Bielenstein 1892). The design of the maps is highly suggestive in the sense that Wenker shows correlations between different linguistic phenomena. Moreover, he demonstrates both the enormous heterogeneity of language and its underlying structure. This is an important characteristic of his early work since, in contrast to his later approaches, Wenker here presents a spatial analysis. By presenting his linguistic results, Wenker underlines the linguistic potential of his method. His subsequent expansion of the project to northern Germany, published in part as the Sprach-Atlas von Nord- und Mitteldeutschland (1881), has a similar linguistic objec-

575

576

VI. Data analysis and the presentation of results tive, even if the cartography has changed. It is less a presentation of data than a collection of interpretations (“Wiedergabe der Ergebnisse”, Wenker 1881: XI). Wenker’s work now became known to the academic world, where it was immediately welcomed with “unclouded pleasure, even with enthusiasm” (“mit ungetrübter Freude, ja mit Begeisterung”; Behaghel 1881: 434). Internationally, the phonetician Ellis considered the project “the greatest, the best-designed, and the best-executed attempt hitherto made to determine the peculiarities of local speech” (Ellis 1882: 30). Finally, a scientific gap seemed to have been partly filled, at least in Germany. In the final step, which led to the Sprachatlas des deutschen Reichs (1889⫺1923; cf. ), Wenker’s principle objective and methodology shifted fundamentally once again. In his earlier approach, every map displayed a relatively homogenous distribution of language. Many exceptions to the main spatial distribution were eliminated, which is why every map needs to be understood as a finished interpretation about linguistic space. The Sprachatlas des deutschen Reichs (along with its small, relatively unknown offshoot, the Pronomina in Nordwestdeutschland, Wenker 1886) is new in the sense that the maps explicitly show every detail of the data captured ⫺ and that for over 45,000 locations. This innovation entailed a stark change in the status of the maps. If they had previously been considered as analytical output, they were now seen as the instruments of future research. The increased diligence may also be due to the fact that the data transcribed were from laymen and thus required careful interpretation. Nevertheless, the linguistic potential of this approach is enormous, and it led to the recognition that, given the existence of sound laws, linguistic features ⫺ following Wenker ⫺ have their own unique spatial distribution, which can vary word for word. This atlas is undoubtedly remarkable, from both a cartographic and a linguistic point of view. Each map that had been finished was officially sent to the Berlin State Library, where it was intended to be publicly available. Further, the maps were accompanied by commentary in the form of individual booklets (Wenker, to appear), which set out to document everything that might be of interest for further analysis. All the first drafts remained in Marburg, where Wenker worked. In the period which followed, this Marburg copy was virtually the only one in use (the commentaries remained unknown). What never became widely known is that the Berlin copy was created as a new type of atlas (cf. Lameli 2008a). The visitor to the library finds a paper map depicting monosyllabic words or root syllables. The word endings (e.g., sounds, suffixes) are mapped out on transparent paper. It thus becomes possible to overlay stem and suffix and to analyze the morphological distribution of related word components, for instance. Where linguistic atlases normally document paradigmatic issues, in this special case ⫺ by looking at different elements of form at the same time ⫺ the user can examine syntagmatic aspects. This technique further underlines the strongly word-based orientation. Wenker had already experimented with transparencies in the Sprach-Atlas von Nord- und Mitteldeutschland (Wenker 1881: XII), making a fundamental characteristic of atlas cartography obvious in the process: comparability.

3.2. Jules Gillie´ron The importance of comparability can also be found in the work of the Swiss linguist Gillie´ron. His first atlas from 1880 is known as the Petit Atlas phone´tique du Valais roman (sud du Rhoˆne). Gillie´ron gathered the necessary data directly during a hike through the

32. Linguistic atlases southern Rhone valley, a region in the contact zone between Switzerland, Italy and Savoy where Romance varieties are spoken. The transcriptions are phonetically detailed and systemized on maps. This fundamental methodological distinction between Gillie´ron (direct data collection) and Wenker (indirect data collection) is a stereotype in the history of linguistics. But in light of the historical period, it is also true that the two scholars were quite close in terms of their linguistic orientation. Gillie´ron emphasizes that his work stands in the tradition of his teacher, Paul Meyer. Quoting Meyer, Gillie´ron underlines to the goal of the atlas right at the opening of his introduction: “Faire en quelque sorte la ge´ographie des caracte`res dialecteaux bien plus que celle des dialectes” (Gillie´ron [1880]: 7). This makes clear what Schmeller had already pointed out (cf. section 2.4). Because of a focus on the regularities of language, in this epoch only microlinguistic features like sounds seemed usable to obtain deeper insights. But this atlas is also interesting from a cartographic point of view. The main method is based on color symbols. Different linguistic types are signified by different colors on a rather schematic map that features the national borders, the Rhone River and some villages. A particular location where a specific type was found is underlined in the appropriate color. Where necessary, the lines are broken or dotted, so that it becomes possible to multiply the number of symbols in use. This method was complicated by the fact that for financial reasons it was not possible to print the colors. Hence, like Wenker, who drafted onto preprinted base maps, Gillie´ron was forced to draw the lines into every printed booklet by hand (Gillie´ron [1880]: 7). Whether this work was really the first printed atlas is thus open to interpretation (cf., e.g., Thun 2000: 76). The principle of documentation (not analysis) aside, what is most striking ⫺ and this again brings Gillie´ron ideologically close to Wenker ⫺ is the need for extralinguistic information. Gillie´ron includes a wealth of information concerning topography, population, language history, unmapped linguistic items and ⫺ not least ⫺ linguistic similarities between the surveyed locations. In so doing, Gillie´ron refers to the high degree of similarity between the dialects. Even if it was not explicitly so intended, this practice is typologically useful. This is remarkable since it is not a common topic in modern atlas cartography but is nevertheless related to later dialectometry to some extent (cf. Goebl 1984). In this regard, Gillie´ron’s work illustrates the desire for a comprehensive and integrative description as well as the creativity of this early period. Even if the influence of this early work has been rather selective, the next stage in Romance dialect geography, Gillie´ron’s and Edmont’s Atlas linguistique de la France (ALF, 1902⫺1914), needs to be seen as the foundation of a linguistic tradition that made another important contribution to linguistic atlas cartography. The main objective of the ALF was to record the diatopic heterogeneity of the dialects in France (cf. Thun 2000: 76⫺80). According to Werlen (1996: 442), it should also be understood as a response to Wenker’s rather workaday methodology. In contrast to Wenker (whose work was known to Gillie´ron), the method of data collection was a direct one: an investigator (Edmond Edmont) traveled all over France making narrow phonetic transcriptions in 639 locations. The instrumental character of the whole work is underlined by the cartographic organization. Each map shows the original transcripts for a specific linguistic variable. The map user’s first impression is that of a “spatial tableau”, i. e., topologically ordered words on a page. A more detailed spatial structuring, like that which Wenker intended, was not sought, which is why the ALF was able to rigorously pursue the documentation principle. In addition, a separate volume includes demographic data on

577

578

VI. Data analysis and the presentation of results the informants and further comments on the transcriptions. It is highly interesting that the authors went as far as possible to provide all information about their material. This is true not only of the very detailed (in comparison to Wenker) description of the informants (e.g., profession, age, sex). The user also finds information on local variation in the ALF maps, in some cases information on the social differentiation of the variants used in a particular location, specified via comments like “chez les jeunes”, “chez certains individus”, “par les vieillards” (Gillie´ron and Edmont 1902: 21). At this early stage, we thus already find a clear consideration of pluridimensional (i. e., social) aspects, which, as an aside, is also evident in the work of Wenker when he compares the language use of older and younger people using a language map in 1889 (cf. Lameli 2008b). With regard to cartography, it also needs to be mentioned that by this stage a more precise base map was in use, showing the departments of France as well as the individual locations. The authors abandon the use of both colors and isoglosses. The latter might be attributable to the selection of only a small subset of potential locations, something the editors themselves saw as a “faible partie” (‘weak point’, Gillie´ron and Edmont 1902: 3). But since, in contrast to Wenker, a phonetic transcription system was used, the documentation is as objective as was possible at the time. The user thus finds a sound data basis for linguistic analysis. From a cartographic perspective, it is also remarkable that the authors combined various linguistic items onto one map. An example is Map 44 from the first ALF volume, where two realizations of the phrase cette anne´e in different sentence positions (Cette anne´e, il y a eu beaucoup de fruit vs. La chaleur a e´te´ tardive, cette anne´e) are contrasted. The criticism might be raised here, as it was of Wenker (1878) for instance, that the authors have exceeded the principle of documentation by suggesting clear interpretations. But on the other hand, firstly ⫺ and in contrast to many modern maps which show combined phenomena ⫺ it is still clear which elements are mapped and, second, the user becomes sensitized to syntactic variation that is less obvious than phonetic variation. This highlights other distinctive features of the ALF, namely its inclusiveness with regard to linguistic sublevels (phonology, morphology, lexicon, syntax) on the one hand, and its orientation to the everyday life of its informants, which results in a concentration on specific semantic or lexical fields (e.g., personal names, plants), on the other. Both became important for subsequent projects.

3.3. Immediate successors Obviously, Wenker’s investigations in Germany and Gillie´ron’s in organizational terms parallel research in Switzerland and later France fundamentally influenced the scientific potential of dialect cartography in and of themselves. The most important factor might be that it was only through them that the empirical basis reached a (provisional) peak in terms of the quantity of data and its linguistic accuracy. In terms of theory, it could be concluded that Romance and German linguists stood in relatively close contact. For instance, the Literaturblatt für germanische und romanische Philologie founded in 1880 by Behaghel and Neumann, has played a significant role. However, in practice, collaborations in the field of language geography remain rare to this day, making it simpler to describe these lines of tradition separately.

32. Linguistic atlases Starting with the German tradition, the following national projects that are associated with the Marburg research center established in the context of Wenker’s work should be mentioned. In Marburg, the Sprachatlas des deutschen Reichs was finally partially published as the Deutscher Sprachatlas (DSA, 1927⫺1956) by Ferdinand Wrede (a former employee of Wenker) and others. The final, methodologically substantiated description of the project is found in a separate handbook that forms part of the DSA (Mitzka 1952). Among other aspects, the analysis of the material gave rise to strategies for the interpretation of linguistic maps. Time and again, the editors found particular patterns of spatial distribution which they summarized into spatial types that are relevant for the interpretation of language change, for instance. In a sense, a summation of these is the classification of German dialects by Wiesinger (1970). In using the Wenker data in combination with more recent studies, Wiesinger demonstrates ⫺ from the perspective of a structuralist ⫺ the spatial structuring as a whole, together with the transition zones between larger dialect areas. All in all, this is not just a continuation of Wenker’s work. When one remembers that Wenker started out with the classification of the Rhenish dialects, Wiesinger’s synthesis also closes the circle. Historical developments meant that only then, 100 years after Wenker, were linguists able to analyze and interpret the cartographic information in more reliable ways and thus go beyond the documentation principle and develop analytical maps of high quality (for a general critique cf. Kirk 2001: 359⫺360). Aside from this, the extralinguistic embedding of Wenker’s data led to the elaboration of the subdiscipline of social linguistics (Soziallinguistik) under Wrede’s direction, whose most important output is the Deutsche Dialektgeographie series (DDG; cf. Schrambke in this volume). Another step in the documentation of language across space can be seen in the various collections of sound recordings, such as the early recordings of Joseph Seemüller (Vienna) from 1905 on, or the collection of Wilhelm Doegen (Berlin) from 1909 on. The latter was carried out in collaboration with the second generation of researchers in Marburg, especially Wrede. Finally, a so-called “talking atlas” (sprechender Atlas) was published by the Phonogrammarchiv in Zurich in the context of the SDS (cf. below). Interestingly, its appearance plays back to the first (early) atlas definition mentioned above: this type of atlas has no maps (on the linguistic relevance of sound atlases, cf. Rabanus 2005; Schmidt and Herrgen to appear). Completely different is the Deutscher Wortatlas (DWA, Mitzka, Schmidt and Hildebrandt 1951⫺1980). This major project (22 volumes), initiated under Walther Mitzka, can be seen as the continuation of Wenker’s national atlas and, as such, it was formally denominated as part of the DSA from volume 11 on. Whereas Wenker focused on sounds and forms, the DWA is dedicated to words. Its onomasiological goal is the exploration of the dialectal synonyms of various concepts. Basically, the DWA is oriented toward agricultural vocabulary. Given its lexical focus, phonological details were redundant and a phonetic transcription became superfluous. Hence, the indirect method already favored by Wenker was seen as adequate for the nationwide collection of data. Using a questionnaire of 200 items (mostly words plus a few sentences), Mitzka and his staff polled over 48,000 localities between 1939 and 1942, representing roughly a hundred percent of the German communes and nearly the same density as that reached by Wenker’s project, with around one location for every fourteen square kilometers. Even though the data was not phonetically transcribed, the editors did not attempt a comprehensive classification of the individual records. At most, a (rather approximate) classification of lexical

579

580

VI. Data analysis and the presentation of results records was made on the basis of a typological system of symbols. Further, the lexical types the authors considered to be most relevant are displayed using exemplars (Leitformen). Criticism of this atlas has been raised, that the extensive integration of original forms via geometric symbols in some cases cancels out or obscures the visual impression. On the other hand, prior to this, the systemic structuring of German dialect space was predicated virtually exclusively on the phonological and in parts the morphological dimension. Against this background, one of the most important merits of the DWA is to have demonstrated the lexical structuring of space and to have shown that, in Germany, lexical heterogeneity and lexical particularity is largely independent of phonological and morphological distribution patterns (cf. Hildebrandt 1987). This finding was only made possible by the inclusion of the original forms. In proceeding thus, the DWA reveals itself to be a continuation of the documentation principle. A comprehensive dicussion of the DWA is given by Wiegand and Harras (1971). In the Romance tradition, the (yet again) Swiss scholars Karl Jaberg and Jakob Jud are regarded as the immediate successors to the ALF. Both studied under Gillie´ron and their Sprach- und Sachatlas Italiens und der Südschweiz (AIS, Jaberg and Jud 1928⫺ 1940) is strongly endebted to their teacher’s methodology. Nearly every innovative aspect in the work of Gillie´ron is rigorously worked out in the AIS, be it the transcription system, the typing of map entries, or the classification of the questionnaire’s items and with it the semantic organization of the atlas maps. The principal motivation behind every organizational decision is strongly dependent upon the special requirements of the investigation area. This is what has made the atlas the standard reference work for many atlas projects in different philological traditions. For anyone interested in the development of language atlases, their introduction (Jaberg and Jud 1928) remains to this day one of the most important starting points. Against this background, the powerful influence of their work on later dialectology is no surprise. Another immediate successor (that also became an inspiration for many other atlases) is the Sprachatlas der deutschen Schweiz (SDS, Hotzenköcherle et al. 1962⫺1997), initiated by Rudolf Hotzenköcherle in 1935. It is by no means coincidental that, once again, the extraordinarily heterogeneous linguistic situation in Switzerland was the focus. Hotzenköcherle, a former student of Jud’s, set out to develop an atlas of a geographically restricted region (“Kleinraumatlas”), and in a sense thus closed the circle, bringing the Romance tradition back to the starting point of Gillie´ron. Hotzenköcherle’s focus on confined regions is a result of intensive discussion, especially within Romance philology (cf. the Atlas linguistiques de la France par Re´gions called for by Albert Dauzat in 1939). For direct data collection, the primary advantage of exploring small areas is the possibility of obtaining a dense network of survey locations. Where, for instance, the ALF polled around two percent and the AIS 5.5 percent of all possible communes, the SDS explored 33 percent. This leads to a density of one location for every 37 square kilometers in the case of the SDS, whereas one location in the ALF and the AIS represents 830 and 765 square kilometers, respectively. The distance between the individual survey locations averages five to seven kilometers in the SDS, whereas it is thirty kilometers for both the ALF and the AIS (Hotzenköcherle 1962: 86⫺87). This type of atlas ⫺ regionally restricted but with directly collected data and comprehensive coverage of linguistic phenomena ⫺ finally matched the masses of data collected in the indirect paradigm and, thanks to its inter-philological orientation, the SDS also influenced subsequent German atlas projects.

32. Linguistic atlases It is worth mentioning that once again the methodological aspect is highly prominent. On the general goals of linguistic atlases, Hotzenköcherle (1962: 1⫺2) summarizes his perspective as follows: 1. A linguistic atlas should present a cross-section of sounds, forms, words, etc. over space and time 2. Building on the spatial structure that results from (1), the goal should be a “languageinternal” analysis leading to a spatial typology 3. Finally, the goal should be to relate the linguistic findings (whether of individual, structural or typological nature) to extralinguistic phenomena in space. While (1) is an expression of the documentation principle, points (2) and (3) refer to application, and thus to the analytical potential of an atlas. In this, the SDS combines the evolving emphases in the history of linguistic cartography documented earlier with the findings of social linguistics. In concentrating on the German language, Hotzenköcherle liaises with German philology, as becomes obvious in the design of the atlas. Even though the data is phonologically valid in detail, the SDS uses symbols, not the original forms. By using (grosso modo) the German paradigm, the atlas offers its readers a clear spatial impression when looking at the mapped phenomena. But Hotzenköcherle’s work should also be seen as a link between the two philologies in terms of its linguistic modeling. In this regard, his third goal is explicitly tied to the epochal work of Aubin, Frings and Müller (1926), which attempted to establish a strong correlation between linguistic and non-linguistic phenomena. As a result, the atlas is also designed to offer answers to folkloristic and social questions insofar as they implicate spatial relations. In the final instance, this is also a consequence of defining language as embedded within a comprehensive network of the specific conditions of human life. The SDS renders the need to integrate additional information and comments acute by linking language and linguistic geography to other human factors. Hence the SDS is extensively connected to other projects such as, e.g., the folkloristic Atlas der schweizerischen Volkskunde (Geiger et al. 1950⫺1995). From a methodological point of view, these crossconnections are highly interesting, since they offer possibilities that exceed the traditional potential of linguistic atlases. For instance, by linking the individual thematic maps to entries in the Schweizerisches Idiotikon, a semasiological approach that goes beyond the usual onomasiological orientation of atlases becomes possible. This maximizes the potential of an atlas, and, in terms of intent, is not that distant from the texts that have augmented atlases since the earliest times. At the same time it becomes obvious that linguistic atlases ⫺ in contrast to earlier views ⫺ are cohesive entities (cf. Hotzenköcherle’s first goal). Only the analysis of every thematic approach an atlas offers enables a complete understanding of the spatial structuring of language. This view represents a fundamental geolinguistic insight. It is the precondition for dialectometric approaches that are based on the analysis of complete atlases, for instance (e.g., Goebl 1984). Another special feature of the SDS are the sound recordings that were collected into a sound library or “Phonotek”. From the outset, these recordings were intended to give a more complete impression of spoken language than a printed atlas can. But, unlike later projects, they were designed not to validate the phonetic transcriptions of the data, but as separate undertakings (1954⫺1959) which sought to supplement the documentation basis of the whole atlas. In this light, the SDS should also be seen as an early example of a sound atlas (cf. Schmidt and Herrgen to appear). In conclusion, it should

581

582

VI. Data analysis and the presentation of results be pointed out that the methodological basis for virtually all activity in the field of linguistic or rather dialect atlases is elaborated (or at least considered) within the first approaches in the German and Romance philological traditions together with their immediate successors (up until the first half of the twentieth century): ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺

the awareness of the potential of linguistic atlases for smaller and larger regions both the documentation principle and analytical approaches the principles of data collection, data arrangement and display the thematic structuring and needs of atlases the organization of atlases as cohesive works in terms of thematic structuring, reference systems, symbolic representation, comment, etc. the treatment of both onomasiological and semasiological material in atlas cartography both paradigmatic and syntagmatic data processing in atlas cartography multimedial extensions of printed atlases the principles and possibilities of data interpretation extralinguistic linkages.

At the same time, the fundamentals of other aspects, like the pluridimensional paradigm, for instance, are elaborated largely outside of the philologies examined here.

4. Further development With the atlases mentioned so far, the pioneers of methodological progress in both German and Romance philology, at least up until the mid twentieth century, have been addressed. There are many other projects that have been broadly influenced by the individual undertakings, such as the regional atlases of the Marburg research center or the continuations of the ALF in many Romanic countries. Also worthy of mention are the recent regional atlases in Germany and Switzerland, which (their computer-based implementation aside) are of high empirical significance and methodological refinement but which generally represent relatively little cartographic progress. For more detail here, the reader is referred to the relevant articles in both this volume and the Language Mapping handbook (Lameli, Kehrein and Rabanus, to appear). Furthermore, it should of course be pointed out that there are numerous other approaches that aim to describe geographical variation in language via atlases besides those of the German and Romance schools. Above all, traditions in the English-speaking and Slavic countries and the Netherlands should be mentioned here. In this regard, the dialectometric approach which initially emerges solely from atlas contexts is highly relevant; it is a point of emphasis in the traditions that is not examined in further detail here. The most recent and probably most complete overview is that offered in the philological chapters in the Language Mapping handbook mentioned above. A solid overview can also be found in Chambers and Trudgill (1998: 13⫺31), Kirk (2001) or Veith (2006). For an exhaustive description of the pre-1950 scene, the reader is referred to Pop (1950). In language cartography in recent years, the investigation of other linguistic levels, especially syntax and morphology, has become more prominent. In contrast, the prosodic dimension has remained a desideratum. The most important developments in more

32. Linguistic atlases recent atlas cartography, however, are related to two aspects: first ⫺ related to the internal design ⫺ there is the combination of linguistic and extralinguistic factors; second ⫺ related to the methodological realization ⫺ there is the increasing use of computers. These aspects are described separately.

4.1. Pluridimensional orientation It has been argued above that, to the extent that they were able to observe it, the authors of the ALF already tried to integrate social variation in language use. With Wenker too, the integration of linguistic variability has been pointed out. However, for these authors a social embedding of linguistic data was not the actual goal of their investigations. From a global perspective, to date the typical language atlas still focuses on just one variety as spoken by just one social type of informants (or social group). Hence, they follow a monodimensional approach, mostly concentrating on the language production of so-called NORMs: “non-mobile, older, rural males” (Chambers and Trudgill 1998: 29). But over the course of the twentieth century, the explication of socially conditioned linguistic variability also became an important and increasingly relevant topic of study. In geolinguistics, one result of this is the pluridimensional dialectology that has become more prominent in recent years. The goal of the pluridimensional orientation is to provide a more comprehensive picture of the social embedding of language. This is also a recognition of the importance of linguistic representativity (cf. Herrgen in this volume). In contrast to monodimensional atlases, their pluridimensional counterparts take account of a set of disparate factors, such as gender and age. Maps and atlases thus become the medium for different types of information which visually enables an understanding of language in space (cf. Mang and Wollin, to appear). Jaberg and Jud’s AIS, which is marginally pluridimensional (Thun 2000: 81, footnote 33), aside, it is above all the work of Hans Kurath, as represented by the Linguistic Atlas of New England (LANE, Kurath et al. 1939⫺1943) that should be mentioned in this context. Given that Gillie´ron’s pupil, Jud, was directly involved in the theoretical and methodological preparation of the atlas (cf. Kurath 1939: xii), there is a sense in which this is yet another continuation of the Romance paradigm in a different typological frame. Kurath ⫺ who displays a thorough knowledge of the history and methodology of language cartography (especially the German and Romance traditions) ⫺ aimed to describe both the regional and the social differentiation of language. Right from the beginning of the project, he considered different social groups of informants, classified into three levels of education (little formal education vs. better formal vs. superior education) and two generations (aged/old-fashioned vs. middle-aged/more modern; cf. Kurath 1939: 44), although these differences are not fully explored at every location (there are a total of 400 informants across 200 communities). Because informants’ backgrounds vary, at least one constant is needed: all informants must have grown up in the particular location under investigation. In the case in point, the local focus is relatively unique: in line with New England’s status as a former colony, special attention is paid to the history of the chosen communities, thus opening up a particular dimension of interpretative access (cf., e.g., Kurath 1939: Plate I). Hence, even though the actual language use is admittedly highly variable, there is still a strong awareness at individual locations of the historical dimension that gives rise to local types of language production. This is impor-

583

584

VI. Data analysis and the presentation of results tant in that even where linguists address the heterogeneity of language enough constancy nonetheless remains (within the individual repertoires of the speakers of local dialects) to be able to define local types, both through time and geographically. In this regard, the pluridimensional approach illustrates particularly well the interest of linguistic atlas makers in documenting both differences and similarities in language use, even if the demonstration of differences usually has priority, thanks to its potential for structuring linguistic space (cf. also Gillie´ron’s similarity approach, section 3.2). But a structural or gradual differentiation of language in geographical, social or situational terms only becomes possible through an (at least indirect) consideration of the constant elements of language. From a cartographic point of view, the implementation of pluridimensional data is rather complicated. The LANE displays the informants’ realizations (narrow phonetic transcription), sorted according to the classification mentioned above, directly at the relevant location. Explicitly a completion but also ⫺ from a pluridimensional point of view ⫺ an expansion of Kurath’s work is the Linguistic Atlas of the Gulf States (LAGS, Pederson 1968⫺ 1992) under the editorship of Pederson. The selection of informants is closely aligned with Kurath’s approach (cf. Pedersen 1968: 16⫺23; 33⫺40). In addition, differences among (1) four generations are explored, further distinguished in terms of (2) regionality (communities, counties, parishes), (3) gender, (4) ethnicity (black vs. white), (5) informant types (old-fashioned and insular, modern and wordly [sic!], etc.). The LAGS thus indicates the enormous potential of the pluridimensional approach. In fact, the atlas reveals many interdependencies linking the individual variables. Cartographically, the LAGS tries to handle its wealth of material by presenting tables and maps dedicated to selected linguistic and nonlinguistic phenomena. A similar project in terms of the goal of documenting the language use of different social groups (differentiated by age) is the Linguistic Atlas of the Seto Inland Sea (LAS, Fujiwara 1974). Another approach to handling the variability of linguistic reality within an atlas project is evident in the methodology of the Mittelrheinischer Sprachatlas (MRhSA) by Bellmann, Herrgen and Schmidt (1994⫺2002; cf. Herrgen in this volume). There, the differences between two social groups (aged and sedentary vs. middle-aged and mobile) are displayed on colored contrast maps. However, due to the particular situation in Germany, the linguistic goal is different to that of the LANE and the LAGS. In Germany, education and age are not the most relevant factors determining the use of dialect features. More important is the amount of regular communication outside (or with people from outside) the home town. In line with this context, in which there is no real marked linguistic difference between sociolects and dialects, the authors of the MRhSA sought to find indicators of the vertical differentiation of language, i. e., on a spectrum between standard language and dialect (Bellmann 1994: 1⫺4) along which people can shift and switch up and down according to situational requirements. The primary goal is thus not the description of the social differentiation of language but the description of existing linguistic varieties (local dialects vs. regional dialects) using sociodemographic factors as extralinguistic correlates. Hence, this bidimensional approach provides fresh insights into the real state of language and language use. At the same time, it demonstrates language change in terms of the widespread replacement of local dialects by regional dialects. In a similar vein, other German pluridimensionally oriented projects are based on correlations between language features and social or contextual factors around and within major cities (cf., e.g., Mang 2004).

32. Linguistic atlases Certainly the most ambitious pluridimensional work is that conducted by Harald Thun in South America, focusing on Spanish, Portuguese and Guaranı´ (for an introduction cf. Thun 2000, to appear). The Atlas lingüı´stico Diato´pico y Diastra´tico del Uruguay (Thun/Elizaincı´n 2000⫺; ADDU) can provide a suitable impression. Thun, who first proposed the term pluridimensionality, structures the atlas and the informants into different dimensions at particular sub-levels (cf. Thun 2000: 82): 1. diatopic a. topostatic (constant domicile) b. topodynamic (recently changed domicile) 2. diastratic (two social groups) 3. diagenerational (two age groups) 4. diasexual 5. diaphasic (three styles) 6. diareferential (contrast between answers and comments) 7. dialingual (Spanish vs. Portuguese) A wide range of both linguistic and extralinguistic factors can thus be controlled and displayed on maps using a special symbolization system (quadrants). Once again, the linguistic goal is quite distinct. In contrast to the LANE, the LAGS and the MRhSA, the ADDU documents the communicative breadth of language. Even if it is not possible to explore all factors in the same depth and if in some locations only certain dimensions can be explored, this approach is at present the most comprehensive. Even compared to sociolinguistic studies that are often confined to a single location but explore multiple extralinguistic factors in depth, such a variety of controlled parameters is exceptional. According to Thun, the approach opens the way to linguistic atlases that display human language within communicative networks, and from a philological point of view this might well be the next stage in linguistic atlas cartography (cf. Thun 2000: 84⫺86). In this regard, Schmidt (to appear) argues that the possibility of such new methodological approaches is necessarily tied to progress in the technical means of realization. The keyword here is “linguistic dynamics”, and its analysis is dependent upon the geographic orientation, inextricably tied firstly to linguistic atlases and secondly to their digital availability.

4.2. Computerization The computerization of linguistic cartography has generated a wide and international variety of undertakings (cf., e.g., Kirk 2001: 351⫺353). During the 1970s, the first projects to use computer methods in linguistic cartography ⫺ especially in the planning of atlases ⫺ emerged, like the Computer Developed Linguistic Atlas of England (CLAE, Viereck and Ramisch 1991⫺1997), the Atlas Linguarum Europae (ALE, Alinei, Viereck and Weijnen 1983) and the Kleiner Deutscher Sprachatlas (KDSA, Veith, Putschke and Hummel 1984⫺1999). More recently, however, virtually every atlas is realized using computers. A differentiation is possible by looking at the specific uses to which computers are put and their potential value for language documentation and linguistic analysis. Whereas some projects use specific graphics software for the design of atlas maps (e.g.,

585

586

VI. Data analysis and the presentation of results in Germany, most volumes of the Bayerischer Sprachatlas series; BSA), others develop tailor-made computer systems to answer particular linguistic questions (e.g., the Digital Wenker Atlas [DiWA], Schmidt and Herrgen 2001⫺2009). While some projects view a digital version as a worthwhile supplement to an edited book (e.g., the World Atlas of Language Structures [WALS ], Haspelmath et al. 2008), others take the other tack by first publishing their databases on the internet as a means of linguistic analysis before producing printed atlas volumes of representative maps (e.g., the Syntactische Atlas van de Nederlandse Dialecten [SAND], Barbiers et al. 2005). Another distinction is possible in terms of how the digital atlases are made available (desktop vs. internet), and so on. A special case is the Atlas of North American English (ANAE ) by Labov, Ash and Boberg (2006). While phenomenal atlases ⫺ especially sound-oriented atlases ⫺ are usually based on phonetic transcriptions made by specially trained investigators, the ANAE places its emphasis on acoustic analysis and thus sets a new cornerstone in atlas cartography. Significantly, in addition to the computer-based telephone interviews, the encoding of spoken language is in many cases purely computer dependent. In total, 439 of the 805 recorded interviews are analyzed with regard to the details of the formants they exhibit. The use of an automatic readout of the acoustic speech signal is a special case of computerization in geolinguistics which indicates some of the potential of the digital era (for a critique cf. Künzel 2001). However, currently, the most common procedure is to digitize the collected data and then analyze it statistically. Even though computer-based methods yield an additional benefit, the data continues to be explored, encoded and mostly classified using traditional techniques. Many atlas projects are currently refurbishing their (sometimes historical) data by providing digital editions with enhanced functions. Occasionally such work is conducted within larger organizational frames, as with the North American atlases (e.g., the Linguistic Atlas of the Middle and South Atlantic States [LAMSAS ], cf. Kretzschmar and Schneider 1996) or the Bavarian atlases (e.g., the Bavarian Database of Dialects [BAYDAT ], cf. Zimmermann 2007). The fact that the LAMSAS gave rise to an Atlas by the Numbers (Kretzschmar and Schneider 1996) shows that in the meantime, and in no small part due to computer techniques, the documentation principle ⫺ which pioneers needed time to become aware of ⫺ also enables promising future analytical approaches. At the same time, the digitization of data makes it possible to cross-connect projects, although this proves rather difficult and complicated in detail. The idea of linking data is not new and it is not necessarily bound to the computerization of linguistics. The example of the print-based SDS, with its attempt to link information from lexica and other sources, already demonstrates the intent to link linguistic material of differing provenience. But the development of modern information systems compels linguistic data linking. In the long term, this will dissolve the at present rather clear distinctions between dictionaries, atlases, sound archives, bibliographies, etc. (cf. Moulin in this volume). Whether such information systems can still be considered atlases is a question of definition that has not been raised. At any rate, they also serve the documentation principle. One of the most ambitious undertakings in this regard is the recently launched regionalsprache.de project (REDE; cf. ) at the Marburg research center. Among other things (e.g., the collection of fresh linguistic data), REDE aims to digitize a wide range of German language atlases and bring them together via the internet along with numerous sound files and bibliographic information (cf. in this

32. Linguistic atlases regard the design of DiWA as well). Further, in collaboration with other institutes and individual scholars, a cross-connection with other information sources (e.g., dictionaries, questionnaires, alignments, etc.) will be implemented. Due to the vast mass of data thus pooled, quantitative analyses must be implemented, and a strong analytical orientation to the goals of information systems (cf. in this regard the goals of the SDS) is a result. From a theoretical point of view the potential of the cross-connection is enormous, and in this specific case, the analysis of linked language data from different periods in the history of linguistics has given rise to the theoretical framework of the linguistic dynamics approach (cf. Schmidt in this volume). Finally, it should be pointed out that, thanks to GIS technology, cartographic quality in linguistics has improved in recent years. Whereas in the 1990s and the beginning of the twenty-first century, linguists often still designed their atlases on their own (leading to some problematic results that can be summarized as the “CorelDraw syndrome”), in future ⫺ via web mapping, etc. (e.g., Nickel 2008) ⫺ both the computational basis and collaborations within and across philologies seem set to become more relevant than ever. Assuming that the cross-linking of data and information systems can be achieved, the result will be an extremely powerful instrument that could well elevate linguistics to another analytical sphere. Nevertheless such intentions are beset with both fundamental and more specific problems. One technical and organizational issue is the problem of how to handle cross-connectivity. A problem of interpretation is the high degree of suggestiveness maps and cartographic instruments possess: steps need to be taken to ensure that the user of the linguistic information system is afforded the support of comparable data and data types. At the same time, the user must be aware of the interpretative traps that are an inevitable consequence of the availability of mass data. Hence, as a future challenge, geolinguists (and others) have the difficult task of ensuring that students and scholars are adequately educated. This, however, is essentially no different to the situations faced and generated by the pioneers described above.

Acknowledgement I am grateful to Jürgen Erich Schmidt, Marburg, for helpful comments on this article and to Mark Pennay, Marburg, for once more correcting my English.

5. Reerences Alinei, Mario, Jacques Allie`res, Ruben I. Avanesov, Terho Itkonen, Wolfgang Viereck and A. A. Weijnen (eds.) 1983⫺ Atlas linguarum Europae (ALE). Assen/Maastricht/Rome: van Gorcum/Istituto Poligrafico e Zecca dello Stato. Anonymous [Schmeller, Johann Andreas] 1842 Review of Sprachkarte von Deutschland; Slovansky´ Zemeˇvid. Gelehrte Anzeigen 69⫺71: 553⫺576. Anonymous [Schmeller, Johann Andreas] 1844 Review of Atlante linguistico d’Europa. Gelehrte Anzeigen 137⫺138: 65⫺80.

587

588

VI. Data analysis and the presentation of results Aubin, Hermann, Theodor Frings and Josef Müller 1926 Kulturströmungen und Kulturprovinzen in den Rheinlanden. Geschichte, Sprache, Volkskunde. Bonn: Röhrscheid. Balbi, Adrien 1826 Atlas ethnographique du globe: ou classification des peuples anciens et modernes d’apre`s leurs langues. Paris: Rey et Gravier. Barbiers, Sjef, Hans Bennis, Gunther De Vogelaer, Magda Devos and Margreet van der Ham 2005 Syntactic Atlas of the Dutch Dialects. Amsterdam: Amsterdam University Press. Behaghel, Otto 1881 Review of Georg Wenker (1880): Sprachatlas von Nord- und Mitteldeutschland. Literaturblatt für germanische und romanische Philologie 12: 434. Bellmann, Günter 1994 Einführung in den Mittelrheinischen Sprachatlas (MRhSA). Tübingen: Niemeyer. Bellmann, Günter, Joachim Herrgen and Jürgen Erich Schmidt 1994⫺2002 Mittelrheinischer Sprachatlas (MRhSA). With the assistance of Georg Drenda, Heiko Girnth and Marion Klenk. 5 vols. Tübingen: Niemeyer. Bernhardi, Karl 1844 Sprachkarte von Deutschland. Als Versuch entworfen und erläutert von Karl Bernhardi. Kassel: Bohne´. [The map itself is dated 1843.] Bielenstein, August Johann Gottfried 1892 Atlas der geographischen Ethnologie des heutigen und des prähistorischen Lettenlandes. Hannover-Döhren: Hirschheydt. Biondelli, Bernardino 1841 Atlante Linguistico d’Europa, vol. 1, part 1. Milan: Chiusi. Bollmann, Jürgen and Wolf Günther Koch 2001 Lexikon der Kartographie und Geomatik in zwei Bänden. Vol. 1. Heidelberg/Berlin: Spektrum. Chambers, Jack K. and Peter Trudgill 1998 Dialectology. 2nd ed. (Cambridge Textbooks in Linguistics.) Cambridge: Cambridge University Press. Ellis, Alexander J. 1882 On dialect, language, orthoepy and Dr. Wenker’s speech atlas. In: Eleventh Annual Address of the President to the Philological Society, Delivered at the Anniversary Meeting, Friday, 19th May, 1882: 20⫺32. London: Trübner. Fischer, Hermann 1895 Geographie der schwäbischen Mundart. Mit einem Atlas von achtundzwanzig Karten. 2 vols. Tübingen: Laupp. Freudenberg, Rudolf 1965 Die Entwicklungsgeschichte der dialektgeographischen Methode. Zeitschrift für Mundartforschung 32: 170⫺182. Fujiwara, Yoichi 1974 A Linguistic Atlas of the Seto Inland Sea. 3 vols. Tokyo: University of Tokyo Press. Geiger, Paul, Richard Weiss, Walter Escher, Elsbeth Liebl and Arnold Niederer 1950⫺1995 Atlas der schweizerischen Volkskunde. 2 parts. Basel: Schweizerische Gesellschaft für Volkskunde. Gillie´ron, Jules [1880] Petit Atlas phonetique du Valais roman (sud du Rhoˆne). Paris: Champion. [Year of publication not shown.] [Gillie´ron, Jules and Edmond Edmont] 1902 [1968] Atlas linguistique de la France. Notice servant a l’intelligence des cartes. Paris: Champion. [Authors not named; available from or as a reprint: Bologna: Forni.] Gillie´ron, Jules and Edmond Edmont 1902⫺1914 Atlas linguistique de la France. Bologna/Paris: Forni/Champion.

32. Linguistic atlases Goebl, Hans 1984 Dialektometrische Studien. Anhand italoromanischer, rätoromanischer und galloromanischer Sprachmaterialien aus AIS und ALF. 3 vols. Tübingen: Niemeyer. Haspelmath, Martin, Matthew S. Dryer, David Gil and Bernard Comrie (eds.) 2008 The World Atlas of Language Structures Online. Munich: Max Planck Digital Library. Available via . Hensel, Gottfried 1741 Synopsis universae philologiae. Nuremberg: Homann Herrgen, Joachim 2001 Dialektologie des Deutschen. In: Sylvain Auroux, E. F. K. Koerner, Hans-Josef Niederehe and Kees Versteegh (eds.), Geschichte der Sprachwissenschaften, 1513⫺1535. (Handbücher zur Sprach- und Kommunikationswissenschaft 18.2.) Berlin/New York: de Gruyter. Hildebrandt, Reiner 1987 Zur Raumtypologie des deutschen Wortschatzes. In: Ulrich Knoop (ed.), Studien zur Dialektologie I, 149⫺162. (Germanistische Linguistik 91/92.). Hildesheim: Olms. Hotzenköcherle, Rudolf 1962 Einführung in den Sprachatlas der deutschen Schweiz. A: Zur Methodologie der Kleinraumatlanten. Berne: Francke. Hotzenköcherle, Rudolf, Robert Schläpfer, Rudolf Trüb and Paul Zinsli 1962⫺1997 Sprachatlas der deutschen Schweiz (SDS). 8 vols. Berne: Francke. Jaberg, Karl and Jakob Jud 1928 Der Sprachatlas als Forschungsinstrument. Kritische Grundlegung und Einführung in den Sprach- und Sachatlas Italiens und der Südschweiz. Halle/Saale: Niemeyer. Jaberg, Karl and Jakob Jud 1928⫺1940 Sprach- und Sachatlas Italiens und der Südschweiz. (AIS). 8 vols. Zofingen: Ringier. Kate, Lambert ten 1723 Anleiding tot de Kennisse van het verhevene deel der Nederduitsche sprake. 2 vols. Amsterdam: Wetstein. Kirk, John M. 2001 Maps: Dialect and Language. In: Rajend Mesthrie (ed.): Concise Encyclopedia of Sociolinguistics. Amsterdam/Oxford: Elsevier, 350⫺362. Klaproth, Julius 1823 Asia Polyglotta. Sprachatlas. Paris: Heideloff & Campe. Knoop, Ulrich, Wolfgang Putschke and Herbert Ernst Wiegand 1982 Die Marburger Schule: Entstehung und frühe Entwicklung der Dialektgeographie. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke and Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, 38⫺92. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.1.) Berlin/New York: de Gruyter. Kretzschmar, William A. Jr. and Edgar W. Schneider 1996 Introduction to Quantitative Analysis of Linguistic Survey Data: An Atlas by the Numbers. (Empirical Linguistics.) London/New Delhi: Sage. Künzel, Hermann J. 2001 Beware of the “telephone effect”: The influence of telephone transmission on the measurement of formants frequencies. Forensic Linguistics 8(1): 80⫺99. Kurath, Hans 1939 Handbook of the Linguistic Geography of New England. With the collaboration of Marcus L. Hansen, Julia Bloch and Bernhard Bloch. Providence: Brown University. Kurath, Hans, Miles L. Hanley, Bernard Bloch, Guy S. Lowman Jr. and Marcus L. Hansen 1939⫺1943 Linguistic Atlas of New England. 3 vols. Providence: Brown University.

589

590

VI. Data analysis and the presentation of results Labov, William, Sharon Ash, Charles Boberg 2006 The Atlas of North American English: Phonetics, Phonology and Sound Change. Berlin/ New York: Mouton de Gruyter. Lameli, Alfred 2008a Was Wenker noch zu sagen hatte … Die unbekannten Teile des “Sprachatlas des Deutschen Reichs”. Zeitschrift für Dialektologie und Linguistik 75(3): 255⫺281. Lameli, Alfred 2008b Sprachkontakt in Norddeutschland. Ein Fundstück aus der Geschichte der Sprachkartographie. Niederdeutsches Jahrbuch 131: 51⫺69. Lameli, Alfred, Roland Kehrein and Stefan Rabanus (eds.) to appear Language and Space: An International Handbook of Linguistic Variation, vol. 2: Language Mapping. (Handbücher zur Sprach- und Kommunikationswissenschaft 30.2.) Berlin/New York: Mouton de Gruyter. Mang, Alexander 2004 Sprachregion Nürnberg. (Sprachatlas von Mittelfranken 6.) Heidelberg: Winter Mang, Alexander and Markus Wollin to appear Sprachraum and socio-demographic variables. In: Lameli, Kehrein and Rabanus (eds.). Mitzka, Walther 1952 Handbuch zum Deutschen Sprachatlas. Marburg: Elwert. Mitzka, Walther, Ludwig Erich Schmitt and Reiner Hildebrandt 1956⫺1980 Deutscher Wortatlas. 22 vols. Gießen: Wilhelm Schmitz. Münster, Sebastian 1572 [1544] Cosmographiae universalis. Basel: Henricpetrus. Nickel, Jost 2008 Das “Informationssystem Sprachgeographie”: Ein Kartographieprogramm für die Variationslinguistik. Germanistische Linguistik 190⫺191: 181⫺196. [Stephan Elspaß and Werner König (eds.), Sprachgeographie digital. Die neue Generation der Sprachatlanten.] Pallas, Peter Simon 1786⫺1789 [1977⫺1978] Linguarum totius orbis vocabularia comparativa. Edited (with an afterword) by Harald Haarmann. 2 vols. Hamburg: Buske. Pederson, Lee 1968⫺1992 Linguistic Atlas of the Gulf States. 7 vols. Athens, GA/London: University of Georgia. Pop, Sever 1950 La Diale´ctologie. Aperc¸u historique et me´thodes d’enqueˆtes linguistiques. 2 vols. Louvain: Duculot. Rabanus, Stefan 2005 Sprachkartographie des Deutschen: Von Schmeller bis zum Digitalen Wenker-Atlas. In: Claudio Di Meola, Antonie Hornung and Lorenza Rega (eds.), Perspektiven Eins, 345⫺ 363. Rome: Istituto Italiano di Studi Germanici. Ritter, Carl 1806 Sechs Karten von Europa. Mit erklärendem Texte. Schnepfenthal: Erziehungsanstalt. Sˇafarˇik, Pavel Jozef 1842 Slovansky´ Zemeˇvid. Prague. [Publisher not named.] Schmeller, Johann Andreas 1821 Die Mundarten Bayerns grammatisch dargestellt. München: Thienemann. Schmeller, Johann Andreas 1956 Tagebücher 1801⫺1852, vol. 2: 1826⫺1852. Edited by Paul Ruf. (Schriftenreihe zur bayerischen Landesgeschichte 48.) München: Beck. Schmidt, Jürgen Erich to appear Pluridimensionale Sprachgeographie und Sprachdynamik. In: Fred Boller (ed.), Die pluridimensionale Sprachgeographie in der Diskussion. Festschrift für Harald Thun. (Dialectologia Pluridimensionalis Romanica.) Kiel: Westensee.

32. Linguistic atlases Schmidt, Jürgen Erich and Joachim Herrgen (eds.) 2001⫺2009 Digitaler Wenker-Atlas (DiWA). Executed by Alfred Lameli, Alexandra Lenz, Jost Nickel and Roland Kehrein, Karl-Heinz Müller, Stefan Rabanus. Marburg: Forschungszentrum Deutscher Sprachatlas. Available via . Schmidt, Jürgen Erich and Joachim Herrgen to appear Sprachdynamik. Eine Einführung in die moderne Regionalsprachenforschung. (Grundlagen der Germanistik.) Berlin: Erich Schmidt. Stanforth, A. W. 1971 The German linguistic atlas. Proceedings of the University of Newcastle-upon-Tyne Philosophical Society 2 (1971/1972): 1⫺14. Thun, Harald 2000⫺ Altes und Neues in der Sprachgeographie. In: Wolf Dietrich and Ulrich Hoinkes (eds.), Romanistica se movet … Festgabe für Horst Geckeler zu seinem 65. Geburtstag, 69⫺89. Münster: Nodus. Thun, Harald to appear Pluridimensional cartography. In: Lameli, Kehrein and Rabanus (eds.). Thun, Harald and Adolfo Elizaincı´n 2000⫺ Atlas Lingüı´stico Diato´pico y Diastra´tico del Uruguay (ADDU). Kiel: Westensee. Veith, Werner H. 2006 Dialects. Early European Studies. In: Keith Brown et al. (eds.), Encyclopedia of Language and Linguistics, 2nd ed., vol. 3, 540⫺560. Oxford: Elsevier. Veith, Werner H., Wolfgang Putschke and Lutz Hummel 1984⫺1999 Kleiner Deutscher Sprachatlas (KDSA). 4 vols. Tübingen: Niemeyer. Viereck, Wolfgang and Heinrich Ramisch 1991⫺1997 The Computer Developed Linguistic Atlas of England (CLAE). 2 vols. Tübingen: Niemeyer. Wenker, Georg 1877 Das rheinische Platt. Den Lehrern des Rheinlandes gewidmet. Düsseldorf: Self-published. Wenker, Georg 1878 Sprachatlas der Rheinprovinz nördlich der Mosel sowie des Kreises Siegen. Nach systematisch aus ca. 1500 Orten gesammeltem Material zusammengestellt, entworfen und gezeichnet. Manuscript (hand-drafted onto printed base maps) held in the archives of the Forschungszentrum Deutscher Sprachatlas. Wenker, Georg 1881 Sprach-Atlas von Nord- und Mitteldeutschland. Auf Grund von systematisch mit Hülfe der Volksschullehrer gesammeltem Material aus circa 30 000 Orten. Text. Einleitung. Strasbourg/London: Trübner. Wenker, Georg 1886 Pronomina in Nordwestdeutschland. Strasbourg: Trübner. Wenker, Georg 1889⫺1923 Sprachatlas des deutschen Reichs. Manuscript (hand-drafted onto printed base maps) held in the archives of the Forschungszentrum Deutscher Sprachatlas and in the Staatsbibliothek zu Berlin. Wenker, Georg to appear Gesamtausgabe der Schriften zum “Sprachatlas des Deutschen Reichs”. Mit einer Dokumentation der Sprachvarianten. Edited by Alfred Lameli. (Deutsche Dialektgeographie.) Hildesheim/New York/Zürich: Olms. Werlen, Iwar 1996 Dialektologie und Sprachgeographie vom 13. bis 20. Jahrhundert. In: Peter Schmitter (ed.), Sprachtheorien der Neuzeit II. Von der Grammaire de Port-Royal (1660) zur Konstitution moderner linguistischer Disziplinen, vol. 5, 427⫺458. Tübingen: Narr.

591

592

VI. Data analysis and the presentation of results Wiegand, Herbert Ernst and Gisela Harras 1971 Zur wissenschaftshistorischen Einordnung und linguistischen Beurteilung des Deutschen Wortatlas. Germanistische Linguistik 1⫺2. Wiesinger, Peter 1970 Phonetisch-phonologische Untersuchungen zur Vokalentwicklung in den deutschen Dialekten. 2 vols. (Studia Linguistica Germanica 2.) Berlin: de Gruyter. Wrede, Ferdinand, Walther Mitzka and Bernhard Martin 1927⫺1956 Deutscher Sprachatlas. Auf Grund des von Georg Wenker begründeten Sprachatlas des Deutschen Reichs. Marburg: Elwert. Zimmermann, Ralf 2007 BAYDAT: Die Bayerische Dialektdatenbank. (Würzburger elektronische sprachwissenschaftliche Arbeiten.) Available from .

Alfred Lameli, Marburg (Germany)

33. Dialect dictionaries - traditional and modern 1. 2. 3. 4. 5. 6.

Introduction Words in space: The origins of dialect-lexicographic endeavor Mapping the German dialect dictionaries: The German language dialect dictionary landscape From index cards to proofs: Data analysis and presentation of results Electronic dialect lexicography and complex lexicographic information systems References

1. Introduction The following article is intended as an overview of the most significant phases in the development of dialect lexicography and of its methods of data analysis and presentation of results. Because space is limited and in light of the current state of research, where a systematic review or comparison of the dialect-lexicographic traditions in individual languages is not yet possible, the German language is taken as an example. However, at times other languages and related philological traditions are referred to and potential links are mentioned. The need for a typological, cross-linguistic examination of dialect lexicography in (and beyond) the European context is made even more apparent by the absence of an appropriate article in the HSK “Dictionaries” handbook. There, the topic of “dialect dictionaries” is dealt with on the basis of the situation in French (Re´zeau 1990).

2. Words in space: The origins o dialect-lexicographic endeavor In the history of how modern European languages are written, collections of words can already be found from the Early Middle Ages, in the form of glossaries for Latin texts or specialized monolingual glossaries for example. These collections can definitely be

33. Dialect dictionaries ⫺ traditional and modern considered as precursors of the “dictionary” genre. However, they are not primarily based on meta-linguistic concepts like the collection of lexicographic units from a level in time, content or space. Over the course of the emergence of popular languages and their gradual displacement of Latin in the written arena, from the late Middle Ages and especially in early modern times, the individual vernaculars deliberately developed descriptive instruments in the form of linguistic theories (grammars, orthographic doctrines) and dictionaries. An interest in the areal dimensions of language, or regional variants, became apparent early on in grammatical theorizing (in the case of German, from roughly the sixteenth century). In particular, linguistic debates about the nature of High German during the baroque era draw attention to the collection and description of diatopic or dialectal variants. Emerging from this tradition, Leibniz designed a lexicographic program that encompassed, alongside an investigation of etymology and related languages, a concerted collection of data on dialects: The fundamental soil of a language is its words, upon which its idioms grow, as it were, as fruits, and hence, one of the principal requirements of the main German language is a survey and examination of all German words, which, once complete, will include not only those used by everyone, but also those that are a feature of particular arts and ways of life. And not just those we call High German, which so dominate in writing, but also Low German, Märkisch, Upper Saxon, Franconian, Bavarian, Austrian, Silesian, Swabian and whatever else is more common among the rural population than in the cities. (Leibniz 1697: § 32; my translation)

Leibniz’ ideas proved fertile appreciated, especially in the Low German area (e.g., Richey 1755; cf. Niebaum 1979: 345⫺346). The dichotomous approach developed by Leibniz, with an emergent overarching High language counterposed by the various dialects (including the rural⫺urban opposition), is one of the factors that played a role in the emergence of idioticography in the eighteenth century (cf. Haas 1994: 1996). According to Kühn and Püschel (1990: 2060), it is primarily found in the Upper German records of this time. Another factor in the emergence of dialect lexicography at this time is the concept of “enrichment” (Kühn and Püschel 1990: 2060), which saw the survey of the complete German language as enhanced by the documentation of individual language areas. It was in this context that the first attempts at a complete survey of the German dialects were made, e.g., in Popowitsch’s (1780) Vereinigung der Mundarten von Teutschland ‘Aggregation of the Dialects of Germany’, a goal which has also reemerged recently in the attempt to create a comprehensive network of dialect dictionaries. The folkloristic preoccupation with dialects and their vocabulary within a more general cultural-cumhistorical context was joined in the late eighteenth and early nineteenth century by scientifically, and in part philologically, motivated attempts at dialect-lexicographic documentation (cf. Stadler 1806⫺1812; Schmeller 1827⫺1837, 1872⫺1877). The scientifically based form of dialect lexicography, which led to the broad-ranging territorial or dialect dictionaries, has its roots in the wider context of the nineteenth-century elaboration of single-language philologies, comparative linguistics and historical linguistics research. The establishment of national-language lexicography and the beginnings of modern dialectology also date from this period; as such they are part of a European language history. In the course of these efforts, particularly towards the end of the nineteenth century, both an intensive dialect-geographic survey of the German language area (featuring both appropriate dialect monographs and linguistic atlas projects) and numerous

593

594

VI. Data analysis and the presentation of results lexicographic projects, also regionally oriented, appeared (cf. Wiegand 1998: 690). These enterprises differ from the dialect dictionaries and lexicons that preceded them in the size of the area surveyed ⫺ usually much larger ⫺ and in their theoretical/methodological approach, which was deliberately based on dialect-geographic and dialect-lexicographic methods (cf. also Kühn and Püschel 1990: 2061). The number of published lexicons (of all types and sizes) of German dialects is currently estimated at over 400, of which roughly a third have been published since 1945 (Wiegand 1990: 2198).

3. Mapping the German dialect dictionaries: The German language dialect dictionary landscape The major dialect-lexicographic enterprises (also known as Territorialwörterbücher ‘territorial dictionaries’ or diatopische Gebietswörterbücher ‘diatopic area dictionaries’), often running over several decades or even longer, cover many parts of the German language area. Map 33.1 illustrates the areal coverage of the major dictionaries; sometimes the survey areas overlap (shown in brown on the map). In detail, the German language area is covered by more than thirty broad-area dictionaries (see Table 33.1; cf. Friebertshäuser 1983; Kühn 1978: 125⫺141; Löffler 2003: 106⫺107; Niebaum and Macha 2006: 37⫺38; Niebaum 1979). Even the titles chosen for

Map 33.1: Wide-area dictionaries of the German dialects plus Frisian and Luxembourgian (König 2007: 138)

33. Dialect dictionaries ⫺ traditional and modern Tab. 33.1: Wide-area dialect dictionaries of the German language area; an asterisk marks those dictionaries with an online component or homepage Badisches WB* Bayerisches WB* Brandenburg-Berlinisches WB Elsässisches WB Hamburgisches WB* Helgogländer WB Hessen-Nassauisches WB* Lothringisches WB Luxemburger WB Mecklenburgisches WB Mittelelbisches WB* Niedersächsisches WB Nordfriesisches WB Nordsiebenbürg.-Sächs. WB Obersächsisches WB Ostfränkisches WB Ostfriesisches WB Pfälzisches WB* Pommersches WB* Preußisches WB Rheinisches WB Saterfriesisches WB Schlesisches WB Schleswig-Holsteinisches WB Schwäbisches WB Schweizerisches Id* Siebenbürgisch-Sächs. WB Siegerländer WB Südhessisches WB* Sudetendeutsches WB* Thüringisches WB* Vorarlbergisches WB Bair. Mundarten in Österr.* Westfälisches WB*

Badisches Wörterbuch 1925⫺ (A⫺Schlenz) Bayerisches Wörterbuch 2002⫺ (A⫺ bitz) Brandenburg-Berlinisches Wörterbuch 1976⫺2001 (4 vols.) Wörterbuch der elsässischen Mundarten 1899⫺1907 (2 vols.) Hamburgisches Wörterbuch 1985⫺2006 (5 vols.) Helgoländer Wörterbuch 1957⫺1968 (A⫺luuwet) Hessen-Nassauisches Volkswörterbuch 1943⫺ (from vol. 2; L⫺Zankdiviensalat) Wörterbuch der deutsch-lothringischen Mundarten 1909 Luxemburger Wörterbuch 1950⫺1977 (5 vols.) Mecklenburgisches Wörterbuch 1937⫺1998 (7 vols. plus index with addendum) Mittelelbisches Wörterbuch 2002 (vol. 2: H⫺O) Niedersächsisches Wörterbuch 1965⫺ (A⫺körperlik, La¯b⫺Ma¯nwessel) Nordfriesisches Wörterbuch (in separate part-publications) Nordsiebenbürgisch-Sächsisches Wörterbuch 1986⫺2006 (5 vols.) Wörterbuch der obersächsischen Mundarten 1994⫺2003 (4 vols.) in progress Wörterbuch der ostfriesischen Sprache 1879⫺1884 (3 vols.) Pfälzisches Wörterbuch 1965⫺1998 (6 vols. and a supplement) Pommersches Wörterbuch 2007⫺ (A⫺K) Preußisches Wörterbuch 1981⫺2005 (6 vols.) Rheinisches Wörterbuch 1928⫺1971 (9 vols.) Seelter Woudebouk (1 vol.); Näi Seelter Woudebouk 1992⫺ (A⫺E); Saterfriesisches Wörterbuch 1980. Schlesisches Wörterbuch 1963⫺1965 (3 vols.) Schleswig-Holsteinisches Wörterbuch 1927⫺1935 (5 vols.) Schwäbisches Wörterbuch 1904⫺1936 (vols. I⫺VI.2) Schweizerisches Idiotikon. Wörterbuch der schweizerdeutschen Sprache 1881⫺ (A⫺W-m, Wa¯n⫺warnen) Siebenbürgisch-Sächsisches Wörterbuch 1924⫺ (A⫺R) Siegerländer Wörterbuch 1938; 2nd ed. 1968 Südhessisches Wörterbuch 1965⫺ (A⫺zäckerig) Sudetendeutsches Wörterbuch 1988⫺ (A⫺Geilstelle) Thüringisches Wörterbuch 1966⫺ (L⫺Z, A⫺Kohlrunkel) Vorarlbergisches Wörterbuch 1960⫺1965 (2 vols.) Wörterbuch der bairischen Mundarten in Österreich 1970⫺ (A⫺ (auf)ge-dunsen; 2 supplements) Westfälisches Wörterbuch 1969⫺ (A⫺Bra˚mbi ere, dä⫺Furcht; 1 supplement)

the projects indicate differing conceptual parameters; for example, the survey boundaries can reflect areas defined geographically, culturally and historically, dialectally, or nationally. For the sake of completeness and because of methodological ties to German dialect lexicography, Table 33.1 also includes dictionaries of Frisian and Luxembourgish, although these are independent languages. The dialect dictionaries of Transylvanian Saxon are included as well, but these lie outside the central European area shown in Map 33.1.

595

596

VI. Data analysis and the presentation of results Many of the undertakings that are still in progress have developed an online presence, with their own homepage providing information on the history, materials, progress or publication reports; these are marked with an asterisk in Table 33.1. In addition to these wide-ranging, usually multiple-volume, dialect dictionaries, there have from the outset of the dialect-lexicographic tradition also been dictionaries devoted to more confined regions, cities, or villages (see the selection in Kühn 1978: 125⫺141; Löffler 2003: 106⫺107).

4. From index cards to proos: Data analysis and presentation o results The individual dictionary undertakings sometimes differ substantially in data acquisition, analysis and presentation of results. This is in part due to differences in their corpuses, in part to their differing theoretical and methodological approaches (for details, see Niebaum 1979; Kühn 1982; Friebertshäuser 1983: 1986; Bauer 1996). The data basis of the major dictionaries may vary substantially over the preparation phase; furthermore, extended periods of data collection imply that some older material is incorporated. Hence heterogeneous layers (differing in time and space) can be expected within the dictionary corpus (see Baur 1986 on data collection). In addition to synchronic/diatopic information, the dictionary can thus provide historic/diachronic details insofar as these are appropriately flagged within the article itself. The Bayerisches Wörterbuch ‘Bavarian Dictionary’, for example, draws on an archive of several million records from questionnaires covering more than three generations and the research literature. It offers both a synchronic component and a diachronic component that extends back to the beginnings of writing in early medieval Bavarian scriptoriums (cf. Rowley and Schmid 1996). The major dialect dictionaries of the German language area generally rely upon a semasiological approach to description, investigating semantics from the starting point of the concrete citation (in contrast to Dutch, where a number of systematic/onomasiological dialect dictionaries are being written, e.g., the Woordenboek van de Limburgse Dialecten [1982⫺] or Das Woordenboek van de Overijsselse Dialecten [2000⫺]); the questionnaires on which they are based are usually of onomasiological design. This approach (fairly typical for lexicography) also reflects practical considerations that facilitate corpus construction, data processing and data presentation: dictionary articles are usually organized and analyzed as alphabetically ordered lists of words, and data are also usually presented in (more-or-less) alphabetical order, too. For practical reasons, many undertakings publish consignments that can eventually be combined into volumes. The unifying features of a semasiological approach and the use of Standard German for descriptions aside, many differences can be observed. These include ⫺ to name only the most important ⫺ the intended audience, the vocabulary covered, headword selection and ordering, cross-referencing, and the nature of the definitions and the grammatical information offered. The heterogeneity of approach can be illustrated taking the West Middle German dictionaries as an example (cf. Table 33.2 and Fournier 2003: 156⫺159). Criteria for the five major dialect dictionaries of this language area differ in virtually all of the aspects mentioned. Thus, the multivolume dictionaries of Palatine German, Rhinelandic and Luxembourgish encompass the entire everyday vocabulary of the areas in question,

33. Dialect dictionaries ⫺ traditional and modern

597

Tab. 33.2: Different layout and data presentation in the dialect dictionaries of West Middle German (based on Fournier 2003: 158; Elsässisches, Lothringisches, Rheinisches, Luxemburger and Pfälzisches WB, respectively; cf. Table 33.1) ElsWb

LothrWb

RheinWb

LuxWb

PfWb

Commenced

1887/1890

1897/1900

1904

1952/1935

1912

1st vol. publ.

1899

1909

1928

1950

1965

Completed No. of vols.

1907 2

⫺ 1

1971 9

1977 4⫹1

1997 6

Approx. no. of lines (2 cols.)

1730

550

12800

1840 ⫹ 230

4700

Region Data and source

south questionnaire; earlier data collections and research

west questionnaire; “Paris conversations”; official documents

north questionnaire; no sources given in the entries Teachers

west (research lit.); additional written sources for C19th

east questionnaire; earlier data collections and research

Target vocabulary

non-standard

non-standard

complete lexicon of the Rhineland

“language” of Luxemburg in C20th, in part also C19th

everyday language

Historical dimension Keyword type

semi-historical dialect

dialect

Keyword order

Schmeller system

alphabetical

Audience

Unique features Crossreferences Modeled on

ElsWb, early LuxWb ElsWb

planned but not executed standard

according to word family lexical maps from vol. 3 on

everyday C20th language; so-called Luxembourgish koine´

esp. for keyword etymology standard

alphabetical

strictly alphabetical figures and maps

RheinWb

ElsWb, LothrWb, RheinWb, LuxWb RheinWb

whilst the one or two-volume dictionaries of Lorraine-Franconian and Alsatian focus on those lexical elements that deviate from the standard language, making them similar to idioticons. Further, because of their choice of target vocabulary, the Alsatian and Lorraine-Franconian dictionaries adapt their lemmas to a dialectal norm while the Rhinelandic and Palatine-German dictionaries adjust headwords in line with the standard language. The Luxembourgish dictionary follows a system that aligns its headwords with a virtual Luxembourgian metasystem, the so-called “common dialect (koine´) of the twentieth century” (Wörterbuchkommission 1950⫺1977: vol. 1: XL; my translation). The ordering of the headwords is different, too: The Lorraine-Franconian, Luxembourgish and Palatine-German dictionaries are strictly alphabetical while the Rhinelandic dictionary sorts entries by word family and then alphabet. The Alsatian dictionary (like the Swiss Idioticon) follows the somewhat more user-friendly “etymological-cum-alphabetical” headword order of Schmeller’s Bavarian dictionary, based on the consonantal structure of the stem syllables.

598

VI. Data analysis and the presentation of results In spite of all the differences, there are some universals of data analysis and presentation which effectively provide a skeletal structure for every article. (These include adherence to the “word-geographic” principle [wortgeographisches Prinzip], the organizational unit of a headword with grammatical information, definitions, illustrative examples, distributional information, and cross-referencing systems; see also Friebertshäuser [1986: 12] and the model analysis in Niebaum and Macha [2006: 114⫺125].) The dialect dictionaries do not usually go into much detail about morphological and syntactic analytics, but this is largely dependent upon whether any grammars of the area in question exist. As a rule, however, details about forms are provided (mostly in smaller type); these include important phonetic and morphological information about a particular lemma. In what follows, dialect-lexicographic data presentation will be discussed using the Wörterbuch der obersächsischen Mundarten ‘Dictionary of the Upper Saxon Dialects’ (vols. I⫺IV, 1994⫺2003) as an example. This dictionary project, hosted by the Saxon Academy of Sciences, and encompassing the dialects of the Free State of Saxony, was started in 1928. Following the loss of the dictionary material during the war, it was reestablished in 1955. The replacement data (approx. 1.5 million records) were collected up until 1977, when work began on the articles. The dictionary was finally published in four volumes that appeared between 1994 and 2003. The material is presented in the form of an alphabetically ordered semantic dictionary, with headwords adapted to Standard or High German orthography. The strict alphabetical order is occasionally disrupted by, among other things, “nests” of cognates. The source corpus in the form of an index-card archive consists of data from direct interviews, indirect questionnaires, material submitted by suitable informants, sound recordings, academic literature and dialectal fiction. Practical (time) constraints meant that no historical material was incorporated. The dictionary has a diverse circle of recipients: Aside from dialectologists and language historians, it also comprises teachers and interested non-professionals. The structure of a lemma article (cf. the example Käulchenlein in Figure 33.1) includes a headword and its grammatical category, followed by a definition in which, as per usual in lexicography, semantic variants (marked off with lower case letters), or distinct meanings (Arabic numerals) are listed. As far as possible, meanings are listed in line with language-historical developments. Where they exist, possible similarities to the standard language area are also reported.

Fig. 33.1: The article for Käulchenlein, taken from Wörterbuch der obersächsischen Mundarten (2003: vol. 2, 516)

33. Dialect dictionaries ⫺ traditional and modern

Fig. 33.2: The article Kerze with onomasiological references, taken from Wörterbuch der obersächsischen Mundarten (2003: vol. 2, 526⫺527)

The textual data presentation also includes information, in abbreviated form, about distribution that aims to describe “the geographic/territorial distribution of specific dialectal features” (Bergmann 1994: XVII; my translation); the relevant survey locations are also listed. Wherever necessary, or possible, information about the age, degree of conventionality, sociolinguistic distribution, diasystematic marking and density of citations for a lemma is provided. In addition, there are text samples illustrating the semantic information or phraseological or formulaic usages. Visibly separated from the body of the article, technical details such as etymology, history of the lemma or about a morphological component can be included. In the Käulchenlein example, morphological and phonetic details (including a transcription) are provided.

Map 33.2: Onomasiological map for Kerze in the Upper Saxon dialects, taken from Wörterbuch der obersächsischen Mundarten (2003: vol. 2, map appendix)

599

600

VI. Data analysis and the presentation of results

Fig. 33.3: Illustrated article for Arschbackenpänert, taken from Wörterbuch der obersächsischen Mundarten (1998: vol. 1, 92)

An elaborate system of cross-references enables the linking of headwords at the levels of phonetics, morphology and content. Through this, onomasiological structures can be revealed, as can be seen in the example Kerze (Figure 33.2). This information is also presented in a word-geographic dot and isarhythmic map (Map 33.2), displaying the distribution of a lemma as a demarcated area and locating deviating forms or heteronyms as symbols. In total, the Wörterbuch der obersächsichen Mundarten contains around 200 lemma maps of this sort. The dictionary maps are an ideal supplement to the individual articles ⫺ sometimes they gather scattered information into a clear, spatially visualized whole; furthermore, they can unburden dictionary articles that would otherwise have excessively long or complex catalogues of variants. Lemma maps in dialect dictionaries are usually drawn to illustrate onomasiological relationships; additionally, phonetic or form maps can be added to the text section. The significance of maps for the presentation of dialect-lexicographic findings in so-called “traditional” print dictionaries has often been a topic of discussion (Berthold 1924/1925; Scheuermann 1978; Friebertshäuser 1983: 1284⫺1286). Additional and innovative possibilities for linking map and dictionary material are opened up by the new instruments of computerized philology (see section 5.1). In addition to map material, many dialect dictionaries contain illustrations of lemmata related to unusual, folkoristic or highly regional objects of material culture (e.g., Figure 33.3), in which the relationship between lexicon development and material research become obvious (Friebertshäuser 1976: 7; Reichmann 1983).

5. Electronic dialect lexicography and complex lexicographic inormation systems The manifold developments in the field of computerized philology have brought about new opportunities for lexicographic undertakings as well, ranging from the use of com-

33. Dialect dictionaries ⫺ traditional and modern puter-based philological procedures in the creation of dictionaries or corpus-based lexicography to the design of new, internet-based lexicographic projects (cf., for instance, Klein and Geyken 2000; Haß 2005; Klosa 2007). In the meantime, the use of computers and computative procedures has become a norm in all (still) current dialect-lexicographic dictionary projects, but the degree to which computer-philological techniques are integrated varies. Some dictionary administrators only take advantage of the technical possibilities in the primary field of text processing and formatting (because of time and financial constraints), others also put cartographic programs and complex databases to use or deliberately develop their own programs for editing articles, administering databases and material and optimizing the information harvest from their articles and dictionary data (including the index-card archives), cf., e.g., the articles in Friebertshäuser [ed.] 1986: Große 1998 and Städtler 2003. The use of computative procedures has been continuously documented in particular detail for the dictionary of Palatine German (Pfälzisches WB), which has deliberately embraced the new technology since the 1980s and developed its own programs which include features like an electronic lemma list, a full-text search function and a cross-referencing system management tool, a generating tool for phonetic or lemma maps, and tools for extracting sources and searching academic literature and questionnaire data (for details see Post 1986: 1998).

5.1. Integration and dynamic linking o existing dialect-lexicographic corpora In addition to using computerized philology to create dictionaries, the idea of also integrating completed and ongoing dialect-lexicographic projects becomes manifest. The (ultimately infeasible) vision of a “central store of German regional lexical data” (Keseling et al. 1970; my translation) was formulated quite early on. The ever more powerful data processing capabilities evoked calls for an “interactive computer dialectology” (Putschke 1994: 246; my translation) or a comprehensive “German dialect dictionary” (Schröder 1997). The technical and philological challenges facing such a network of dialect dictionaries have been outlined in detail by Fournier (2003), who prototypically played out the scenario taking the dictionaries of the (south)western German language area as an example. Since then, the procedures he described are being realized in the University of Trier project, Digitaler Verbund von Dialektwörterbüchern (‘Digital network of dialect dictionaries’, DWV; ). The dialect landscapes of the Rhineland, the Palatinate and the SaarLorLux Euroregion are covered by the following completed broad-scale dialect dictionaries: the ‘Rhinelandic’ (Rheinsches WB), the ‘Palatine German’ (Pfälzisches WB), the ‘Lorraine’ (Lothringisches WB), the ‘Alsatian’ (Elsässisches WB) and the ‘Luxembourgian’ (Luxemburger WB) dictionaries. Since the form and order of the headwords differs in the print dictionaries (cf. Table 33.2), the lexical interweaving of the dialects of these conjoining regions can only be examined in detail using digital versions of the dictionaries. This makes possible complex enquiries, such as a full-text search, a narrow search for specific pieces of information in the dictionary; explicit links in the dictionary can be rendered as hyperlinks as can further crossreferences which first become apparent through various statistical and computer-philological procedures (Burch and Fournier 2004; Burch and Rapp 2007). Additional cross-

601

602

VI. Data analysis and the presentation of results

Fig. 33.4: Digitaler Verbund von Dialektwörterbüchern (DWV): internet user interface for the ‘Rhinelandic Dictionary’ with cross-referencing (right-hand frame)

references to the digital edition of the Grimm brothers’ Deutsches Wörterbuch (‘German Dictionary’, ) and the network of Middle High German dictionaries () taken from the dialect dictionaries also allow dialect vocabulary to be accessed via standard language and permit diachronic research. The digital network of dialect dictionaries was developed in several stages. A machine-readable full-text version of each dictionary in the network was created. The maps and illustrations contained in the dictionaries have also been digitalized and linked to their respective headwords. In the machine-readable version, the individual elements of each article have been (semi)automatically and uniquely labeled in SGML/XML using TUSTEP scripts applied to complex layout information (e.g., headwords, grammatical information, phonetic transcriptions, semantic information, citations, survey locations, etc.). This facilitates both the onscreen display of the dictionaries and the targeted accessing of the specific information types. The markup conforms to the internationally accepted guidelines of the Text Encoding Initiative (TEI). All in all, the aim is to ensure long-term, cross-platform availability of the data. A graphical user interface enables access to the digitized data and allows searching in both the individual dictionaries and the network. As shown in Figure 33.4, dynamically generated and classified network information can be called up in the right-hand frame: both the lemma on which the cross-reference is based and the catalogue number of the reference target are displayed. Different types of cross-reference are distinguished: explicit cross-references which appear in the printed dictionaries, so-called symmetrical cross-references (A⫺B, B⫺A), transitive cross-references (A⫺B⫺C, A⫺C) and automatically generated (untested) cross-references.

33. Dialect dictionaries ⫺ traditional and modern

Fig. 33.5: Map module in the Digitaler Verbund von Dialektwörterbüchern (DWV). The map displays the survey locations named in the respective article. Depending on the zoom level, further information is available (rivers, additional place names, etc.)

An additional step is the implementation of a map module ⫺ initially in the internet version of the ‘Palatine German Dictionary’ (Pfälzisches WB). This map module enables the user to visualize the distribution of an article using dynamically generated maps of the dictionary domain. Searching in the reverse direction is also possible: the user can call up all the articles for a specific survey location from a map. In a further step, the linking of selected dictionary entries and existing georeferenced or newly generated isoglossic maps with the Digital Wenker Atlas (DiWA) is planned; here too, reverse searches will be possible. The dictionary network can be gradually expanded to include further dialect dictionaries or other components, e.g., corpus data or dialect works of fiction. Interlinking the dictionaries, as in the conversion of existing cross-references between the works to hyperlinks and the generation of new cross-references at a semantic level, should prove onomasiologically revealing (cf. Figure 33.6) ⫺ enabling us to take a step

603

604

VI. Data analysis and the presentation of results

Fig. 33.6: Screenshot of an onomasiological search (Aalraupe) in the Digitale Verbund der Dialektwörterbücher (DWV)

in the direction suggested by Reichmann (1986: 176; my translation), “to render semasiological dictionaries onomasiologically useful”. A content-based cross-linking of lemmata from dialect dictionaries via the table of contents is also a feature of the printed work, Wörterbuch Deutscher Dialekte (‘Dictionary of German Dialects’, Knoop 2001), in which selected terms from a total of ten dialects are presented in a short synoptic overview via a Standard German lemma. In comparison with printed projects, the advantage of the internet-based linkage of the dialect network lies in the fact that all existing dialectlexicographic materials can be included completely and dynamically and can always be extended with further components. The next step is obvious, the integration of the network of dialect dictionaries into an even larger dictionary network. Taking the headword Butterblume ‘buttercup’ (Figure 33.7) as an example; the Trierer Wörterbuchnetz (‘Trier Dictionary Network’) can be used to illustrate how such a network of linguistic information can be formed after using computer-philological procedures to edit dictionary information. Here, Butterblume is linked to semantically corresponding headwords from different dictionaries, dialectal and other, such as Ankenbluem or Wibelewick, and to the expressively similar but orthographically different Butter⫽Blume from the Krünitzsche Enzyklopädie ‘Krüntiz’s Encyclopedia’. The clear distinction between the different types of cross-reference and their qualitative classification needs to be taken into consideration; this is a problematic area

33. Dialect dictionaries ⫺ traditional and modern

Fig. 33.7: Cross-references for the Butter-blume entry in the ‘Palatine German Dictionary’, sorted by type

that ought to be tackled with comprehensive, dynamically generated meta-lemma lists for instance (cf. Figure 33.7). Such a dictionary network, if it is to be effective and established in a feasible time frame, can only be set up with the aid of algorithmic methods drawn from the field of information retrieval (on this, see Burch and Rapp 2007). Ideally, the philological appraisal of automatically generated references could be conducted by active dictionary projects in collaboration: as a rule, dictionary editors consult other (dialect) dictionaries during their work. If they were to do this using an electronic version, they could confirm existing cross-linkages or establish new ones and thus thicken the net. Another essential instrument for the documentation of dialectal variation and research into variant texts is the concept of a “meta-lemma list”, which enables standardized access to the material. Preparatory work towards a meta-lemma list of German was conducted in the TextGrid project (); this is being intensively pursued further by an internationally oriented project consortium. One result of this preparatory effort is the TextGrid lemma search service implemented in the Trierer Wörterbuchnetz (), which makes it possible to search globally in the existing digital dictionaries using WebService technology.

605

606

VI. Data analysis and the presentation of results

5.2 Computer-based inormation systems and dialect lexicography in the past and present The electronic interlinking of dictionaries not only provides the opportunity to relate (in this case dialect-lexicographic) works with a similar content to each other and to thus achieve an areal, preferably comprehensive collection of dialectal or regional vocabulary; it also makes possible the collection of recent and historical lexicographic data for a language or a dialect. The LexicoLux project follows such a twofold path in making the existing dictionaries of Luxembourgish available electronically and linking and capturing them dynamically. Luxembourgian lexicography can be traced back to the nineteenth century. The codificatory/lexicographic literature thus arises at almost the same time as Luxembourgish emerges as, from a sociolinguistic point of view, an independent language (cf. Moulin 2006). The first dictionary, published in 1847, was entitled Lexicon der Luxemburger Umgangssprache (‘Lexicon of the Luxembourgian Vernacular’, Gangler 1847). This monograph, with its collection of interesting facts, is completely in keeping with the style of the conversational dictionaries of the time: its primary objective is not just to capture Luxembourgish; other languages are also included. In 1897, a state commission was established to create a new Luxembourgian dictionary, which was published in 1906 in one volume under the title Wörterbuch der luxemburgischen Mundart ‘Dictionary of the Luxembourgish Dialect’. The dictionary was conceived as a component in a more comprehensive project; like its predecessor, it essentially describes the vocabulary of the capital and the lower Alzet valley. In the context of a new, official undertaking dating from 1935 the most extensive lexicographic documentation of the Luxembourgian language to date was created: the Luxemburger Wörterbuch ‘Luxembourgian Dictionary’ (LWB), published between 1950 and 1977 in the tradition of the broad-scale dialectal dictionaries. The LWB data basis is in line with that of the earlier dictionaries; the data collection is primarily drawn from nineteenth-century fiction and other vocabulary as well as surveys of its own. Furthermore, it documents diatopic variation in detail. Predominantly, endogenous vocabulary (including many proper names) was included, with additional borrowings from High German and French insofar as they had been phonetically or morphologically integrated (cf. Rinnen 1976). In creating a dynamic and multi-directional dictionary network, the LexicoLux project not only documents the Luxembourgian language from its nineteenth-century beginnings, it also provides a basis for further linguistic comparison and analysis. Structural analyses can thus be performed on existing materials, focusing on diatopic variation in the vocabulary, Romanic elements, phraseological structures or literary elements, for example. The headwords of different reference works can be consolidated using a standardized “hyperlemma list” of Luxembourgish (see above), so as to create an integrated access point for the differently treated lemmas (cf. Figure 33.8). In the case of the Luxembourgian dictionaries, not only are the etymology and semantics of the headwords captured, the contents of the individual articles are also analyzed in order to be able to relate the differing treatments of lemmas to each other correctly, especially since the dictionaries also embody three distinct orthographic systems. In addition, the language-internal components of analysis are augmented by the linking of the Luxembourgish lexicographic data with that on neighboring varieties held in the network of dialect dictionaries. National borders (which often hinder dialectological

33. Dialect dictionaries ⫺ traditional and modern

Fig. 33.8: Online version of the Lexikon der Luxemburger Umgangssprache (Gangler 1847), with lemma list (left), individual entries (middle) and automatically generated cross-references to related dialect dictionaries and others (right)

projects) can thus be surmounted and common vocabulary structures and culturalhistorical relationships better detected. The principle of interoperability also extends to external resources. In the case of the LWB, the survey locations can be used to determine a precise areal distribution, to generate a cartographic visualization, or to link to the Digitaler Luxemburgischer Sprachatlas (‘Digital Luxembourgish Linguistic Atlas’, ; cf. Gilles and Moulin 2008). Alongside its connection to existing dictionary and linguistic atlas resources, LexicoLux can also be accessed from external electronic documents. Firstly, a permanent web address can be obtained via the icons created for each word article, and secondly, the project is happy to supply the complete lemma lists with the relevant addresses on request. Finally, in the context of such a dictionary network, there is the opportunity for productive links from the dictionary entries to yet other texts, such as electronic text editions of primary sources. Moreover, all other kinds of primary source (audio, video, maps, notes, diagrams etc.) can be easily linked with lemmas or stretches of text. The possibilities outlined taking the Luxembourgish dictionaries as an example can be applied prototypically to other dictionary networks.

5.3 The dictionary administration as a source and store o knowledge Traditional dictionary administrations (Wörterbuchkanzleien), with their archives that have often been accumulating over decades or even longer, are a rich dialectological and

607

608

VI. Data analysis and the presentation of results linguistic hoard ⫺ not just from a narrowly lexicographical point of view. The data collected, including citation cards, questionnaires, (annotated) text corpora, maps and much more, some of it unpublished, represents an accumulation of knowledge that, with regard to the documentation of individual dialects, is in part unique. These archives, which form the heart of every dictionary project, are usually set up for internal use within the individual projects and are easily accessed by the editors. To an extent, the core data for the creation of corpora and dictionary articles is electronically stored and processed. In recent times, some projects have also set out to document, structure and offer online their material (or part of it) for further scholarly use by third parties. The online presence of the Bayerische Wörterbuch (‘Bavarian Dictionary’, BWB), for example, includes a map archive featuring a digitalized selection of Eberhard Kranzmayer’s hand-drawn linguistic maps from the 1920s and 1930s (). The BWB in Munich possesses more than 2000 such dialect maps, the digitalization of which would definitely be of interest, either linked with the BWB or in the framework of the dialect-cartographic information systems of existing linguistic atlas projects. The Wörterbuch der bairischen Mundarten in Österreich (‘Dictionary of the Bavarian Dialects in Austria’, WBÖ), historically related to the BWB, has since 1993 been building an electronically sortable dialect database (DBÖ), which represents a comprehensive dialectal documentation system for the Austrian language (). The database (so far only accessible from within the institute, but to be offered across the internet in the future course of the project) provides above and beyond the publication, exhaustive information about the phonetic appearance, individual definitions (both in light of regional, temporal and social factors), onomasiological interrelations within the semantic field (information on related meanings) and usage of the lemma in dialectal syntagmatic contexts […]. Furthermore, it contains citations and figures of speech and idioms. Factual information, illustrations and detailed descriptions about instruments or occupational procedures, traditions, popular beliefs, traditional medicine and the like are retrievable from the database in a structured way. (Städtler 2003: 335; my translation; see also Geyer 2000; Wandl-Vogt 2006)

Thus prepared, the dictionary databases and materials can, as well as providing an archive, become the starting point for further research, not just in linguistics or dialect geography. Theoretically, the advantages offered by computer-philological procedures and online publication can also be put to use in dictionary publication itself. For example, it is possible to make dictionary articles, which are often published in an abbreviated form in the printed version, available in their full length online (where they can also be updated regularly). Such a procedure, freed from the constraints of space, is probably most relevant for the newer (and seldom) large-scale lexicographic projects within the classical tradition (see, for example, Sappler [2000: 387⫺396] on the data-processing design of the new Middle High German dictionary). Ideally, arising from and building upon a dictionary project, a dynamic dialect geography and lexicography information system can be established, opening up myriad interdisciplinary opportunities for linguists, computer philologists, cartographers, ethnologists, etc. The fundamental preconditions for such complex information systems are the quality and interoperability of the data, ensured by the development and implementation of international standards, plus a willingness to cooperate within the academic community.

33. Dialect dictionaries ⫺ traditional and modern

6. Reerences Bauer, Werner 1996 Die Etymologie in der deutschen Dialektlexikographie. In: Bremer and Hildebrandt (eds.), 219⫺229. Baur, Gerhard W. 1986 Quellen und Corpora. Zur Materialbasis deutschsprachiger Dialektwörterbücher. In: Friebertshäuser (ed.), 75⫺91. Berthold, Luise 1924/1925 Die Wortgeographische Forderung und die Programme der modernen deutschen Mundartwörterbücher. Teuthonista 1: 222⫺226. Bergmann, Gunter 1994 Einführung. In: Wörterbuch der obersächsischen Mundarten, vol. 3, IX⫺XVIII. Berlin: Akademie. Besch, Werner, Ulrich Knoop, Wolfgang Putschke and Herbert Ernst Wiegand (eds.) 1983 Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 2. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.2.) Berlin/New York: de Gruyter. Bremer, Ernst and Reiner Hildebrandt (eds.) 1996 Stand und Aufgaben der deutschen Dialektlexikographie, vol. 2: Brüder-Grimm-Symposion zur Historischen Wortforschung. Beiträge zu der Marburger Tagung vom Oktober 1992. Berlin/New York: de Gruyter. Burch, Thomas and Johannes Fournier 2004 Vom Buch zur elektronischen Publikation ⫺ Textdigitalisierung auf der Basis von SGML/XML. In: Alexander Mehler and Henning Lobin (eds.), Automatische Textanalyse. Systeme und Methoden zur Annotation und Analyse natürlichsprachlicher Texte, 265⫺ 283. Wiesbaden: VS Verlag für Sozialwissenschaften. Burch, Thomas and Andrea Rapp 2007 Das Wörterbuch-Netz: Verfahren ⫺ Methoden ⫺ Perspektiven. In: Geschichte im Netz: Praxis, Chancen, Visionen. In: Daniel Burckhardt, Rüdiger Hohls and Claudia Prinz (eds.), Beiträge der Tagung .hist 2006, 607⫺627. (Historisches Forum 10: I.) Available online from the edoc-server of the Humboldt Universität zu Berlin: . Fournier, Johannes 2003 Vorüberlegungen zum Aufbau eines Verbundes von Dialektwörterbüchern. Zeitschrift für Dialektologie und Linguistik 70: 155⫺176. Friebertshäuser, Hans 1976 Relevante Aspekte der Dialektlexikographie. In: Friebertshäuser (ed.), 5⫺10. Friebertshäuser, Hans 1983 Die großlandschaftlichen Wörterbücher der deutschen Dialekte. Areale und lexikologische Beschreibung. In: Besch et al. (eds.), 1283⫺1295. Friebertshäuser, Hans 1986 Zu Geschichte und Methoden der deutschen Dialektlexikographie. In: Friebertshäuser (ed.), 1⫺13. Friebertshäuser, Hans (ed.) 1976 Dialektlexikographie. Berichte über Stand und Methoden deutscher Dialektwörterbücher. Wiesbaden: Steiner. Friebertshäuser, Hans (ed.) 1986 Lexikographie der Dialekte. Beiträge zu Geschichte, Theorie und Praxis. Tübingen: Niemeyer.

609

610

VI. Data analysis and the presentation of results Gangler, Jean-Franc¸ois 1847 Lexicon der Luxemburger Umgangssprache (wie sie in und um Luxemburg gesprochen wird). Luxemburg: Hoffman. Geyer, Ingeborg 2000 Die digitale Dialektdatenbank Österreichs (DBÖ) und das Wörterbuch der bairischen Mundarten in Österreich (WBÖ). In: Protokoll des 78. Kolloquiums über die Anwendung der Elektronischen Datenverarbeitung in den Geisteswissenschaften an der Universität Tübingen vom 5. Februar 2000. Available online from: . Gilles, Peter and Claudine Moulin 2008 Der digitale luxemburgische Sprachatlas (LuxSA). Stand und Perspektiven. In: Stephan Elspaß und Werner König (eds.), Sprachgeographie digital. Die neue Generation der Sprachatlanten, 133⫺147. (Germanistische Linguistik 190⫺191.) Hildesheim: Olms. Große, Rudolf (ed.) 1998 Bedeutungserfassung und Bedeutungsbeschreibung in historischen und dialektologischen Wörterbüchern. Beiträge zu einer Arbeitstagung der deutschsprachigen Wörterbücher, Projekte an Akademien und Universitäten vom 7. bis 9. März 1996. Stuttgart/Leipzig: Hirzel. Haas, Walter 1994 “Die Jagd auf Provinzial-Wörter”. Die Anfänge der wissenschaftlichen Beschäftigung mit den deutschen Mundarten im 17. und 18. Jahrhundert. In: Klaus Mattheier and Peter Wiesinger (eds.), Dialektologie des Deutschen. Forschungsstand und Entwicklungstendenzen, 329⫺365. (Reihe Germanistische Linguistik 147.) Tübingen: Niemeyer. Haas, Walter 1996 Über die Idiotismensammlungen des 18. Jahrhunderts. In: Bremer and Hildebrandt (eds.), 175⫺190. Haß, Ulrike (ed.) 2005 Grundfragen der elektronischen Lexikographie: elexiko ⫺ das Online-Informationssystem zum deutschen Wortschatz. Berlin/New York: de Gruyter. Hausmann, Franz Josef, Oskar Reichmann, Herbert Ernst Wiegand and Ladislav Zgusta (eds.) 1990 Wörterbücher: Ein internationales Handbuch zur Lexikographie / Dictionaries: An International Encyclopedia of Lexicography / Dictionnaires: Encyclope´die internationale de lexicographie, vol. 2. (Handbücher zur Sprach- und Kommunikationswissenschaft 5.2.) Berlin/ New York: de Gruyter. Keseling, Gisbert, Bernd-Ulrich Kettner, Wolfgang Kramer, Wolfgang Putschke, Monika RössingHager and Ulrich Scheuermann 1970 Richtlinien zur Ablochung und zentralen Speicherung mundartlichen Wortmaterials des Deutschen. Germanistische Linguistik 2: 179⫺242. Klein, Wolfgang and Alexander Geyken 2000 Projekt Digitales Wörterbuch der deutschen Sprache des 20. Jahrhunderts (DWDS). In: Jahrbuch 1999 / Berlin-Brandenburgische Akademie der Wissenschaften, 277⫺289. Berlin: Akademie. Klosa, Annette 2007 Korpusgestützte Lexikographie: besser, schneller, umfangreicher? In: Werner Kallmeyer and Gisela Zifonun (eds.), Sprachkorpora. Datenmengen und Erkenntnisfortschritt, 105⫺ 122. Berlin/New York: de Gruyter. Knoop, Ulrich 2001 [1997] Wörterbuch deutscher Dialekte. Eine Sammlung von Mundartwörtern aus zehn Dialektgebieten im Einzelvergleich, in Sprichwörtern und Redewendungen. Köln: Parkland [München: Bertelsmann]. König, Werner 2007 dtv-Atlas Deutsche Sprache. 16th ed. Munich: Deutscher Taschenbuch Verlag (dtv)

33. Dialect dictionaries ⫺ traditional and modern Kühn, Peter 1978 Deutsche Wörterbücher. Eine systematische Bibliographie. Tübingen: Niemeyer. Kühn, Peter 1982 Typen lexikographischer Ergebnisdarstellung. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke and Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 1, 703⫺723. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.1.) Berlin/New York: de Gruyter. Kühn, Peter and Ulrich Püschel 1990 Die deutsche Lexikographie vom 17. Jahrhundert bis zu den Brüdern Grimm ausschließlich. In: Franz Josef Hausmann et al. (eds.), 2049⫺2077. Leibniz, Gottfried W. 1697 [1983] Unvorgreifliche Gedanken, betreffend die Ausübung und Verbesserung der deutschen Sprache. In: Uwe Pörksen (ed.), Unvorgreifliche Gedanken, betreffend die Ausübung und Verbesserung der deutschen Sprache. Zwei Aufsätze, 5⫺46. Stuttgart: Reclam. Löffler, Heinrich 2003 Dialektologie. Eine Einführung. Tübingen: Narr. Moulin, Claudine 2006 Grammatisierung und Standardisierung des Le¨tzebuergeschen. Eine grammatikographisch-sprachhistorische Annäherung. In: Claudine Moulin and Damaris Nübling (eds.), Perspektiven einer linguistischen Luxemburgistik. Diachronie und Synchronie, 305⫺339. Heidelberg: Winter. Niebaum, Hermann 1979 Deutsche Dialektwörterbücher. Deutsche Sprache 7: 345⫺373. Niebaum, Hermann and Jürgen Macha 2006 Einführung in die Dialektologie des Deutschen. 2nd ed. Tübingen: Niemeyer. Popowitsch, Johann Siegmund Valentin 1780 Versuch einer Vereinigung der Mundarten von Teutschland: als eine Einleitung zu einem vollständigen Wörterbuche mit Bestimmungen der Wörter und beträchtlichen Beiträgen zur Naturgeschichte. Wien: von Kurzböck. Post, Rudolf 1986 Zettel und EDV. Methodische und praktische Probleme beim Einsatz von EDV in einem laufenden Dialektwörterbuchunternehmen. In: Friebertshäuser (ed.), 115⫺123. Post, Rudolf 1998 Möglichkeiten der elektronischen Strukturierung, Vernetzung und Verfügbarmachung von lexikographischen Daten bei der Arbeit am Pfälzischen Wörterbuch. In: Große (ed.), 211⫺220. Putschke, Wolfgang 1994 Überlegungen zur Konzeption eines computerdialektologischen Arbeitsplatzes. In: Klaus Mattheier and Peter Wiesinger (eds.), Dialektologie des Deutschen. Forschungsstand und Entwicklungstendenzen, 245⫺255: Tübingen: Niemeyer. Re´zeau, Pierre 1990 Le dictionnaire dialectal: l’exemple franc¸ais. In: Hausmann et al. (eds.), 1467⫺1475. Reichmann, Oskar 1983 Untersuchungen zur lexikalischen Semantik deutscher Dialekte: Überblick über die theoretischen Grundlagen, über die Sachbereiche und den Stand ihrer arealen Erfassung. In: Besch et al. (eds.), 1295⫺1325. Reichmann, Oskar 1986 Die onomasiologische Aufbereitung semasiologischer Dialektwörterbücher. Verfahrensvorschlag und Nutzen. In: Friebertshäuser (ed.), 173⫺184. Richey, Michael 1755 Idioticon Hambvrgense oder Wörter-Buch, Zur Erklärung der eigenen, in und um Hamburg gebräuchlichen Nieder-Sächsischen Mund-Art. Hamburg: Koenig.

611

612

VI. Data analysis and the presentation of results Rinnen, Henri 1976 Luxemburger Wörterbuch (1950⫺1975). In: Friebertshäuser (ed.), 65⫺68. Rowley, Anthony R. and Hans Ulrich Schmid 1996 Grundsätzliche Überlegungen zu historischen Quellen und zum historischen Material des neuen Bayerischen Wörterbuchs. In: Bremer and Hildebrandt (eds.), 259⫺264. Sappler, Paul 2000 Prinzipien des EDV-Konzepts. In: Kurt Gärtner and Klaus Grubmüller (eds.), Ein neues Mittelhochdeutsches Wörterbuch. Prinzipien, Probeartikel, Diskussion, 387⫺396. (Nachrichten der Akademie der Wissenschaften in Göttingen. I. Philologisch-Historische Klasse, 8.) Göttingen: Vandenhoeck & Ruprecht. Scheuermann, Ulrich 1974 Die Sprachkarte im Dienste des Dialektwörterbuches. Niederdeutsches Wort 18: 70⫺90. Schmeller, Johann Andreas 1827⫺1837 Bayerisches Wörterbuch. Stuttgart: Cotta. Schmeller, Johann Andreas 1872⫺1877 Bayerisches Wörterbuch. 2nd ed. Munich: Rudolf Oldenbourg. Schröder, Martin 1997 Brauchen wir ein neues Wörterbuchkartell? Zu den Perspektiven einer computerunterstützten Dialektlexikographie und eines Projekts “Deutsches Dialektwörterbuch”. Zeitschrift für Dialektologie und Linguistik 64: 57⫺66. Stadler, Franz Joseph 1806⫺1812 Versuch eines Schweizerischen Idiotikon mit etymologischen Bemerkungen untermischt. Samt einer Skizze einer Schweizerischen Dialektologie. Aarau: Sauerländer. Städtler, Thomas 2003 Wissenschaftliche Lexikographie im deutschsprachigen Raum. Heidelberg: Winter. Wandl-Vogt, Eveline 2006 Von der Karte zum Wörterbuch ⫺ Überlegungen zu einer räumlichen Zugriffsstruktur für Dialektwörterbücher. Dargestellt am Beispiel des Wörterbuchs der bairischen Mundarten in Österreich (WBÖ). In: Elisa Corino, Carla Marello and Cristina Onesti (eds.), Atti del XII Congresso internazionale di Lessicografia. Torino, 6⫺9 Settembre 2006, vol. 2, 721⫺732. Alessandria: Edizioni dell’ Orso. Wiegand, Herbert E. 1990 Die deutsche Lexikographie der Gegenwart. In: Hausmann et al. (eds.), 2100⫺2246. Wiegand, Herbert E. 1998 Historische Lexikographie. In: Werner Besch, Anne Betten, Oskar Reichmann and Stefan Sonderegger (eds.), Sprachgeschichte. Ein Handbuch zur Geschichte der deutschen Sprache und ihrer Erforschung, 643⫺715. 2nd ed. (Handbücher zur Sprach- und Kommunikationswissenschaft 2.1.) Berlin/New York: de Gruyter. Wörterbuchkommission (ed.) 1950⫺1977 Luxemburger Wörterbuch (LWB). Im Auftrage der Großherzoglich Luxemburgischen Regierung. 5 vols. [Reprinted 1995 in 2 vols]. Luxembourg: Linden.

Claudine Moulin, Trier (Germany)

34. Community-based investigations

34. Community-based investigations: From traditional dialect grammar to sociolinguistic studies 1. Introduction: Genealogical and positivistic perspectives on dialect and dialect variation 2. Dialect grammars: The speakerless, homogeneous speech community 3. Sociolinguistic correlation: The heterogeneous stratified speech community and the group grammar 4. Beyond the group grammar: The speaker’s language 5. References

1.

Introduction: Genealogical and positivistic perspectives on dialect and dialect variation

1.1. Objective Current sociolinguistic research based on realistic models of a language has led to significant progress toward a fuller understanding of how language is used in society and how it changes. However, despite appearances, this trend has upheld a conservative line of dialectological thought based on the old idea of a therapeutic defense of the genuine variety against modernization, on organic solidarity and on the standard variety. In fact, in the main research trend at least, variation is considered as the natural context of authenticity alone, i. e., the vernacular variety closely related to each speaker. Contemporary critical (interpretive, microsociolinguistic, etc.) vindication of the concepts of mechanical solidarity and speaker emancipation, however, has revived the original nonconformist stance of sociolinguistics and uncovered a hidden line of historical development in dialect research. Alternative perspectives on dialect variation reject positivism and interpret nonstandard varieties as a source of emancipation rather than one of alienation and obstacles to progress. In this article we seek to gain a clearer understanding of the present-day approach to dialect and dialect variation by considering its origins and development out of comparative historical and traditional dialectological research. We retrace a non-chronological development of social dialectology (section 1), focusing on the evolution of traditional dialect grammars through sociological and linguistic monographs to sociolinguistics (section 2). From this standpoint, current variationist correlational research is seen as an up-to-date quantitative version of earlier models (section 3). Criticism of Labovian theory and methods suggests the need for a comprehensive reinterpretation of its entire development and for a proposed critical social dialectology (section 4).

1.2. Genesis o social dialectology 1.2.1. Language variation: The contemporary version o an old idea The current commonsense view of a language is that of an orderly heterogeneous system of varieties along four dimensions: diachronic, diatopic, diastratic and diaphasic (Cos¸e-

613

614

VI. Data analysis and the presentation of results riu 1981; Rona 1970: 204⫺206; Klein 1974: 18⫺71; Kubczak 1979: 11⫺53; see Berruto in this volume). However, acceptance of this view is relatively recent among linguists. The main endeavor of nineteenth and twentieth-century (pre-)dialectological researchers into diachronic variation had been to prove the existence of a grammar of exceptionless sound laws (the so-called Ausnahmslosigkeit der Lautgesetze) that characterized each standard language and was capable of explaining the division between languages in terms of a genealogical tree (Stammbaumtheorie; see Murray in this volume). Comparativist historical linguists and Neogrammarians carried out solid empirical work based largely on earlier written data from the major Indo-European languages. However, as soon as certain scholars born into the same tradition (Sievers, Paris), began to look into living spoken diatopic varieties, they realized the need for a comparison between geographical varieties of the same language. The aim was to prove that “sound laws” were regular and applied without exception and, hence, that clearly delimited linguistic borders could be drawn even between these varieties. However, early work on speech use revealed (1) that there is an interplay of language dimensions in each area or local speech community such that diachronic stratification may be perceptible within the geographical space (Gillie´ron 1918), (2) that the homogeneity of the local speech community is very likely to be a myth (Rousselot 1891; Gauchat 1905). Both findings are decisive for the evolution of linguistics and of social dialectology in particular, and respectively represent two different but complementary objectives: the first, the study of language use and its correlation with extralinguistic factors (stratification, ways of communication, accommodation, etc.); the second, the definition of varieties on explicitly empirical foundations.

1.2.2. Key stages: Historical linguistics, traditional and social dialectology The development briefly described above may be tentatively analyzed into three distinct key stages as shown in Figure 34.1, from pre-dialectological schools (I) through traditional dialectology (II) to social dialectology (III). Traditional and contemporary social dialectology are outgrowths of the same historical trend (nineteenth and early twentiethcentury historical comparative linguistics including Schleicher Darwinism and the Neogrammarian school), but they differ in their attitude (decay versus progress) toward the same object of work: 1. Traditional dialectology adopts a pessimistic stance based on the idea that modern urban life leads to dialect attrition and death (ideology of decay) and projects its work as a program of defense against convergence toward the standard language (Berthele 2004: 723⫺724); 2. Variationist sociolinguistics exhibits a more optimistic stance, accepting evolution as progress, not decay (ideology of progress); it views variation as a natural attribute of languages, capable and deserving of empirically based description. Differences between the two perspectives are thus primarily ideological, then theoretical, and finally methodological. On the one hand, traditional dialectological schools (despite their theoretical and methodological differences) have carried out essential work either on dialect geography (direct representation of local variation using vocabularies and maps, from Wenker and Gillie´ron through to the contemporary projects) or on dialect

34. Community-based investigations monographs (reconstructing the prototypical local dialect, from Ascoli and Wegener through to present-day sociolinguistics), which aimed to show that laws of linguistic change are exceptionless and that le patois authentique (the genuine dialect of a local speech community) is the purest form of a language, worth preserving from external contamination and studying as a reservoir of original purity. On the other hand, social dialectology rejects the idea of an allegedly genuine dialect or sociolect and states that its real object of research (variety) needs to be built up from a set of linguistic traits that all correlate with the same extralinguistic factors (speakers’ variables, such as status, sex, age, background, etc.). This required the introduction of quantitative techniques of analysis and, hence, triggered what would later be called a “new paradigm” (Weydt and Schlieben-Lange 1981: 118; Shuy 1990: 198⫺201).

1.2.3. Transitional stages: The role o the predecessors Nevertheless, attention had long been paid to the effect of the speakers’ variables on dialect use (Labov 1972d; Koerner 1988: 1991; Dittmar 2004), from the time dialectologists became aware that there were likely to be isoglosses similar to those encountered in geographical space in social space (status, age or sex markers), as well as in its reflection in the linguistic performance (register) of individual speakers. A line of development connecting dialectological schools and practitioners that became aware of the importance of social factors may be traced as a non-chronological path leading to the contemporary social sciences of language (Bellmann 1986; Berthele 2004). This line includes every school or trend challenging the apriority of a homogeneous local speech community as theorized by the Neogrammarians. The essential role played by the nineteenth and early twentieth-century schools (Marburg, French School), as well as contributions by individual dialectologists (Schuchardt, Rohlfs, Wagner, Rousselot, Gauchat), need no emphasis (see Iordan 1967: 76⫺127, 251⫺389, 507⫺672; Pop 1950: I, xxiii⫺lv, 1⫺156, 337⫺434, 467⫺618, II: 737⫺782, 792⫺834 et passim). In fact, the argument of an avant la lettre sociolinguistic stance underlies every reconstruction of national sociolinguistic idiosyncrasy where a synthesis between each country’s tradition and quantitative models imported from the United States is proposed (see Malkiel 1976: 1984; Calvet 2003), at

Fig. 34.1: Genesis of social dialectology

615

616

VI. Data analysis and the presentation of results least in Europe (Alvar 1969; Cortelazzo 1968; Marcellesi and Gardin 1974; SchliebenLange 1976; Dittmar 1976; etc.). However, early socially conditioned studies on dialects, like those designed by Gauchat or Rousselot, never abandoned the traditional dialectological leitmotif: the conservation of the purest form of nonstandard varieties. Only as dialectologists began to realize that a speech community has to be thoroughly described with attention to every speaker and every source of variation can a certain theoretical disquiet be perceived. In addition, the influence of structural dialectology (Martinet, Weinreich) completed the emergence of a new way of doing dialect research.

2. Dialect grammars: The speakerless and homogeneous speech community 2.1. Concept: The study o the genuine community dialect There are two ways of analyzing diatopic variation (and hence two corresponding tasks for dialectology): firstly, through idealized models of the existing dialect continua, devised for the description of dialect diversification (dialect geography, dynamic models, models of spatial diffusion, etc.); secondly, by defining dialects as syntopic systems (dialect grammars). Since the second half of the nineteenth century, the study of geographical variation in language has focused on the drawing of onomasiologically conceived linguistic maps. Dialect geography continues to function as a model for representing diatopic variation, and linguistic atlases act as a complementary source of data for the study of linguistic variation and change (Milroy and Gordon 2003: 15⫺22; Chambers and Trudgill 1998: 89⫺123, 166⫺186). However, dialect maps are not the only means of studying dialect variation. Traditional dialectology also produced complete and precise phonetic descriptions of dialects as well as models of their phonological organization. In their dialect grammars [Ortsgrammatiken], dialectologists described the relationship between phonology and lexicon in full detail, worked on suprasegmental phonology, morphology and syntax, and obtained the oldest, most reliable, and most comprehensive dialectal data available to us today (Petyt 1980: 37⫺67; Chambers and Trudgill 1998: 13⫺31). Dialect monographs emerged as a means of solving some of the problems and insufficiencies that had been discovered in dialect geography since Wenker’s and Gillie´ron’s seminal ventures. The need for a deeper insight into the speech community’s patois authentique and the desire to thoroughly analyze each dialect’s grammar gave rise to a long tradition of specialized monographs in both Germanic (Bach 1969: 214⫺226) and Romance domains (Pop 1950: 45⫺56, 303⫺306, 409⫺427, 525⫺530 et passim; Iordan 1967: 55⫺60; 408⫺423; Malkiel 1976: 61⫺62). The local base dialect [Ortsmundart] was to be the smallest and most regular unit with its own independent history. The canonical form for dialect monographs had been well defined by Wegener and Ascoli, the former following Sievers and Winteler, and the latter in line with Gaston Paris (Pop 1950: II, 740⫺741; I, 45⫺50, 525⫺526; see Murray in this volume). The catalog of tasks includes the following:

34. Community-based investigations 1. Data gathering: a diachronic phase of the mother tongue is taken as a comparative reference base, since the aim is to relate the diversity evident in every phonetic divergence back to the relevant phonetic laws (Pop 1950: I, 525⫺530; II, 740⫺741). 2. To make a diachronic study possible, dialect monographs needed to include precise analyses of phonetics and grammar; this entailed a comprehensive collection of words covering every consonant and vowel and required either close acquaintance with the local dialect or the use of well-selected bona fide informants. 3. The idea was to obtain the complete grammar for each dialect (including every component) and the historical basis for interdialectal comparison (i. e., the source form and the corresponding Lautgesetze). 4. On this basis it was very unlikely that serious consideration would be given to underlying questions such as the informants’ representativity (Berthele 2004) or the method’s atomism (Weydt and Schlieben-Lange 1981).

2.2. Evolution: Sociological and structural monographs Early sociological criticism of traditional skewed samples of informants and linguistic criticism of atomism was voiced and developed quickly. Both trends have been labeled above (section 1.2.3) as transitional and are retrospectively understood as antecedents of the new sociolinguistic era (Petyt 1980: 101⫺116; Chambers and Trudgill 1998: 32⫺ 44). Works by Rousselot (1891) and Gauchat (1905), among others (Wegener, Raumer, etc.) are considered representative of the former, while Martinet (1939) and Weinreich [1954] lead the latter (see Figure 34.1). Traditional diachronic dialect grammars then evolved toward both social-centered or sociological monographs (section 2.2.1) and language-centered or structural (linguistic) monographs (section 2.2.2).

2.2.1. Sociological monographs In fact, as pointed out by Chambers and Trudgill (1998: 30), the real and “greatest single source of disaffection” with traditional dialectology among contemporary social dialectologists is “the narrow choice of informants” (Malkiel 1976: 70⫺73). As the aim of the research was the oldest stage of the genuine dialect, its data source excluded a large proportion of the community’s population. Consequently, the traditional ideal informant would be “une personne pour qui le parler des aı¨eux constitue encore le seul moyen de communiquer non seulement avec sa propre famille, mais aussi avec les gens du meˆme pays” (Pop 1950: II, 1158). Unrepresentative groups of non-mobile older rural male speakers (NORMS), in combination with the dialectologist’s own intuition, were thus the only source of data. Such data collection procedures are conceivable only on the basis of an understanding of “space” that implies strong idealization, i. e., a belief that the other linguistic dimensions do not affect the homogeneity of dialects. Traditional dialect monographs should thus be considered in relation to a particular historical period, a particular type of community, and even a predominant ideology, though this does not mean being blind to the effect of society on speech use. As mentioned in section 1.2.3, schools opposed to the Neogrammarians (the French School, Marburg, etc.) did consider the inclusion of social parameters (such as sex, education,

617

618

VI. Data analysis and the presentation of results mobility, etc.) as factors likely to constrain dialect use, but only to demonstrate the threat of dialect alienation. This utopist ideology upholds the implicit assumption that progress and urbanization cause dialect alienation (Bellmann 1986: 22⫺42; Berthele 2004: 732⫺735; see section 3.1). However, homogeneous traditional rural local speech communities, based on strong network links between speakers, mechanical solidarity and behavioral similarity (including linguistic behavior), were evolving into contemporary stratified large urban speech communities where the division of labor and functions is reflected in stratified linguistic use. This would force community grammar research to accept social science standards as theoretical and methodological foundations and become, properly speaking, social dialectology (see section 3). Nevertheless, the acceptance of such standards does not entail the rejection of the mountains of rich data these monographs contain; on the contrary, the historical perspective outlined above should be used to assess this traditional work positively, rather than disqualify it on the basis of contemporary sociolinguistic advances. Growing acceptance of the key sociological procedures among dialect monograph researchers began to change the whole scenario. The myth of the homogeneous speech community (as well as the corresponding idea of a consistent idiolect) was exploded and by the second half of the twentieth century the idea of heterogeneity was fairly well established. A series of monographs appeared in an attempt to adapt to the new requirements; in fact, the transition from dialect monographs to sociolinguistic studies was relatively smooth until Labov’s groundbreaking research on Martha’s Vineyard [1963] and New York City [1966] appeared. In sum, this qualitative shift should be taken as the final product of a long chain of quantitative changes affecting theory and, above all, method. Two main aspects should be borne in mind: 1. The reliability and validity of the findings. Gradually, contaminated informants tended to be replaced by representative samples of speakers. Increasing attention paid to variation led to a detachment from both unrepresentative groups of NORM speakers and, apparently, from genuine dialects. The introduction of speaker variables made samples more complicated and forced researchers to be more accurate in their fieldwork strategies (Petyt 1980: 110⫺116) and correlational analyses (Chambers and Trudgill 1998: 45⫺53; Petyt 1980: 132⫺170). 2. The context. Dialect grammarians’ attention began to focus on urban research. At first, urban monographs were conducted just as in traditional dialectology, selecting subject and informants, as in Sivertsen’s (1960) study of Cockney phonology, De Camp’s (1958/1959) research on San Francisco speech or Viereck’s (1966) research on Gateshead “pure” dialect (see Chambers and Trudgill 1998: 46⫺47; Milroy and Gordon 2003: 15⫺22). However, most of the features that characterize contemporary variation research gradually emerged in these monographs. On the one hand, European urban research by Se´guy [1950] or socially-oriented geographical studies by Bach [1950] and Va`rvaro [1972⫺1973] cannot be separated from Gauchat’s and Rousselot’s legacy. On the other hand, American contributions by Kurath et al. (1939) and McDavid (1948) owe a great deal to Jaberg and Jud’s Sprach- und Sachatlas Italiens und Südschweiz (AIS, 1928⫺1960). All these lines of development converge in the essential work by Weinreich (Weinreich, Labov/Herzog 1968). This pre-sociological work is being reassessed by contemporary interpretive researchers (Sankoff, Williams, Auer), in line with more comprehensive criticism of correlational sociolinguistics’ shortcomings (see section 4).

34. Community-based investigations

2.2.2. Linguistic monographs As Malkiel (1976: 67) has pointed out, the cleft between Geneva structuralism on the one hand, and Zurich and Bern/Marburg dialect research on the other, “developed into an ‘isogloss’ of the first magnitude”. As already discussed, Neogrammarian interest in the linguistic (as opposed to cultural) foundations of historical comparative research developed into diachronic atomistic description. The idea of describing the underlying system of units for each dialect without any regard to the strict diachronic relation to the overarching standard language, let alone the unequal social prestige associated with both, seems originally to have been applied by Martinet (1939) in his research on the Franco-Provenc¸al dialect of Hauteville, a commune in Savoy. Work by Lausberg (1939), Weinreich [1954]: Va`rvaro [1972⫺1973] and, among others, Fourquet, Pulgram, Francescato, Catala´n, and Rona (see Alvar 1969: 35⫺53; Malkiel 1976: 67⫺73; Petyt 1980: 117⫺131; Chambers and Trudgill 1998: 32⫺44) reveals the advantages and disadvantages of this method of work. Unfortunately, technical difficulties (how to describe closely related systems in the simplest way?) and ideological obstacles (why consider stigmatized varieties as autonomous systems?) made it very hard to combine linguistic and social interest, as the development of social dialectology would show (see section 3).

3.

Sociolinguistic correlation: The heterogeneous stratiied speech community and the group grammar

3.1. Social dialectology: Signiicance and objective 3.1.1. A new way o analyzing dialects: Linguistic and social depth Social dialectology is the result of an adaptation of traditional dialectology to the new conditions of the twentieth century (particularly to urban stratified communities). The Durkheimian opposition between mechanical and organic solidarity (Durkheim 1902; Berthele 2004: 732⫺734) is an excellent starting point for exploring the transition from traditional to social dialectology. Examples of a mostly rural, closely knit community, functioning through mechanical solidarity and characterized by the relative uniformity of the actors’ e´tats de conscience and activities were vanishing and the idea of them was becoming out-of-date. This increasingly led to the concept of an endangered homogeneous genuine dialect closely related to this type of community proving to be inadequate. A new form of community appeared to be emerging, as a new division of labor and an organic solidarity became established; sharply stratified loose-knit urban communities do not show unified patterns of behavior but instead present an image of heterogeneity which only extralinguistic correlates (occupation, education, age, sex, etc.) could help to organize. The aim therefore became to grasp the regular patterns of linguistic use in the community with the help of these extralinguistic factors and try to describe its varieties. The ideal objective of social dialectology embraces both what we could call social depth (i. e., representative samples both of speakers and linguistic data) and linguistic

619

620

VI. Data analysis and the presentation of results depth (i. e., analysis of the linguistic structure underlying surface data). This means, on the one hand, the redefinition of social dialectology as a realistic discipline, which implies its taking on of an empirical (i. e., statistical and sociological, as opposed to a descriptive [traditional dialectology] or intuitive [formal linguistics]) character. On the other hand, Weinreich’s heritage brought together social and linguistic depth, but Labov’s early work did not pay equal attention to both aspects of linguistic variation, since he assumed that insufficiencies of method (particularly, fieldwork and analysis) came before all else (Labov 1972c, 1984). The early development of American social dialectology therefore showed a strong tendency to correlate large amounts of atomistic (mainly phonetic) data with well-structured social patterns. This may explain the later evolution of variation research.

3.1.2. Objectives: Academic and social Social dialectology aims to describe collective patterns of linguistic behavior (in both large and small speech communities) and how they are reflected in the behavior of individual speakers. Two separate arguments support this aim. The first is scientific or academic: that a better understanding of linguistic performance in its social context requires an expansion of the object so as to capture previously neglected aspects of language; this requires interdisciplinary collaboration (self-liquidating prophecy; see Fishman 1972: 8⫺9; Dittmar 1976: 131). The second argument is a social one: that acquaintance with patterns of linguistic use in a speech community can help to improve the social conditions (wealth, education and development) of speakers. In both arguments, what could be called a therapeutic stance supported by the ideology of social progress based on science and technology is apparent (see Kjolseth 1971: 10⫺22; Dittmar 1989: 23⫺55). Accordingly, virtual equality in society will be achieved if progress and technology manage to tame the world and its dangers. On this basis, mainstream research on social variation in language has been directed toward quantitative sociolinguistic correlation, largely supported by empirical sociological theories and statistical techniques. As pointed out (section 3.1.1), the history of social dialectology may be seen as a statistically and sociologically modified extension of traditional dialectology in which Labov’s (1966) study of the social stratification of English in New York City is taken to be the main reference work. Contemporary correlational research emerges from Labov’s synthesis of, on the one hand, traditional European and American trends of socially-conditioned dialect research brought together by Weinreich and, on the other hand, techniques of fieldwork and analysis from the social sciences and statistics. This is true at least for the core developmental trend (see Figure 34.2). However, this is not the only feasible historical interpretation. As Sankoff (1988) pointed out, an alternative, less overt development is likely to have taken place: instead of considering Labov’s research on New York as the starting point, his work in Harlem (1972b) should be taken as the beginning of something quite different. Social dialectology would thus be seen to have adopted a critical stance toward positivism and scientism, and even social progress. As Habermas (1981) has pointed out, since it is based on technology as a means of dominating the world, progress produces alienation and actually hinders freedom. Real progress for society would be emancipation. Hence, social science should treat science and technology as a means, not an end, and aim above

34. Community-based investigations all to contribute to the speaker’s emancipation. On this basis, the alternative, mainly interpretive, trend should be related to an effectively hidden line of dialectological thought (Schmeller, Schirmunski), sharply opposed to the dominant trend (Berthele 2006: 721⫺722). This tension between social progress and emancipation helps us to understand the further development of the discipline (see section 4 and Figure 34.2). What is interesting and even curious is the contradictory relationship between the positivistic therapeutic position and the concepts of authenticity and progress. As already mentioned, nonstandard varieties should be investigated as a contribution to a better understanding not only of language, but of a community’s social stratification and the consequences of this. Nevertheless, the underlying ideological representation is that nonstandard varieties represent an obstacle to social progress itself, since they affect communication, especially at school (Bernstein, Badura). This would make vernacular varieties the contemporary counterpart of the traditional genuine dialect. In contrast, the interpretive position accepts nonstandard varieties as a means of social emancipation and rejects the idea of an illusory equality of varieties (Bourdieu 1984; Williams 1992).

3.2. Correlational sociolinguistics More than four decades of research into language stratification have proved that there is no reason to maintain a theoretical and methodological separation between dialectology and sociolinguistics. Social dialectology begins when the social nature of space is assumed and social prestige (whatever it might be) is taken as the key concept in a speech community’s dynamics. On the one hand, inter- and intralinguistic variation is then considered to be normal and the object becomes heterogeneous: space is perceived as a variation space (section 3.2.1). On the other hand, heterogeneity also means stratification: varieties elicited from NORM speakers should be replaced with samples from real speakers who represent the whole community (section 3.2.2).

3.2.1. The concept o variation space The task of social dialectology involves describing (1) the different varieties in a certain variation space; (2) relationships between these varieties and (3) relationships between varieties and extralinguistic factors (Klein 1989: 106). The object of social dialectology is no longer assumed to be a natural, homogeneous, well-delimited, genuine variety. On the contrary, varieties must be built up empirically: “we may define a variety of language as a set of linguistic items with similar social distribution” (Hudson 1996: 22). Linguistic features frequently associated with extralinguistic factors, such as sex, social class, ethnicity, regional background or age, turn out to be marked by these extralinguistic variables and so convey or “mean” their respective content (e.g., “female”, “lower working class”, “African American”, “rural”, “middle aged”, etc.). Thus, varieties are understood to be composed of linguistic features which frequently co-occur with the same extralinguistic variants and hence have the same social connotation (Bierwisch 1988). The speech community’s repertoire is thus made up of coexistent hierarchized varieties that are organized with regard to mainstream social prestige in many different ways and according to particular sociohistorical conditions. The overarching standard variety

621

622

VI. Data analysis and the presentation of results unifies the whole configuration (Auer 2005). The idea is that social stratification is likely to be used as a basis for explaining, and even predicting, linguistic variation, since it is assumed that similar social conditions will produce similar linguistic performance. In summary, this perspective implies that (1) division and heterogeneity in society must be accepted and integrated into theory and methodology; (2) the idea of a community’s genuine dialect is a myth and can therefore no longer be the default dialect; (3) the overarching standard variety is the putative default variety in a stratified community; (4) the only way to understand linguistic reflections of social diversity is to use speakers’ social stratification as the independent variable; this is why social dialectology becomes correlative.

3.2.2. A new paradigm: Social stratiication and quantitative analysis Social dialectology grew out of socially oriented dialect monographs (section 2.2.1), as researchers acknowledged the fact that dialects are part of a more complex object that includes social and stylistic variables as well as geographical features. Research into linguistic varieties can no longer be one dimensional, since each variety (say a dialect) may be analyzed into several social (sociolects) and stylistic (registers) varieties. Thus, new models are needed to account for multidimensional data (Bellmann 1986: 1⫺5, 14⫺22 et passim). To this end it has become necessary to construct what Weydt and Schlieben-Lange (1981: 118) have called a “new paradigm”. According to them, this emergent paradigm is based on the following four common features: 1. Realism. This implies the rejection of the Homogenitätsannahme (Klein 1974: 18⫺40), i. e., the assumption of a homogeneous speech community. Rejection of homogeneity entails, on the one hand, a thorough revision of data collection, including sample representativity and current fieldwork criteria and strategies and, on the other hand, the idea of an interpersonal and intrapersonal grammar capable of covering a speech community’s diversified performance. Accurate protocols of methodological strategies (Milroy and Gordon 2003: 23⫺87) have been developed, and well-acknowledged linguistic models (e.g., variable rules) have been proposed (Fasold 1990: 244⫺257). Nevertheless, representative random aggregates of speakers do not actually represent anything that really exists. They help us to observe mean tendencies, but lack social depth (section 3.1.1). The same may be said of linguistic models: the need to obtain a general picture of the speech community’s social and stylistic variation at the expense of a more detailed insight into the intralinguistic system of each variety (Weydt and Schlieben-Lange 1981: 125⫺130) obliges the analyst to concentrate on the variation of single and isolated sounds (substantialism) without relating them to their counterparts within the system (atomism). 2. Empirical research. Intuition and other “narrow” sources of data are said to be replaced by a larger corpus of real linguistic performance by ordinary speakers who represent the whole speech community. To cope with mountains of data from the real community, social dialectology needs interdisciplinary collaboration: from statistics (quantitative analysis) and the social sciences, especially sociology (sampling procedures, theories of stratification). Though this represents one of the paradigm’s most widely acknowledged achievements, it is also its main obstacle, given its mechanism, scienticism and acceptance of the ideology underlying social sciences (Williams 1992).

34. Community-based investigations 3. Grammar. Early steps by Labov and others included variation research within the framework of generative grammar (Dittmar 1976: 132⫺160), although most of this work until very recently lacked solid theoretical foundations (Hinskens, van Hout and Wetzels 1997). 4. Quantitative analysis. Statistical and probabilistic models of representation and analysis have characterized this trend up until now. Aggregate scores are assumed to be the best means of accounting for orderly heterogeneity. By the end of the twentieth century, nothing seemed to be beyond the reach of this optimistic new way of analyzing linguistic variation. Focusing on the analysis of complex patterns of interaction between linguistic and extralinguistic (mainly social) variables generated very successful results in a climate of euphoria. Even today variation continues to represent one of the most innovative trends in linguistics. However, in spite of the optimism generated by more than forty years of incontestable findings, interpretive criticism has centered on the fact that correlational analysis based on mean scores obscures small group and individual variation behind social classes and other stratificational structures (see section 4).

4. Beyond group grammar: The speakers language 4.1. Emancipation and progress: An interpretive turning point or social dialectology The main concern in community-centered research is how to study dialects in their social context. Correlation between linguistic and stratificational variables has proved to be successful but insufficient in presenting a realistic account of patterns of behavior in the speech community. Interpretation (Verstehen) of the speaker’s performance in his or her social-network context is regarded as the key concept for renewal and even as a turning point in the development of social dialectology (Auer 1989). The reaction against mechanism and social therapeutics by microsociolinguistic practitioners has uncovered the hidden trend of emancipatory social dialectology commented on in section 3.1.2. As shown in Figure 34.2, the hypothesis is that mainstream correlational sociolinguistics (Labov 1966) has carried forward the original trend of traditional community studies, as adapted by European and American pre-sociolinguists (not to mention the French School or significant scholars such as Schuchardt). The emergence of social network methodology (Milroy [1980]) was initially understood as a complementary technique of correlation (Milroy and Margrain 1980). As this research trend began evolving toward a critical interpretive stance (Labov and Harris 1986; Milroy and Milroy 1992; Eckert 2000), an alternative emancipatory research trend was revealed and retrospectively traced back to the work of certain traditional alternative dialectologists (see section 3.1). This line of emancipatory sociolinguistics can be related to Labov (1972b) and has been influenced by several movements (interactional and interpretive sociolinguistics, sociolinguistic criticism, etc.). Such an analysis reveals that the present-day conflict between ideologies of therapeutic authenticity and emancipatory realism reaches right back to the origins of dialectology. At the same time it sheds fresh light on the history of dialectology itself.

623

Fig. 34.2: Directions of social dialectology: Therapeutics and emancipation. Continuous black arrows indicate a fundamental line of development. Discontinuous arrows represent trends of influence

624 VI. Data analysis and the presentation of results

34. Community-based investigations

4.2. Social networks Communities are not made up of ideal randomly selected subjects resulting from and representing the interaction of several extralinguistic factors, but of real people living in their social networks of kin, friends, neighbors and workmates (Milroy and Gordon 2003: 116⫺135). Even in urban contexts, actors tend to integrate into these informal networks of relations, which function as intermediate structures between speakers and social stratification. Thus, knowledge of a speaker’s status, sex, age, ethnicity or background is not enough to predict his or her linguistic performance. Social, ethnic and age groups are all heterogeneous; even within the same social class, clear differences may be perceived as people tend to integrate into the personal networks where most of their ordinary life takes place. As these networks keep very similar people together, and an inner consensus of norms tends to develop (Milroy 1987a; Dittmar and Schlobinski 1985), it is very likely that they speak alike (Villena-Ponsoda 2005). Social networks tend then to reflect the individual’s degree of integration into or isolation from the local speech community and hence can predict linguistic variation (Trudgill 1996: 3⫺4).

4.2.1. Correlational social network analysis Social dialectology seems to have found what it had essentially been searching for: an objective way of accessing the speaker’s natural (i. e., unobserved) behavior and a means of correlating their degree of integration into the community with their linguistic behavior. Hence, social networks have been defined both as a well-accepted bona fide instrument for fieldwork and as a successful way of complementing stratification analysis (Milroy and Margrain 1980; Milroy and Milroy 1992). As reflected in Figure 34.2, the evolution of the concept of social network shows one interpretive and at least three correlative steps, which correspond to four different hypotheses about the significance of the concept for social dialectology. The first (and weakest) hypothesis suggests that social networks are only a methodological tool for solving the observer’s paradox (Milroy 1987b: 57⫺67; Labov 2001: 325⫺ 365), but are unable to explain or predict linguistic variation. This trend accepts the influence of ethnography and has helped greatly to improve sociolinguistic methodology. The second (and strongest) hypothesis considers social networks as independent variables capable of explaining or predicting linguistic variation through correlation between linguistic variables and network markers. The idea is that strong-tie (dense) networks favor local loyalty and hence correlate with the use of strong vernacular varieties, whereas weak-tie (non-dense) networks tend to facilitate convergence toward mainstream norms and the use of the standard variety. As network structure properties (density, multiplexity, centrality, etc.) have proved easy to quantify (Milroy 1987a: 108⫺176; Bortoni-Ricardo 1985: 69⫺97; Dittmar and Schlobinski 1985: 163⫺185; Chambers 2000: 68⫺81; Labov 2001: 325⫺365), network markers can be used as independent variables in the correlation analysis. Measures of the speakers’ strength of ties and frequency of contacts, as well as their personal network size, have been applied. Researchers have built scales adapted to their objectives (Milroy 1987a: 152⫺176; Lippi-Green 1989: 218⫺ 220 and 223⫺231; Bortoni-Ricardo 1985: 162⫺169, etc.), but the results have not always been successful (Villena-Ponsoda 2005: 317⫺322).

625

626

VI. Data analysis and the presentation of results The third (intermediate or interactive) hypothesis emphasizes the explanatory capacity of social networks, i. e., prediction of vernacular maintenance/shift, but considers the interactive effect of status and other macrosociological variables (Milroy and Margrain 1980). Given that correlation between strong vernacular features and network markers is often weak or produces unexpected results, interactive categories of stratification and social networks need to be proposed. Evidence from different speech communities reveals significant differences in the effect of network on linguistic use depending on the speaker’s status, sex or age (see Villena-Ponsoda 2005: 323⫺329). An interactive model including three types of interaction of network and social class (life-modes) has been proposed by Milroy and Milroy (1992). Since empirical evidence shows that the network parameter (dense/non-dense) frequently interacts with the community’s external contact parameter (high/low level), as stated by Trudgill (1996, 2002), a model of interaction between social class, network structure and level of external contact could be proposed. This model would be the best way of capturing the contemporary version of the genuine dialect, i. e., the vernacular.

4.2.2. Interpretive social network analysis It is apparent that linguistic behavior is conditioned by interactive categories including status, network and external contact, but the locus of homogeneity ⫺ if it exists at all ⫺ lies beyond this threshold. After more than twenty years of intensive research into the correlation between social network markers and vernacular variables, results have not always turned out to be strong or significant: speakers living within the same social network are not as similar nor is their corresponding linguistic behavior as alike as might be expected. This does not mean that social networks are of no use for sociolinguistic analysis but they do need to be understood in a different way. Quantitative network analysis allows the analyst to understand norms underlying connected speakers but, as a consequence of its dependence on standardized correlation, it usually points to broad trends to the detriment of details. Thus, a fourth (interpretive) hypothesis must be formulated: social networks should be understood as a framework that is indispensable in interpreting the speaker’s motivations, but not as an end in themselves. Results of correlational network analysis are merely a way of arriving at the subsets of relations between speakers that usually lie beyond the analyst’s reach. These relations emerge as people engage in shared social enterprises (Eckert 2000: 171⫺212). The study of the speaker’s social life (Labov and Harris 1986; Labov 2001; Marshall 2004) allows the analyst to interpret his or her speech behavior. However, this interpretation is only within a participant observer’s grasp, since only an insider can access and understand quantitative results of network analysis beyond the standardized measures of network links. This is why analysis of both stratification and network structure should only be considered as a necessary but preliminary step toward the study of a speaker’s language use. This hypothesis should not be seen as a nostalgic backward step in line with some kind of philological, or indeed dialectological, rejection of the quantitative methods of sociolinguistics. On the contrary, it aims to offer an imaginative solution to a methodological impasse in social dialectology and a significant advance toward reclaiming the originally critical and emancipatory position of social dialectology.

34. Community-based investigations

4.3. An integrated theory o social dialect variation The evolution of social dialect research invites a retrospective reconsideration of traditional dialectology perspectives from two angles. 1. Firstly, the linguistic question: since linguistic variables are not atoms to be correlated with well-defined social structures, but belong to systems, inventories, etc., the exhaustive qualitative analyses common in dialect grammar research should be recovered through a thorough scrutiny of the systematic relationships of each variable under study (see section 2.2). 2. Secondly, the speaker question: stratification and social network correlation are not sufficient to explain an individual’s linguistic behavior. Results from quantitative stratification and network analyses (including interaction between status and social networks) should be complemented by an interpretive analysis of the speaker’s social life. This produces more satisfactory results but requires changes in social dialectology theory and methods in line with those described in this article.

5. Reerences Alvar, Manuel 1969 Estructuralismo, geografı´a lingüı´stica y dialectologı´a actual. Madrid: Gredos. Auer, Peter 1989 Re´fle´xions et e´tudes pre´paratoires pour une linguistique reconstructive de la variation. In: Pierre Cadiot and Norbert Dittmar (eds.), La socio-linguistique en pays de langue allemande, 163⫺190. Lille: Presses Universitaire de Lille. Auer, Peter 2005 Europe’s sociolinguistic unity; or, a typology for European dialect/standard constellations. In: Nicole Delbecque, Johan van der Auwera and Dirk Geeraerts (eds.), Perspectives on Variation: Sociolinguistic, Historical, Comparative, 7⫺42. Berlin/New York: Mouton de Gruyter. Bach, Adolf 1969 [1950] Deutsche Mundartforschung. Ihre Wege, Ergebnisse und Aufgaben. 3rd ed. Heidelberg: Winter Bellmann, Günter 1986 Zweidimensionale Dialektologie. In: Günter Bellmann (ed.), Beiträge zur Dialektologie am Mittelrhein, 1⫺55. (Mainzer Studien zur Sprach- und Volksforschung 10.) Stuttgart: Steiner. Berthele, Raphael 2004 Dialektsoziologie ⫺ Soziolinguistische Aspekte der Dialektologie. In: Ulrich Ammon, Norbert Dittmar, Klaus J. Mattheier and Peter Trudgill (eds.), Sociolinguistics: An International Handbook of the Science of Language and Society, vol. 1, 721⫺738. 2nd ed. (Handbücher zur Sprach- und Kommunikationswissenschaft 3.1.) Berlin/New York: de Gruyter. Bierwisch, Manfred 1988 Language varieties and connotation. In: Ulrich Ammon, Norbert Dittmar, Klaus J. Mattheier and Peter Trudgill (eds.), Sociolinguistics: An International Handbook of the Science of Language and Society, vol. 2, 1108⫺1118. (Handbücher zur Sprach- und Kommunikationswissenschaft 3.2.) Berlin/New York: de Gruyter.

627

628

VI. Data analysis and the presentation of results Bortoni-Ricardo, Stella Maris 1985 The Urbanization of Rural Dialect Speakers: A Sociolinguistic Study in Brazil. Cambridge: Cambridge University Press. Bourdieu, Pierre 1984 Capital et marche´ linguistiques. Linguistische Berichte 90: 3⫺24. Calvet, Louis-Jean 2003 Reflections on the origins of sociolinguistics in Europe. In: Christina Bratt Paulston and G. Richard Tucker (eds.), Sociolinguistics: The Essential Readings, 17⫺25. Oxford: Blackwell. Chambers, J. K. 2000 [1995] Sociolinguistic Theory. Linguistic Variation and its Social Significance. Oxford: Blackwell. Chambers, J. K. and Peter Trudgill 1998 [1980] Dialectology. 2nd ed. Cambridge University Press: Cambridge. Cortelazzo, Manlio 1968 Avviamento critico allo studio della dialettologia italiana. I: Problemi e metodi. Pisa: Pacini. Cos¸eriu, Eugeniu 1981 Los conceptos de “dialecto”, “nivel” y “estilo de lengua” y el sentido propio de la dialectologı´a. Lingüı´stica Espan˜ola Actual 3(1): 1⫺32. DeCamp, David 1958/1959 The pronunciation of English in San Francisco. Orbis 7: 372⫺391, 8: 854⫺867. Dittmar, Norbert 1976 Sociolinguistics: A Critical Survey of Theory and Application. London: Arnold. Dittmar, Norbert 1989 Variatio delectat. Le basi della sociolinguistica. Galatina: Congedo. Dittmar, Norbert 2004 Forschungsgeschichte der Soziolinguistik (seit Verwendung dieses Ausdrucks). In: Ulrich Ammon, Norbert Dittmar, Klaus J. Mattheier and Peter Trudgill (eds.), Sociolinguistics: An International Handbook of the Science of Language and Society, vol. 1, 698⫺720. 2nd ed. (Handbücher zur Sprach- und Kommunikationswissenschaft 3.1.) Berlin/New York: de Gruyter. Dittmar, Norbert and Peter Schlobinski 1985 Die Bedeutung von sozialen Netzwerken für die Erforschung von Ortssprachen. In: Werner Besch and Karl J. Mattheier (eds.), Ortssprachenforschung. Beiträge zu einem Bonner Kolloquium, 158⫺188. Berlin: Erich Schmidt. Durkheim, E´mile 1902 [1893] De la division du travail social. E´tude sur l’organisation des socie´te´s supe´rieures. 2nd ed. Paris: Alcan. Eckert, Penelope 2000 Linguistic Variation as Social Practice. Oxford: Blackwell. Fasold, Ralph W. 1990 The Sociolinguistics of Language. Oxford: Blackwell. Fishman, Joshua A. 1972 Preface. In: Joshua A. Fishman (ed.), Advances in the Sociology of Language, vol. 2, 7⫺ 12. The Hague/Paris: Mouton. Gauchat, Louis 1905 L’unite´ phone´tique dans le patois d’une commune. In: Ernest Bovet, Ernst Brugger Wilheim Degen, Arturo Farinelli, Adolf Fluri, Louis Gauchat, Jakob Jud, Jules Jeanjaquet, Emil Keller, Martha Langkavel, Marie Helene Johanna Minckwitz, Käthe Schirmacher, Ernst Tappolet and Louis Paul Betz, Aus romanischen Sprachen und Literaturen. Festschrift Heinrich Morf, 175⫺232. Halle: Niemeyer.

34. Community-based investigations Gillie´ron, Jules 1975 [1918] Ge´ne´alogie des mots qui de´signent l’abeille d’apre`s l’Atlas Linguistique de la France. Paris: Slatkine Reprints. Guy, Gregory 1988 Language and social class. In: Frederick J. Newmeyer (ed.), Linguistics: The Cambridge Survey, vol. 4, 37⫺63. New York: Cambridge University Press. Habermas, Jürgen 1981 The Theory of Communicative Action. London: Beacon Press. Hinskens, Frans, Roeland van Hout and Leo Wetzels 1997 Balancing data and theory in the study of phonological variation and change. In: Frans Hinskens, Roeland van Hout and Leo Wetzels (eds.), Variation, Change and Phonological Theory, 1⫺33. Amsterdam/Philadelphia: John Benjamins. Hudson, Richard A. 1996 [1980] Sociolinguistics. 2nd ed. Cambridge: Cambridge University Press. Iordan, Iorgu 1967 [1932] Lingüı´stica roma´nica. Evolucio´n, corrientes, me´todos. (Reprint edited by Manuel Alvar.) Madrid: Alcala´. Jaberg, Karl and Jakob Jud (eds.) 1928⫺1960 Sprach- und Sachatlas Italiens und der Südschweiz (AIS). 8 vols. Zofingen: Ringler. Kjolseth, Rolf 1971 Die Entwicklung der Sprachsoziologie und ihre sozialen Implikationen. Kölner Zeitschrift für Soziologie und Sozialpsychologie, Sonderheft 15 (Rolf Kjolseth and Fritz Sack [eds.], Zur Soziologie der Sprache): 9⫺32. Klein, Wolfgang 1974 Variation in der Sprache. Ein Verfahren zu ihrer Beschreibung. Kronberg: Scriptor. Klein, Wolfgang 1989 La variation linguistique. In: Pierre Cadiot and Norbert Dittmar (eds.), La socio-linguistique en pays de langue allemande, 101⫺124. Lille: Presses Universitaire de Lille. Koerner, Konrad 1988 [1986] Aux sources de la sociolinguistique. In: Actes du XVIIIe`me CILFR (Universite´ de Trier), vol. 5: Linguistique pragmatique et sociolinguistique, 16⫺34. Tubingen: Niemeyer. Koerner, Konrad 1991 Toward a history of modern sociolinguistics. American Speech 66: 57⫺70. Kubczak, Hartmut 1979 Was ist ein Soziolekt. Überlegungen zur Symptomfunktion sprachlicher Zeichen unter besonderer Berücksichtigung der diastratischen Dimension. Heidelberg: Winter. Kurath, Hans, Marcus L. Hansen, Bernard Bloch and Julia Bloch 1939 Handbook of the Linguistic Geography of New England. Providence: Brown University Press. Labov, William 1972a [1963] The social motivation of a sound change. In: William Labov, Sociolinguistic Patterns, 1⫺42. Oxford: Blackwell. Labov, William 1972b Language in the Inner City. Studies in the Black English Vernacular. Oxford: Blackwell. Labov, William 1972c Some principles of linguistic methodology. Language in Society 1: 97⫺120. Labov, William 1972d The social setting of linguistic change. In: William Labov, Sociolinguistic Patterns, 260⫺ 325. Oxford: Blackwell. Labov, William 1982 [1966] The Social Stratification of English in New York City. Washington, DC: Center for Applied Linguistics.

629

630

VI. Data analysis and the presentation of results Labov, William 1984 Field methods of the Project of Linguistic Change and Variation. In: John Baugh and Joel Sherzer (eds.), Language in Use: Readings in Sociolinguistics, 28⫺53. Englewood Cliffs, NJ: Prentice-Hall. Labov, William 2001 Principles of Linguistic Change: Social Factors. Blackwell: Oxford. Labov, William and Wendell A. Harris 1986 De facto segregation of Black and White vernaculars. In: David Sankoff (ed.): Diversity and Diachrony, 1⫺24. Amsterdam/Philadelphia: John Benjamins. Lausberg, Heinrich 1939 Die Mundarten Südlukaniens. Halle: Niemeyer. Lippi-Green, Rosina L. 1989 Social network integration and language change in progress in a rural alpine village. Language in Society 18: 213⫺234. Malkiel, Yakov 1976 From Romance philology through dialect geography to sociolinguistics. Linguistics 177: 59⫺84. Malkiel, Yakov 1984 Revisionist dialectology and mainstream linguistics. [Review article.] Language in Society 13: 29⫺66. Marcellesi, Jean-Baptiste and Bernard Gardin 1974 Introduction a` la sociolinguistique. La linguistique sociale. Paris: Larousse. Marshall, Jonathan 2004 Language Change and Sociolinguistics: Rethinking Social Networks. Basingstoke: Palgrave Macmillan. Martinet, Andre´ 1939 La description phonologique, avec application au parler franco-provenc¸al d’Hauteville (Savoie). Geneva: Droz. McDavid, Raven I. 1948 Postvocalic /-r/ in South Carolina: A social analysis. American Speech 23: 194⫺203. Milroy, Lesley 1987a [1980] Language and Social Networks. 2nd ed. Oxford: Blackwell. Milroy, Lesley 1987b Observing and Analysing Natural Language. A Critical Account of Sociolinguistic Method. Oxford: Blackwell. Milroy, Lesley and Matthew Gordon 2003 Sociolinguistics: Method and Interpretation. Oxford: Blackwell. Milroy, Lesley and Sue Margrain 1980 Vernacular language loyalty and social networks. Language in Society 9: 43⫺70. Milroy, Lesley and James Milroy 1992 Social network and social class: Toward an integrated sociolinguistic model. Language in Society 21: 1⫺26. Petyt, K. M. 1980 A Study of Dialect: An Introduction to Dialectology. London: Deutsch. Pop, Sever 1950 La dialectologie. Aperc¸u historique et me´thode d’enqueˆte linguistiques. 2 vols. Louvain (Leuven): Duculot. Romaine, Suzanne 1982 On the epistemological status of sociolinguistic theory. In: Suzanne Romaine, Sociohistorical Linguistics: Its Nature and Status, 239⫺289. Cambridge: Cambridge University Press.

34. Community-based investigations

631

Rona, Jose´ Pedro 1970 La concepcio´n estructural de la sociolingüı´stica. In: Paul L. Garvin and Yolanda Lastra (eds.), Lecturas de sociolingüı´stica, 203⫺217. Me´xico: UNAM 1974. Rousselot, Pierre-Jean 1891 Les modifications phone´tiques du langage e´tudie´es dans le patois d’une famille de Cellefrouin (Charente). Paris: H. Welter. Sankoff, David 1988 Sociolinguistic and syntactic variation. In: Frederick J. Newmeyer (ed.), Linguistics: The Cambridge Survey, vol. 4: Language: The Sociocultural Context, 140⫺161. New York: Cambridge University Press. Schlieben-Lange, Brigitte 1976 Soziolinguistik. Eine Einführung. Stuttgart/Berlin: Kohlhammer. Se´guy, Jean 1978 [1950] Le franc¸ais parle´ a` Toulouse. 3rd ed. Toulouse: Privat. Shuy, Roger W. 1990 A brief history of American sociolinguistics. Historiographia Linguistica 17: 183⫺209. Sivertsen, Eva 1960 Cockney Phonology. Oslo: Oslo University Press. Trudgill, Peter 1996 Dialect typology: Isolation, social network and phonological structure. In: Gregory R. Guy, Crawford Feagin, Deborah Schiffrin and John Baugh (eds.), Towards a Social Science of Language, vol. 1: 3⫺22. Amsterdam/Philadelphia: John Benjamins. Trudgill, Peter 2002 Linguistic and social typology. In: J. K. Chambers, Peter Trudgill and Natalie SchillingEstes (eds.), The Handbook of Language Variation and Change, 707⫺718. Oxford: Blackwell. Va`rvaro, Alberto 1984 [1972⫺1973] Storia della lingua: passato e prospettive di una categoria controversa. In: Alberto Va`rvaro, La parola nel tempo. Lingua, societa` e storia, 9⫺77. Bologna: Il Mulino. Viereck, Wolfgang 1966 Phonematische Analyse des Dialekts von Gateshead-upon-Tyne, Co. Durham. Hamburg: Cram/de Gruyter. Villena-Ponsoda, Juan Andre´s 2005 How similar are people who speak alike? An interpretive way of using social networks in social dialectology research. In: Peter Auer, Frans Hinskens and Paul Kerswill (eds.), Dialect Change: Convergence and Divergence in European Languages, 303⫺334. Cambridge: Cambridge University Press. Weinreich, Uriel 1977 [1954] Is a structural dialectology possible? In: Joshua A. Fishman (ed.), Readings in the Sociology of Language, 305⫺320. 4th ed. The Hague: Mouton. Weinreich, Uriel, William Labov and Marvin I. Herzog 1968 Empirical foundations for a theory of language change. In: W. P. Lehmann and Y. Malkiel (eds.), Directions for Historical Linguistics: A Symposium, 97⫺195. Austin: University of Texas Press. Weydt, Harald and Brigitte Schlieben-Lange 1981 Wie realistisch sind Variationsgrammatiken? In: Wolf Dietrich and Horst Geckeler (eds.), Logos Semantikos. Studia Linguistica in Honorem E. Coseriu, vol. 5, 117⫺145. Madrid/ Berlin/New York: Gredos/de Gruyter. Williams, Glyn 1992 Sociolinguistics: A Sociological Critique. London/New York: Routledge.

Juan Andre´s Villena-Ponsoda, Ma´laga (Spain)

VII. Exemplary studies 35. Untying the language-body-place connection: A study on linguistic variation and social style in a Copenhagen community o practice 1. 2. 3. 4. 5. 6.

Introduction Language and place ⫺ in the wake of globalization Place and body ⫺ the community as a social construct Body and language ⫺ deconstructing the authentic speaker Conclusion References

1. Introduction It is a deep-rooted conviction that languages, speakers and places constitute a unity. Languages belong to specific speakers, and speakers belong to specific places. In spite of the fact that many people in Denmark (especially in the larger cities) experience daily that one place may contain different speakers and more than one language, the belief that a language and its speakers belong to a specific place is profound. Danish is thought to belong to Danes, and Danes are thought to live in Denmark. The tying together of places, languages and people is not only characteristic of most lay people’s conceptualizations of languages. A tight connection between place, body and language also lies at the basis of most sociolinguistic and dialect studies, whether linguistic variation is described horizontally, as linked to geography, or vertically, as linked to social stratification. Languages and dialects have traditionally been treated as belonging to a place; so have speakers. The ideal informant lives in the place in which he and his parents were born. Furthermore, the speaker is believed to incarnate one, and only one, authentic language (the vernacular of the Labovian tradition). Although both dialectology and sociolinguistics have been harshly criticized for their rigid treatments of speaker categories (gender, class, age and ethnicity), there has in general been little focus on the insufficiently thought-out connection between body, language and place. However, in today’s large cities, where mobility and heterogeneity seem to be the norm rather than the exception, it is difficult, if not impossible, to uncritically maintain the language-body-place connection as the starting point for a linguistic description. In a globalized world where people move and communicate around the globe, it is easy to problematize a taken-for-granted connection between language, body and place. For the sociologist Zygmunt Bauman (1998: 12⫺18), the notion of globalization entails a contraction of time and space, which follows in the wake of late modern technologies. Bodies are able to move and communicate over long distances at greater and greater speeds at less and less cost. Consequently, bodies (the privileged ones) are less tied to specific geographical places ⫺ and, drawing on Paul Virilio, Bauman even talks about

35. Untying the language⫺body⫺place connection “the end of geography” (1998: 12). Anthony Giddens likewise (1991: 21⫺23) claims that globalization reorganizes time and space with fundamental human consequences in that a “globalisation of social activity” collides with established traditional practices. Giddens and Bauman thus argue that the consequences of globalization influence and restructure social organization, practices and identities (Bauman 1998: 18⫺26; Giddens 1991: 14⫺ 34). Such deep-seated changes also impact the study of language and communication, perhaps especially the sociolinguistic study of language variation in cities. The populations of larger cities become ethnically and linguistically more heterogeneous as migration increases (at least this has been the case in most of Europe; Extra and Gorter 2001). New communication technologies, mobile telephony and the World Wide Web reorganize communicative spaces and bring, so to speak, globalization into the core of speakers’ daily communicative practices. A perspective on language that involves treating one place as having one language or one dialect is a considerable simplification. The aim of this article is to show how the body-language-place connection shatters when the study focuses on language variation in ethnically mixed urban communities. I will in turn critically scrutinize connections between language and place, place and body, and body and language. This will be done through a presentation of findings from a study of language variation and social style in a Copenhagen community of practice (Quist 2005). Notions of style and practice will be developed as the main means needed to grasp the complexity of “language, body and place” in the study of linguistic variation.

2. Language and place - in the wake o globalization The Copenhagen speech community, in variationist studies, has traditionally been described as composed of two sociolects: low Copenhagen (used by working-class speakers) and high Copenhagen (used by middle and upper-class speakers; Brink and Lund 1975; Jørgensen 1980; Gregersen and Pedersen 1991b). Jørgensen (1980) and Gregersen and Pedersen (1991b) describe studies designed in line with the Labovian paradigm, defining a “Copenhagen speaker” as somebody who was born in Copenhagen (and preferably whose parents were also born in Copenhagen). In the Gregersen and Pedersen study, however, the criterion of being native to Copenhagen was difficult to maintain, as it turned out that only ten percent of the inhabitants had lived in the neighborhood under study for more than ten years (Gregersen 1989: 49; Gregersen and Pedersen 1991a). These proportions may even be lower today. We still find speakers of the traditional sociolects, low and high Copenhagen, but they hardly complete the picture of the linguistic variation in the city. If we today made second-generation nativeness to Copenhagen (excluding all second-generation immigrants from other countries) the criterion for delimiting Copenhagen speakers, we would end up with very few and certainly unrepresentative speakers. Thus, if we wish to gain substantial insight into the development, use and change in Copenhagen speech, we will have to employ methods other than those traditionally used (Brink and Lund 1975; Jørgensen 1980; Gregersen and Pedersen 1991b). The geographical place that we call Copenhagen contains a multitude of languages and language varieties and speakers who live with and use this multitude. The study

633

634

VII. Exemplary studies Tab. 35.1: Highest level of education attained by parents. Approximately 53 students attended the two high school classes during the five months of observation (some dropped out during the study, others arrived later). We have information on parents’ educational backgrounds from 47 students Highest level of education attained

Number of students

Extended tertiary study / university Mid-level higher education (skilled workers, teachers, etc.) Basic schooling (unskilled workers)

14 22 11

Tab. 35.2: Parents’ migration/linguistic backgrounds Parents’ backgrounds

Number of students

Both parents have a Danish ethnic background One parent has a non-Danish ethnic background

27

Both parents have a nonDanish ethnic background

7 (whereby all mothers have Danish backgrounds; fathers: 2 of Moroccan descent, 1 Iraqi, 1 British, 1 South African, 1 Greek, 1 French) 19 (5 students of Pakistani descent, 3 Turkish, 2 Moroccan, 2 Lebanese, 1 Iraqi, 1 Syrian, 1 Bosnian, 1 Faroese, 1 Jordanian, 1 Kuwaiti, 1 Sudanese)

described in this article illustrates this. It was conducted in 2002 in central Copenhagen, in the Nørrebro neighborhood. It was based on five months of ethnographic research conducted with two high-school classes at Metropolitanskolen, an ethnically mixed school. Besides participant observation, data consisting of 52 interviews, ten group recordings, self-recordings and questionnaires was collected (Quist 2005: 85⫺118). Traditionally, Nørrebro has been a working class area and stereotypically considered a stronghold of the low Copenhagen sociolect. However, during the last twenty to thirty years, the population has become ethnically and socially more diverse. Immigrants from Turkey, the former Yugoslavia, the Middle East and Pakistan settled in Nørrebro in the 1970s and 1980s. Later, during the 1990s, (parts of) Nørrebro, like other areas of inner Copenhagen, experienced a process of gentrification, with an influx of people with middle-class incomes, especially artists and academics. These demographic developments are reflected in the composition of the student body at Metropolitanskolen. In Table 35.1, we can see that 22 of the students in the two classes included in the study have parents with mid-level higher education (i. e., three to four years of training after nine or ten years of basic schooling). Fourteen students have parents with university degrees, and eleven students have parents with only basic school education (by and large, immigrant and Danish parents are evenly distributed across all three categories). Table 35.2 shows that about half of the students have Danish ethnic backgrounds, which means that they grew up in monolingual Danish homes. The other half grew up in families where Danish is the second or third language. Seven of the students have fathers of foreign descent and Danish mothers (incidentally, in this study there are no Danish fathers who married foreigners). Nineteen students have parents who themselves immigrated to Denmark. Most of them came in the 1970s.

35. Untying the language⫺body⫺place connection Taken together, Tables 35.1 and 35.2 show that the students in the two classes constitute a variegated group of individuals of different social, ethnic and linguistic backgrounds. Students with Danish ethnic backgrounds form the largest group. Still, it is no exaggeration to consider Metropolitanskolen to be demographically heterogeneous. Half of the students have mother tongues other than Danish. From time to time, some of them codeswitch between L1 and L2 (viz. those who speak Turkish, Punjabi/Urdu and Arabic). The linguistic picture is much more complex than that which the earlier studies of low and high Copenhagen sociolects were able to capture (Quist 2006). Danish is spoken in a variety of ways, for example, the use of multi-ethnolect is common (Quist 2000, 2008). There is no one-to-one correspondence between the physical place, here the Metropolitanskolen, and (one variety of) the Danish language (or the low Copenhagen sociolect). That different languages and varieties are present is due not only to the different backgrounds of the students. The communication technologies (Internet and mobile phones) that students integrate into their daily communicative practices bring in languages and varieties from other places. In some Danish high schools today, students are provided with a personal laptop, which they are also allowed to take home. This was the case in the two classes under study. At Metropolitanskolen, the students were able to connect their personal laptops wirelessly to the Internet at all times and places in the school. And so they did! They were online from the beginning of the school day until leaving school in the afternoon (and most of them were online again when they got home). During lunch breaks, for instance, the classrooms were filled with sounds from the computers. The students visited both Danish and foreign (mostly youth-oriented) websites where they watched short videos and listened to music. They also used the computers for instant messaging, e-mailing, and website-based chatting. The school premises, chiefly the classroom, constituted the physical place in which the students created their communicative spaces. Among other things, these spaces were characterized by (1) the virtual presence of interlocutors who were physically to be found in other places, in other cities and countries for example; and (2) the virtual presence, through websites, of different languages, mainly Danish and English (including different varieties of these languages), but also Turkish, Arabic, Hindi and other minority languages. In sum, the consequences of globalization ⫺ people’s mobility and new communication technologies ⫺ are evident at Metropolitanskolen: (1) the composition of students reflects the social and ethnic heterogeneity of the neighborhood, thus bringing in different languages and varieties; and (2) the use of laptops and mobile phones means that virtual interlocutors, languages and varieties are present and integrated into daily communication practices. Metropolitaskolen is not exceptional. In Copenhagen we find many communities of practice (this notion is described below) characterized by a diversity of languages and forms of communication that exceed the physical place. This reality is a challenge that a sociolinguistic study needs to face. It is not a new critique that disciplines like geographical dialectology and variationist sociolinguistics seem to treat places as more homogenous than is justifiable. Le Page, Gumperz and others (Gumperz 1962: 133; Le Page 1980: 336) have argued that a definition of a speech community must include all its speakers and languages (codes), because, in Gumperz’ words, “the criterion for inclusion of a code in a study of a linguistic community is that its exclusion will produce a gap in the communication matrix” (1962: 134). What is new in this discussion, however,

635

636

VII. Exemplary studies is the possibility for speakers to create virtual communicative spaces ⫺ not parallel to face-to-face interactions, but integrated into ongoing face-to-face interactions. (We shall look at an example of this later in this article.) Interlocutors, languages and varieties are in contact in and across Metropolitanskolen as a physical place. If we were to leave out the virtual communication, there would indeed be a “gap in the communication matrix”. The question is how the virtual communication and linguistic diversity can be included in a sociolinguistic study. One suggestion is to implement the notion of “community of practice” and to base the study in ethnography (see the next section for further discussion). The complex heterogeneity at Metropolitanskolen does not result in linguistic randomness or in an absence of patterns in language variation, choice and use. Nor does it result in different groups of students using different distinct styles without sharing and borrowing from each other. In the next section, we take a closer look at an example of how a traditional sociolinguistic variable distributes in unexpected ways and forms part of stylistic practices.

3. Place and body - the community as a social construct The diversity of speakers in Copenhagen, and at Metropolitanskolen, is largely an effect of the mobility of speakers, who move across country borders, city borders and between different communities within the city. Hence, if we wish to capture the linguistic diversity of a specific place, we need a concept of community that does not treat the connection between body and place as static. Community of practice is such a concept. It captures the dynamism of peoples’ connections to places by considering the community as a social construct. One advantage of approaching the speech community as consisting of communities of practice is that it allows for a multiplicity of linguistic codes and (mobile) speakers. Penelope Eckert and Sally McConnell-Ginet (1992) suggest that the notion of community of practice should be useful in the sociolinguistic study of language and gender. They borrowed the term from Lave and Wenger’s theory of situated learning (1991), the main idea of which is that learning takes place in active engagement in practices. Eckert and McConnell-Ginet argue that the social meaning of linguistic variation is actively produced and reproduced through practices within the communities in which speakers engage. According to Wenger (1998: 73), a community of practice is characterized by three dimensions: mutual engagement, joint enterprise and a shared repertoire. All three dimensions are characteristic of the two high school classes at Metropolitanskolen, and the classes are thus defined as a community of practice (Quist 2005: 53⫺57). The students are mutually engaged in the daily activities that take place within the framework of the school. This means going to school every day from Monday to Friday, attending classes on specific topics at specific times of the day, all in anticipation of completing the final studentereksamen (upper secondary examination). So, going to school, taking part in lessons, being in the right place at the right time is part of what we may call the joint enterprise. Within this framework, which is largely defined for them by others, the students mutually engage in daily activities such as hanging out in classrooms during breaks, playing basketball in the school yard, smoking under the roofed porch, doing

35. Untying the language⫺body⫺place connection

Fig. 35.1: Smokers and non-smokers

assignments in groups in the lab, helping each other out with the latest math homework, playing games or sharing music on their computers, etc. Through all such activities the students develop a shared repertoire of signs and resources that over the course of time become socially meaningful to them, i. e., become part of stylistic practices (Quist 2005: 76⫺78). The students were observed over five months, beginning with their first day at school. Drawings indicating each person’s position in the classrooms, hallways, yard, etc. were collected in a field diary. After a few weeks, everybody moved around in regular patterns (as illustrated in Figures 35.1 and 35.2). In other words, the same people spent their breaks in the school yard; the same people stayed in the classroom; the same people sat next to each other and worked together in group sessions during lessons. Through their routine movements, the students connect their daily practices to the physical place of Metropolitanskolen (or maybe better, in interaction with the physical place). The movements/positions overlap different kinds of practices. As a brief example, we can take a look at smoking. In Figure 35.1, the grey rectangles signify those who smoke, the white rectangles those who do not smoke and the dotted rectangle signifies a person who only smokes at parties. The sociogram illustrates that there is a group of smokers in class B who hang out together (Olav, Mads, Philip, Max, Jakob), a group of smokers in class A (Alexander, Ian, Malou ⫹ Malte and Doran) and a few more smokers distributed across the space.

637

638

VII. Exemplary studies The students’ names are also plotted on the chart, which presents us with even more information, namely the gender identity of the students. In general, boys move around in proximity to other boys, and girls close to other girls (Malou and Adam are the only exceptions). Although the sociograms present a static picture of something which is inherently dynamic, they do give us an idea of patterns and regularities in the practices of the two high school classes. From Figure 35.1, for instance, it is clear that there are at least two groups of boys who move in proximity to each other, and who smoke. These boys have more things in common. In fact, they perform a series of practices that taken together can be analyzed as a style cluster. A style cluster is defined in Quist (2005: 199) as a collection of signs that over the course of time cluster together, i. e., they are performed in regular patterns by the members of the community of practice. The notion of sign is defined broadly here, to include body signs (gendered and racial), signs of school attitude, physical positioning in the school area, preferences for music, clothing, and so on. It is argued that it is the regularity of the clustering that makes the style clusters socially meaningful as part of a shared repertoire of the community of practice. Space does not allow a detailed presentation of the style clusters. Instead, we shall focus on one linguistic element, a Copenhagen sociolinguistic variable that varies across the two high-school classes as part of the style clusters: the affrication of t in onset position. Brink and Lund (1975: 353⫺355) report that an affricated t in onset position is characteristic of low Copenhagen speech in the entire period of their study (they base their description of the Copenhagen sociolects on recordings from the Danish National Radio Archive, which includes speakers who were born in the years 1840⫺1955). The absence of affrication of t was a characteristic feature of speakers of high Copenhagen born in the nineteenth century; for example the number to ‘two’ was pronounced [dho1] (1 marks the Danish stød). This complete absence of affrication disappears from Copenhagen speech at the beginning of the twentieth century, but affrication of t remains a sociolinguistic variable in that long/strong affrication is associated with low Copenhagen and short/weak affrication with high Copenhagen. Parallel to the findings of Brink and Lund, Fisher-Jørgensen reports from an acoustic investigation that young Copenhagen speakers aspirate the unvoiced stop consonants /p/, /t/ and /k/ for longer than do older speakers, namely between 90 and 100 milliseconds, compared to the older speakers’ approximately 70 milliseconds (FisherJørgensen 1980; Fischer-Jørgensen/Hutters 1981: 79). Since low Copenhagen in general has been associated with young and modern speech (Brink and Lund 1974), it is of no surprise that Fisher-Jørgensen reports that younger Copenhageners pronounce a longer t than do older Copenhageners. (She did not look at differences in socioeconomic background.) If we were to follow a traditional sociolinguistic model, in line with Brink and Lund’s description of the two Copenhagen sociolects, we would predict the affrication of t to distribute in accordance with the socioeconomic backgrounds of (the parents of) the students at Metropolitanskolen. Alternatively, if we were to follow the implicit language change predictions in Fisher-Jørgensen’s study, we would expect no significant variation in the affrication of t at all, since the students in the two classes belong to the same age group (they are between 15 and 16 years old). However, an analysis of word initial t among the students at Metropolitanskolen shows that neither of the two predictions holds. There is systematic variation in the affrication of t, but the distribution correlates with neither socioeconomic background nor, obviously, age.

35. Untying the language⫺body⫺place connection

639

Tab. 35.3: Duration of t-affrication in four groups according to gender and ethnic background Group Boys ⫺ Danish Girls ⫺ other Girls ⫺ Danish Boys ⫺ other

Number of tokens

Average

F

p-value

135 103 111 53

0.109 0.099 0.098 0.092

9.910

0.0001

The ethnographic study and an explorative study of the recorded data indicated that the strong/long t was mostly used by some of the “white” boys, those called danskerne ‘the Danes’ in this community of practice. So the duration of the affrication of word initial t was measured for 52 speakers. All in all, 402 tokens were measured using the PRAAT software; that is approximately eight tokens per person. An analysis of variance of the duration of the ts indicated that there was a statistically significant difference, overall, between the four groups: boys with Danish ethnic backgrounds, girls with Danish ethnic backgrounds, boys with an ethnic background other than Danish, and girls with an ethnic background other than Danish. The results are shown in Table 35.3. Boys with Danish ethnic backgrounds sustained their affricated ts the longest (109 milliseconds on average), while boys with a different ethnic background had the shortest affrications (92 milliseconds on average). Post hoc tests showed that boys with Danish ethnic backgrounds had longer lasting affrications than did all other groups (all p ⫽ 0.0001). The difference between the two groups of girls was not significant, while the boys with a non-Danish ethnic background had significantly shorter t-affrications than did the girls overall. The ts were measured in two contexts: interviews and group sessions. The group pattern of variation was the same in both situations. The variation in t-affrication does not correspond to the socioeconomic backgrounds of the parents (categorized according to the amount of education as shown in Table 35.1). The variable correlates in this community of practice first and foremost with gender and ethnicity (Quist 2005: 222⫺225). Interestingly, however, a closer look at the boys with Danish ethnic backgrounds reveals that the boys who perform style cluster 1 are the ones who use the longest ts. Style cluster 1 is primarily performed by boys with white skin; i. e., those who are recognized as “Danes”. They show little interest in official school activities (they rarely do their homework and are relatively passive during lessons). They use their personal laptops primarily for games, music and communication (e-mail, chat and messaging). On the computer screens they typically display images of half-naked women (e.g., pinups). They also use their computers to search for and share music, mostly hip-hop and rock music. These boys smoke cigarettes, so they go to the outdoor smoking areas during the breaks, and they leave the school premises during lunchtime. They drink alcohol and talk about alcohol at school. The clothes they wear are a sort of hip-hop style with baggy jeans and large T-shirts. As for the linguistic resources analyzed, these boys had the longest affrication of word initial t, they used more slang and more swear words than did the others, and they never used multi-ethnolect, except in instances of parody. (For details about the specific style features, see Quist 2005: chapters 9 to 14.) A style cluster should not be confused with a group of individuals (“group” in the sense used in theory of social identity; Hogg and Abrams 1998), but rather understood

640

VII. Exemplary studies

Fig. 35.2: Style clusters and the twelve boys with the longest t-affrications

as a collection of signs and practices that some individuals perform more than others. Members of a group may perform resources from different clusters. In Figure 35.2, the seven style clusters that were identified at Metropolitanskolen are illustrated with a circle around their prime performers. Style cluster 1 is illustrated with a solid line, indicating that Philip, Max, Mads, Jakob, Olav, Alexander and Ian are the core performers of this cluster. The twelve boys who have the longest t-affrications in this community of practice are marked in gray. From Figure 35.2, we see that the style cluster 1 performers are all also users of long t-affrication. Malou (female) hangs out with Alexander and Ian, and sometimes with Adam, Vibeke and Otella. She does not perform all practices that characterize style cluster 1; e.g., she wears different clothes to the boys, she displays other types of images on her laptop, but she smokes, and she exhibits a reluctant attitude to school, just like Alexander and Ian. That is why she is portrayed on the line, a little inside and a little outside of the solid circle. Malou exemplifies the fact that some students do not perform all practices of a style cluster; that is, it is not possible to situate her as a core performer of a cluster. Figure 35.2 also shows that there are boys who do not perform all elements of style cluster 1: but who still use long affrication of t, namely Malte, Johan, Adam, Nikolaj and Kristoffer. Some of them perform other clusters (Nikolaj, style cluster 3; Johan and Malte, style cluster 2), and Adam and Kristoffer perform practices that do not make up a systematic pattern. These exceptions and the foreshortened expose´ of the style clusters may leave us with the feeling that there is little of coherence going on as far as sociolinguistic variation is concerned. The methods of analysis employed in the study do not provide a clear-cut coherent and complete picture of language use. The study of the

35. Untying the language⫺body⫺place connection community of practice at Metropolitanskolen gives us something else: it gives us insight into patterns and regularities of performance by Copenhagen speakers in daily practice. We find regularities and exceptions to the regularities. The variation of t in connection with the style clusters shows us that the variation forms part of the social and linguistic landscape in a much more complex way than a more traditional approach to the Copenhagen speech community is able to capture. In the context of this article, it is especially important to stress that the heterogeneity of speakers in late-modern Copenhagen does not result in complete muddled variation. We do find regularities in linguistic variation and the performance of practices. The linguistic variation is connected to stylistic practices, and practices connect the speakers to the locus of Metropolitanskolen.

4. Body and language - deconstructing the authentic speaker We now turn to the last of the three pairs of connections treated in this article, namely the connection that is typically made between a body and a language (or a body and a dialect/variety). It is an implicit assumption of traditional dialectology and sociolinguistics that “one body holds one authentic language”. This assumption is parallel to the ideas (or ideologies, see Bucholtz 2003) of the authentic speaker and authentic speech, as discussed and problematized in sociolinguistics in recent years (by, for instance, Eckert 2003; Bucholtz 2003; and Coupland 2003). The sociolinguistic search for authentic speakers and authentic speech is in general parallel to the connections made between language, body and place. Authentic speakers are considered by the researcher to be such precisely because they belong to a specific geographic place. The same holds for authentic speech. Eckert (2003: 392) puts it this way: Authenticity is constructed in relation to particular locations such as the traditional peasant in an isolated community (Holmquist 1985), the street kid in the inner city (Labov 1972[b]), and the burned-out burnout in a Midwestern high school (Eckert 2000). Locally located and oriented, the Authentic Speaker produces linguistic output that emerges naturally in and from that location.

Bucholtz argues that the idea of authentic speech and authentic speakers is reproduced in an ideology which she describes in terms of the linguist as an obstacle to authenticity: “even if we find authentic speakers, they may not produce authentic speech in our presence” (2003: 406). Labov’s notion of the observer’s paradox (Labov 1972a) reflects this ideology. The paradox is that the researcher should aim at “capturing” speech as it is used without the presence of the researcher, the assumption being that the speaker possesses an authentic way of speaking, the collection of which ⫺ by way of mere presence ⫺ can be hindered by the researcher. To Labov, the most authentic speech is the vernacular in its most casual form. The vernacular can be “captured” in a “sociolinguistic interview” or in “the field” at local locations, where people are engaged in leisure time activities: As we enter the city we look for preadolescent and adolescent peer groups engaged in sports or hanging-out; we encounter family groups at tea or after dinner; we join old men at bowls; in pubs; or sitting at pensioner’s benches (Labov 1972b: 256).

641

642

VII. Exemplary studies Labov clearly preferred non-institutionalized settings as his research ground, i. e., he would try to avoid work places and schools for instance. Bucholtz calls such favoring of quotidian language and settings an ideology of “linguistic mundaneness” (Bucholtz 2003: 405). Although sociolinguistics represents an important repudiation of linguistics’ taking of introspection and context-free sentences as the only evidence of language, there has been a tendency to restrict the study of language use to “a narrow subset of all language use” (Bucholtz 2003: 406). In contrast to the speakers in Labov’s study of language in the inner city (Labov 1972b), young people in Copenhagen do not hang out together much outside of institutions. A young Copenhagener spends most of his or her time by far in school, recreational classes or at institutionalized leisure time activities in venues such as sports clubs and youth clubs. Therefore, it would represent a crude truncation of the reality of these youngsters to restrict a study of their speech to times in which they are not in an institutionally organized setting. Arguably, a lunch break at Metropolitanskolen is just as casual and relaxed as the types of settings Labov refers to in the quotation above. The vernacular in its most casual and unconscious form is thought to be the most authentic representation of a person’s language. If we take the observed linguistic practices at Metropolitanskolen as our case in point, we will see that it is difficult, if not impossible, to decide where and when in the flow of language use the most authentic representation is found. The students in the two classes constantly make use of a broad repertoire of languages and varieties; they cite, imitate, make parodies, perform (in Bauman’s [1986] sense of the word), sing, do language crossing (Rampton 1995) ⫺ and they do all of this much more than they do not. Let us look at an extract from a conversation between three girls at Metropolitanskolen. The extract is taken from self-recordings done by Catrine, a girl with a Danish ethnic background, who is a typical performer of style cluster 6 (Quist 2005: 198⫺200). She has, as usual, her laptop on. She is playing music so loudly on it that it almost drowns out the music played on the other computers in the room. To Catrine’s right is Helena, who also has her computer on. Both of them are searching the Internet as they talk to each other. Helena is looking for a specific mp3 file of a Turkish song, and she is trying to get Catrine involved in the search (lines 6, 22, 25 and 27). But Catrine does not want to help Helena. She is busy looking for sites that will inspire her in her search for a new instant messaging alias. Catrine is tired of her existing name, which is a compound of Catrine and the Muslim name Aisha: Caisha. Right behind Catrine and Helena is Amina. She is looking at Catrine’s computer screen trying to help her find a new alias. [A modified CA convention is used for this transcription: (.) micro pause; (2) pause two seconds long; _ syllable emphasis; Capital letters: high volume; ºtxtº low volume; slow speech; >txt< fast speech; : the sound is prolonged; [txt…] overlap; ((txt)) comments inserted by researcher; xxx incomprehensible speech.] 1 Cat: 2 Ami: 3 Cat: 4 Ami: 5 Cat:

jeg har fa˚et at vide at jeg skal skifte navn (.) fordi det her er for somebody told me that I should change my name (.) because this one is too hvad hedder du what’s your name Caisha okay det er lidt for: (.) hvad skal jeg hed:de (.) help me. it’s a bit too: (.) what shall I ca:ll myself (.) help me

35. Untying the language⫺body⫺place connection 6 Hel: 7 Cat: 8 (2) 9 Cat: 10 Ami: 11

Cat:

prøv lige og send den der please send this one what’s my name say my name you bitch what’s my name say my name you bitch ºCai:shaº (.) ej men hvad skal jeg he:dde. ºCai:shaº (.) no but what shall I ca:ll myself du kan ogsa˚ godt hedde (.) Caisha. you can just as well call yourself (.) Caisha (>prøv lige og se< ) fandt vi inde i den der (>take a lookhvad skal jeg hedde< >what shall I call myself< 18 ((they search the Internet for names)) (7) 19 Hel: xxx ((sounds like a suggestion for a name)) 20 Cat: det er lidt krukket ikke. it’s a bit affected isn’t it ((about 15 seconds omitted)) 21 Cat: give me a name: give me a name: 22 Hel: ga˚ ind i den der xxx [jeg kan ikke finde den ] enter that one xxx [I can’t find it ] 23 Cat: [NE:J JEG GIDER IKKE ] [NO I WON’T] 24 Cat: men ellers sa˚ kald (0.5) ved du hvad [det det det her kalder] det kalder Maja mig but otherwise you could (0.5) you know what [this name] Maja calls me this 25 Hel: [okay men sa˚ send den til mig sa˚ ga˚r jeg xxx] [okay then send it to me xxx] 26 Ami: ej helt ærlig lad være oh no honestly don’t 27 Hel: hallo jeg ga˚r ind og finder et eller andet o [kay] hey I enter and find something o [kay] 28 Ami: [eller] ga˚ ind og a˚bn den der hjemmeside et sted ligesom xxx og fa˚ et nyt navn [or] enter and open that homepage a place like xxx get a new name 29 Cat: ne:j det er ikke noget med det hun har kaldt mig det hele tiden (.) det hedder jeg i hendes telefonbog (.) no: it’s not about that she has called me that all the time (.) she has that name in her phone book (.) 30 bellahøjmafia

This extract exemplifies three points that are important in the context of this article. Firstly, as mentioned in the section on language and place, the laptops play a central

643

644

VII. Exemplary studies role in this community of practice, just as communication technologies do in late modern societies in general. The conversation between Catrine, Amina and Helena is an example of how the laptops are integrated into their communicative practices. The extract represents a typical way for students to spend their breaks between lessons (Quist 2005: 180⫺ 183) ⫺ their activities and their conversational topics are centered on the computers (for more on this, see Quist 2005, chapter 15, 16). Catrine and Helena are simultaneously conversing with each other, and with Amina, as they search the Internet for a new alias and the Turkish music file. Both of them ask for help and thereby they include others around them in their ongoing activities on the computers. To conduct a complete analysis of the role of the computers in this conversation, we need video recordings or other types of data that indicate what the participants are doing on the computers as they speak. We do not have such data in the current study. However, the audio recording does lend much support to the point that is important to us here, namely that computers as a means of communication form an integral part of the speakers’ communicative practices. The second important point that should be emphasized here was also addressed in the section on language and place: speaker heterogeneity in terms of linguistic and ethnic backgrounds. Catrine’s parents are of Danish ethnic descent; Amina has a Danish mother and a Moroccan father; and Helena has a Greek father and a Danish mother. Although Catrine is blond, and by appearance would not be associated with immigrants, she uses an immigrant speech style, or multi-ethnolect, as part of her linguistic repertoire. For instance, in lines 29⫺30 she changes intonation into an unmistakable multi-ethnolectal pronunciation (Quist 2000, 2008). She does this unremarkably, without breaking the flow of speech. There is nothing ironic about it, and no signs to signal that she is not “speaking as herself ”. This type of shift is very common in Catrine’s speech, and cannot be understood as crossing in the sense that it is used “by people who are not accepted members of the group associated with the second language they employ” (Rampton 1995: 280). Multi-ethnolect is part of Catrine’s own language. There is thus, in the case of Catrine (and many others at Metropolitanskolen), no given one-to-one correspondence between the particular speaker category (Danish, blond) and the language used. This is linked to the third and last important point, which is the problem of extracting “authentic speech” and the (dis)connection of body and language. Catrine’s speech, from line 1 to line 30: is constantly changing, using different voice qualities, different languages and language varieties. Although the situation is casual and relaxed ⫺ and the event typical ⫺ it is not possible to judge when Catrine is speaking authentically as “herself”. She sings (line 14), speaks loudly, even shouts (line 23) and whispers (line 9); in lines 5, 7 and 21 she switches from Danish into English (italicized in the gloss); and in lines 29⫺30 she uses a multi-ethnolectal pronunciation. In line 5: she says with an almost importunate voice “what shall I call myself”, as if she was begging the others for vital help. The begging is underlined by the proceeding switch into English, “help me”, which she enunciates in such a way as to sound as if she were crying. Nobody thinks she is crying of course. The utterance is produced with a kind of ironic distance, as if it was a direct quote from an American soap or mainstream Hollywood movie. The other two shifts into English (lines 7 and 21) also appear in the form of imitation or quotation, but they are enunciated very differently. “What’s my name say my name you bitch” is said with a rhythm and intonation suggesting Catrine is copying a line from a rap song. There is something insistent, even chant-like, about this

35. Untying the language⫺body⫺place connection utterance that makes it very different from “help me” in Catrine’s previous turn, so it cannot be understood as a continuation of the former (English) voice. Catrine’s hip-hop voice in line 7 resembles the type of language used in verbal hip-hop battling (Bejder and Holt 2005: 6⫺7). In hip-hop, it is common for the rapper to call on others to say his (because it is typically a male) name as a way to show respect. Catrine is a big fan of the rapper TuPac. She writes his name everywhere and she has pictures of him on her laptop screen. In a rap song, TuPac sings, for example, “Scream my name / cause baby it’s delicious / got a weak spot for pretty bitches”; “What’s my motherfucking name nigga?”. These are examples of requests to say (or “scream”) the name of the rapper and the use of the invective “bitch”. Catrine’s “what’s my name say my name you bitch” appears in line 5 after she has asked her friends for help to find a new name. Helena does not take up this request. Instead, she keeps on doing her own thing, searching for music on her computer, and asking Catrine in line 6 to “send this one” (“this one” being a music file). Then, as if Catrine is offended by Helena’s disregard, she answers Helena in the voice of a rapper, just as a rapper would answer an attack in a verbal battle. In this way, Catrine simultaneously restates her request for help to find a name and ignores Helena’s request. Space does not allow a detailed analysis of every switch in language and enunciation in this conversation. The point to be stressed here ⫺ which the switches into English in lines 5 and 7 demonstrate ⫺ is that the uses of different voices and languages are not “empty” playing or fooling, around with language without a purpose. They are not superficial performances that cover Catrine’s “own”, “authentic” language. The many different voices of Catrine are part of her linguistic repertoire, which she uses with pragmatic and interactional intents related to the activity she is engaged in and the goals she wishes to pursue. What Catrine does in this extract is not uncommon in the community of practice at Metropolitanskolen. In the words of Jørgensen (2004), she is doing languaging, i. e., she is a languager: “speakers employ whatever linguistic features are at their disposal” to achieve their communicative goals as best they can. Robert Le Page made a parallel argument in a discussion with Labov about the vernacular (Le Page 1980; Labov 1980) ⫺ his words shall be the last of this section: [e]ach individual’s competence subsumes partial knowledge of many socially-marked systems, and each individual’s performance reflects choice among those systems, constrained by the social and psychological factors operating upon him at any given moment (Le Page 1980: 336).

5. Conclusion The goal of this article was to examine and problematize the often taken-for-granted ties between language and place, place and body, and body and language. The community of practice at Metropolitanskolen in Copenhagen makes the case in point by demonstrating the difficulty of maintaining fixed ties between language, body and place. These ties are more or less dissolved in the late modern city, where mobility, new communication technologies and heterogeneity in terms of speaker backgrounds reorganize communicative spaces and practices. Treating a place like Copenhagen as having one language or two sociolects would be a considerable oversimplification. It would treat speakers as

645

646

VII. Exemplary studies incarnating only one language or dialect, even though what speakers do in real daily life is to constantly employ a broad repertoire of linguistic codes. We saw in the section on body and place that the students at Metropolitanskolen occupy the physical location in regular patterns linked to their practices and social styles. The boys who stay in the courtyard during breaks, who sit next to each other at the back of the classroom during lessons, and who leave the school premises at lunchtime also share a series of other practices that in the course of daily repetition cluster together in a style cluster. Characteristic of the speech of these boys is, among other things, the long affrication of t in onset position. The long affrication of t in earlier studies was found to correlate with working-class Copenhagen speech, i. e., the low Copenhagen sociolect. This is not the case at Metropolitanskolen (the social class of the parents does not seem to matter). Instead, we find that the long affricated t is used in stylistic practices as a linguistic resource that combines with “white” masculinity and an apparent antischool attitude. In the section on language and body, we saw how problematic it is to maintain a “one body ⫺ one language” ideology in the community of practice at Metropolitanskolen. Catrine and her friends mix all kinds of linguistic resources in their communicative activities. Just as with the boys who use longer t-affrications than everybody else, Catrine’s practices, including her use of multi-ethnolect and her fascination with North American hip-hop, form part of a style cluster. When Catrine, for instance, is able to successfully employ multi-ethnolect in a conversation, it is because she ⫺ in the course of stylistic practice ⫺ has created a position where multi-ethnolect becomes accepted as part of her language. Thus, mixing and switching do not happen at random, but are parts of social practices in which it is possible to act against expectations (expectations like “a blond Danish-looking girl does not use multi-ethnolect”; see Quist and Jørgensen [2007] for a discussion of who is and who is not able to use multi-ethnolect as part of his or her language.). The intention of this article was not to show that geographical place, the body and language (language defined as a linguistic system) do not matter; they do. But nonetheless, they exist as social constructs with complex and dynamic connections, not as given entities with fixed ties and social meanings. If we wish to understand why they matter and how they work together, we need operationalizable notions that grasp the dynamism and the processual connections between them. In this article, I have suggested that style, practice and community of practice constitute such notions.

6. Reerences Bauman, Richard 1986 Story, Performance and Event: Contextual Studies of Oral Narratives. Cambridge: Cambridge University Press. Bauman, Zygmunt 1998 Globalization. Cambridge: Cambridge University Press. Bejder, Peter and Kim Boje Holt 2005 Hiphop de´r [Hip-hop Right]. Skødstrup: Man˜ana. Brink, Lars and Jørn Lund 1974 Udtaleforskelle i Danmark [Variation in Pronunciation in Denmark]. Copenhagen: Gjellerup.

35. Untying the language⫺body⫺place connection Brink, Lars and Jørn Lund 1975 Dansk Rigsma˚l. Copenhagen: Gyldendal. Bucholtz, Mary 2003 Sociolinguistic nostalgia and the authentication of identity. Journal of Sociolinguistics 7(3): 398⫺416. Coupland, Nikolas 2003 Sociolinguistic authenticities. Journal of Sociolinguistics 7(3): 417⫺431. Eckert, Penelope 2000 Linguistic Variation as Social Practice. Malden: Blackwell. Eckert, Penelope 2003 Elephants in the room. Journal of Sociolinguistics 7(3): 392⫺397. Eckert, Penelope and Sally McConnell-Ginet 1992 Think practically and look locally: Language and gender as community-based practice. Annual Review of Anthropology 21: 461⫺490. Extra, Guus and Durk Gorter (eds.) 2001 The Other Languages of Europe. Demographic, Sociolinguistic and Educational Perspectives. Clevedon: Multilingual Matters. Fisher-Jørgensen, Eli 1980 Temporal relations in Danish autosyllabic CV sequences with stop consonants. Annual Report of the Institute of Phonetics 14: 207⫺261. Fischer-Jørgensen, Eli and Birgit Hutters 1981 Aspirated stop consonants before low vowels, a problem of delimitation ⫺ its causes and consequences. Annual Report of the Institute of Phonetics 15: 77⫺102. Giddens, Anthony 1991 Modernity and Self-Identity. Cambridge: Polity Press. Gregersen, Frans 1989 Hvordan undersøger man Københavnsk? [How can Copenhagen speech be studied?]. Tijdschrift voor Skandinavistiek 10(1⫺2): 38⫺58. Groningen: Scandinavisch Instituut. Gregersen, Frans and Inge Lise Pedersen 1991a Copenhagen as a speech community. In: Kjell Lars Berge and Ulla Britt Kotsinas (eds.), Storstadsspra˚k och Storstadskultur i Norden: Föredrag fra˚n ett Forskarsymposium [Urban Language and Culture in Scandinavia: Presentations from a Research Symposium], 57⫺ 69. (MINS 34.) Stockholm: University of Stockholm, Department of Nordic Languages and Literature. Gregersen, Frans and Inge Lise Pedersen 1991b The Copenhagen Study in Urban Sociolinguistics. 2 vols. Copenhagen: Reitzels. Gumperz, John J. 1962 [1993] Types of linguistic communities. Anthropological Linguistics 4(1): 28⫺40. [Reprinted in Anthropological Linguistics 35 (Special issue: A Retrospective of the Journal of Anthropological Linguistics: Selected Papers, 1959⫺1985): 130⫺142.] Hogg, Michael A. and Dominic Abrams 1998 Social Identifications. London: Routledge. Holmquist, Jonathan 1985 Social correlates of a linguistic variable: A study in a Spanish village. Language in Society 14: 191⫺203. Jørgensen, Jens Normann 1980 Det flade a vil sejre. En undersøgelse pa˚ sociolingvistisk grundlag af visse københavnske sprogforhold [The flat a will prevail. A sociolinguistic study of some features of Copenhagen speech]. Skrifter om Anvendt og Matematisk Lingvistik (SAML) 7: 67⫺124. Jørgensen, Jens Normann 2004 Languaging and languagers. In: Jens Normann Jørgensen and Christine Dabelsteen (eds.), Languaging and Language Practices, 5⫺23. (Copenhagen Studies in Bilingualism 36.) Copenhagen: University of Copenhagen, Faculty of Humanities.

647

648

VII. Exemplary studies Labov, William 1972a Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Labov, William 1972b Language in the Inner City. Philadelphia: University of Pennsylvania Press. Labov, William 1980 Is there a creole speech community? In: Albert Valdman and Arnold Highfield (eds.), Theoretical Orientations in Creole Studies, 369⫺388. New York: Academic Press. Lave, Jean and Etienne Wenger 1991 Situated Learning. Legitimate Peripheral Participation. Cambridge: Cambridge University Press. Le Page, Robert 1980 Theoretical aspects of sociolinguistic studies in pidgin and creole languages. In: Albert Valdman and Arnold Highfield (eds.), Theoretical Orientations in Creole Studies, 331⫺ 367. New York: Academic Press. Quist, Pia 2000 Ny københavnsk “multietnolekt”. Om sprogbrug blandt unge i sprogligt og kulturelt heterogene miljøer [New Copenhagen “multiethnolect”. Language use amoung youth in linguistic and cultural heterogeneous environments]. Danske Talesprog 1: 143⫺212. Quist, Pia 2005 Stilistiske praksisser i storbyens heterogene skole. En etnografisk og sociolingvistisk undersøgelse af sproglig variation [Stylistic practices in the urban heterogeneous school. An ethnographic and sociolinguistic study of language variation. Unpublished Ph.D. dissertation. Nordisk Forskningsinstitut, afd. for dialektforskning [Dialectology Section, Faculty of Humanities], Copenhagen University. Quist, Pia 2006 Sproglig variation i storbyen⫺tre senmoderne udfordringer [Language variation in the city ⫺ three challenges of late modernity]. In: Dorthe Duncker (ed.), Studier i Nordisk 2004⫺2005. A˚rsskrift for Selskab for Nordisk Filologi [Studies in Scandinavian 2004⫺ 2005. Annual of the Society of Scandinavian Philology], 158⫺170. Copenhagen: Selskab for Nordisk Filologi. Quist, Pia 2008 Sociolinguistic approaches to multiethnolect: Language variety and stylistic practice. International Journal of Bilingualism 12 (1⫺2): 43⫺61. Quist, Pia and J. Normann Jørgensen 2007 Crossing ⫺ negotiating social boundaries. In: Peter Auer and Li Wei (eds.), Handbook of Multilingualism and Multilingual Communication, 371⫺389. Berlin: Mouton de Gruyter. Rampton, Ben 1995 Crossing. Language and Ethnicity among Adolescents. London: Longman. Wenger, Etienne 1998 Communities of Practice. Cambridge: Cambridge University Press.

Pia Quist, Copenhagen (Denmark)

36. A study on areal diffusion

36. A study on areal diusion 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Dynamic interpretation of static maps The case studies Case 1: Short vowel lowering A diffusion model Case 2: Long vowel raising Language-internal constraints and the diffusion of innovations Variability Isoglosses ⫺ the limits of diffusion Regions ⫺ the outcome of diffusion References

1. Dynamic interpretation o static maps This study examines the areal diffusion of linguistic features using examples from Swiss German. The Sprachatlas der deutschen Schweiz (SDS; Hotzenköcherle et al. 1962⫺ 2003) provides the material basis. Traditional linguistic maps are static, usually atomistic snapshots of the geography of a single “caracte`re dialectal” (Meyer 1875: 295). In contrast, diffusion describes a dynamic process across time and space in which an increasing number of speakers add a linguistic feature to their repertoire. The distribution of a feature on a linguistic map is the frozen geographic reflex of a shift in linguistic behavior ⫺ a change in language. Studying the diffusion of linguistic features on the basis of static maps requires a theory of language change in order to bridge “the chasm between the study of language in time and the study of language in space” (Ogura 1990: 4, 15). Speakers with their attendant psychological, social and linguistic preconditions, linguistic features, time, and space must all be brought together into a systematic relationship capable of “explaining” the spatial patterns discovered, that is, able to suggest reasons for the patterns being as they are. One such theory was the concept of “radiation” (Strahlung), which long dominated the dialect-geographical discussion (Bach 1969: §§ 79, 95⫺101). It focused on “active” innovators, rather than “passive” adopters. The opposition “active” vs. “passive” distinguishes between the emergence of an innovation and its diffusion. Following Labov (e.g., 1972: 277), sociolinguistic theories equate the emergence of a language change with its diffusion: emergence describes a process in which a deviation from former regular usage becomes the common property of a group. Only when several speakers exhibit the same innovation, that is, when it “diffuses” to others, does it become a linguistic fact and evidence of language change. In this sense, one can agree with Labov. Nevertheless, the equation is dissatisfying. The initial transfer of a new linguistic fact into the linguistic repertoire of a group is the most problematic phase of any change, and it requires an extra intellectual effort to think it through at all (Keller 1994). The concepts of “active linguistic landscape” and “radiation center” attempt to differentiate this first phase from a phase of diffusion, during which the speakers of related lects adopt an innovation already established elsewhere as a linguistic fact. Although this process is not “trivial”, as Lüdtke (1980: 9) would have it, it appears conceivable in terms of familiar concepts like imitation and adaptation (Auer and Hinskens 2005; Haas 1998: 844⫺845).

649

650

VII. Exemplary studies Radiation theory was based on rudimentary social psychological concepts and attempted to derive dialect-geographical maps directly from those preferred (thanks to their origin) usages which accrue around new linguistic forms. Sociolinguistics captured this “added value” with the dazzling catch-all category of prestige. As early as the 1920s, Walther Mitzka counterposed the concept of radiation with that of alignment (Anschluss). He insisted that the issue was less the diffusion of an innovation than the speakers’ adoption or rejection of an innovation (1952: 164). This touched on a central aspect of the problem: in any linguistic process, the only agent is the speaker, and every action captured on linguistic maps as the diffusion of an innovation is enacted by speakers who have adopted that innovation, not those who had already integrated it into their repertoire. What is meant by the “diffusion of an innovation” and, even more so, what might be counted as material for a case study of diffusion, cannot be decided independently of theoretical considerations. In the following, maps of a series of linguistic changes are interpreted as static reflexes of dynamic diffusion processes based on Haas’s (1978) theory of language change (see Domaschnev, Stirnickaja and Najditsch 1980; Mattheier 1984; Moulton 1980; Keller 1982; Reed 1982; Reiffenstein 1986; Wiesinger 1984 for critical responses to Haas).

2. The case studies Two wide-ranging qualitative reconfigurations of the Middle High German (MHG) vowel system, seen as the historical precursor to the Swiss German vowel systems, can serve as examples for the diffusion of linguistic innovations. The presentation is confined to spontaneous lowering and raising, that is, to changes in vowel height. Rounded front vowels generally behave in the same as their unrounded analogues and are not mapped here. Tab. 36.1: The Middle High German vowel system. Vowel series of equal tongue height are labeled (a)⫺(e). Examples and map numbers are taken from SDS vol. 1.a short

long

(a) (b) (c) (d) (e) a

i Schlitten 48 e Bett 15 e¨ Speck 21 ä Wespe 19

ü Füchse 52 ö “Götti” 47

ˆı Eis 105

iu Mäuse 107

uˆ Maus 106

eˆ Schnee 95 æ “strääle” 73

œ “Chööl” 101

oˆ Brot 99 aˆ Abend 61

u “Chuchi” 50 o “Gotte” 41

a Achse 11

See Haas (1978: 107⫺113) for a justification of the system. The assumption of long open midvowels is philologically justified and supported by the observation that the way from diphthongs (e.g., WGerm. *ai, *au) to closed monophthongs (e.g., OHG/MHG eˆ, oˆ) is often via long open vowels (Labov, Yaeger and Steiner 1972: 225).

36. A study on areal diffusion The phonemic system taken is that of classical structuralist phonology, as applied to Swiss German by W. G. Moulton (1960) and others. The vowel system of any Swiss German dialect can be related to the system shown in Table 36.1. In historical terms, it can be postulated that the vowel system of the current dialects have evolved from this system via language change. The vertical relations can essentially be derived from two sound shift complexes: (1) a “chain” of short vowel lowering; (2) a “chain” of long vowel raising.

3. Case 1: Short vowel lowering 3.1. Presentation If the short vowel system of the Swiss German local dialects is set in relation to the postulated Middle High German system, it becomes clear that, compared to Middle High German, many dialects exhibit lowered short vowels. The number of lowered short vowels in each dialect declines from west to east: western dialects have many lowerings, eastern ones none. Map 36.1 shows three regions of lowering, each characterized by a specific combination of spontaneous vowel lowerings. The geographic constellation of the various types of lowering reveals a typical “wave formation” (Wellenbild ): the region with the most changes of a particular kind (III) borders on a region with fewer changes (II) which in turn borders on the region with the fewest (I). The easternmost dialects are not lowered at all. The structure of the geographic neighborhood thus reflects a wavelike spread of innovations. The regions of lowering are geographically coherent. This finding has to be interpreted as evidence that the individual dialects have not undergone the various innovations independently of one another. Structuralist dialectology explains the development of each local dialect on the basis of the conditions prevailing in each “closed” system; but this does not help us understand how coherent regions with identical developments can arise. For this we must make reference to adoption, accommodation, alignment ⫺ in short with the diffusion of an innovation. It is of secondary importance for this study which “sociogeographical” path the diffusion takes; whether it progresses from location to neighboring location (contagion diffusion, Britain 2004: 623) or from one larger population center to the next (parachutage linguistique, Strahlung, Bach 1969: § 98), and there from higher social classes down to lower ones (hierarchical path, Britain 2004: 622⫺627) or up from below (Baumgartner 1940: 44⫺48). The parachutage ‘air drop’ pattern is presumably the norm in mobile societies, but it does not preclude simple areal spread (Herrgen 2005: 299). Usually both modes eventually lead to homogenous regions. They are therefore less easy to identify in atlases of traditional and stable dialects than Britain (2004: 623) would have us expect. The true complexity of the processes first becomes apparent in sociolinguistic investigations; Siebenhaar (2000) is a good example for the region under study. The size of the region in which a particular lowering has occurred can be interpreted as a reflection of the age of the change. From this standpoint, lowering stage 1 (cf. Map 36.1) can be seen as the oldest, 3b as the most recent, act in the overarching process.

651

652

VII. Exemplary studies

Map 36.1: Short vowel lowering in Swiss German

36. A study on areal diffusion The geographical overlap of all of the lowerings in Zone III can be viewed as evidence that all of the lowerings here are old; the point of origin of the change complex can thus be located in Zone III ⫺ the shift must have “arisen” here. Early Generative Grammar saw the adoption of a change by other speakers as the root of its generalization; the dialect with the strictest implementation was therefore believed to be the dialect which underwent the shift last and hence in its most thoroughly generalized form (that would be III here). This interpretation overlooks the fact that language is also handed on within the original dialect. The internal (linguistic) structure of the change complex is of special interest. The lowerings do not simply affect ever more vowels at random; rather they proceed in a clear order. First, the most open vowel, Gmc /e¨/, is lowered, then follow the mid vowels, finally the close vowels (the uppermost rows in Table 36.1); within each series the front vowels shift before back vowels. Thus, in consecutive stages, ever more neighboring vowels in the system seem to be “affected” by the same change: a lowering by one degree of openness. This internal and implicative coherence of the innovations fills the until now purely quantitatively defined neighborhood structure with linguistic content. The wave form proves itself to be a “Baileyan” wave, distinct from older implication models in that it only admits features that belong together within the system into an implication scheme (Bailey 1973: 65⫺109). The lowering of short vowels conforms to one of the often demonstrated “natural” tendencies of change (Labov 1994: 116). This tendency may explain the successful diffusion of the change, but it also renders the similarity between neighboring lowering dialects as natural.

3.2. Some reinements Anton Pfalz observed that language shift affects all vowels of the same height (in the same “row” of the vowel quadrilateral) in the same manner: vowels shift in series (Reihenschritten; 1983a [1918]: 50). In a system of phonological parameters this can be expressed as the change in a single parameter, viz. tongue height. Short vowels from the same Middle High German series must therefore continue to form a series after a lowering. According to Weinreich, Labov and Herzog, this is a practically universal condition of sound change (1968: 175). Lowering 1 eliminates MHG e¨ by merging it with MHG ä; the chain-shift principle plays no role for this isolated phoneme. For the vowels of series (c) (MHG e…) and (b) (MHG i…) in Table 36.1, this principle applies without restriction in Zone III: after lowering, both series once again each form a series of vowels of equal height. But this principle is not a “law”. The dialects of Zone II are defined precisely by the fact that there has been no chain shift in vowel series (c). In the local dialects of Zone II, the front vowel MHG e has been lowered to /e/, while the back vowel MHG o has maintained the closed quality of /o/. If innovations diffuse in waves which progress from phoneme to “neighboring” phoneme, then it can be expected that the change must begin with one of the phonemes of a series, but only so as to then immediately affect the next one. In light of the chain-shift principle, it is unexpected that change can manifestly freeze in a “pre-progressive” state for an extended period of time.

653

654

VII. Exemplary studies This finding is evidence for the assumption that lowering first affects the front vowels and then the back ones, which can in turn be traced back to differing articulatory contexts (Weinreich, Labov and Herzog 1968: 175). The temporal precedence of frontvowel lowering (Table 36.1, 36.2a) over back-vowel lowering (2b) is reflected in its larger area of diffusion, which of course includes all of Zone III. But that the generalization of the process to the back vowels is the “aspirational” norm is evident in the fact that Zone III, with both lowerings (2a and 2b) is far larger than Zone II with just one lowering. A wave model that takes adequate account of the chain-shift theory also needs to specify that lowerings in series (b) can only occur when all of the vowels of series (c) have been lowered. This assumption is to a large extent confirmed by the distribution map. This is remarkable, since the lowering of (for example) all front vowels yet only front vowels would not lead to the collapse of any phonemic distinctions. The vertical dash symbols on Map 36.1 stand for a conditioned lowering of MHG e before r (⫹C) (based on contrastive data from SDS vol. 1, Map 18 “hert”). They are found in a contiguous region to the east of Zone II. In terms of the proposed theory of diffusion, the conditioned lowering could be interpreted as a precursor to the spontaneous lowering of (2a), thus explaining its territorial distribution: the conditioned development is older than the spontaneous one and has therefore attained a greater degree of spread.

4. A diusion model The vowel-lowering map can be interpreted as a reflection of the manner in which innovations diffuse through the linguistic landscape. It lends itself to the following conclusions: 1. Innovations spread through space from speaker to speaker, and hence from dialect to dialect (extensive diffusion). 2. This process of extensive diffusion requires time, during which the innovation is also transmitted to new speakers within the group of innovators-in-progress. 3. During extensive diffusion, the innovation process can encroach upon ever more morphemes containing the same original vowel and also become generalized to similar vowels (intensive diffusion). This is also to be expected in intergenerational transmission within the group. 4. Intensive diffusion is subject to the structural parameters of a language; the colloquial paraphrase of “encroaching on neighboring phonemes” refers to this process, which a dynamic theory of phonological naturalness must be able to explain. These interconnections can be rendered schematically as in Figures 36.1 and 36.2. The idealized linguistic map displays the diffusion of an innovation at a particular time (here t3). The difference in the age of the innovations that belong together (A, B, C….) is reflected geographically in the varying size of the areas over which they have diffused by t3. The idealized map neglects virtually all social and extralinguistic factors. It takes no account of the fact that different innovations can diffuse at different rates and hence occupy territories of differing size precisely because of these factors. It also essentially

36. A study on areal diffusion

Fig. 36.1: Diffusion of an innovation (a sound change) across time and space; α⫺η… represent local dialects

Fig. 36.2: An idealized linguistic map: A, B, C symbolize an increasing number of contexts in which the change occurs, or an increasing number of words containing the original phoneme affected by the change; I, II, III represent regions in which the local dialects exhibit the same stage of change

fails to capture the fact that all real processes of diffusion come to a halt (at least temporarily). This is particularly problematic for sound change that must be seen as incomplete from a language-internal/theoretical perspective, such as the chain shift in Zone II. An idealized map is even less suitable for predicting either the point in the

655

656

VII. Exemplary studies landscape at which a change will be discontinued or the reason (Siebenhaar 2000: 62⫺ 64, 241 is informative about the current situation). It should thus be self-evident that the idealized map and real diffusion correspond only in principle. A real linguistic map supports the theoretical assumptions about the diffusion of innovations when it confirms the predictions of the idealized map. This is the case when the neighborhood relations between the real regions of innovation reflect the implicational relations between the postulated stages of innovation: Geographical reflection: Stages of innovation

Zone III | Zone II ABC AB

| Zone I A

In contrast, the geographical reflection of the diffusion process would be disrupted and in need of explanation if a neighborhood relation III | I | II were found. This appears to be the situation between UW and LU on Map 36.1. The dialect boundary between these two is one of the most robust in German-speaking Switzerland; the innovation in UW must be expected to come from the east. What we have in the west is not a disruption of the implications but a “compression” of the regions of lowering. Here, Zone III abuts directly on the French⫺German border and Zones II, I and Ø are all absent (cf. the schematic depiction in Figure 36.3). This geographical pattern is compatible with the diffusion model. A language boundary represents a major hurdle for the diffusion of a sound change. Because the lowering cannot diffuse across the language boundary, over the course of time several stages of lowering pile up on top of one another, so to speak, and their isoglosses converge at the language boundary. It is important to note that the implicational relations between the individual stages are in principle not disturbed by the pileup of isoglosses. This explanation can also be applied to other multistage changes, e.g., the second Germanic sound shift. Here too, the regions characterized by the most thoroughgoing change lie on the alpine language boundary. It has often been doubted that such a successful chain shift could have spread from the most remote alpine valleys. But of course the process need not have arisen at the very edge of the most intensively affected

Fig. 36.3: Pileup of isoglosses from related stages of change at a robust linguistic boundary

36. A study on areal diffusion zone ⫺ the lowering discussed above did not begin at the language boundary but at the center of what is now Zone III. In summary, it can be said that the linguistic map of the actual lowering process is highly congruent with the predictions of the idealized map. This can be taken as an indication that the theory of diffusion sketched here may not be all that far from the truth. But it is of course an informal and incomplete theory. Whilst no attempt is made in the following to amend its informality, its incompleteness does need to be addressed. The theory postulates a close relationship between extensive (i. e., relative to the speaker and space) and intensive (relative to the language system) diffusion. Following Mitzka’s suggestion, the speakers themselves, as actors in the process who adopt or reject an innovation, have to be considered here. Sociolinguistic constructs such as acts of identity or the evaluation of innovations shape the extensive aspect of diffusion. But the speaker is equally important in the explanation of the intensive aspect of diffusion. The adoption of an innovation is the learning of a new linguistic fact by a speaker, diffusion the learning of a new linguistic fact by ever more speakers. The “generalization” of a rule is a cognitive act, which plays an important role in learning. An innovation does not necessarily have to be generalized in the course of its diffusion, but generalization without transmission to new speakers is hardly conceivable: extensive and intensive diffusion are thus inseparably linked. Finally, positing so-called linguistic preconditions for sound change only makes sense when speakers are seen as the locus of the process: “phonemic system”, “chain shift”, “natural tendencies” etc. are viable concepts only when they are located in speakers’ minds (no matter what theoretical model is applied). Only via speakers’ minds can cognitive objects like phonemic systems function as supportive, impeding, or formative factors in the learning of linguistic innovations. How they so act is little known; but they do it with such regularity that the consequences simply cannot be overlooked on dialectologists’ maps. This diffusion model requires a more elaborate theory of language change than that of the neogrammarians or the structuralists, according to whom a sound change affects the entire vocabulary simultaneously and without exception, and variability is effectively excluded. But even a dialectological countertheory, which incorporates two different types of change, an exceptionless sound shift and substitution via word-for-word adoption (Pfalz 1983b: 87) is not enough. Labov is a proponent of this division into two; his neogrammarian change is, however, not exceptionless, but rather one which can affect every relevant morpheme of the vocabulary simultaneously but at first variably (Labov 1994: 440⫺444). Wang and his school render sound substitution an absolute, the sole path for a sound change (Ogura 1990). In opposition to this, a point of view proposed by Bremer (1893: x⫺xvii; cf. Haas 1995: 334⫺335) is mooted here. According to this, the different types of change can be understood as phases of a single process, which begins with lexical diffusion. On this basis, an initially variable and restricted soundchange rule can be abstracted. In the course of extensive diffusion it can generalize its field of effect to include more and more relevant morphemes and spread to further phonemes. The loss of variability can also proceed step by step. “Exceptionlessness” can be posited as the endpoint of the process, with not every change reaching this goal.

657

658

VII. Exemplary studies

5.

Case 2: Long vowel raising

5.1. Presentation If the long-vowel systems of the Swiss German dialects are contrasted with the putative Middle High German system (cf. Table 36.1), it becomes clear that many dialects feature long vowels that have been “raised” compared to Middle High German. The number of raised long vowels in each dialect declines from east to west: the eastern dialects have many raisings, the western none. The basic features of Map 36.2 match the depiction of vowel lowering, but the direction of diffusion is reversed. Once again the region with the most changes of the same kind (III) conjoins an area with fewer changes (II), itself adjoining an area with the fewest changes (I), and then finally there is a region without any raising. Once again, neighboring vowels within the system are affected by the “same” change (raising by one level) in a particular order (closed vowels before open ones, back vowels before front vowels). The raising of the long vowels also accords with an often documented “natural” tendency (Labov 1994: 116). In all three regions of raising, the reflexes of MHG eˆ and oˆ each exhibit the same tongue height. The perfect chain shift is reflected in the practically point perfect correspondence between the regions of diffusion for the raising of each vowel. This can also be interpreted as an indication for the contemporaneous diffusion of both sound changes. In contrast, the regions in which aˆ and æ have been raised are of different size; they must have diffused at different times in two separate waves. The raising of aˆ has diffused more widely than the raising of æ. Where aˆ is not raised, neither is æ. This implicational relation is rigorously valid, which is evidence that the two shifts belong together, and that the two long low vowels function phonologically as a single vowel series, although phonetic arguments could support a different interpretation. In light of the perfect chain shift (the raising of MHG eˆ and oˆ), it can be argued that the “imperfect” chain shift linking æ and aˆ is attributable to this phonetic-phonological problem (cf. Wiesinger 1970: vol. 1, 32). As with the lowering, the linguistic map of the actual raising complex corresponds closely to the predicted implications of the theoretical map. Isoglosses pile up in this change process as well: in the east, the intermediary Type II is often missing between Types I and III; in the west, the intermediary Area I is absent from between Area II and the region without raising. Since there is no language boundary which can be deemed responsible for the pileup in this instance, it is worth searching for other social boundaries (see section 8). In terms of the proposed fundamental mechanisms, raising and lowering can be described as parallel processes. This strengthens the suspicion that the model captures essential facets of the process of diffusion of linguistic innovations.

5.2. Diusion o new change on the basis o previous change In three regions on Map 36.2, the relations between the vowels can not be explained via the generalization of a phonological process to ever more neighboring phonemes. These relations are therefore of particular theoretical interest.

36. A study on areal diffusion

Map 36.2: Long vowel raising in Swiss German

659

660

VII. Exemplary studies In the mottled Jurassic region MHG æ and aˆ manifest as closed /e:/ and /o:/, which have merged with the raised reflexes of /eˆ/ and /oˆ/. The products of raisings (2a), */c:/, and (2b), */ε:/, have thus been raised further. This is not the transfer of the same change to neighboring phonemes, it is a renewed application of the same procedure to phonemes that have already been changed. In Labovian terminology, one might talk of a “hypercorrection”; but there are several indicators (not least the phonemic merger) that this is an independent change coming out of the northwest which has transformed the reflexes of raisings 2a and 2b. In the westernmost of the two hatched areas, the level of raising was initially the same as in the Jura Mountains, but then, under Middle Bernish influence, partially wound back (see section 7.1). In the hatched region around Zurich, aˆ was initially raised to /c: ~ o:/, but then lowered back to /A:/, whereby the spread of an urban/written pronunciation of MHG aˆ might be implicated. There is precious little historical evidence for this interpretation, but the almost ideal linguistic geographic image of radiation out from an urban centre is enticing. In both hatched regions, more recent, independent change can be seen to have acted upon earlier states; this preserves the historical implicational relations and consequently the geographical embedding. Also common to both regions is the fact that the newer change is regressive, in that it leads back to older linguistic states.

6. Language-internal constraints and the diusion o innovations If we accept that speakers are also influenced by their own linguistic systems in their choice of which innovations to take up or reject, then every expansion of an innovation (every change made) alters the systemic conditions and thus the chances that a subsequent innovation will be adopted. An important (albeit not necessary) condition for the adoption of an innovation appears to be that it does not provoke the collapse of a phonemic contrast. Within German-speaking Switzerland, it appears to be a further systemic condition that a change contribute to symmetry between the two subsystems of short and long vowels. In principle, in the modern dialects every stressed phoneme in the inventory of short vowels corresponds to a qualitatively identical long vowel; in contrast to Standard German both short and long vowels of the “same” height can have both lax and tense forms. The symmetry between the subsystems ought to have first emerged in the post⫺Middle High German period. Although the waves of lowering and raising have led via counterposed processes to different systemic states in the east and west, they have also contributed to the realization of this symmetry. The articulatory, perceptive, and cognitive mechanisms that lie behind such trends are little known, but their effects, e.g., the symmetry of the systems, are beyond doubt (Haas 1998: 840). It is also not known why the New High German pairing of shortness and laxity on the one hand, length and tenseness on the other, can be observed in so many of the world’s languages that it can be regarded as “natural”, nor why languages such as Swiss German nevertheless pursue, virtually systematically, a “more unnatural” symmetry between the subsystems. The two great chain shifts in the vocalic system had their origins at opposite ends of the language area and the diffusion waves have overlapped in the center (Figure 36.4).

36. A study on areal diffusion

Fig. 36.4: Overlapping waves

This means that the systemic preconditions for the adoption of newly arriving subwaves are also in flux, dependent upon the relative “arrival time” of each wave. That the short vowel lowering began in the west might well be connected to the retention there of the older open quality of the long vowels. The lowering of the short vowels was not only natural, it also contributed to the symmetry between the short and long vowel subsystems. If the long vowels are supposed to have had an influence on the lowering of the short vowels then it will have been via the system-defining characteristic of symmetry. The lowering of the short vowels in the absence of long vowel raising led in the western regions without overlapping waves to systemic states like those in Figure 36.5. In the eastern regions without overlapping waves, the raising of the long vowels in the absence of short vowel lowering led to states like those in Figure 36.6.

Fig. 36.5: Lowering and raising in the west

Fig. 36.6: Lowering and raising in the east

661

662

VII. Exemplary studies

Fig. 36.7: Lowering and raising in the center

The raising of the long vowels was adopted into the east of the language area. It followed a natural trend and at the same time contributed to the establishment of symmetry with the unlowered short vowel subsystem. This in turn could be one of the factors that inhibited the natural tendency of the short vowel system to move lower. In the far east, the long vowel raising could even have encouraged an (unnatural!) raising of the close vowel series (b), MHG i > [i≈]. With the overlapping waves, the preconditions for diffusion of the second wave can be altered by the changes to the system generated by the adoption of the first wave. In the region of overlap, the dialects have, as in the west, lowered the short vowels but also raised the long vowels, as in the east (see Figure 36.7). The development in the center (see Figure 36.7) led to phonologically symmetrical but etymologically “disrupted” relations, since the MHG “partners” i ⫺ ˆı … now stood in a different relationship: *i ⫺ *eˆ. This is a problem for historicizing dialect orthography, yet following the monosyllable lengthening which refilled vowel series (d), the systems are functionally well balanced; only the short vowel series (b) has a low functional load. Precisely how the encounter between the opposing diffusion waves may have unfolded can hardly be reconstructed without apparent time data.

7. Variability One of sociolinguistics’ most indisputable findings is that language change includes a variable stage. Zones into which innovations are in the process of diffusing when surveyed must be characterized by variation between the old and new variants of the element that is changing and hence interpretable using the apparent-time approach. Variability contradicts the assumption of exceptionless sound change ⫺ variable change is not without exception. Further, according to the neogrammarian dogma, the isoglosses for an innovation must coincide for all words that feature the affected sound in the same context. Early dialectology was at pains to meet the expectations of the leading linguistic schools. In order to obtain clear maps, it protected itself from undesired empirical variability by, say, surveying just a few informants about just one example for each sound law and determining the course of the isogloss on the basis of this sole example (cf. Table 36.1). Atlases are thus freer of variation than the languages they map. On the other hand, local dialects are not that bad at meeting the “neogrammarian” expectations. Once the variable stage has been passed through, the relations between sounds in the local dialects in many cases resemble the products of an exceptionless law: an etymologically identical sound displays the same transformation in most of the relevant morphemes and the isoglosses around an innovation do indeed coincide for the majority of the morphemes.

36. A study on areal diffusion Of course, there are always “exceptions”, about which there is less worry these days: it has become clear that language can never be completely free of variation, particularly not at the boundaries of lingistic features. This makes those cases in which even the static maps display variation, and hence dynamics, all the more welcome.

7.1. Variation as a transitional phase As mentioned in section 5.2, in the shaded part of Raising Zone II (Map 36.2), MHG aˆ > */c:/ > /o:/ and MHG æ > */ε:/ > /e:/ were initially raised further, erasing the distinction between them and the descendants of MHG oˆ and eˆ (also /o:/ and /e:/). A new diffusion of the still open Middle Bernish long vowels into this region led to a “regressive” change, in the course of which the merged phonemes æ and eˆ were once again split, etymologically correctly, into /æ:/ and /ε:/. This indicates lexical diffusion (Lautersatz). In contrast, aˆ and oˆ continue to be represented by a single phoneme /c:/, which ranges between [c: ~ o:]. They have made their way into the atlas because the change was in progress at the time of the recordings; the variation indicates change in apparent time.

7.2. Long-standing variation According to Map 36.1, a region between Luzern and Zurich, the so-called “Freiamt”, can be assigned to Lowering Zone II. But it displays a series of anomalies: second stage lowering appears to have been confined to stage 2a whilst, in contradiction of the implicational relations, stages 3a and 3b have already been implemented. The region also breaches the symmetry between the short and long vowel subsystems, since MHG o is represented by the closed and unlowered short vowel /o/, while the counterpart to MHG aˆ is the open and raised long vowel /c:/. Checking the atlas data reveals, however, that there is in fact variation in the reflexes of MHG o, aˆ, and oˆ. But this variability is probably not a phase in a diffusion process playing out in apparent time, but rather a long-standing variation. It may well be typical for a clearly politically/socially defined region positioned between two strong centers but without a center of its own. Variation found here is less likely to be evidence of developmental dynamics and much more likely to be performing a “mediating” function. The problematicity of complicated variability in dialects between strong centers has been clearly demonstrated by Siebenhaar (2000).

8. Isoglosses - the limits o diusion At the language boundary, the outer limit of the region of related lects, the diffusion of innovations generally comes to a standstill (see section 4). But usually the innovation never reaches the linguistic boundary; it comes to a halt within the region of related lects. This is where isoglosses arise. Traditional dialectology looked beyond language for the reasons diffusion might stop: innovations diffuse across areas in which the inhabitants are linked to one another

663

664

VII. Exemplary studies through intensive “intercourse” (Verkehr); they come to a standstill at societal boundaries which inhibit social intercourse (cf. Mitzka 1952: 144⫺148). Dialectology’s highest mission was to synchronize linguistic and extralinguistic boundaries (“The root of the tongue and the velum have ceded the leading role to the historical atlas”, Wrede 1963 [1919]: 337). The correlation of several isoglosses with one extralinguistic boundary was taken as evidence of the divisive power of the extralinguistic boundary. In contrast, it was of little import whether or not congruent isoglosses were linguistically (internally) connected. Objections were also not raised that the linguistic boundaries of a “dialect area” followed varying extralinguistic boundaries in varying coalitions. Rather than concluding that the boundary determination method could be at fault, it was posited that there were no dialects. Jules Gillie´ron was one of the first to propound the idea that the diffusion of innovations could also be curtailed by language-internal boundaries. Structuralist dialect geography then went on to systematically establish regions based on the isoglosses of systemically connected linguistic features, e.g., the phonemes of a subsystem. Extralinguistic boundaries were used not to define boundaries, but at best to explain the specific location of a boundary. Structuralist dialectology was in fact able to establish relatively clearly defined, contiguous dialect regions (Goossens 1969). The case study here proceeded structuralistically insofar as it too selected the isoglosses of phenomena related within the language system. But incorporation into a process breaks this relationship down into stages that follow, and follow from, one another: it is assumed that a shift can begin with one phoneme and then proceed to a neighbor within the system. A consequence of this is that the relative position of the isoglosses in the landscape must conform to the implications of the transformational process. Adherence to the same path is of less significance. Nonetheless, systematic constraints (trend towards chain shifts, symmetry) generally lead several isoglosses of connected innovations to merge. Hence, in LU (around Luzern) for instance, at the border between Raising Zone II and the non-raising zone, the isoglosses of the reflexes of MHG eˆ, oˆ, and aˆ follow the same path. The partial coincidence of linguistic and political boundaries is not uncommon on Maps 36.1 and 36.2. It is true that with isarithmic maps (i. e., those with areas circumscribed by isoglosses), the cartographic method is often responsible, especially in the uninhabited High Alps. Nevertheless, it is clear that the political boundaries of the Old Swiss Confederacy have often staked out the locations at which an innovation came to a halt. What is important in this diffusion model is not the boundary as such, but rather the combination of features it encloses. Thus the combination of features that define Lowering Zone III also includes the features of Lowering Zones I and II, whose boundaries run elsewhere. The linguistic coherence of a region is more important than its boundary.

9. Regions - the outcome o diusion The first linguistic geographers hoped to be able to delineate homogeneous dialect regions with their isoglosses. Dialectology had for the first time used empirical methods to raise scholarly awareness of people’s linguistic reality. In so doing it exposed the

36. A study on areal diffusion fiction of homogeneous regional varieties ⫺ a service which, against the background of the dominant ideologies, dialectologists at first regarded as a fault that they attempted to overcome with great methodological effort (cf. Knoop, Putschke and Wiegand 1982). Using features characterized by the closest possible match between their isoglosses and political boundaries to define dialects is linguistically arbitrary; combining them on a map inevitably leads to a confusion of isoglosses. Localized dialect is the largest homogeneous variety that lacks explicit standardization, and even it is far from logical homogeneity: “L’unite´ du patois […] est nulle” (Gauchat 1905: 222). All varieties that are more than the dialect of small networks are types, which have to be constructed by their speakers and by linguists. Interlinked, language-internal, systemic characteristics are more important to the process than isolated features (of the lexicon for instance). Phonemic systems are part of this core area, and they can be used to bundle local dialects into regional groups with relatively clear boundaries ⫺ such as groups with identical lowering and raising relationships. This is possible because the diffusion of innovations is not a random process, but is subject to (among others) language-internal constraints, which appear to be decisive for the consistency of the regional types; conversely, coherent regions can only be defined using features that are linguistically connected. An alternative way of establishing types is offered by dialectometry (Goebl 1982). It is true that it does not use any linguistically connected features to do so, but nor does it ignore any of them, because, say, they do not follow the desired borders or do not fit with the chosen linguistic theory. Naturally, it can hardly elicit precisely which linguistic and extralinguistic factors have given rise to a particular dialectometric geotype and its geographic extent. But the law of large numbers could lead to regions that are also real for their speakers and which codetermine the further development of the language. It is therefore unlikely to be a coincidence that an explorative dialectometric computation of the SDS has produced type zones (Typenareale) that correspond quite well with the Raising Zones of this diffusion study (Kelle 2001: 29A).

10. Reerences Auer, Peter and Frans Hinskens 2005 The role of interpersonal accommodation in a theory of language change. In: Peter Auer, Frans Hinskens and Paul Kerswill (eds.), Dialect Change: Convergence and Divergence in European Languages, 335⫺357. Cambridge: Cambridge University Press. Bailey, Charles-James N. 1973 Variation and Linguistic Theory. Arlington: Center for Applied Linguistics. Baumgartner, Heinrich 1940 Stadtmundart ⫺ Stadt- und Landmundart. Beiträge zur bernischen Mundartgeographie. Bern: Herbert Lang. Bremer, Otto 1893 Deutsche Phonetik. Leipzig: Breitkopf & Härtel. Britain, David 2004 Space and spatial diffusion. In: J. K. Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.), The Handbook of Language Variation and Change, 603⫺637. Oxford: Blackwell. Domasˇnev, Anatolij. I., S. V. Stirnickaja and Larissa E. Najdicˇ 1980 [Review of Haas 1978]. Voprosy jazykoznanija 3: 129⫺132.

665

666

VII. Exemplary studies Gauchat, Louis 1905 L’unite´ phone´tique dans le patois d’une commune. In: Aus romanischen Sprachen und Literaturen ⫺ Festschrift Heinrich Morf, zur Feier seiner fünfundzwanzigjährigen Lehrtätigkeit von seinen Schülern dargebracht, 175⫺232. Halle/Saale: Niemeyer. Goebl, Hans 1982 Dialektometrie. Prinzipien und Methoden des Einsatzes der numerischen Taxonomie im Bereich der Dialektgeographie. Vienna: Österreichische Akademie der Wissenschaften. Goossens, Jan 1969 Strukturelle Sprachgeographie. Heidelberg: C. Winter. Haas, Walter 1978 Sprachwandel und Sprachgeographie. Untersuchungen zur Struktur der Dialektverschiedenheit am Beispiele der schweizerdeutschen Vokalsysteme. Wiesbaden: Steiner. Haas, Walter 1995 Wenker contra Bremer: Empirie und Theorie des Dialekts. In: Jose´ Cajot, Ludger Kremer and Hermann Niebaum (eds.), Lingua Theodisca: Beiträge zur sprach- und Literaturwissenschaft. Jan Goossens zum 65. Geburtstag, 331⫺340. Münster: LIT. Haas, Walter 1998 Ansätze zu einer Theorie des Sprachwandels auf lautlicher Ebene. In: Werner Besch, Anne Betten, Oskar Reichmann and Stefan Sonderegger (eds.), Sprachgeschichte. Ein Handbuch zur Geschichte der deutschen Sprache und ihrer Erforschung. 2nd ed., 836⫺850. (Handbücher zur Sprach- und Kommunikationswissenschaft, 2.1.) Berlin: de Gruyter. Herrgen, Joachim 2005 Sprachgeographie und Optimalitätstheorie. Am Beispiel der t-Tilgung in Auslaut-Clustern des Deutschen. Zeitschrift für Dialektologie und Linguistik 72: 278⫺317. Hotzenköcherle, Rudolf, Heinrich Baumgartner and Rudolf Trüb 1962⫺2003 Sprachatlas der deutschen Schweiz (SDS). 10 vols. Berne: Francke. Kelle, Bernhard 2001 Zur Typologie der Dialekte in der deutschsprachigen Schweiz: Ein dialektometrischer Versuch. Dialectologia et Geolinguistica 9: 9⫺34. Keller, Rudi 1994 Sprachwandel. Von der unsichtbaren Hand in der Sprache. 2nd ed. Tübingen: Francke. Keller, Rudolf E. 1982 [Review of Haas 1978]. Beiträge zur Geschichte der deutschen Sprache und Literatur 104: 89⫺93. Knoop, Ulrich, Wolfgang Putschke and Herbert Ernst Wiegand 1982 Die Marburger Schule: Entstehung und frühe Entwicklung der Dialektgeographie. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke and Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, 38⫺92. (Handbücher zur Sprach- und Kommunikationswissenschaft, 1.1.) Berlin: de Gruyter. Labov, William 1972 Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Labov, William 1994 Principles of Linguistic Change, vol. 1: Internal Factors. Oxford: Blackwell. Labov, William, Malcah Yeager and Richard Steiner 1972 A Quantitative Study of Sound Change in Progress. Philadelphia: US Regional Survey. Lüdtke, Helmut 1980 Sprachwandel als universales Phänomen. In: Helmut Lüdtke (ed.): Kommunikationstheoretische Grundlagen des Sprachwandels, 1⫺19. Berlin: de Gruyter. Mattheier, Klaus J. 1984 [Review of Haas 1978]. Rheinische Vierteljahrsblätter 48: 324⫺329. Meyer, Paul 1875 [Review of I. Ascoli, Schizzi francoprovenzali]. Romania 4: 294⫺296.

36. A study on areal diffusion

667

Mitzka, Walther 1952 Handbuch zum Deutschen Sprachatlas. Marburg: Elwert. Moulton, William G 1960 The short vowel systems of northern Switzerland. Word 16: 155⫺162. Moulton, William G. 1980 [Review of Haas 1978]. Journal of Linguistics 16: 275⫺288. Ogura, Meiko 1990 Dynamic Dialectology. A Study of Language in Time and Space. Tokyo: Kenkyusha. Pfalz, Anton 1983a [1918] Reihenschritte im Vokalismus. In: Peter Wiesinger (ed.), Die Wiener dialektologische Schule. Grundsätzliche Studien aus 70 Jahren Forschung. 43⫺63. Vienna: Halosar. Pfalz, A. (1983b [1925]) Grundsätzliches zur deutschen Mundartenforschung. In: Peter Wiesinger (ed.), Die Wiener dialektologische Schule. Grundsätzliche Studien aus 70 Jahren Forschung, 85⫺ 100. Vienna: Halosar. Reed, Carroll E. 1982 [Review of Haas 1978]. Journal of English and Germanic Philology 81: 73⫺75. Reiffenstein, Ingo 1986 [Review of Haas 1978]. Zeitschrift für Dialektologie und Linguistik 53: 374⫺379. Siebenhaar, Beat 2000 Sprachvariation, Sprachwandel und Einstellung. Der Dialekt der Stadt Aarau in der Labilitätszone zwischen Zürcher und Berner Mundartraum. Stuttgart: Steiner. Weinreich, Ulrich, William Labov and Marvin I. Herzog 1968 Empirical foundations for a theory of language change. In: Winfred P. Lehmann and Yakov Malkiel (eds.), Directions for Historical Linguistics, 95⫺195. Austin, TX: University of Texas Press, Wiesinger, Peter 1970 Phonetisch-phonologische Untersuchungen zur Vokalentwicklung in den deutschen Dialekten. Berlin: de Gruyter. Wiesinger, Peter 1984 [Review of Haas 1978]. Kratylos 29: 203⫺205. Wrede, Ferdinand 1963 [1919] Zur Entwicklungsgeschichte der deutschen Mundartforschung. In: Luise Berthold, Bernhard Martin and Walther Mitzka (eds.), Ferdinand Wrede: Kleine Schriften, 331⫺ 344. Marburg: Elwert.

Walter Haas, Fribourg (Switzerland)

668

VII. Exemplary studies

37. The Linguistic Atlas o the Middle Rhine (MRhSA): A study on the emergence and spread o regional dialects 1. 2. 3. 4. 5.

The Linguistic Atlas of the Middle Rhine ⫺ Research context and goals Bidimensional dialectology and linguistic dynamics The methods of the MRhSA Results References

1. The Linguistic Atlas o the Middle Rhine - Research context and goals Although interdisciplinarity has, for good reason, been required of the social sciences for some time, research praxis continues to be shaped by disciplinary boundaries. This tendency runs so deep that a complex, multivariant research field can become dissolved into separate, discipline-specific objects, each investigated using a unique methodological arsenal. In the end, those involved barely recognize that they are working in an intrinsically unitary field of study, and instead reify their methodologically constituted and differentiated objects of study into separate segments of reality. The empirical linguistic investigation of language across space provides an example of such a situation. On the one hand, in the process of becoming an independent and established linguistic discipline, traditional dialectology increasingly blocked out social and contextual factors in linguistic variation in the quest for a scientifically sound methodology that exclusively targeted the areal dimension of variation. A similar development is also true of sociolinguistics and pragmatics, which emerged later. Since the 1960s, the communicative-pragmatic turn has occasioned an expansion of the traditional dialectological methods through various new linguistic subdisciplines, each with its own theoretical-methodological approach. Although sociolinguistics, pragmalinguistics, discourse linguistics, etc., cover other aspects than traditional dialectology, this has not resulted in a more complex penetration of the research field, but rather in the constitution of new objects of study. Where dialectology excluded the social-pragmatic dimension, these new disciplines shut out the areal dimension of variation. To this day it is widely believed, within the relevant disciplines, that it is impossible to simultaneously investigate areal and social/situational dimensions of linguistic variation, and geolinguistic research praxis thus continues to be shaped by reductionist impulses. Such a methodologically conditioned reduction of the object of geolinguistic research is particularly awkward when change in regional dialects is investigated. For an adequate description and explanation of the dynamics of language across time and space is only possible when areal and social factors are explored simultaneously and their interdependence is accurately captured (cf. Schmidt in this volume). In the following, the example of the Linguistic Atlas of the Middle Rhine (Mittelrheinischer Sprachatlas, MRhSA) is used to demonstrate that an integrative data collection, analysis, and presentation of simultaneously both the areal and the social dimensions of

37. The Linguistic Atlas of the Middle Rhine language is possible and that important insights into the dynamics of regional dialects can thus be won. The idea behind and explicit goal of the MRhSA was to suspend the traditional yet inadequate subdivision of variation linguistics for a particular linguistic subregion. The MRhSA was the first linguistic atlas in Europe to set itself the dual goal of recording and analyzing the linguistic structures within its survey area bidimensionally, i. e., in terms of both areal extent and social complexity. To this end, the MRhSA systematically relates the diatopic/horizontal and the diastratic/vertical dimensions of linguistic variation to one another: While linguistic atlases usually only document spoken language relative to space, the MRhSA’s object of study is the areal (diatopic) and to some extent the social (diastratic) dimension. Accordingly, it pursues two goals: (1) The discovery and documentation of the areal structure of the least standard domain of spoken language, the base dialects. (2) The discovery and documentation of diastratically determined contrasts in local language forms and thence the changing areal structure in a further non-standard domain of spoken language, the regional dialects. (Bellmann, Herrgen and Schmidt 1989: 285⫺286; my translation)

Through an integrative, albeit partial, simultaneous investigation of the social and areal dimensions of linguistic variation, the MRhSA was supposed to contribute to the integration of sociolinguistics and dialectology. A similar approach had previously been adopted by Hans Kurath in the Linguistic Atlas of New England (1939⫺1943) and by Yoichi Fujiwara in the Linguistic Atlas of the Seto Inland Sea (1974). Eberhard Zwirner’s first series of tape recordings (Lautbibliothek der deutschen Mundarten, later Deutsches Spracharchiv [DSAv], Mannheim, cf. Zwirner 1956) also belongs to this tradition of multidimensional variation linguistics. Yet for various reasons, these projects evoked only limited international resonance, so that the initiator role in bidimensional or pluridimensional dialectology was effectively played by the MRhSA (1994⫺2002) and later the Atlas Lingüı´stico Diato´pico y Diastra´tico del Uruguay (ADDU; Thun and Elizaincı´n), which began to appear in 2000. The model of multidimensional dialectology thus established has since become quite productive (cf., e.g., Mang 2004), even though projects around the world also continue to follow the earlier tradition of division by discipline (cf., e.g., Labov 2006).

2. Bidimensional dialectology and linguistic dynamics The Linguistic Atlas of the Middle Rhine’s integrative investigation of social and areal factors in linguistic variation makes it possible to explore linguistic dynamics in space and time. In particular, stabilizing and innovative processes (the stasis, spread, and shrinkage of language areas) can thus be systematically analyzed. Here too, a decisive theoretical/methodological innovation of the MRhSA can be seen. In the past, geolinguistics was primarily concerned with the least standard areal varieties, the dialects. The conspicuous dynamic of the areal varieties was regarded by dialectologists as troublesome at best, since they saw in it a tendency for dialects to disappear, taking with them the widely acknowledged focus of their discipline. The geolinguistic research tradition is

669

670

VII. Exemplary studies therefore dominated by studies which more or less deliberately ignore the dynamics of areal varieties. In contrast, the MRhSA views such dynamics not as interference to be eliminated via suitable collection and analysis methods, but rather as a constitutive feature of the object of study. Here, areal varieties are regarded as constantly changing sectors (of restricted register and area) of the entire complex of a single language. This article will examine the methods used in the MRhSA to explore linguistic dynamics across space and the research findings these methods have generated.

3. The methods o the MRhSA 3.1. The region studied The MRhSA investigates areal linguistic structures in West Middle (Central) German, more precisely, in those parts of the federal States of Rhineland-Palatinate and Saarland that lie to the west of the Rhine River, i. e., in an area bounded to the east by the Rhine, to the west by the border to France, to the south by the Wieslauter River and to the north by the Ahr (see Map 37.1). In linguistic terms, the domain of the MRhSA is clearly divided into two: the Rhine-Franconian dialects of the southwest and the Moselle-Franconian dialects as part of the Middle Franconian grouping in the northwest. The two regions are connected by a broad transition zone running from Saarbrücken in the direction of Koblenz (cf. Wiesinger 1983: 846⫺859). What makes the domain of the MRhSA interesting for linguistic studies, aside from the dialectal structure across space, is its high sociodemographic diversity: in part progressive (modernized), in part conservative;

Map 37.1: Domain of the MRhSA (cf. Schmidt 1986: 229)

37. The Linguistic Atlas of the Middle Rhine

671

very different linguistic dynamic impulses can be expected in the different subregions. A distinction can be drawn between (1) conservative rural regions like the Eifel and the Hunsrück, (2) stagnating former industrial regions like Saarland, and (3) modern conurbations like the Frankfurt Rhine Main Region and the area around Ludwigshafen and Mannheim.

3.2. Data collection and analysis The MRhSA investigates the phonology and morphology of the dialects within the chosen domain. Lexis and syntax are not within its ambit, but the lack of a word-geographic section is compensated for by the otherwise good lexicographic coverage of the area (cf. Pfälzisches Wörterbuch [Palatinate Dictionary], Südhessisches Wörterbuch [South Hessian Dictionary], Rheinisches Wörterbuch [Rhenish Dictionary]). An investigation of the spatial distribution of syntax will remain a desideratum for some time, however. The phonetic-phonological and morphological data of the MRhSA were collected directly using a questionnaire procedure that combined traditional and innovative methods: The basis of the informant survey was a questionnaire developed by Günter Bellmann (cf. Bellmann 1994a: 165⫺207), which featured the target lemmas within sentences. The elicitation of the linguistic data proceeded via oral translation of the questionnaire sentences into the appropriate local dialect; this occurred in group interviews of three to five informants per location. This thoroughly traditional survey technique cannot aim to elicit current performance data from the informants; rather, it was intended as a measure of dialectal competence. The informants’ inevitable corrections of one another about typical dialect features which arose in group interviews and discussions were seen not as interference but rather as welcome data for the assessment of competence. What distinguishes the MRhSA from traditional regional atlases in this data collection phase is, firstly, its bidimensionality. In a sense the MRhSA is ⫺ a novelty in the history of linguistic atlases ⫺ two atlases in one. In the first place, an initial phase pursued traditional data collection methods. At each survey location, teams of informants who had lived their lives in the one locality, as had their parents before them, who were more than 70 years old and had worked in a manual occupation, were selected and interviewed. The MRhSA’s key innovation lay in a second, additional series of recordings. In these, the same methods were employed with a younger set of informants. A maximum of two years elapsed between the collection of the two data series. The informants for the second survey also had blue-collar jobs and their families had lived in the one locality for at least two generations. But in contrast to the first series they were between 30 and 40 years old and commuted to work daily. The “typical” informant for the MRhSA was thus an old farmer for series 1, and a commuting tradesperson or laborer of the middle generation for series 2 (cf. Table 37.1). Tab. 37.1: MRhSA: Sociodemographic data about the informants Data series 1 1. 2. 3. 4.

Data series 2

<Mobility: ⫺> <Mobility: ⫹>

672

VII. Exemplary studies On the basis of such an admittedly arduous bidimensional survey, the effects of differences in the informants’ age and geographical mobility on the areal structures of the dialects can be precisely determined. It was important that the MRhSA survey aimed to capture the deepest achievable active dialectal competence of both demographic groups ⫺ the methodological consistency rendered the two series directly comparable. Focusing on the deepest active dialectal competence meant that a well-known fallacy that often crops up in intergenerational linguistic comparisons could be avoided, a fallacy which, on the basis of differences in usage alone, has in the past wrongly been used to predict dialect change, even dialect death. Data were elicited between 1978 and 1988. A total of 2,510 informants from 841 localities were interviewed. The results of the survey (sound recordings and IPA transcriptions) have been archived. This archive consists of 841 logbooks, in which the informants’ responses (only the lemmas) are transcribed using the International Phonetic Alphabet, and a collection of around 2,500 hours of sound recordings. The MRhSA is the first linguistic atlas to permanently record the data collected on tape and phonetically (on-location transcription). The MRhSA sound recordings are currently being digitalized, so that they will soon be available to the general public (cf. ; ).

3.3. Mapping and publication For a linguistic atlas, the primary tool of data analysis and presentation is the map. The linguistic maps of the MRhSA present the linguistic information in point symbol format on a geographically precise base map. The linguistic maps of the atlas were created in a semi-computative process in keeping with the information technology of the time: without automatic map generation from a database, but with computer plotting tools that allowed the inclusion of hand-drawn maps and elements and diverse options for manipulating symbols. As a result of the bidimensional data collection there are many bidimensional linguistic maps in the MRhSA. In line with its dual research goal (the documentation of the areal structure of the base dialects and the documentation of diastratically determined local contrasts), two kinds of maps are employed: (1) the full view (at a scale of 1 : 600,000) and (2) the contrast view (1 : 1,000,000). The full view serves to display the base dialectal areal structure as completely as possible in terms of, first, the language system including the historical distribution of phonemes, and second, the density of survey locations, reaching 23 percent here. The contrast views are paired maps which present the dialects of the older informants (always on the left) and the younger (on the right) across an identical selection of locations. No attempt is made to depict the full depth of data. Firstly, the network of locations is less dense than in the complete view (around 12 percent), and secondly, contrast views are only produced when the information they present is not redundant. The assignment of symbols is systematic and uniform across all maps: (1) monophthongs are always displayed as lines, diphthongs as shapes; (2) the thickness of the lines symbolizes phonetic duration: thin lines correspond to phonetic shortness, thick lines to length (in the case of diphthongs, overlength); (3) intraseries variants are always mapped and separated from one another by a comma; (4) the legend is consistent for each phoneme or morphological category. Contrasting information is also systematically encoded:

37. The Linguistic Atlas of the Middle Rhine

Map 37.2: Contrast map (Bellmann, Herrgen and Schmidt 1994⫺2002: vol. 4, Map 332/1⫺2 beißen)

(1) in cases where both data series match (“zero contrast”), the symbol remains black in the contrast maps, but where the dialect of the younger generation contrasts with that of the older generation, the symbol is set in red on the right-hand-side contrast view map (dialect of the younger generation); (2) if a variant or a diacritic is absent, the contrast map features a red zero. This visualization method is designed to allow the atlas user to analyze the dialectal contrasts by location and assess their implications for a constant or changing areal distribution. This visualization method thus provides an adequate depiction of the interdependence of diatopic and diastratic content and permits the detection of linguistic dynamic tendencies from the maps (see Map 37.2 for a detail from a contrast map). Various pieces of supplementary information found on the map sheet also serve to facilitate interpretation of the atlas maps. (1) The base maps include a selection of original records (IPA transcriptions), which provide a context for the segments displayed on the map. (2) A relevant vowel or consonantal context is sketched on the point symbol maps using isoglosses to delineate certain systematic distributions. (3) The contrast maps include bar graphs to support a quantitative analysis of the linguistic features mapped. Publication of the MRhSA in its full linguistic scope (phonetics/phonology and morphology) as originally planned is complete; five volumes of 1073 maps and an introduction are available. From the outset, the MRhSA was also a linguistic experiment: its novel theoretical/methodological approach instigated a series of pilot studies, accompanying analyses and then data evaluations that tested its analytic potential. The following studies must be mentioned: Bellmann 1982, 1983, 1985, 1986a, 1986b, 1987, 1989, 1990, 1994a, 1994b, 1995, 1996a, 1996b, 1997, 1998; Bellmann, Herrgen and Schmidt 1989; Drenda 2000; Girnth 2000; Herrgen 1986, 1994, 2000, 2001, 2006a, 2006b; Herrgen and Schmidt 1985, 1986, 1989; Kehrein, Lameli and Nickel 2005; Lameli 2004; Lenz 2003, 2004, 2007; Orlovic-Schwarzwald and Schmidt 1986; Rabanus 2004, 2005, 2008; Schmidt

673

674

VII. Exemplary studies 1986, 1988, 1992, 1993, 1998, 2005; Schmidt and Herrgen (to appear); Schmitt 1986, 1992; Smazal 1986; Steiner 1994; Thinnes 1981. More complete bibliographic details can be found in the reference sections of the atlas volumes and at .

4. Results 4.1. Structural boundaries in West Middle German Its specific methodological innovations aside, the MRhSA has generated a whole series of research findings. For a start, the MRhSA represents a full regional atlas of the phonology and morphology of West Middle German. As such, it provides systematic linguistic information on a German region that was formerly in large part (especially the Moselle-Franconian region) only very sporadically covered (cf. Bellmann, Herrgen and Schmidt 1994⫺2002: Maps 13, 479a). Even without the bidimensional methods, the comprehensive treatment of, first, phonology (including difficult to access phenomena like the Rhenish accentuation, centralization, quantity-based distinctions in the consonantal system, etc.) and second, the systematic geographic depiction of inflectional and word-formation morphology represents an impressive scholarly contribution. These results cannot be discussed here in detail. But the summarizing analysis of the spatial structure of dialect distribution to be found in volume 4 of the MRhSA is noteworthy. It classifies the atlas area along structural boundaries distilled from an aggregate of maps. Adopting Wiesinger’s (1970, 1983) approach, the MRhSA team (Bellmann, Herrgen and Schmidt 1994⫺2002: vol. 5, IX⫺XI and Map 479b) show that the area of the MRhSA is strikingly characterized by such dialectal isolines, which delineate structural differences between subregions. These “structural boundaries” are each indicated by clear bundles of isoglosses and have proved highly significant for, on the one hand, the division of the atlas area into dialectality regions (section 4.2.2) and for detecting trends in dialect change (see sections 4.2.3, 4.3) on the other. The MRhSA establishes the following structural boundaries (cf. Map 37.3) as decisive for the areal structure of the base dialects. Foremost is the tone accent boundary; this is the most important structural boundary in the West Middle German area and separates the Moselle-Franconian (with a tone accent distinction) from the Rhine-Franconian (no tone accent distinction) dialects. The significance of this boundary is a result of the prosodic difference itself, its effects on vowel quality (phonological split; cf. Wiesinger 1970: I, 65, 127 and II, 178⫺179) and quantity (cf. Schmidt 1986: 185⫺191). Second in terms of relevance is the “unrounding” boundary, which forms the southern border of the Ripuarian⫺Moselle-Franconian transition zone. It separates the southern dialects without vowel rounding from the northern dialects, which feature vowel rounding. Further system boundaries include that of the diphthongization of MHG eˆ-(œ)-oˆ and that of the phoneme merger of MHG ei-(öu)-ou. Together, these two separate the southern Palatinate relict zone off from the rest of the Rhine-Franconian region. Yet other structural boundaries demarcate the southern and northern boundaries of the Rhine-Franconian⫺Moselle-Franconian transitional zone ⫺ to the south, the boundary of the maintenance of the Rhine-Franconian distinction between MHG e¨ and e/ä; to the north, the <s>-voic-

37. The Linguistic Atlas of the Middle Rhine

Map 37.3: Structural boundaries in the MRhSA (Bellmann, Herrgen and Schmidt 1994⫺2002: vol. 5, Map 479b)

ing boundary. A final structural boundary defines the eastern extent of the Eifel relict zone: this delimits the retention of long obstruents in medial syllable boundaries first documented in the MRhSA. Taken as a whole, the structural boundaries delineate the relevant subregions within the atlas area. An indicator of the high diatopic relevance of

675

676

VII. Exemplary studies the structural boundaries is their covariation with the degree of interdialectal intelligibility: Schmitt (1992: 118⫺194) was able to show that it is the language regions delimited by structural boundaries which determine the extent of interdialectal intelligibility. This finding can now also be explained in terms of a theory: as Schmidt demonstrates in this volume, structural regions are also synchronization regions and structural boundaries are synchronization boundaries. It can no longer come as a surprise that structural regions are also regions of mutual intelligibility, nor that processes of regional language change are contained by structural boundaries.

4.2. The yield o the bidimensional method 4.2.1. The interseries comparison The decisive innovative aspect of the MRhSA is the bidimensional method which enables a systematic investigation of the connection between the diatopic and diastratic dimensions and thus an exploration of trends in dialect change. During the final phases of data collection for the MRhSA, Günter Bellmann analyzed the initial results recorded in the survey protocols and draft maps (cf. Bellmann, Herrgen and Schmidt 1989: 301⫺ 305). It became clear that with data series 1 (the older generation), the MRhSA had accessed a “language state of extremely high dialectality” (Bellmann, Herrgen and Schmidt 1989: 301; my translation). As far as data series 2 (the younger generation) is concerned, the MRhSA documents a state of spoken language that is still highly dialectal yet increasingly locally undifferentiated. Bellmann interprets the contrasts between the two data series diachronically and diatopically: “The more coarse-grained areality, which is associated with reduced dialectality, can be understood as diatopic change” (Bellmann, Herrgen and Schmidt 1989: 303; my translation). All together, Bellmann claims, “the target variety of this development may be an emergent regional dialect” (Bellmann, Herrgen and Schmidt 1989: 305). Thereby, the dialectal competence of the younger informants is increasingly oriented to the region, not the locality. Where new regional dialects emerge, an independent development can be observed, in parts characterized by innovations that even lead away from the New High German Standard. The Linguistic Atlas of the Middle Rhine documents numerous cases showing that the regional dialectal form is preferred over both the local dialectal form and that of the standard language, which might have been expected to be especially supported by the media and school. The prevailing trend documented by the MRhSA can thus be best encapsulated under the rubric of “dialectal regionalization”. It is also evident in the frequency counts that accompany the MRhSA maps: the predominant trend is for rare and highly localized features to be replaced by more common and widespread ones. Yet differentiated analyses reveal that there is no simple and linear reduction in dialectal features. What can instead be observed is a dialectal redistribution (Umdialektalisierung) that takes different forms depending on the base dialectal areal structure, including both the stability and the replacement of base dialectal features. Bellmann describes different forms of buildup and decline in linguistic variability. He typologizes the interseries contrasts as follows: Type 1: substitution (A > B); Type 2: devariabilization (A, B > B); Type 3: variabilization (A > A, B); Type 4: zero contrast (A > A; Bellmann 1994a: 302⫺303). Exploring consonantal systems, Herrgen (2000) could show that these

37. The Linguistic Atlas of the Middle Rhine developments unfold in accordance with locally differentiated structural preconditions: Depending upon the base dialectal structure, a single map can chart the dominance of standard-conform substitution in one subregion and dialectal regionalization in another.

4.2.2. Dialectality and change In light of the variation in substitution and (de-)variabilization processes across space, the MRhSA authors emphasized the need to achieve an overall estimate. This was intended to show, firstly, how dialectal features are grouped at a location and across a region, so that the total dialectality typical of a location or a region could be ascertained. Secondly, such an analysis ought to show the extent to which these aggregate values differ for the two data series, and hence whether there is any “dialect decline” in general. To this end, separate dialectality measurements for both data series were undertaken (cf. Bellmann, Herrgen and Schmidt 1994⫺2002: vol. 4, VII and Map 314; Herrgen and Schmidt 1989; Bellmann 1998). The phonetic measurement technique used (cf. Herrgen and Schmidt 1989) counted the number of phonetic features per survey location that deviated from moderate standard pronunciation. The dialectality scores in the MRhSA range from a minimum of 201 (or two nonstandard phonetic features per word) to a maximum of 349.5 (or an average of three and a half nonstandard phonetic features per word). The following results are noteworthy: (1) four clear subregions of differing dialectality became apparent (West Eifel mountains, Moselle⫺eastern Eifel, Saarpfalz, southern Palatinate); (2) the dialectality in data series 2 is approximately five percent lower than in data series 1; (3) an internal leveling of dialectality can be seen within all of the subregions. The analysis of these measurements permits an initial response to the question of dialect decline ⫺ for the MRhSA area. Contrary to the expectations of nonexperts and the dominant view in the research literature, within the population segment that the Linguistic Atlas of the Middle Rhine investigated, namely two generations of locally born blue-collar workers, there is no evidence of a general or a partial disappearance of the dialects. The dialect of the younger generation is indeed closer to the spoken standard than the dialect of the more senior generation, but only by about five percent overall. To put it another way: ninety-five percent of the dialectality is retained. Standard convergence in the dialect of the younger generation is thus evident from the sum of the phenomena, yet their dialect can still be classed as strongly dialectal. It must be stressed that material from a linguistic atlas like the one presented here can only offer a narrow view of trends in dialectal change, since a linguistic atlas survey is characterized by regional, survey contextual, and social restrictions. But for use in conjunction with other dialectological projects, such as dialect census projects and linguistic variation studies for example, it provides an indispensable empirical basis, from which recent diatopic trends in dialect change can be tackled. A comparison of the dialectality measurements (Bellmann, Herrgen and Schmidt 1994⫺2002: vol. 4, Map 314) with the analysis of structural boundaries already mentioned (Bellmann, Herrgen and Schmidt 1994⫺2002: vol. 5, Map 479b) is instructive. This reveals that the quantitative and qualitative (systemic) differences correlate: the structural regions are also dialectality regions. But this does not mean that the differences in dialectality are simply a product of the mapped structural distinctions. In con-

677

678

VII. Exemplary studies trast, seen in isolation, some of the structural boundaries ought even to generate diametrically opposed dialectality contrasts (cf. vowel rounding or the voiced/unvoiced opposition). If a structural region displays a higher total dialectality than neighboring structural regions, this is because structural regions act as regions of synchronization. Internally, they possess a higher degree of consistency than they do across the structural boundary. Incorporating the findings on interdialectal intelligibility described above, we could say that structural boundaries are dialectality boundaries, which in turn are intelligibility boundaries.

4.2.3. The spread o regional dialects The central questions for geolinguistics ⫺ whether, how and to what extent dialects spread ⫺ can only be answered if an adequate empirical basis is available. The Linguistic Atlas of the Middle Rhine provides such a basis. The limited broadscale or local decline in areal features aside, a relatively comprehensive inventory of areal phenomena has been discovered in the MRhSA survey data and its analysis. Particularly interesting are those cases of regional language diffusion in which local or regional language features spread, i.e., divergence from the standard. In these cases, we can safely assume original regional linguistic processes are at work. In the following, three examples of nonstandard areal diffusion that have emerged from analyses done for the MRhSA are discussed. The first two examples are phonetic/phonological, the third morphological.

The diusion o /s/-voicing An example of standard-divergent spread of a linguistic phenomenon can be seen on MRhSA Map 332, beißen ‘to bite’, where in the Rhine-Franconian dialect of the younger generation a highly frequent replacement of the voiceless dental spirant by a voiced one can be observed (cf. Map 37.2). The function of the voiced/unvoiced opposition in Rhine Franconian lies behind this. There is a tendency, especially among the younger generation, but also already among the older generation, to treat voiced and unvoiced dental spirants as combinatory variants. In the beißen map, unvoiced [s] dominates in the base dialect, as in the standard language, but in data series 2 voiced [z] dominates across the entire Rhine-Franconian area (except for the southern Palatinate relict zone). In the regional dialect of the broad area between the southern Palatinate relict zone, in which voiced [z] is generally absent, and the Moselle-Franconian area, in which there is a phonological opposition between /z/ and /s/, voiced [z] thus becomes an allophonic variant in voiced contexts. In this case we are thus dealing with a sonorization of the [s] in voiced environments supported by phonetic naturalness. The spread reaches its limits at the northern and southern structural boundaries (northwest boundary of the RhineFranconian⫺Moselle-Franconian transition zone, boundary of the southern Palatinate relict zone).

Coronalization o [c¸ ] Coronalization (cf. Herrgen 1986; Schmidt and Herrgen to appear) is the replacement of a dorsal [c¸] with a coronal [C], which phonetically lies between [ s ] and [c¸]. The coronalization, currently very widespread in the Middle German dialects, first emerged polygenetically in the major cities of the Middle German region in the mid-nineteenth century.

37. The Linguistic Atlas of the Middle Rhine

Map 37.4: Coronalization in the MRhSA (based on Bellmann, Herrgen and Schmidt 1994⫺2002: vol. 4, Map 350/1⫺2)

The MRhSA maps (cf. Bellmann, Herrgen and Schmidt 1994⫺2002: vol. 4, Maps 349, 350) show coronalization to have already been areally dominant in data series 1 (cf. Map 37.4), with no sign of this development halting in the dialect of the younger generation. Outside of the relatively small southern Palatinate relict zone, the coronal spirant is now practically universal in the broader area. It should be stressed that we are dealing with a regional language innovation that continues to diverge away from both the spoken standard and the traditional dialects. A whole series of factors have been invoked to explain this phenomenon (cf. Herrgen 1986). (1) As a result of a number of historical linguistic processes, the set of voiceless fricatives is overfull in German, making a distinction difficult. A neutralization of the /c¸/⫺/s/ distinction can thus be classified as an articulatory simplification. (2) The functional load of the /c¸/⫺/s/ opposition is light; there are only a handful of minimal pairs (Kirche/Kirsche ‘church’/‘cherry’, selig/seelisch ‘blessed’/‘mental/psychic’, Männchen/Menschen ‘male’/‘people’). It therefore does not prevent the opposition from being neutralized. (3) The “melting-pot” situation in the rapidly growing cities during the nineteenth-century modernization phases brought with it a relatively high degree of norm tolerance. This meant that the /c¸/⫺/s/ opposition could be rapidly abandoned in Middle German (in contrast to Low or Upper German, where other historical linguistic conditions prevailed). (4) In subsequent generations the coronalization was also adopted in the villages. Here, the above factors continued to apply, with the addition of the prestige of the cities from which it originated. With the coronalization process, the MRhSA maps again illustrate an impressive regional linguistic diffusion process that diverges from both the spoken standard and the local dialects. But the significance of the structural boundaries discussed above is also evident here ⫺ the southern Palatinate relict zone has until now remained unaffected by the coronalization (cf. Map 37.4).

679

680

VII. Exemplary studies A third example of the spread of destandardizing linguistic features is the distribution of the past participle of bringen ‘to bring’, which Schmidt (in this volume) analyzes from a linguistic dynamic perspective. As further evidence for the fertility of the bidimensional material and for the high systematicity of the morphological areal contrasts we could also cite, e.g., Girnth (2000), who analyzes the MRhSA data from the point of view of an empirical investigation of grammaticalization and is able to demonstrate a high degree of areal regularity.

4.3. Future research perspectives: Areal structure and linguistic dynamics A polydimensional atlas like the MRhSA gains additional linguistic dynamic explanatory power when it is systematically related to comparable geolinguistic surveys undertaken at different points in time. A double advantage can be gained from such a comparison: on the one hand, a linguistic dynamic investigation like the MRhSA receives additional temporal depth; on the other hand, the validity of a comparison of linguistic data from differing times is significantly improved when one of the data sources has been designed along linguistic dynamic lines, in that it can then be unequivocally determined whether differences in the areal distribution at various times really do represent linguistic dynamic processes or whether they are merely artefacts reflecting differences in the survey methods. Thanks to its integration into trendsetting linguistic dynamic research platforms like the Digital Wenker Atlas (DiWA; ) or the “regionalsprache.de” project (REDE; ), the MRhSA can once again serve as an example. These research platforms aim to thoroughly investigate the structure and dynamics of German by integrating diverse data collections into a single online analysis system. Although they are fundamentally similar in approach, DiWA is confined to material and results that are directly comparable to Georg Wenker’s Sprachatlas des deutschen Reichs ‘Linguistic Atlas of the German Empire’, while REDE will encompass a significantly larger collection of material, including new, precise surveys of the nearstandard regional registers (regional accents). The MRhSA’s incorporation into DiWA reveals the opportunities thus opened up: systematic comparison of the Wenker atlas (data collected 1876⫺1887) with the MRhSA (data collected 1978⫺1988) produces a real-time difference of around 100 years; the interseries comparison of the MRhSA data adds a further “apparent-time” difference of 30 years. For linguistic dynamic purposes, at least three different points in time can be contrasted ⫺ the future integration of additional comparable data (sound recordings of different age, monographs, etc.) will bring a further refinement of the linguistic dynamic perspective. Already, the combination of as yet relatively few atlases within a single research platform is rendering highly systematic linguistic dynamic processes visible (cf. Schmidt 2005; Rabanus 2004, 2005, 2008). The future possibilities already intimated can once again be demonstrated using an example of the spread of standard-divergent dialectal features, this time over the period of a century.

Deletion o -t in hast The deletion of -t in the verb form hast (2sg of haben ‘to have’) occurs in different regions (cf. Goeman 1999; Herrgen 2006a, 2006b; Rabanus 2008). Forms such as has,

37. The Linguistic Atlas of the Middle Rhine

Map 37.5: Linguistic dynamics of t-deletion (based on Herrgen 2006a: 135)

hos, hasch, hosch are widely found in Middle and Upper German dialects. Map 37.5 compares the situation recorded in the Wenker Atlas, the MRhSA (older generation) and the MRhSA (younger generation). A circle is used to mark t-deletion; t-retention is not indicated. The map shows t-deletion even in the Wenker Atlas (for Moselle Franconian and the southern Palatinate). Comparison with the MRhSA reveals a similar picture with the extent of deletion slightly altered: t-deletion has visibly expanded in both the Rhine-Franconian and the Moselle-Franconian regions. It is remarkable, then, that the trend has continued in the dialect of the younger generation (progressive t-deletion). Hence, a dynamic areal trend in Middle German over 100 years can be confirmed: the spread of t-deletion. That speakers abandon local dialectal features in the course of this development is of little surprise. But what is very surprising is the long-term independent diffusion of a regional linguistic feature that is diverging away from the norm of the spoken standard. (Those who would interpret the dynamics of regional speech as a onedimensional decline of the dialects and convergence on the standard need to be confronted with this finding.) Here, as in the other cases of areal linguistic dynamics discussed above, the factors that motivate the development can be identified. Optimization processes are at work in the morphological system (cf. Goeman 1999: 157), especially involving the factor of “contrastivity”. For those Middle and Upper German inflectional paradigms in which t-deletion occurs, t-deletion in the second-person singular enhances contrastivity, since the morphological contrasts between the inflectional forms are more marked. The spread of the innovative forms has been in progress for over 100 years via (meso-)synchronization between interactants in the affected German regions. What makes it attractive for the speakers to adopt forms that are supported by neither the base dialectal nor the standard language structures is a type of morphological optimization, namely their increased paradigmatic contrastivity.

5. Reerences Bellmann, Günter 1982 Deskriptive Sprachgeographie in der Gegenwart. Zu Konzept und Praxis des Mittelrheinischen Sprachatlasses. Rheinische Vierteljahrsblätter 46: 271⫺287.

681

682

VII. Exemplary studies Bellmann, Günter 1983 Probleme des Substandards im Deutschen. In: Klaus J. Mattheier (ed.), Aspekte der Dialekttheorie, 105⫺130. (Reihe Germanistische Linguistik 46.) Tübingen: Niemeyer. Bellmann, Günter 1985 Substandard als Regionalsprache. In: Georg Stötzel (ed.), Germanistik ⫺ Forschungsstand und Perspektiven. Vorträge des deutschen Germanistentage 1984, vol. 1, 211⫺218. Berlin/ New York: de Gruyter. Bellmann, Günter (ed.) 1986a Beiträge zur Dialektologie am Mittelrhein. (Mainzer Studien zur Sprach- und Volksforschung 10.) Stuttgart: Steiner. Bellmann, Günter 1986b Zweidimensionale Dialektologie. In: Günter Bellmann (ed.), Beiträge zur Dialektologie am Mittelrhein, 1⫺55. (Mainzer Studien zur Sprach- und Volksforschung 10.) Stuttgart: Steiner. Bellmann, Günter 1987 Der Mittelrheinische Sprachatlas und das Pfälzische (mit einer Karte). In: Wolfgang Kleiber (ed.), Symposion Ernst Christmann. Veranstaltet von der Pfälzischen Gesellschaft zur Förderung der Wissenschaften, Speyer. (Kaiserslautern 8./9. November 1985). Vorträge zur Dialektlexikographie, Sprachgeographie und Volksforschung des Westmitteldeutschen, 75⫺ 87. (Mainzer Studien zur Sprach- und Volksforschung 11.) Stuttgart: Steiner. Bellmann, Günter 1989 Variation und Devariation. In: Albrecht Greule and Uwe Ruberg (eds.), Sprache. Literatur. Kultur. Studien zu ihrer Geschichte im deutschen Süden und Westen. Wolfgang Kleiber zu seinem 60. Geburtstag gewidmet, 203⫺213. Stuttgart: Steiner. Bellmann, Günter 1990 Ältere und neuere Arbeiten zur rheinischen Dialektologie. In: Rudolf Große (ed.), Sprache in der sozialen und kulturellen Entwicklung. Beiträge eines Kolloquiums zu Ehren von Theodor Frings (1886⫺1968), 60⫺64. (Abhandlungen der Sächsischen Akademie der Wissenschaften zu Leipzig. Philol.-hist. Klasse 73: 1.) Berlin: Akademie-Verlag. Bellmann, Günter 1994a Einführung in den Mittelrheinischen Sprachatlas (MRhSA). Tübingen: Niemeyer. Bellmann, Günter 1994b Multidimensionale Dialektgeographie? In: Klaus J. Mattheier and Peter Wiesinger (eds.), Dialektologie des Deutschen. Forschungsstand und Entwicklungstendenzen, 165⫺169. (Reihe Germanistische Linguistik 147.) Tübingen: Niemeyer. Bellmann, Günter 1995 Hans Reis und der Beginn dialektologischer Forschung am Mittelrhein. In: Winfried Dotzauer, Wolfgang Kleiber, Michael Matheus und Karl-Heinz Spiess (eds.), Landesgeschichte und Reichsgeschichte. Festschrift für Alois Gerlich zum 70. Geburtstag, 417⫺432. (Geschichtliche Landeskunde 42.) Stuttgart: Steiner. Bellmann, Günter 1996a Arealität und Sozialität? Avec un re´sume´ en Franc¸ais. In: Edgar Radtke and Harald Thun (eds.), Neue Wege der romanischen Geolinguistik. Akten des Heidelberger und Mainzer Kolloquiums zur empirischen Dialektologie. 21.⫺24. Oktober 1991, 50⫺77. (Dialectologia pluridimensionalis Romanica 1.) Kiel: Westensee. Bellmann, Günter 1996b Grenzen und Linien als Darstellungsmittel und als Probleme der Dialektologie. In: Hildegard Frieß-Reimann and Fritz Schellack (eds.), Kulturen, Räume, Grenzen. Interdisziplinäres Kolloquium zum 60. Geburtstag von Herbert Schwedt, 59⫺78. (Studien zur Volkskultur in Rheinland-Pfalz 19.) Mainz: Gesellschaft für Volkskunde in Rheinland-Pfalz. Bellmann, Günter 1997 Zur Technik und Aussagefähigkeit zweidimensionaler Dialekterhebung und Dialektkartographie am Beispiel des Mittelrheinischen Sprachatlasses. In: Gerhard Stickel (ed.),

37. The Linguistic Atlas of the Middle Rhine Varietäten des Deutschen. Regional- und Umgangssprachen, 271⫺290. Berlin/New York: de Gruyter. Bellmann, Günter 1998 Between base dialect and standard language. Folia Linguistica 32: 23⫺34. Bellmann, Günter, Joachim Herrgen and Jürgen Erich Schmidt 1989 Der Mittelrheinische Sprachatlas (MRhSA). In: Werner Veith and Wolfgang Putschke (eds.), Sprachatlanten des Deutschen. Laufende Projekte, 285⫺313. (Studien zum Kleinen Deutschen Sprachatlas 2.) Tübingen: Niemeyer. Bellmann, Günter, Joachim Herrgen and Jürgen Erich Schmidt 1994⫺2002 Mittelrheinischer Sprachatlas (MRhSA). Unter Mitarbeit von Georg Drenda, Heiko Girnth und Marion Klenk. Tübingen: Niemeyer. Drenda, Georg 2000 Zentralisierung. Probleme der Vokalentwicklung im Westmitteldeutschen. (Mainzer Studien zur Sprach- und Volksforschung 23.) Stuttgart: Steiner. Fujiwara, Yoichi 1974 A Linguistic Atlas of the Seto Inland Sea. 2 vols. Tokyo: University of Tokyo Press. Girnth, Heiko 2000 Untersuchungen zur Theorie der Grammatikalisierung am Beispiel des Westmitteldeutschen. (Reihe Germanistische Linguistik 223.) Tübingen: Niemeyer. Goeman, Antonie Cornelis Maria [Ton] 1999 T-deletie in Nederlandse dialecten. Kwantitatieve analyse van structurele, ruimtelijke en temporele variatie. (LOT international series 26.) The Hague: Holland Academic Graphics. Herrgen, Joachim 1986 Koronalisierung und Hyperkorrektion. Das palatale Allophon des /CH/-Phonems und seine Variation im Westmitteldeutschen. (Mainzer Studien zur Sprach- und Volksforschung 9.) Stuttgart: Steiner. Herrgen, Joachim 1994 Kontrastive Dialektkartographie. In: Klaus J. Mattheier and Peter Wiesinger (eds.), Dialektologie des Deutschen. Forschungsstand und Entwicklungstendenzen, 131⫺163. (Reihe Germanistische Linguistik 147.) Tübingen: Niemeyer. Herrgen, Joachim 2000 Dialektgeographie und Dialektwandel. Zu rezenten konsonantischen Entwicklungstendenzen im Westmitteldeutschen. In: Dieter Stellmacher (ed.), Dialektologie zwischen Tradition und Neuansätzen. Beiträge der internationalen Dialektologentagung, Göttingen, 19.⫺ 21. Oktober 1998, 48⫺64. (Zeitschrift für Dialektologie und Linguistik. Beiheft 109.) Stuttgart: Steiner. Herrgen, Joachim 2001 Die Dialektologie des Deutschen. In: Sylvain Auroux et al. (ed.), Geschichte der Sprachwissenschaften. Ein internationales Handbuch zur Entwicklung der Sprachforschung von den Anfängen bis zur Gegenwart, vol. 2, 1513⫺1535. (Handbücher zur Sprach- und Kommunikationswissenschaft 18.2.) Berlin/New York: de Gruyter. Herrgen, Joachim 2006a Die Dynamik der modernen Regionalsprachen. In: Joachim Gessinger and Anja Voeste (eds.), Dialekt im Wandel. Perspektiven einer neuen Dialektologie, 119⫺142. (Osnabrücker Beiträge zur Sprachtheorie 71.) Bremen: Redaktion OBST. Herrgen, Joachim 2006b Sprachgeographie und Optimalitätstheorie. Am Beispiel der t⫺Tilgung in Auslaut-Clustern des Deutschen. Zeitschrift für Dialektologie und Linguistik 72: 277⫺317. Herrgen, Joachim and Jürgen Erich Schmidt 1985 Systemkontrast und Hörerurteil. Zwei Dialektalitätsbegriffe und die ihnen entsprechenden Meßverfahren. Zeitschrift für Dialektologie und Linguistik 52: 20⫺42.

683

684

VII. Exemplary studies Herrgen, Joachim and Jürgen Erich Schmidt 1986 Zentralisierung. Eine phonetisch-phonologische Untersuchung zu Konstanz und Wandel vokalischer Systeme. In: Günter Bellmann (ed.), Beiträge zur Dialektologie am Mittelrhein; 56⫺100. (Mainzer Studien zur Sprach- und Volksforschung 10.) Stuttgart: Steiner. Herrgen, Joachim and Jürgen Erich Schmidt 1989 Dialektalitätsareale und Dialektabbau. In: Wolfgang Putschke, Werner H. Veith and Peter Wiesinger (eds.), Dialektgeographie und Dialektologie. Günter Bellmann zum 60. Geburtstag von seinen Schülern und Freunden; 304⫺346. (Deutsche Dialektgeographie 90.). Marburg: Elwert. Kehrein, Roland, Alfred Lameli and Jost Nickel 2005 Möglichkeiten der computergestützten Regionalsprachenforschung am Beispiel des Digitalen Wenker-Atlas (DiWA). In: Georg Braungart, Peter Gendolla and Fotis Jannidis (eds.), Jahrbuch für Computerphilologie 7, 149⫺170. Paderborn: Mentis. Kurath, Hans 1939⫺1943 Linguistic Atlas of New England. 3 vols. (Linguistic Atlas of the United States and Canada.) Providence, RI: Brown University. Labov, William et al. 2006 Atlas of North American English: Phonetics, Phonology and Sound Change. Berlin: de Gruyter. Lameli, Alfred 2004 Dynamik im oberen Substandard. In: H. Scheuringer and S. Gaisbauer (eds.), Linzerschnitten. Beiträge zur 8. Bayerisch-österreichischen Dialektologentagung, zugleich 3. Arbeitstagung zu Sprache und Dialekt in Oberösterreich, in Linz, September 2001, 197⫺208. (Schriften zur Literatur und Sprache in Oberösterreich 8.) Linz: Adalbert-Stifter-Institut des Landes Oberösterreich. Lenz, Alexandra N. 2003 Struktur und Dynamik des Substandards. Eine Studie zum Westmitteldeutschen (Wittlich/ Eifel). (Zeitschrift für Dialektologie und Linguistik. Beiheft 125.) Stuttgart: Steiner. Lenz, Alexandra N. 2004 Zur Interpretation des Intendierten Ortsdialekts. In: Alexandra N. Lenz, Edgar Radtke and Simone Zwickl (eds.), Variation im Raum. Variation and Space. 113⫺131. (VarioLingua 20.) Frankfurt/Main: Lang. Lenz, Alexandra N. 2007 Zur Grammatikalisierung von geben im Deutschen und Le¨tzebuergeschen. Zeitschrift für germanistische Linguistik 35(1/2): 52⫺82. Mang, Alexander 2004 Sprachregion Nürnberg. (Sprachatlas von Mittelfranken 6.) Heidelberg: Winter Orlovic-Schwarzwald, Marija and Jürgen Erich Schmidt 1986 Gastarbeiterdeutsch und deutscher Dialekt. In: Günter Bellmann (ed.), Beiträge zur Dialektologie am Mittelrhein; 230⫺255. (Mainzer Studien zur Sprach- und Volksforschung 10.) Stuttgart: Steiner. Rabanus, Stefan 2004 Morphological change in German Dialects: Two cases of plural verbs in Alemannic. In: Britt-Louise Gunnarsson, Lena Bergström, Gerd Eklund, Staffan Fridell, Lise H. Hansen, Angela Karstadt, Bengt Nordberg, Eva Sundgren and Mats Thelander (eds.), Language Variation in Europe. Papers from the Second International Conference on Language Variation in Europe, ICLaVE 2. Uppsala University, Sweden, June 12⫺14, 2003, 339⫺ 352. Uppsala: Uppsala Universitet. Rabanus, Stefan 2005 Dialektwandel im 20. Jahrhundert. Verbalplural in Südwestdeutschland. In: Eckard Eggers, Jürgen Erich Schmidt, and Dieter Stellmacher (eds.), Moderne Dialekte ⫺ neue Dialektologie. Akten des 1. Kongresses der Internationalen Gesellschaft für Dialektologie

37. The Linguistic Atlas of the Middle Rhine des Deutschen (IGDD) am Forschungsinstitut für deutsche Sprache “Deutscher Sprachatlas” der Philipps-Universität Marburg vom 5.⫺8. März 2003, 267⫺290. (Zeitschrift für Dialektologie und Linguistik. Beiheft 130.) Stuttgart: Steiner. Rabanus, Stefan 2008 Morphologisches Minimum. Distinktionen und Synkretismen im Minimalsatz hochdeutscher Dialekte. (Zeitschrift für Dialektologie und Linguistik. Beihefte 134.) Stuttgart: Steiner. Schmidt, Jürgen Erich 1986 Die mittelfränkischen Tonakzente (Rheinische Akzentuierung). (Mainzer Studien zur Sprach- und Volksforschung 8.) Stuttgart: Steiner. Schmidt, Jürgen Erich 1988 Der moselfränkische Dialekt. In: Jahrbuch 1989 für den Kreis Bernkastel⫺Wittlich, 200⫺ 205. Monschau: Weiss. Schmidt, Jürgen Erich 1992 Tendenzen dialektalen Wandels im Westmitteldeutschen. In: Andreas Weiss (ed.), Dialekte im Wandel. Referate der 4. Tagung zur bayerisch⫺österreichischen Dialektologie, Salzburg, 5. bis 7. Okt. 1989, 67⫺80. (Göppinger Arbeiten zur Germanistik 538.) Göppingen: Kümmerle. Schmidt, Jürgen Erich 1993 Zweidimensionale Dialektologie und eindimensional⫺vertikale Analyse. Ein exemplarischer Vergleich. In: Wolfgang Viereck (ed.), Verhandlungen des Internationalen Dialektologenkongresses. Bamberg, 29. 7.⫺4. 8. 1990, vol. 2, Historische Dialektologie und Sprachwandel. Sprachatlanten und Wörterbücher, 454⫺467. Stuttgart: Steiner. Schmidt, Jürgen Erich 1998 Moderne Dialektologie und regionale Sprachgeschichte. Zeitschrift für deutsche Philologie 117: 163⫺179. Schmidt, Jürgen Erich 2002 Die sprachhistorische Genese der mittelfränkischen Tonakzente. In: Peter Auer, Peter Gilles and Helmut Spiekermann (eds.), Silbenschnitt und Tonakzente, 201⫺233 (Linguistische Arbeiten 463.) Tübingen: Niemeyer. Schmidt, Jürgen Erich 2005 Sprachdynamik. In: Eckhard Eggers, Jürgen Erich Schmidt and Dieter Stellmacher (eds.), Moderne Dialekte ⫺ Neue Dialektologie. Akten des 1. Kongresses der Internationalen Gesellschaft für Dialektologie des Deutschen (IGDD) am Forschungsinstitut für deutsche Sprache „Deutscher Sprachatlas” der Philipps-Universität Marburg vom 5.⫺8. März 2003, 15⫺44. (Zeitschrift für Dialektologie und Linguistik. Beiheft 130.) Stuttgart: Steiner. Schmidt, Jürgen Erich and Joachim Herrgen to appear Sprachdynamik. Eine Einführung in die moderne Regionalsprachenforschung. Berlin: Erich Schmidt. Schmitt, Ernst Herbert 1986 Syn- und diachronische Untersuchungen zum Vokalismus von Nackenheim. In: Günter Bellmann (ed.), Beiträge zur Dialektologie am Mittelrhein, 101⫺131. (Mainzer Studien zur Sprach- und Volksforschung 10.) Stuttgart: Steiner. Schmitt, Ernst Herbert 1992 Interdialektale Verstehbarkeit. Eine Untersuchung im Rhein- und Moselfränkischen. (Mainzer Studien zur Sprach- und Volksforschung 18.) Stuttgart: Steiner. Smazal, Lothar 1986 Zur Implikationsanalyse. Ihre Möglichkeiten bei der Beschreibung dialektaler Sprachvariation. In: Günter Bellmann (ed.): Beiträge zur Dialektologie am Mittelrhein, 192⫺229. (Mainzer Studien zur Sprach- und Volksforschung 10) Stuttgart: Steiner. Steiner, Christiane 1994 Sprachvariation in Mainz. Quantitative und qualitative Analysen. (Mainzer Studien zur Sprach- und Volksforschung 19.) Stuttgart: Steiner.

685

686

VII. Exemplary studies Thinnes, Norbert 1981 Untersuchungen zur Variation nasaler Vokale. Ein soziolinguistischer Beitrag zum Rheinfränkischen. (Mainzer Studien zur Sprach- und Volksforschung 5). Wiesbaden: Steiner. Thun, Harald and Adolfo Elizaincı´n (eds.) 2000⫺ Atlas lingüı´stico diato´pico y diastra´tico del Uruguay (ADDU). Kiel: Westensee. Wiesinger, Peter 1970 Phonetisch-phonologische Untersuchungen zur Vokalentwicklung in den deutschen Dialekten, vol. 1: Die Langvokale im Hochdeutschen; vol. 2: Die Diphthonge im Hochdeutschen. Berlin: de Gruyter. Wiesinger, Peter 1983 Die Einteilung der deutschen Dialekte. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke and Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 2, 807⫺900. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.2.) Berlin/New York: de Gruyter. Zwirner, Eberhard 1956 Lautdenkmal der deutschen Sprache. Zeitschrift für Phonetik und Allgemeine Sprachwissenschaft 9: 3⫺13.

Joachim Herrgen, Marburg (Germany)

38. Challenging the homogeneity assumption in language variation analysis: Findings rom a study o multilingual urban spaces 1. 2. 3. 4. 5. 6.

Introduction Multilingual urban spaces in Sweden Some results of the SUF project in Stockholm, Gothenburg and Malmö Discussion Conclusions References

1. Introduction The close tie between language and space ⫺ the idea that “in country/region/neighborhood/community X people speak language/variety X” ⫺ is a powerful conception in popular as well as professional discourse about language. Its influence is not confined to that of the ideal of one language⫺one nation, routinely refuted in most textbooks on sociolinguistics and multilingualism, but can also be detected in contemporary perceptions of and approaches to language variation and varieties. Linguistics has been, and to a large extent still is, permeated by the conception of languages and varieties as bounded in space and tied to local, homogeneous speech communities. More generally, like classical sociological research, sociolinguistics within the quantitative paradigm has

38. Challenging the homogeneity assumption had as its point of departure a homogeneity assumption: that groups of speakers who are sociologically similar tend to be linguistically similar (Romaine 1982: 11; Wolfram and Thomas 2002: 160). While this would at first glance appear to be a necessary step in hypothesis formulation, it tends to have the effect that not only (perceived) groups of speakers, but also (perceived) ways of speaking are commonly “homogenized” and essentialized ⫺ both by lay persons and linguists (for further discussions see, e.g., Le Page 1977, 1988; Pratt 1987; Leung, Harris and Rampton 1997; Bucholtz 2003). Many sociolinguists continue to work as if individual variation or intragroup variation is of secondary importance (cf. Rampton 1997: 330; Wolfram and Thomas 2002: 160⫺165; Wolfram 2007). The language and space tie and the homogeneity assumption are coupled with the tendency among both laymen and linguists to uncritically apply dichotomous categorizations of language users such as native/non-native speakers, first/second language users and speakers/non-speakers (of language X) ⫺ implying that any individual is either a native first language user or a non-native second language user, and that any individual either speaks or does not speak language X. The default assumption is that individuals are monolingual and that, if not, they are either balanced bilinguals (i. e., two monolinguals in one) or bilinguals with one language clearly dominant (i. e., monolinguals with clear first and second languages). This monolingual bias or norm for linguistics has been pointed out by many others before us (e.g., Cook 1992; Kachru 1994; Backus 1999; Block 2003). The tie between language and space disavows the realities of widespread and multifaceted multilingualism and the diversity of language practices involving multitudes of languages and varieties in “the same” or contiguous spaces, global interaction involved in long distance travel or various forms of mass media and computer-mediated communication, similarities in patterns of variation over non-contiguous areas, and the transnational nature of migrant languages and non-territorial minority languages, as well as the sort of individual variation in micro-groups that Dorian (1994) among others has studied. The dichotomies of nativeness fail to account for the linguistic realities of many speakers, excluding forms of language competences and practices involving, e.g., partial knowledge of languages (e.g., Dorian’s [1981] semi-speakers), code-switching, or the mere use of a limited number of indexical words or phrases from another language for various purposes usually studied with regard to the concept of identity (e.g., Childs and Mallinson 2006). All these, often implicit (and therefore potentially even more damaging), assumptions have attracted increasing criticism for at least three decades. The criticism has in particular been rooted in experiences from multilingual practices and contexts, something which to some extent might explain why it has not reached the linguistic society at large. But even in countries like Sweden, where a monolingual ideology has been dominant for a long time, things are changing. Today, the complex and diverse linguistic realities of, in particular, young people of the modern multilingual city and the variation in the ways languages are acquired and used forcefully challenge the language⫺space tie, the homogeneity assumption and the dichotomous notions connected with nativeness. This complexity has also been the challenge of the Gothenburg-Lund-Stockholm project, Spra˚k och spra˚kbruk bland ungdomar i flerspra˚kiga storstadsmiljöer (SUF) ‘Language and language use among adolescents in multilingual urban settings’. The aim of the project, which was funded by The Bank of Sweden Tercentenary Foundation, is to describe and analyze language as it is used in such settings.

687

688

VII. Exemplary studies This article will briefly summarize some results and implications from five of the sub-studies within the SUF project. It will describe where a journey starting with the preconceived notion of a new variety of Swedish called Rinkeby Swedish (hereafter RS) has taken members of this project. In this article, we will argue that while varieties such as RS (here also used to refer to its equivalents in Gothenburg and Malmö) certainly have a valid existence as social constructions (among lay people as well as linguists), they cannot be adequately described as varieties in the traditional sense of a set of linguistic features connected to a specific speech community (cf. Hudson 1996). Language and language use (as well as speakers) in contemporary multilingual settings in Sweden (and possibly elsewhere) exhibit a variation that we can only hope to be able to account for by avoiding essentializing categorizations of speakers and their ways of speaking, and by using an analysis integrating various linguistic, psycholinguistic and sociolinguistic approaches.

2. Multilingual urban spaces in Sweden After its long history as a country which sent migrants to other parts of the world, Sweden has become a country which receives immigrants (cf. Statistics Sweden, ; Boyd 2001). Since the mid 1960s relatively large numbers of labor market migrants, refugees and their families and adopted children have come to the country. Sweden having no history of major colonies, its foreign-born population is unusually diverse. Furthermore, the country of origin with the highest number of new arrivals changes every few years in line with changing conditions in different parts of the world and shifting immigration policies in Sweden. Currently, national groups from outside of Europe dominate, particularly persons born in Iraq, but during the 1990s for example, refugees from former Yugoslavia dominated the new arrivals. Fluctuations of this kind in immigration to Sweden have resulted in a very diverse population. In addition to historical minorities, ten percent of the total population and twenty to thirty percent of the population of the three largest cities have a “foreign background”, i. e., are either born abroad or are children of two parents born abroad. The proportion of young people of school age with a foreign background (using this definition) has recently been estimated to be fourteen percent in Sweden as a whole (SOU 2008: 153). New arrivals in Sweden tend to live in the three largest cities of Sweden, although many refugees are at least initially directed to live in other municipalities. In the cities, they typically settle in publicly owned apartments in neighborhoods on the outskirts which were built in the 1960s or early 1970s. The populations of these suburbs include not only new arrivals, but also earlier ones who have stayed in the area, as well as working or lower class native-born persons and their families. The neighborhoods are therefore quite diverse, typically lacking a single dominant national or ethnic origin group; in other words, there are few, if any, “little Helsinkis”, “Bosnian neighborhoods” or the like in Sweden. The number of heritage languages in neighborhood schools is often claimed to be over 50 and may sometimes be as many as 100. The majority of the young people studied in the SUF project grew up at least partly in such settings. Primary and lower secondary public education is typically organized by

38. Challenging the homogeneity assumption neighborhood, so the schools the young people have attended have also been diverse in this respect. When our project was carried out, however, the young people studied were attending upper secondary school, which in some cases implied that they had left their neighborhoods to attend school with peers from other parts of the city. All schools included not only a diversity of pupils of foreign background, but also pupils with a monolingual Swedish background. In all three cities, attending upper secondary school involved a broadening of the young people’s contacts with pupils from neighborhoods and backgrounds other than their own. There were certainly strong similarities between the spaces (neighborhoods and schools) the students had grown up in and the spaces (i. e., upper secondary schools) they now moved in, even though the range of movement for almost all students had increased significantly.

3. Some results rom the SUF project in Stockholm, Gothenburg and Malmö The overarching aim of the SUF project has been to describe, analyze and compare language and language use among young people from multilingual urban settings in Stockholm, Gothenburg and Malmö, with a focus on the majority language, Swedish. The sub-studies encompass analyses of phonetics and phonology, syntax, lexicogrammar and pragmatics as well as of perceptions of varieties, identity development and ethnographies of the negotiation of identity and power in the multilingual classroom (see Källström and Lindberg to appear). The project staff consisted of two post-doctoral and five senior researchers and several graduate students. The participants in the project have been 222 young people from eight classes at eight upper secondary schools. All classes but one followed the social science program, a broad academic program which provides basic qualification for certain programs or courses at university; this was chosen because it was offered at all the selected schools and had a good balance of male and female students. The aim was to include schools with differing proportions of pupils with foreign background; in the selected classes the proportion varies between 33 and 100 percent. Almost three-fifths (59 percent) of all participants were either themselves born abroad or both their parents were born abroad. About thirty percent have no foreign background, while the remaining eleven percent have one parent born abroad. The project team gathered data from the young people in a wide variety of settings, both in and outside of school. The participants were first interviewed by a researcher about their backgrounds and language use, then recordings (the bulk of them audio) were made of semi-directed and non-directed group discussions, individual presentations to the class, and in a number of informal circumstances. Many informal recordings were self-recordings, where the participants borrowed equipment and recorded themselves in various everyday situations. The individual graduate students also made more directed recordings, such as interviews, focus group discussions, film-retellings and picture series descriptions, in order to elicit speech of a specific type for their particular research questions. Samples of the young people’s writing, in the form of the essay component of the national examination in Swedish, have also been collected and analyzed.

689

690

VII. Exemplary studies Four of the sub-studies described below focus on different linguistic features that had been observed to vary in the speech of young people in Sweden today. The features differ with regard to their association with, on the one hand, alleged new varieties such as RS, and/or, on the other hand, learner language. As will be shown, they represent all four possible combinations of these associations. The fifth study investigates young people’s perceptions of contemporary language variation, reflecting the diversity of social meanings that are attributed to different ways of speaking.

3.1. Phonetic/phonological variation In addressing the question of potentially conventionalized phonetic features of the language of young people in multilingual urban spaces in Sweden, Petra Bode´n (formerly Hansson) carried out a preliminary series of listener tests (Hansson and Svensson 2004; Bode´n and Grosse 2006). The aim was to see if young people had similar ideas about which of a number of selected speech samples could be labeled as examples of RS. The speech samples were extracted from the total of about 300 hours of recordings in the SUF project. Each listener group only listened to samples from their own city. It turned out that young people in all three cities chose rather consistently between “RS” and “not RS” for a number of the samples, although there was also some disagreement between listeners. (As we will see in section 3.5, another sub-study indicates that different listeners appear to apply these labels to rather different entities.) Interestingly, different stimuli from the same speaker could be labeled differently, and the “RS” label was not applied solely to the speech of young people with a foreign background or of those who were active multilinguals, nor was the “not RS” label applied only to monolingual young people with no foreign background. An interesting question is what phonetic characteristics led most listeners to apply the RS label. In her main study, Bode´n carried out phonetic analyses of the speech samples identified by a majority of listeners as RS. Bode´n found both segmental and prosodic features characteristic of the identified samples. A segmental feature found to vary in all three cities was the use of an affricate or a fricative in loan words such as checka ‘check out’, chilla ‘chill out’ or names like Charles and Jesus. In standard Swedish phonology, these affricates are typically replaced by simple fricatives, but in the speech samples studied the young people sometimes used an affricate. Bode´n could find no “foreign accent” explanation for this replacement, as it occurred even among young people without affricates in their heritage language. Prosodic features that varied in the material for the listener test included the typical Swedish stress pattern in phrases. In Swedish, the basic pattern is that the last content word in a phrase receives the greatest prominence; occasionally, however, Bode´n found examples of phrases among young people from all three cities, where a non-content word receives such prominence. In Malmö, this more consistent tendency for stress on the last item in a phrase or sentence was coupled with a flat or slightly rising F0, where the pattern for standard Swedish is a lowering of F0 through the sentence. This variation in the basic F0 pattern was, however, not found in the other two cities. Bode´n has furthermore found many examples where young people use rising F0 within the stressed syllable of a word to signal prominence, a feature which is unmarked in languages of the world, but which is not used in most dialects of Swedish, restricted as it is by its

38. Challenging the homogeneity assumption

691

distinctive word accents. This final prosodic feature was found in recordings in all three cities. These prosodic features help to contribute to the impression of a distinctive “staccato” rhythm that the speech of some young people gives to listeners. Bode´n concludes that one segmental and several prosodic features of the language of young people in these multilingual urban spaces can be found in all three cities; others are specific to one city only. Thus, the speech recorded in multilingual urban spaces shows both characteristics shared with other such spaces in the same country and characteristics of the local environment, e.g., diphthongs in Malmö. Many listeners describe the prosody as “bumpy” or “staccato”. Interestingly, similar terms have been used about the prosody of Danish and German spoken in multilingual contexts in Copenhagen (Quist 2000) and Berlin (Kern 2007), respectively, as well as in Nuuk-Danish (Jacobsen 2000, 2001). These similarities should be studied with care, in order to avoid hasty conclusions about prosodic features common to a certain type of contact variety and about their possible sources, but the similarity in characterization is nonetheless interesting and worth investigating further.

3.2. Instances o grammaticalization: san such, a sant and such and helt totally Lena Ekberg has carried out research within the SUF project on particular uses of three lexical items that appear to be characteristic of the speech of some of the young people in Malmö. One is the pronoun sa˚n, a spoken form of the written form sa˚dan ‘such (a)’, which is beginning to acquire the functions of a determiner, specifically an indefinite article. (Note that the names below and in other sub-studies of the project are pseudonyms). (1) Gorda: nej de e sa˚n journalfilm # sa˚n färgfilm # ja # sa˚na fem kakor ‘No, it’s such news film, such color movie, yes, such five cakes’ (Ekberg 2007: 52) The second is the tag expression a˚ sa˚nt ‘and such’ (cf. English and stuff), an extremely frequent discourse particle in the Malmö material, which also seems to be broadening functionally to be not only a modifier, but also a terminal marker of reported speech or a general boundary marker. (2) Aurora: han ba kan du inte komma hit a˚ sa˚nt ‘He just [said] can’t you come here and stuff.’

(Ekberg 2007: 68)

The third is the adverb helt ‘totally’, which is beginning to be used as an emphasizer also with unbounded adjectives, such as benig ‘boney’ and ful ‘ugly’. (3) Jing:

han var helt svettig ‘He was totally sweaty.’

(Ekberg 2007: 71)

692

VII. Exemplary studies The material Ekberg bases her (primarily cognitive semantic) analysis on consists of recordings of two groups of girls who were close friends, from the project sample in Malmö (see Svensson 2009: 90⫺94). The groups attend different schools in the city and take two different programs of study. The four girls in one group (Cgr1) have a variety of foreign backgrounds, are actively bilingual and are (based on information from interviews and various other contacts with them) ambitious, good students; the three girls in the other (Egr2) do not have foreign background, are monolingual and are not as ambitious or interested in succeeding in school as the first group. The student body of both classes includes students with foreign background to about the same degree (60 to 65 percent). All four of the students in the first group were included in Bode´n’s listener test (see section 3.3.1); two were judged by listeners to be “speakers of RS”, the third not, and the fourth ended up in between. Stimuli from two of the girls in the second group were also included in the listener test: both samples were considered “not RS”. From a sociolinguistic viewpoint, what is interesting about Ekberg’s results is that there is no indication of differing usage of the three lexical variables between the girls in the two groups. The monolingual, less ambitious girls in the second group, who were judged not to be speakers of RS, use the forms studied to the same degree as the multilingual, ambitious girls (at least two of whom were judged to be speakers of RS) of the first group. The variation does not therefore seem to be something particular to girls with a foreign background or to a particular multiethnic style or variety, as some girls were judged to be speakers of RS and others not. Since all the speakers in this sub-study were girls attending schools in multilingual environments in Malmö, and the material the study was based on was very informal speech, further research is needed in order to find out to what extent this usage is found among other young speakers in other places, for example, among boys and also in more monolingual environments in general (as Ekberg believes might be the case), as well as in more formal speech. Ekberg suggests that at least the second variable, a˚ sa˚nt, is probably a current, local Malmö phenomenon. Comparisons with data from other studies in other parts of Sweden suggest that frequent use of this tag could be specific either to Malmö, to more recent times, or to both (Ekberg 2007). Nevertheless, it is important to note that we have here three variables whose use is partly local, perhaps more common in casual speech and among girls, and possibly linked to multilingual environments in Malmö but not specifically to multilingual speakers.

3.3. Variation in relexive/personal pronoun usage Sofia Tingsell’s work (2007) with a sub-sample of young people from the full project sample, as well as with a small sample of adults, has been on the use of reflexive versus personal pronouns. Tingsell has based her findings primarily on directed speech and writing tasks, which generate 35 to 40 examples per person. Variation in choice of pronouns has existed in the language for some time and is described to some extent in the literature of (native) Swedish linguistics (Teleman, Hellberg and Andersson 1999), but Tingsell sees signs that the variation she has recorded involves relatively new contexts, at least as compared with those described by Teleman and coauthors. Since it is a pattern which is difficult for many learners, a connection with learner language must be considered. In contrast to the XSV word order studied by Ganuza (2008, see section 3.4), we

38. Challenging the homogeneity assumption

693

believe the choice of pronoun to have a relatively low salience, not being heavily stigmatized. In Swedish, the basic pattern is that reflexive pronouns are used when possessive or object pro-forms are coreferential with the subject within the same domain, usually the same clause, as in (4): (4) Anna gick pa˚ bio med sin (refl) bror. ‘Anna went to the movies with her (i. e., Anna’s) brother.’ When the pro-form is not coreferential with the subject in the same domain, a personal pronoun is used, as in (5): (5) Anna gick pa˚ bio med hennes (pers poss) bror. ‘Anna went to the movies with her (i. e., someone else’s) brother.’ In Tingsell’s material, a certain amount of variation or deviation from the norm as described above occurred. In some of these cases (examples [6] and [7]), personal pronouns were used instead of reflexive pronouns; in other contexts (example [8]), it was the other way around. (6) Interviewer: … vem tror du att hon e sur pa˚? ‘… who do you think she’s mad at?’ B: pa˚ hennes (pers poss) mor ‘her mother’

(Tingsell 2007: 102)

(7) a˚ sa˚ kollade hon i hennes ⫹ i hennes väska da˚ ‘and then she checked in her (pers poss) ⫹ in her (pers poss) bag’ (Tingsell 2007: 100) (8) Hon ger gubbeni sini pla˚nbok och ga˚r hem. ‘She gives the guy his (refl) wallet and goes home.’ (Tingsell 2007: 96; written data) Since the choice between reflexive and personal pronouns is a problem for learners of Swedish, Tingsell tested the hypothesis that the variation would be related to the age at which a young person began to acquire Swedish. It turned out, however, that the pattern of variation was more complicated than that. The number of deviations from the norm was relatively low overall, but variation occurred in the speech of a sizeable proportion of (at least by heritage) monolingual young people as well as over half of the multilingual young people (Tingsell 2007: 152). (Tingsell’s characterizations of participants as “monolingual” or “multilingual” are based on analyses of several variables in the database.) At the same time, 49 percent of the monolingual and 40 percent of the multilingual young people had no such variation at all. Only a small number of young people produced more than a few examples of deviant pronoun use. Although the proportion of deviations was significantly higher among multilingual young people there was no such clear effect for the variable “age of onset”, operationalized as the age the young person reported having begun to learn Swedish. In an earlier paper (Fraurud and Boyd 2006), we

694

VII. Exemplary studies discuss some of the problems with operationalizing “age of onset” as well as with the concept of nativeness as applied to young people such as those in this study. Despite the connections of this kind of variation with multilingualism and learner errors, this variable would seem not to be simply a feature of interlanguage. Rather, the variation proved more significantly linked to the young people’s multilingualism and social networks at the time of the study. There was for example a difference in the amount of variation between young people who reported actively using their languagesother-than-Swedish (at the time of the study) and those that did not; the former deviated in the use to a slightly higher degree than the latter. An important factor, according to Tingsell, seems to be the kind of micro-environment the individual young people can be found in as well as their choice of social network. Her findings indicate that multilingual young people in more “monolingual” schools (i. e., schools with fewer multilingual students) vary less than their counterparts in more “multilingual” schools. There were also interesting differences in the micro-environments of the monolingual young people. For them, variation seemed to be unaffected by the proportion of young people in their school with foreign background. However, the by heritage monolingual young people who have multilingual friends and who speak languages-other-than-Swedish with their friends vary more (i. e., follow the norm less closely) than monolingual young people who do not. This would seem to indicate that going to the same school as other speakers who vary their speech in this way isn’t sufficient; the young person also needs to socialize with multilingual young people in order to have this resource at hand. It also suggests a complex connection between this variable and multilingualism.

3.4. Variation in subject-verb word order Natalia Ganuza’s work within the project has concerned a highly stigmatized syntactic variable that is common in at least lower stages of Swedish as a second language, but which also is commonly associated with the Swedish of multilingual young people more generally: the XSV word order or “non-inversion” (Ganuza 2008). Swedish is a typical V2 language, which requires that if a sentence begins with a constituent other than the subject (e.g., an adverbial or a fronted object), the order of subject and first auxiliary or main verb is inverted, here exemplified by the participant Bushra: (7) SVX order: (8) XVS order:

han började springa du vet. ‘he started running you know.’

(Ganuza, pers. comm.)

sa˚ började han kuta. ‘then he started running.’ (lit.: ‘then started he running’) (Ganuza, pers. comm.)

What are found in typical “learner Swedish” and variably in the Swedish of young people in multilingual environments are constructions of the type XSV, as in this example from the same participant: (9) XSV order:

a˚ sen dom kutar ut ‘and then they run out’

(Ganuza, pers. comm.)

38. Challenging the homogeneity assumption Ganuza analyzed both spoken and written data from one larger sub-sample comprising 126 participants, whose data include a film re-telling, a written essay and a grammaticality judgment test, and a smaller sub-sample comprising 20 participants, who contributed several additional data types for analysis, including self-recordings, group discussions and presentations in class. Contrary to her expectations, clear cases of XSV constructions such as the one in (9) occur in only about four percent of possible environments (i. e., clauses with constituents other than the subject in initial position) in her large sample. In the smaller sample, which included more recordings from spontaneous speech, the proportion increased, but only to ten percent. The SV inversion rule (or V2 rule) constitutes a problem for learners of Swedish as a second language. It is a very frequent construction, which is acquired early in L1 Swedish (Ha˚kansson 1998). Ha˚kansson (2003), using Pienemann’s processability hierarchy, considers mastery of the V2 rule as an important milestone in acquisition of Swedish as a second language. That this variable originates in learner Swedish seems unquestionable, a conclusion supported by the fact that similar variable word orders are characteristic of the V2 languages Norwegian and Danish of young speakers in multilingual settings in Oslo and Køge (outside Copenhagen) (Aasheim 1997; Quist 2000). It is interesting to note that XSV word order, which is used extremely frequently in literary versions or stylizations of these varieties ⫺ cf. Källström’s (2003, 2005) studies within the SUF project ⫺ and in various attempts to re-create it in other forms of art and entertainment, occurs to such a limited extent in our corpus. Even the most consistent users of XSV word order in Ganuza’s small sample use it less often than the standard XVS order. However, when discussing RS, some of the young people in the Malmö sample exemplified the variety using a sentence with this distinctive word order (as well as a distinctive prosody). In other words, despite its low frequency in our corpus, the feature has high salience and is strongly associated with these perceived varieties. The rate of inversion or non-inversion turns out to covary with a number of constraints, both internal and external. The internal constraints include the length and type of the fronted constituent. For example, the adverbials sen ‘then/after that’ and da˚ ‘then/ at that time’, frequent connectors in narratives, tend to favor non-inversion more than other adverbials ⫺ at least for some speakers in some situations, but non-inversion also occurs with (the less frequent) fronted subclauses and prepositional phrases. In addition to the internal factors, there are a large number of external constraints that appear to be in operation. Although XSV word order, like pronoun choice, would seem to have its origin in interlanguage, there turned out to be no clear relationship between this variable and age at which the young people began to acquire or learn Swedish (“age of onset”). The more frequent users of XSV did not differ significantly from those who never used it in terms of the age at which they report having begun to learn Swedish; further, some young people with a monolingual Swedish background did use the feature. Looking at the results from a geographical perspective, the proportion of inversion did not differ significantly between the three cities. If one looks at the results on an individual level, however, it turns out that a few young people in the Stockholm sample have relatively many XSV sentences compared to the rest of the project sample as a whole. Still, the similarities in the pattern of variation among the three cities are greater than the differences, although there are also indications that link this variation particularly to Stockholm. It was difficult to isolate a category of young people who would be likely to be frequent users of XSV.

695

696

VII. Exemplary studies Even more interesting for our purposes here are a number of important factors that seem to have to do with the immediate context of the speech situation. Ganuza has shown that speakers who otherwise tend to follow the V2 rule consistently can use XSV if they are talking in a group with one or more speakers who more frequently use this word order or when the conversation is about certain topics, such as in group discussions about a literary text partly using features associated with RS. The number of non-inversions also tends to increase when the style is very involved or the speaker holds the floor for a longer period, e.g., with a narrative. Male speakers use non-inversions slightly more often than female speakers, but the difference is not significant; both sexes use non-inversion and one of the female speakers, Bushra, is among those with the highest number of non-inversions. Ganuza finds those speakers with a strong identification with their neighborhood and the language use of their communities tend to non-invert somewhat more often. These observations might lead a traditional variationist to consider XSV to be a feature of the vernacular of these young people, quite simply, when speakers pay minimal attention to how they speak (Labov 1972). But to Ganuza and to us, the pattern of variation in word order seems instead to be a resource actively used by young people to create a certain style of speech, which expresses solidarity, strong identification with one’s own community and one’s multilingual network. But even here, there is not a simple relationship between use of the variable and a particular meaning. The meaning attached to it also seems to vary with the particular situation in which the XSV variable is used (cf. Auer 2005). Like the non-standard pronouns studied by Tingsell, XSV word order seems to be used sparingly in very specific contexts by young people in all three cities (indeed, in similar contexts, even in other Scandinavian countries). Unlike Tingsell’s pronouns, this feature is, as noted, strongly stigmatized.

3.5. Perceptions o variation within the linguistic space o young people rom Stockholm As may already be evident from the brief reports of some results from four of the SUF sub-studies, much of the linguistic variation found in contemporary multilingual contexts cannot be accounted for simply in terms of new varieties or language acquisition. Still, reified entities such as RS play an important role in people’s thinking and in debates about language and education. This is the focus of another sub-study within the SUF project. In an on-going series of listener experiments, Ellen Bijvoet and Kari Fraurud investigate lay peoples’ perceptions and constructions of young Stockholmers’ ways of speaking (Bijvoet and Fraurud 2008, to appear). Their main research question is how people of different linguistic and social backgrounds conceptualize the linguistic space of (young) Stockholm, e.g., how language users with different sociolinguistic experiences divide this linguistic space, as reflected both in their labeling and description of different ways of speaking and in their attitudes towards speakers. An important point of departure for this series of studies was the observation ⫺ from interviews and informal discussions as well as from media discourse ⫺ that labels such as RS have very different extensions and connotations for different language users (including linguists). This observation is the reason for the exploratory nature of the experiments and motivated a choice

38. Challenging the homogeneity assumption of open questions about labeling and descriptions of the speech stimuli (rather than multiple choice or yes/no- questions). The results reported here come from a pilot study conducted within the SUF project, and involve some of the participants in the Stockholm part of the project as speakers or listeners. Seven speech stimuli of about 30 seconds were judged by several listener groups with different backgrounds, two of which are focused on here: (i) sixteen participants in the larger project, adolescents attending a suburban senior high school with a large proportion of bilingual students (BIL); and (ii) 24 monolingual first-year language students at Stockholm University of a median age of 28 (MON). Subjects listened to the speech samples and were asked to judge the speakers using semantic differential scales and to label and describe the speech samples and make guesses about the speakers’ background (place of residence in Stockholm, length of residence in Sweden and mother tongue). The open question about how the listeners label the various ways of speaking produced a wide diversity of labels and characterizations, the well-known label RS being only one of numerous suggestions. What interests us here is the way this label is applied very differently by different listeners, i. e., to how many and which of the speech samples it was applied. The speech samples that most frequently attracted the RS label come from two girls in lively discussions with their friends. But, rather than only considering these conceived “typical cases” of RS, it is instructive to compare listeners’ labeling of two of the speakers for whom listeners diverge most: Bobby and Ashur, who were very dissimilar from each other in background and lifestyle or attitude. Bobby came to Sweden at the age of 15. He is an ambitious learner of “proper” Swedish with a negative attitude towards RS, which for him represents “bad” language mainly because it includes slang words, so he tries to avoid using it. Ashur is born in Sweden and was pointed out by classmates as a proficient speaker of RS. But in this particular speech sample he gives a presentation in class demonstrating his high proficiency in Swedish along with certain non-standard features mainly at the phonetic level. Despite the differences between these two speech samples, both are labeled RS by some listeners. One listener considers Bobby’s speech to be RS “proper” (in contrast to other speakers’ label of RS “light”), while another remarks that his use of RS is “unconscious”. That Ashur’s way of speaking is not always seen as typical RS is acknowledged by modifications such as in “intellectual” RS, but it is nevertheless still RS for listeners who use such modifications. Not too surprisingly, the tendency for such broader constructions of RS is stronger among the listeners in the MON group, with less experience of multilingual contexts. Listeners in the BIL group tend to use the label RS more restrictively. First, they more often distinguish RS from “broken”, “learner”, “new arrival” or “immigrant” Swedish, here represented by Bobby ⫺ thus discriminating between (involuntary) learner language and the (voluntary) use of linguistic features associated with multilingual youth. Secondly, Ashur’s speech is more often characterized as ordinary or good Swedish, e.g., “Standard Swedish; ‘normal’ pronunciation etc.”, “Stockholmian; he tries to speak clear Swedish without using slang words” (in sharp contrast to one of the MON listeners’ characterization of Ashur’s speech as “immigrant Swedish; strong accent, large vocabulary but [it is] used ‘incorrectly’ according to the Swedish standard model”). Similar lines of differences between the two listener groups in dividing up examples drawn from the linguistic space of (young) Stockholm can be seen in the data from other

697

698

VII. Exemplary studies parts of the study as well. But there was also divergence within each listener group that might be attributed to individual preferences and attitudes towards language use, as well as different degrees of mobility in the city. The results suggest the existence of many diverse constructions of different ways of speaking associated with different combinations of geographical, social, age related and ethnic spaces ⫺ and sometimes with particular speech practices within these spaces. The label RS is for example alternatively tied to multilingual suburbs, immigrants, learners, youth, informality, intimacy and humor, or a combination of some of these elements. When it comes to the guesses about where in Stockholm the speakers live, however, listeners tended to show a higher degree of agreement, although somewhat modified by their different local experiences. The results of this pilot study thus offer an empirical illustration both of the notion that varieties/languages are social constructions that ⫺ albeit to different degrees ⫺ may take different shapes for different language users as well as linguists, and also illustrate our contention in the introduction to this article that lay people (and linguists?) are inclined to identify a speech sample with a place, such as in our case the well-known multilingual suburb of Rinkeby.

4. Discussion Sociolinguistics and dialectology strive to describe and analyze variation in language within and across various social settings or spaces. Traditionally, sociolinguistics has focused on urban settings and vertical variation and change, dialectology on rural settings and horizontal variation and change. Both strands of research tend to concentrate their studies on monolingual speakers of (varieties of) the majority language of the country or region under study, commonly excluding speakers considered non-native from the sample. (In his classic study of language variation in New York City, Labov [1966: 174⫺175, 187⫺188] excluded nearly half of his original random sample of Lower East Side residents on the grounds that they were not clearly native speakers of English). Multilingualism, both as a social and psychological phenomenon, has traditionally been treated separately ⫺ primarily in studies of language maintenance and shift (e.g., Fishman, Cooper and Ma 1971), language contact and change (e.g., Thomason 2001) and second language acquisition (SLA, e.g., Doughty and Long 2003). The homogeneity assumption and the related language⫺space tie and dichotomies of nativeness criticized in the introduction to this article are deeply rooted, even within fields of linguistics devoted to the study of heterogeneity. Despite the existence of both early and more recent critical work within these and neighboring fields, we believe that the influence of classical studies such as those of Labov (1966) and Fishman, Cooper and Ma (1971) is still strong in much contemporary research ⫺ including at least the early stages of the SUF project reported here. Speakers in multilingual urban spaces have often either been excluded from samples of informants as non-native or non-authentic speakers, or studied solely in ascribed roles such as language shifters, language learners, or speakers of alleged new varieties of majority languages. These roles tend to be contrasted with native speakers as norms or controls. (Conversely, in studies of multilingualism, individuals assumed to be monolingual native speakers tend to be either excluded or used as a control group.)

38. Challenging the homogeneity assumption The importance of handling such essentializing categorizations of speakers with care was emphasized in an earlier sub-study in the SUF project, based on analyses of the background interviews (Fraurud and Boyd 2006). This study clearly showed that a large majority of the 222 participants did not fit neatly into dichotomous categorizations such as native/non-native speakers or first/second language users. Most of these young people are in a broad sense multilingual in terms of both their background and their language proficiency and language use, but few of them would, according to prevailing definitions, qualify as native speakers of Swedish nor of their heritage language, nor would they be considered typical non-native speakers or learners. This observation is in itself an important challenge to paradigms presupposing comparison between learners, shifters or bilinguals and a native control group. In the SUF project, the results of our studies led us to begin to see the variables associated with nativeness as parameters that did not necessarily have the expected effects on the language variation studied. It turned out that for several of the studies within the project it did not make sense to conceive of a selected group of “native” participants as a “control group”. Neither was it possible to isolate a number of homogeneous groups within a “non-native” segment of our sample. The different sub-studies of the SUF project each contribute a piece of insight into the overall picture of language and language use among young people in contemporary Swedish urban spaces. In addition, an evaluation of both the results of these studies and the journeys that the individual researchers and students have made during their research can also be instructive for an assessment of what theoretical and methodological frameworks suggested by earlier studies of language and language use have had to offer in relation to these multilingual contexts. When the SUF project was initiated in 1999⫺2000, previous studies of the language and language use of young people similar to those in our project had either taken an psycholinguistic/SLA perspective, focusing on advanced or near-native second language use of Swedish (e.g., Stroud 1988; Hyltenstam 1992; Ekberg 1997), or a descriptive one, focusing on alleged new ”foreign sounding” varieties of Swedish not necessarily spoken only by young people with immigrant background (e.g., Kotsinas 1988, 2000). Both of these perspectives were also present at the outset of SUF, but most of the project members felt a need after a time to integrate several approaches, in some cases also including folk linguistic or ethnographic ones. Four of the sub-studies of contemporary language variation briefly described above focused on particular features at different linguistic levels, examining how these variable features are located in time and space. Even these brief glimpses may provide at least an impression of the complexity of the variation found. As mentioned, the features studied represent four different possible combinations of associations with alleged new varieties and/or learner language. Both the word-order (section 3.4) and the phonetic features studied (section 3.1) are ⫺ in contrast to the other features ⫺ highly stigmatized and indexical of “foreign-sounding Swedish”, often labeled RS. Variation in word order, as well as in pronoun choice (section 3.3), is also found in second language acquisition, which is not the case with the other features studied. The grammaticalized lexical items (section 3.2) cannot be connected with either perceived new varieties or learner language. As regards variation in space, some linguistic features turned out to be confined to “micro” or local spaces, others appear at a national level (although with possible differences in frequencies in different cities), and some might even be perceived as having

699

700

VII. Exemplary studies transnational character (although we do not yet know enough to be able to distinguish structurally motivated similarities from contact phenomena possibly conveyed by media). As regards variation related to parameters associated with nativeness, none of the features could be ascribed solely to language acquisition, and no features were found exclusively among young people with multilingual backgrounds, or exclusively in multilingual settings ⫺ although all these factors influenced the frequencies of use of some features. It is furthermore important to note that even linguistic features with a similar (if not identical) distribution in our sample may have widely differing meaning potentials and are likely to be more or less available as resources for speakers and listeners due to, among other things, different degrees of salience. For all the linguistic features studied there was also a considerable intra-individual variation, affected by a number of sociopragmatic factors in ways that we have only begun to understand. Importantly, the linguistic variation found among the young people in the SUF project cannot be reduced to manifestations of near-nativeness, nor could it adequately be accounted for in terms of homogeneous and delimited speech communities or varieties. The latter conclusion is further emphasized by the fifth sub-study described above, delineating the widely diverging constructions of speech communities and language varieties among language users. For example, while some listeners distinguish RS from learner language, others perceive all foreign-sounding Swedish as one and the same thing, RS. In fact, for a speaker to be judged to speak RS, just one manifestation of a single salient stigmatized feature (e.g., use of certain slang words, “staccato” prosody, non-inversion), or some other “foreign-sounding” feature, or even just careless speech may suffice. To the extent that, in working with our data, we need to speak in terms of language varieties (or styles or practices), we believe that these notions can best be approached as social constructions conceptualized as “pools” of linguistic resources (cf. Eckert 2000), which are employed to different extents by different speakers in different situations for different purposes and that are accessible at different levels of awareness. This said about the need to acknowledge heterogeneity among speakers and in their language, we want to add a word about the need also to recognize a pragmatic and political aspect of language discourses. While criticizing the essentialization of speech communities and language varieties, Bucholtz (2003) suggests that a total rejection of essentialism is not always desirable. Despite her criticism, Bucholtz suggests that strategic essentialism can function as an important intellectual and social tool in certain contexts and at certain points in time, in particular ”when the group under study is seen by the dominant groups as illegitimate or trivial, or when a stigmatized group forms an oppositional identity to counter such negative ideologies” (Bucholtz 2003: 400⫺401). The Swedish linguist Ulla-Britt Kotsinas’ “defense” ⫺ which she has advanced since the 1980s ⫺ of RS as a youth language rather than as some popular opinion would have it, just “bad” Swedish, is perhaps an illustration of this dilemma. Given the potential intellectual and social usefulness of strategic essentialism, many researchers studying new ways of speaking may feel the need to use various descriptive labels such as “multiethnolect” (Quist 2000), “multiethnic youth language” (Fraurud and Bijvoet 2004) or “suburban slang” (Bijvoet and Fraurud 2006). But, again, it is important to remember that, when using the tool of strategic essentialism, “researchers must remain mindful of the assumptions it brings along with it concerning ‘real’ language and ‘authentic’ speakers”

38. Challenging the homogeneity assumption (Bucholtz 2003: 403, cf. also Auer 2005; Jaspers 2007; Wolfram 2007: 16). We now believe we have good reason to be very restrictive and aware if and when using strategic essentialism in talking about the language of young people in these environments. Neither the groups of speakers nor the variation fulfill the criteria required by the traditional notions of speech community and language variety. To continue to talk about a category of young people (however defined) using a particular variety (however described) tends to reify their ways of speaking, hiding important complexity.

5. Conclusions We would like to conclude by making three specific practical points about the sampling and methods used in the SUF project, which have helped us approach the complexity of our object of study. First, the sample of participants is all-inclusive; in our case a cluster sample consisting of all (willing) students in the selected school classes. We did not select or exclude individuals according to (often dubious and always difficult) distinctions such as monolingual/multilingual and/or first/second language speakers of Swedish; neither did the notion of a control group make sense in our study. It should however be noted that our participants do not constitute a random sample of young people in the cities as a whole, nor in the selected schools or their neighborhoods. The choice of schools and of the program of study were made in order to increase comparability between cities, not to provide the possibility for broad generalizations about the language of various categories of young people in Sweden. Second, our database includes a broad range of settings for recording the young people and a number of different genres, both spoken and written. No particular style of speech is assumed beforehand to be more authentic or genuine than another (cf. Bucholtz 2003, Eckert 2003). Furthermore, we expected there to be interesting and important style shifting in the young people’s language in different situations, and we were also curious as to the relationship between the use of the perceived spoken varieties such as RS and contemporary literary versions or stylizations of it (see Källström 2003, 2005). Our assumption that important differences exist between the spoken languages of these urban settings and the literary versions or stylizations was borne out. Third, a multi-methodological approach was employed, motivated both by the complexity of our object of study and the exploratory character of much of our research. Some of us used rather traditional sociolinguistic and sociophonetic methods, others followed more closely methods used within second language studies, or analyzed transcriptions with the help of functional grammar or cognitive semantic approaches; one of the graduate students carried out an ethnographic study, etc. Both quantitative and qualitative methods have been used, sometimes by the same researcher. At least some of this variety in approach may reflect the fact that we had somewhat different ideas about what the object of study was when we started out. Importantly, we believe that all the project members have benefited from working in a context where several different approaches were represented. We hope to have shown that the complexity of the linguistic practices among young people in contemporary multilingual urban spaces invites a number of different research approaches, all possibly contributing to but not providing the whole picture. We are

701

702

VII. Exemplary studies reminded of the saying “as the question, so the answer”. If you approach contemporary urban spaces as a sociolinguist looking for variation along class lines, you will find social stratification (among the informants not excluded from your sample for being “nonnatives”). If you are interested in on-going language change as reflected in contemporary language variation (possibly more vivid in multilingual contexts), you are likely to find some tendencies toward that. If you approach it as a sociologist of language, you see “second generation immigrants” carrying out language shift. If you approach it as an SLA researcher, you see second language learners at different levels of proficiency. If you are looking for young people’s language or slang, you’ll find evidence for that as well. We believe that all of these aspects, and perhaps more, are notable in the language of these young people. To put a single label on these linguistic practices, however, essentializes both the language and the speakers, and simplifies their complex linguistic realities. A multitude of theoretical and methodological approaches will be necessary to do justice to the linguistic experiences of these young people we would like to argue.

6. Reerences Aasheim, Stine C. 1997 “Kebab-norsk” ⫺ fremmedspra˚klig pa˚virkning pa˚ ungdomsspra˚ket i Oslo [Kebab-Norwegian ⫺ foreign language influence on youth language in Oslo]. In: Ulla-Britt Kotsinas, Anna-Brita Stenström and Anna-Malin Karlsson (eds.), Ungdomsspra˚k i Norden [Youth Language in the Nordic Region], 235⫺242. (MINS 43.) Stockholm: Department of Scandinavian Languages, Stockholm University. Auer, Peter 2005 A postscript: Code-switching and social identity. Journal of Pragmatics 37: 403⫺410. Backus, Ad 1999 Mixed native languages: A challenge to the monolithic view of language. Topics in Language Disorders 4: 11⫺22. Bijvoet, Ellen and Kari Fraurud 2006 “Svenska med na˚got utländskt” [“Foreign sounding Swedish”]. Spra˚kva˚rd 2006/3: 4⫺10. Bijvoet, Ellen and Kari Fraurud 2008 Svenskan i dagens flerspra˚kiga storstadsmiljör: En explorativ pilotstudie av unga stockholmares perceptioner av variation och varieteter [Swedish in today’s multilingual urban settings: An explorative pilot study of young Stockholmers’ perceptions of variation and varieties]. Nordisk tidskrift for andrespra˚ksforskning, Nordand 3(2): 7⫺38. Bijvoet, Ellen and Kari Fraurud to appear Rinkeby Swedish in the mind of the beholder. Studying listener perceptions of language variation in multilingual Stockholm. In: Pia Quist and Bente A. Svendsen (eds.), Multilingual Urban Scandinavia: New Linguistic Practices. Clevedon: Multilingual Matters. Block, David 2003 The Social Turn in Second Language Acquisition. Edinburgh: Edinburgh University Press. Bode´n, Petra and Julia Grosse 2006 Youth language in multilingual Göteborg. In: Proceedings from Fonetik 2006. Lund, June 7⫺9, 2006, 17⫺20. Lund: Center for Languages and Literature, Department of Linguistics and Phonetics, Lund University. Boyd, Sally 2001 Immigrant languages in Sweden. In: Guus Extra and Durk Gorter (eds.), The Other Languages of Europe, 177⫺192. Clevedon: Multilingual Matters.

38. Challenging the homogeneity assumption Bucholtz, Mary 2003 Sociolinguistic nostalgia and the authentication of identity. Journal of Sociolinguistics 7(3): 398⫺416. Childs, Becky and Christine Mallinson 2006 The significance of lexical items in the construction of ethnolinguistic identity: A case study of adolescent spoken and online language. American Speech 81(1): 3⫺30. Cook, Vivian 1992 Evidence for multicompetence. Language Learning 42(4): 557⫺591. Dorian, Nancy 1981 Language Death: The Life Cycle of a Scottish Gaelic Dialect. Philadelphia: University of Pennsylvania Press. Dorian, Nancy 1994 Varieties of variation in a very small place: Social homogeneity, prestige norms and linguistic variation. Language 70: 631⫺696. Doughty, Catherine J. and Michael H. Long (eds.) 2003 The Handbook of Second Language Acquisition. Oxford: Blackwell. Eckert, Penelope 2000 Linguistic Variation as Social Practice. The Linguistic Construction of Identity in Belten High. Oxford: Blackwell. Eckert, Penelope 2003 Elephants in the room. Journal of Sociolinguistics 7(3): 392⫺397 Ekberg, Lena 1997 Diskursiva och syntaktiska mönster i a˚terberättelser hos invandrarbarn i Rosenga˚rd [Patterns of discourse and syntax in retellings by immigrant children]. In: Gisela Ha˚kansson et al. (eds.), Svenskans beskrivning 22. Förhandlingar vid Tjugoandra sammankomsten för svenskans beskrivning [The Description of Swedish 22. Proceedings of the Twenty-second Meeting for the Description of Swedish], 97⫺110. Lund: Lund University Press. Ekberg, Lena 2007 “… sa˚n svensk a˚ blond a˚ sa˚nt du vet”. Lexiko-grammatiska drag i Malmöungdomars talspra˚k [“… such a Swede and blond and such you know”: Lexico-grammatic features of the speech of Malmö young people]. In: Lena Ekberg (ed.), Spra˚ket hos ungdomar i en flerspra˚kig miljö i Malmö. Rapporter fra˚n projektet Spra˚k och spra˚kbruk hos ungdomar i flerspra˚kiga storstadsmiljöer. [Language among Young People in a Multilingual Environment in Malmö. Reports from the Project Language and Language Use among Young People in Multilingual Urban Environments]. Lund: Centre for Languages and Literature, Lund University. Fishman, Joshua A., Robert Leon Cooper and Roxana Ma 1971 Bilingualism in the Barrio. Bloomington: Indiana University Publications. Fraurud, Kari and Ellen Bijvoet 2004 Multietniska ungdomsspra˚k och andra varieteter av svenska i flerspra˚kiga miljöer [Multiethnic youth language and other varieties of Swedish in multilingual settings]. In: Kenneth Hyltenstam and Inger Lindberg (eds). Svenska som andraspra˚k [Swedish as a Second Language], 377⫺405. Lund: Studentlitteratur. Fraurud, Kari and Sally Boyd 2006 The native⫺non-native speaker distinction and the diversity of linguistic profiles of young people in Swedish multilingual urban contexts. In: Frans Hinskens (ed.), Language Variation ⫺ European Perspectives. Selected papers from the Third International Conference on Language Variation in Europe (ICLaVE 3), Amsterdam, June 2005, 53⫺69. Amsterdam: John Benjamins. Ganuza, Natalia 2008 Syntactic variation in the Swedish of adolescents in multilingual urban settings. Subjectverb order in declaratives, questions and subordinate clauses. Ph.D. dissertation, Centre

703

704

VII. Exemplary studies for Research on Bilingualism, Stockholm University. Available from . Hansson, Petra and Gudrun Svensson 2004 Listening for “Rosenga˚rd Swedish”. In: Peter Branderud and Hartmut Traunmüller (eds.), Proceedings FONETIK 2004: The Swedish Phonetics Conference, May 26⫺28 2004, 24⫺27. Stockholm: Department of Linguistics, Stockholm University. Ha˚kansson, Gisela 1998 Spra˚kinlärning hos barn [Language Learning among Children]. Lund: Studentlitteratur. Ha˚kansson, Gisela 2003 Tva˚spra˚kighet hos barn i Sverige [Bilingualism among Children in Sweden]. Lund: Studentlitteratur. Hudson, Richard. A. 1996 Sociolinguistics. Cambridge: Cambridge University Press. Hyltenstam, Kenneth 1992 Non-native features of near-native speakers. In: Richard J. Harris (ed.), Cognitive Processing in Bilinguals. Amsterdam: Elsevier. Jacobsen, Birgitte 2000 Sprog i kontakt. Er der opsta˚et en ny dansk dialekt i Grønland? En pilotundersøgelse [Languages in contact. Has a new dialect of Danish emerged in Greenland? A pilot study]. In: Grønlandsk kultur- og samfundsforskning 1998/1999 [Greenlandic Cultural and Social Research 1998/1999], 37⫺50. Nuuk: Ilisimatusarfik/Forlaget Atuagkat. Jacobsen, Birgitte 2001 The emergence of a new variety in Greenlandic urban societies due to the language contact situation. Report from an ongoing project. Unpublished paper presented at the 8th Nordic Conference on Bilingualism, Stockholm, November. Jaspers, Jürgen 2007 In the name of science? On identifying an ethnolect in an Antwerp secondary school. (Working Papers in Urban Language and Literacies 42: Kings College London.) Available from . Kachru, Yamuna 1994 Monolingual bias in SLA research. TESOL Quarterly 28(4): 795⫺805. Källström, Roger 2003 Ett spra˚kligt genombrott? [A linguistic breakthrough?]. In: Hans Landqvist, Sven-Göran Malmgren and Kerstin Nore´n (eds.), Texten framför allt. Festskrift till Aina Lundqvist pa˚ 65-a˚rsdagen den 11 september 2003 [Text above All. Festschrift for Aina Lundqvist on her 65th Birthday, the 11th of September 2003], 94⫺102. Gothenburg: Department of Swedish Language, University of Gothenburg. Källström, Roger 2005 Litterärt spra˚k pa˚ tvärs. Lite om spra˚ket hos Leiva Wenger och Hassen Khemiri [Literary language at odds. On the language of Leiva Wenger and Hassen Kemiri]. In: Boel De Geer and Anna Malmbjer (eds.), Spra˚k pa˚ tvärs. Rapport fra˚n ASLA:s höstsymposium, Södertörn, 11⫺12 november 2004 [Language at Odds. Report from ASLA’s Fall Symposium, Södertörn, 11⫺12 November 2004], 147⫺158. (ASLA:s skriftserie 18.) Uppsala: ASLA. Lindberg, Inger and Roger Källström (eds.) to appear Language and language use among adolescents in multilingual urban settings. Kern, Friederike 2007 Rhythm in Turkish German talk-in-interaction ⫺ Forms and functions of a contextualization device. Paper presented at the 10th International Pragmatics Conference, Gothenburg, Sweden, 8⫺13 July. Kotsinas, Ulla-Britt 1988 Rinkebysvenska ⫺ en dialekt? [Rinkeby Swedish ⫺ a dialect?]. In: Per Linell, Viveka Adelswärd, Torbjörn Nilsson and Per A. Petersson (eds.), Svenskans beskrivning 16 [The Description of Swedish 16], 264⫺278. (SIC 21a.) Linköping: Linköping University.

38. Challenging the homogeneity assumption Kotsinas, Ulla-Britt 2000 Pidginization, creolization and creolids in Stockholm, Sweden. In: Norval Smith and Tonjes Veenstra (eds.), Creolization and Contact, 125⫺155. Amsterdam: John Benjamins. Labov, William 1966 The Social Stratification of English in New York City. Washington: Center for Applied Linguistics. Labov, William 1972 The isolation of contextual styles. In: William Labov, Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Le Page, Robert 1977 Processes of pidginization and creolization. In: Albert Valdman (ed.), Pidgin and Creole Linguistics, 222⫺255. Bloomington: University of Indiana Press. Le Page, Robert 1988 Some premises concerning the standardization of languages, with special references to Caribbean Creole English. International Journal of the Sociology of Language 71: 25⫺36. Leung, Constant, Roxy Harris and Ben Rampton 1997 The idealised native speaker, reified ethnicities, and classroom realities. TESOL Quarterly 31(3): 543⫺576. Pratt, Marie Louise 1987 Linguistic utopias. In: Nigel Fabb, Derek Attridge, Alan Durant and Colin McCabe (eds.), The Linguistics of Writing, 48⫺66. Manchester: Manchester University Press. Quist, Pia 2000 Ny københavns “multietnolekt”. Om sprogbrug blandt unge i sprogligt og kulturelt heterogene miljøer [A new Copenhagen “multiethnolect”. On language use among young people in linguistically and culturally heterogeneous environments]. Danske Talesprog 1: 143⫺211. Rampton, Ben 1997 Second language research in late modernity: A response to Firth and Wagner. The Modern Language Journal 81(3): 329⫺333. Romaine, Suzanne 1982 Socio-Historical Linguistics: Its Status and Methodology. Cambridge: Cambridge University Press. SOU 2008 Värna spra˚ken. Förslag till spra˚klag. Betänkande fra˚n Spra˚klagsutredningen. Statens offentliga utredningar (SOU) 2008: 26 [Protect Languages. Proposal for a Language Law. Remit of the Inquiry into a Language Law. Swedish Government Official Reports 2008: 26]. Stockholm: Regeringskansliet [The Government Offices of Sweden]. Svensson, Gudrun 2009 Diskurspartiklar hos ungdomar i ma˚ngspra˚kiga miljör i Malmö [Discourse particles in youth talk in multilingual settings in Malmö]. Ph.D. dissertation, Lundastudier i nordisk spra˚kvetenskap [Centre for Languages and Literature], Lund University. Stroud, Christopher 1988 Literacy in a second language: A study of text construction in near-native speakers of Swedish. In: Anne Holmen, Elisabeth Hansen, Jørgen Gimbel and Jens Norman Jørgensen (eds.), Bilingualism and the Individual, 235⫺251. (Copenhagen Studies in Bilingualism 4.) Clevedon. Multilingual Matters. Teleman, Ulf, Staffan Hellberg and Erik Andersson 1999 Svenska akademiens grammatik [The Grammar of the Swedish Academy]. Stockholm: Norstedts Ordbok. Thomason, Sarah G. 2001 Language Contact: An Introduction. Edinburgh: Edinburgh University Press.

705

706

VII. Exemplary studies Tingsell, Sofia 2007 Reflexivt och personligt pronomen. Anaforisk syftning hos ungdomar i flerspra˚kiga storstadsmiljöer [Reflexive and non-reflexive pronouns. The use of anaphors among adolescents in multilingual urban settings]. (Göteborgsstudier i nordisk spra˚kvetenskap 8.) Ph.D. dissertation, Department of Swedish Language, University of Gothenburg. Available from . Wolfram, Walt 2007 Sociolinguistic myths in the study of African American English. Language and Linguistics Compass 1(4): 292⫺313. Wolfram, Walt and Erik Thomas 2002 The Development of Afro-American English. Malden, MA: Blackwell.

Sally Boyd, Gothenburg (Sweden) Kari Fraurud, Stockholm (Sweden)

39. Variety complexes in contact: A study on Uruguayan and Brazilian Fronterizo 1. 2. 3. 4. 5. 6. 7. 8.

Introduction: Language contact in pluridimensional dialectology A short history of studies on Uruguayan and Brazilian Fronterizo A schematic view of the variety complexes entering into contact Some characteristics of our pluridimensional cartography Conservatism in the Fronterizo varieties Innovations Conclusion References

1. Introduction: Language contact in pluridimensional dialectology Pluridimensional dialectology (explained in more detail in Lameli in this handbook) means not only the methodological fusion of traditional dialectology and sociolinguistic principles. It additionally involves aspects and techniques of language contact analysis. This expansion seems particularly necessary in the New World, where the linguistic landscape has, since the arrival of the Europeans, been shaped by manifold contacts between cultures and languages. Unlike classic studies on linguistic contact (like Weinreich 1970), which tend to reduce the contact configuration to the mutual influence of two languages considered as homogeneous systems, we propose analyzing the contact configuration as an approximation of two or more variety complexes, each of them an architecture of more than one more-or-less homogenous system (see Cos¸eriu 1967). The following survey is intended to demonstrate the utility of this complex approach in the case of Portuguese varieties spoken on both sides of the border separating Uruguay and Brazil.

39. Variety complexes in contact

2. A short history o studies on Uruguayan and Brazilian Fronterizo Northern Uruguay, which makes up nearly a third of the national territory, is a bilingual zone in which Portuguese varieties are spoken together with Spanish. The bilingualism is nowadays collective and unidirectional. This means that nearly all speakers whose mother tongue is Portuguese are also relatively competent in Spanish, but there are many speakers of Spanish who do not speak Portuguese (although the proximity of the two languages means a basic, passive comprehension is possible). During the investigations for our linguistic atlas of Uruguay between 1989 and 1992: we found only one Portuguese-speaking informant who declared that “unfortunately” he did not speak the language of his country, i. e., Spanish. In Uruguay, Portuguese is principally an oral variety, traditionally banned from schools and spoken mostly by the lower classes. On the other side of the frontier, the situation is different. No compact Spanish-speaking community exists in southern Rio Grande do Sul and hence a corresponding collective bilingualism is not found, only speakers of Portuguese varieties ⫺ Portuguese being, of course, the language taught in Brazilian schools. Lusophone zones beyond the borders of Brazil exist in many neighboring countries. The Portuguese-speaking region in northern Uruguay is probably the oldest and the most stable of these. Furthermore, it seems that Portuguese and then Brazilian settlement preceded Spanish and Uruguayan population of the zone. It is northern Uruguay, too, which has been studied more intensely in terms of linguistic aspects than the other frontier zones. The first systematic and the most influential study is that of the Uruguayan linguist Jose´ Pedro Rona, who published a small monograph entitled El dialecto >fronterizo< del Norte del Uruguay in 1965 (based on a paper presented at a conference in Salvador da Bahı´a in 1959). The results from this study have entered the manuals of Hispanic and Romance Studies, together with its empirical limitations and interpretative errors. In Rona’s eyes, this frontier dialect (“dialecto fronterizo”) is essentially the product of a mixture of the Castilian spoken in Uruguay and the Portuguese spoken in the southern Brazilian state of Rio Grande do Sul. The characterization of this mixture as “chaotic” and “instable” does not prevent Rona from subdividing it into two principal dialects, one with a Portuguese base, the other with a Spanish one. Each of them is made up of four subdialects. The phonological systems that Rona attributes to each variety are the manifestations of an internal contradiction. Furthermore, they seem to be structuralist constructions rather than the results of solid data. Rona’s method of investigation was correspondence; his material has been lost. For almost thirty years his work restricted linguists’ attention to the perspective of mixture. This orientation did not change even when sociolinguistics, contact linguistics and applied linguistics (concerned with alphabetization problems in northern Uruguay) entered the scene. Fritz Hensey (1972), who reexamined Rona’s data from a generativist transformational perspective, focused like his precursor on the aspect of mixture. This is also true of the numerous studies by Adolfo Elizaincı´n, which show a strong interest in the consequences of mixture in Uruguayan schools. In his monograph Dialectos en contacto, Elizaincı´n (1992, see also Elizaincı´n, Behares and Barrios 1987) develops an interesting method in order to predict the degree of mixture in the localities he studied empirically. But this method fails to avoid the following interpretative pitfall: experience shows that the responses of northern Uruguayan informants depend on the stimulus. Because Span-

707

708

VII. Exemplary studies ish is the language used with strangers, the first reaction is usually an answer in Spanish, mixed of course with Portuguese elements. To take these mixed Spanish/Portuguese utterances for the Portuguese variety used in internal communication would be an error. It is necessary to give stimuli in Portuguese with great insistence in order to position the informant in a context more likely to evoke his or her mother tongue. The investigations conducted within the Uruguayan-German project, the Atlas lingüı´stico Diato´pico y Diastra´tico del Uruguay (ADDU and ADDU-Norte), directed by Thun and Elizaincı´n (2000a, 2000b), offered the possibility of a more comprehensive view of the linguistic situation. This involved the abandonment of the exclusively mixture-focused perspective and openness for the independent characteristics of the northern Uruguayan varieties. As for the Brazilian side of the frontier, there is only one monograph dedicated to the linguistic situation (Blaser 1995). There is no generally accepted name for the Uruguayan and Brazilian frontier varieties. In order to avoid terms used only locally (like Carimba˜o, of unclear origin) or negative names (like Portun˜ol, a blend of Portugueˆs and Espan˜ol ), we will use the neutral denomination Uruguayan and Brazilian Fronterizo (from frontera ‘frontier’), which is not unknown in the zone. The speakers themselves name their variety Portugueˆs or Brasileiro, adding specifications like fronteiric¸o or da fronteira (‘of the frontier’; see Blaser 2003).

3. A schematic view o the variety complexes entering into contact On the Spanish side, we have to distinguish at least three varieties: Uruguayan Spanish (also known as Castellano), Uruguayan Popular Substandard and Uruguayan Rural Substandard. Uruguayan Spanish is somewhat vaguely defined because it lacks a commonly accepted codification. The Peninsular Spanish Standard is still present in school teaching, although more in precepts than in real speech. As there is no direct connection between the north Uruguayan varieties and the Peninsular Standard, we do not include a separate category for this Standard and instead identify potential correspondences under the Uruguayan Standard, which is thought to be the transmitter in modern times. This does not exclude the possibility that users, dependent upon their level of formal education, occasionally refer to the Peninsular Castilian norm or to the Brazilian or even Portuguese lı´ngua padra˜o. The Brazilian Standard enjoys the benefit of being more Tab. 39.1: A schematic overview of the variety complexes entering into contact Spanish

Fronterizos

Uruguayan Spanish

Uruguayan Popular Substandard Uruguayan Rural Substandard

Portuguese Brazilian Standard (“lı´ngua padra˜o”)

Uruguayan Fronterizo (1. Proximate zone) (2. Remote zone)

Brazilian Fronterizo Brazilian Popular Substandard Riograndense Substandard

39. Variety complexes in contact rigorously codified than Uruguayan Spanish. It is, of course, present in the frontier zone, through its domains of the school and the media. As with Spanish, we distinguish three varieties on the Portuguese side as well. Table 39.1 offers an overview.

4. Some characteristics o our pluridimensional cartography In addition to the traditional areal distribution of the points of inquiry (“diatopical dimension”), in our standard pluridimensional maps we add two more dimensions: the diagenerational and the diastratic. The first contrasts younger informants (18⫺36 years ⫽ GI [“generation I”]) with older ones (60 years old and over ⫽ GII [“generation II”]), the second contrasts informants with less formal education (from the illiterate to those who finished nine years of school ⫽ Cb [“clase baja” ‘lower sociocultural class’ ) with better educated informants (from secondary school to university ⫽ Ca [“clase alta” ‘higher sociocultural class’]). The two dimensions are combined and represented in a cross, with the upper section reserved for the higher sociocultural class, the lower for the lower sociocultural class and the left section for the elder, the right section for the younger generation: CaGII

CaGI

CbGII

CbGI

So, CaGII means higher sociocultural group, older generation; CbGI is lower sociocultural group, younger generation. As a rule, we add statistics to our maps. In the examples provided here, we have refrained from reproducing the informants’ commentaries and linguists’ observations which usually accompany the maps in our atlas.

5.

Preservation in the Fronterizo varieties

5.1. Where Uruguayan Fronterizo is more conservative than Brazilian Fronterizo The Portuguese Fronterizo of northern Uruguay is an uninterrupted continuation of the linguistic and cultural Brazilian “lusitanidade” in Uruguayan political territory. In terms of classic linguistic geography, we can say that this zone fits into the areal typology described by Bartoli (1945). In relation to the very distant centers of innovation of the great Brazilian cities on the Atlantic coast, northern Uruguay can be considered a “lateral area”, as is confirmed by the characteristic preservation of older forms that are obsolete in the more dynamic central varieties. This may be demonstrated using a phonetic example taken from the consonantal domain. Map 39.1 shows the treatment of final -l (“implosive l” at the close of a syllable or word) in words like algue´m ‘someone’, salvo ‘save, except’, lenc¸ol ‘sheet’ or anel ‘ring’. Twenty-one words with final -l appear in the “Parable of the Lost Son” text that our informants were asked to read in order to realize the “lecture style” within our diaphasic

709

710

VII. Exemplary studies

Map 39.1: Implosive -l, from the Atlas lingu¨istico Diato´pico y Diastra´tico del Uruguay (Norte)

program. The other two styles were free or directed conversation and the responses to the long list of questions contained in the questionnaire. The map shows a clear contrast between Brazil and Uruguay. In our Brazilian area, 26 of the 27 groups distributed among the eleven survey locations realize the final -l at the back of the mouth, only one group of informants (BC1) at the front. Six groups from the majority vocalize the final -l, eleven velarize it and nine show a semi-velarization. In Uruguay, on the contrary, only two of the thirty-nine groups manifest vocalization, two other groups exhibit velarization, fifteen semi-velarization and twenty, i. e., the majority, retain the anterior articulation (palatal or prepalatal). Vocalization is generally progressive in Brazilian Portuguese; velarization is still very common in Riograndense Regional Portuguese (Riograndense Substandard). Our map thus reflects a very characteristic pattern of graduation: Brazilian Fronterizo is more conservative than the rest of the Riograndense Substandard; Uruguayan Fronterizo is even more conservative than its Brazilian counterpart.

5.2. Bipartition o Uruguayan Fronterizo due to conservatism in the remote zone Remaining with the areal dimension, there is, furthermore, clear evidence that Uruguayan Fronterizo can be divided into two. The “proximate zone” has adopted the back articulation of -l, while the “remote zone” overwhelmingly retains the frontal pronunciation. The diatopic dimension of our pluridimensional program already enables us to

39. Variety complexes in contact classify the posterior articulation as the result of a process of innovation which has spread from a remote innovative center in Brazil, reached and crossed the political boundary, but then lost force and failed to make the distant reaches of the Uruguayan lusitanidade. This conclusion is confirmed by the diagenerational dimension. In both Brazilian and Uruguayan Fronterizo, posterior articulations are more frequent among the younger generation than among the older. Thus, “apparent time” (Labov 1994), also provides evidence of innovation. As our topic here is conservatism, we should point out that there is a massive block of retained older forms underlying the innovation zone and emerging into daylight in the remote zone of Uruguayan Fronterizo. In Peninsular Portuguese, velarization of the -l is the rule. It is presumed, moreover, that the Portuguese varieties exported to Brazil from the sixteenth century on were velarizing ones. The results from our “lateral area” indicate that this opinion must be revised (for a more detailed discussion see Thun 2000).

5.3. Conservatism in both Fronterizos 5.3.1. Phonetics Conservatism in the vocalic domain may be illustrated by the maintenance of the diphthong [ou] and even [oi] in words like tesoura ‘scissors’, documented in Map 39.2, which is based on the response data. Reduction of this diphthong to the closed and sometimes

Map 39.2: Phonetic map: tesoura ‘scissors’, from the Atlas lingu¨istico Diato´pico y Diastra´tico del Uruguay (Norte)

711

712

VII. Exemplary studies lengthened vowel [oø ] is common in Portugal and Brazil. As can be seen, the diphthong is well-preserved all over our territory, although not with identical vigor. The older generation of the socioculturally upper class retain the diphthong without exception and it is still the principal form among the corresponding younger generation (CaGI). In the socioculturally lower class, the monophthong clearly prevails among the older generation (CbGII), whereas the younger generation (CbGI) exhibit results similar to those from the equivalent upper-class generation. This seems to indicate, in the intergenerational succession, a restorative impulse favoring the more traditional diphthong pronunciation.

5.3.2. Grammar and pragmatics Both Brazilian and Uruguayan Fronterizo preserve, albeit with a clear tendency to give it up, consigo ‘with you’ in non-reflexive, formal, distant and polite use. This form is usual in Portugal, but rare and even criticized as incorrect in Brazil. Map 39.3, based on the construction O doutor quer que eu va´ consigo ao hospital? (roughly ‘Do you want, doctor, that I go with yourself to the hospital?’), shows that nine of our forty-nine groups still continue to use this construction, fifteen groups accept it, whereas a narrow majority of twenty-five groups reject it (preferring constructions like com o senhor, roughly ‘with the Sir’). The diatopical dimension does not indicate a clear areal distribution. This argues against the possible influence of Castilian, where this construction is unknown.

Map 39.3: Use of consigo, from the Atlas lingu¨istico Diato´pico y Diastra´tico del Uruguay (Norte)

39. Variety complexes in contact Formal allocutive consigo is clearly a regressive construction. The histogram gives evidence that not even among the older upper-class group (CaGII) do a clear majority use it. In the other three groups, consigo is a minority choice, even taking into consideration significant levels of use or acceptance. The diagenerational dimension nevertheless allows the conclusion to be drawn that formal allocutive consigo is in the process of falling into disuse.

5.4. Where Brazilian Fronterizo is more conservative than Uruguayan Fronterizo Map 39.4, which is complementary to Map 39.3, does not authorize the conclusion that consigo, while losing its formal, nonreflexive use, finds refuge in reflexive constructions. Like Portuguese and Castilian as popularly spoken, but progressively in other varieties too, and like other Romance languages (the evolution from avec soi to avec lui has been completed in French), Brazilian and Uruguayan Fronterizo are shifting to a preferential use of the nonreflexive form in prepositional third-person constructions (com ele, not consigo in O Joa˜o levou o livro com ele ‘John took the book with him’, not ‘with himself’). Nevertheless, and in spite of its massive decline, there are remnants of the use of reflexive, delocutive consigo, with exclusive use in just one group (Cl1), preferred use in another unique group (A1 CaGII), alternation with com ele in three groups (A5, R1J MV,

Map 39.4: Preposition before pronoun (consigo/com ele), from the Atlas lingu¨istico Diato´pico y Diastra´tico del Uruguay (Norte)

713

714

VII. Exemplary studies

Map 39.5: Erma˜o ‘brother’, from the Atlas lingu¨istico Diato´pico y Diastra´tico del Uruguay (Norte)

R1 CaGI) and tolerated (accepted) use in fourteen groups, among them all the Brazilian groups. This map shows that conservatism in the Fronterizo varieties must be considered, on the whole, as a graduated phenomenon. In this specific case, the many groups which reject consigo and exclusively use the com ele construction underline the more popular character of Uruguayan Fronterizo. The unanimously tolerant attitude of the Brazilian groups may be a result of the influence of high level varieties transmitted by the Brazilian school system. Note, nevertheless, that this hypothetical effort was not strong enough to alter the popular habit of using the com ele construction.

6. Innovations 6.1. Independent innovation in Uruguayan Fronterizo Exclusive innovation interpreted as an independent creation seems to be rare in the two Fronterizos. If the form erma˜o in place of Portuguese irma˜o (and Galegian irma´n) is not a continuation of a hitherto unidentified dialectalism of the Peninsula, it could well be an original Uruguayan Fronterizo crossing of Portuguese irma˜o with Spanish hermano, both meaning ‘brother’. The phenotypical Map 39.5, which records whether or not even a single token is recorded at a particular location, shows that this form has only been

39. Variety complexes in contact

Map 39.6: Pluridimensional phenotypical map: *lh+ pronounced [j], from the Atlas lingu¨istico Diato´pico y Diastra´tico del Uruguay (Norte)

found at Uruguayan locations and at no Brazilian ones. The areal distribution includes localities in both the remote and the proximate zone. Erma˜o is only used among the socioculturally lower class. The ten groups who produced this form are equally divided among the younger and the older generation of Cb. This fact indicates stability. The absence of erma˜o among the socioculturally higher class ⫺ which in Uruguay is to be found in urban locations near the border ⫺ and in Brazil may be explained by the greater permeability of this social stratum to the influence of standard language.

6.2. Innovation due to the inluence o the standard varieties 6.2.1. Innovation in Brazilian Fronterizo There are nonetheless cases where the high-level varieties are successful in bringing Brazilian Fronterizo back to a more conservative form. This is the case with the pronunciation of as a lateral palatal [ Y] in Brazilian Fronterizo and as palatal [ j] in Uruguayan Fronterizo. At the liminal level of a single recorded occurrence, the phenotypical Map 39.6 indicates that the lateral palatal pronunciation is much more frequent and diatopically more densely distributed in Brazil than in Uruguay. There are only five Uruguayan groups which do not delateralize the palatal; all of them are situated in the

715

716

VII. Exemplary studies

Map 39.7: Use of cujo ‘whose’, from the Atlas lingu¨istico Diato´pico y Diastra´tico del Uruguay (Norte)

zone closest to the Brazilian frontier and three represent the younger generation. On the Brazilian side, [Y] is the form for the majority of the groups. Those who use [ j] either live in rural locations (like B1 and B2) or belong to the older generation of the socioculturally lower class. In Brazilian urban locations (like BC2 Quarai, BC4 Santana do Livramento, BC8 Bage´ or BC10 Chui), the usual form is [Y]. Considering that delateralization is common in popular Brazilian Portuguese, and looking at the areal, generational and social distribution of the [Y] form, we conclude that this form is replacing the older popular variant [ j] under the influence of the Brazilian school system, which is less effective in rural zones and absent in Uruguay (where Portuguese has been excluded from formal instruction). We propose an interpretation of the [Y] pronunciation in Brazil as the result of a recursive process. The older [Y], which is still common in Portugal, has been replaced by [ j] and this delateralized form is now receding under pressure from the lı´ngua padra˜o variant, [Y]. If this interpretation is correct, [Y] is, in its actual use in Brazilian Fronterizo, not an example of conservatism, but one of innovation.

6.2.2. Innovation in Uruguayan Fronterizo As Dorotea Frank Kersch points out in her recent study (Frank Kersch 2006), the use of the relative pronoun cujo ‘whose’ is almost exclusively confined to erudite written prose in modern Brazilian Portuguese and it has probably never been a popular language

39. Variety complexes in contact form. Map 39.7 provides evidence of a clear contrast between Brazilian and Uruguayan Fronterizo here. Cujo is completely absent from the linguistic competence of our Brazilian speakers. In Uruguayan Fronterizo, two of the high-status groups of the older generation (in A1 and R1) and one of the younger generation (in R1) claim to use the form. Still more surprising are the following two results: the older lower-class group at the rural location CL2 attests to using this relative and not less than sixteen lower sociocultural status groups together with two higher level younger generation groups (A1 and R1 J MV) claim to know the form, even if they do not use it. Frank Kersch has supplemented our linguistic map based on the questionnaire responses with an analysis of the conversational style, including radio transmissions. It turns out that nobody ever uses cujo in spontaneous style, not even the high-class speakers. Nevertheless, even if there is no evidence for its real use in spontaneous speech (where the polyfunctional que is preferred), cujo does exist at the level of passively known forms. We assume that this is due to the influence of Uruguayan Standard Spanish, where the relative pronoun cuyo (pronounced [1kuzo] like the Portuguese cujo, if not [1kuso]) is used somewhat more often in the spoken variety (see the comparison in Frank Kersch 2006: 167⫺169) and relatively frequently in written text, to which Uruguayan Fronterizo speakers have become acquainted while at school, where teaching is conducted exclusively in Castilian.

6.3. Innovation rom the substandard varieties 6.3.1. In Brazilian Fronterizo Brazilian Fronterizo participates actively in the popular Brazilian reduction of the verb paradigm. Uruguayan Fronterizo is much more resistant to this radical change. As Map 39.8 shows, the present third-person plural subjunctive falem or even the singular fale (which is the most advanced stage in this change) have completely replaced the traditional form falemos ‘that we speak (sbjv)’ for the first-person plural in the Brazilian locations. This tendency is still countervailed in Uruguayan Fronterizo, where good rates for the use of the conservative falemos are to be found. The third-person forms are nevertheless progressive in Uruguay too, as is shown by the high values in the histogram for the third-person solution. This applies to all groups, with a considerable increase especially among the lower-class and the younger generation. We suppose that the absence of this form reduction in the Castilian varieties in Uruguay helps to preserve the traditional form of the first-person plural. From this perspective, Map 39.8 could also be considered under the heading of “Preservation of older Portuguese forms” (cf. section 5). Note that in other cases Uruguayan and Brazilian Fronterizo show similar resistance to the influence of popular substandard Brazilian. The Riograndense construction tu veio ‘you came’, which combines the second-person singular pronoun with the thirdperson singular verb form, in line with the common Brazilian construction voceˆ veio but maintaining the pronoun tu , is still rare on both sides of the border. Both Fronterizos prefer tu vieste or tu viestes which retain the second-person form of both pronoun and verb.

717

718

VII. Exemplary studies

Map 39.8: Reduction of a verbal paradigm: falemos ‘that we speak’, from theAtlas lingu¨istico Diato´pico y Diastra´tico del Uruguay (Norte)

6.3.2. In Uruguayan Fronterizo Map 39.9 illustrates a case of preponderant influence of Uruguayan rural Spanish substandard on Uruguayan Fronterizo with lesser effect on Brazilian Fronterizo. The form equivalent to the adverb so is assim in Standard Portuguese and ası´ in Standard Spanish. Our Brazilian informants recognize the form ansim [a˜’sı ˜ ], but only use the standard form assim [a’sı ˜ ]. Not so in Uruguayan Fronterizo: the exclusive use of ansim is particularly frequent in the western half of the remote zone, with decreasing use in the proximate zone. We conclude that ansim is strengthened by forms like ansı´n and ansina which are known from the Spanish of the Uruguayan rural interior.

6.3.3. In both Brazilian and Uruguayan Fronterizo Since the Brazilian Portuguese substandard and the Uruguayan Spanish substandard coincide in a considerable number of deviations from their respective standard forms, it is not very surprising that both Brazilian and Uruguayan Fronterizo also feature these popular constructions. Map 39.10 offers evidence for the deep roots that the concordance between the adverb and its corresponding adjective possesses across the entire region and among all of our groups. The construction A minha irma˜ e´ meia boba ‘My sister is rather stupid’ is clearly preferred everywhere. There is only one group (CaGII

39. Variety complexes in contact

Map 39.9: Presence of ansim ‘so’, from the Atlas lingu¨istico Diato´pico y Diastra´tico del Uruguay (Norte)

in A1) which prefers the standard form meio boba, and even they do not reject the deviant form meia boba. Similar cases are the popular verbal concordance in impersonal constructions like fazem anos instead of faz anos or tiveram festas alegres in place of teve festas alegres. In these cases, Standard Portuguese and Spanish, unlike English or German, demand the singular third-person form (lit. ‘there has been merry festivities’).

6.4. Innovation due to either standard or substandard varieties (Uruguayan Fronterizo) As Uruguayan Fronterizo is much more exposed than Brazilian Fronterizo to the influence of a different historical language, i. e., Castilian, there are numerous examples of Spanish influence at all structural levels of Uruguayan Fronterizo. It is because of these Castilian elements that Uruguayan Fronterizo is, as we have seen, commonly considered to be a mixed variety. Integrated into the Uruguayan Fronterizo system, these borrowed forms must be considered as innovations when compared with the stock of vocabulary from the Portuguese varieties. We will not, in order to illustrate this important phenomenon, choose lexical loan words like the names of the days of the week (lunes instead of segunda feira ‘Monday’), of the month (enero instead of janeiro ‘January’) or numbers (uno instead of um ‘one’), which are all in general use in the Portuguese of northern

719

720

VII. Exemplary studies

Map 39.10: Adverbial concordance, from the Atlas lingu¨istico Diato´pico y Diastra´tico del Uruguay (Norte)

Uruguay. Morphological or categorical influence is certainly a more profound and thus more convincing proof of the Castilian impact on Uruguayan Fronterizo. We propose the example of gender problems with mass or abstract nouns. One of the most interesting cases is that of ordem which is always feminine in Portuguese and means, like the English word order, either ‘command’ or ‘normal state of things’. Castilian, on the contrary, makes a difference between the two senses by distinguishing between the feminine la orden ‘command’ and the masculine el orden ‘normal state or methodical disposition’. We can infer from Map 39.11 that there is harmony between Uruguayan and Brazilian Fronterizo as to the preference for the masculine gender of ordem when it means ‘command’. Nevertheless, we note some indecision about the gender among eight Uruguayan groups, distributed across the territory, the generations and the sociocultural rankings. The next map, Map 39.12, shows that many Uruguayan groups ⫺ but similarly also the Brazilian CbGII group at the distant B1 location ⫺ have introduced the morphological distinction of using the feminine article for ordem when it means ‘a normal state of things’. This differentiation is certainly attributable to the Castilian model. The hesitation commented upon in the preceding map may be an indicator that the morphological refinement is not fully established. Our linguistic atlas of northern Uruguay contains nineteen cases of similar nouns (like sangue ‘blood’, ac¸u´car ‘sugar’ or leite ‘milk’). Summing up the results from these, we can say that there is less divergence from Portuguese in the gender these words have in Brazilian Fronterizo than in Uruguayan Fronterizo. In Uruguayan Fronterizo, the

39. Variety complexes in contact

Map 39.11: The gender of ordem ‘order’ (command), from the Atlas lingu¨istico Diato´pico y Diastra´tico del Uruguay (Norte)

gender problem usually occurs in contexts in which Portuguese and Castilian feature different solutions. But Uruguayan Fronterizo does not always follow the Castilian example. There are significant degrees of oscillation: in the case of mel ‘honey’ for instance, more than 26 percent of the groups use the feminine article, following the Castilian example of la miel. But 25 percent of the groups are fluctuating between the two genders and almost half of the groups retain the Portuguese masculine. There is no clear preference for these words whose gender is not motivated by natural sex to be classified as either feminine or masculine. All we can say is that the contact between the two source languages, which is particularly intense in Uruguayan Fronterizo, contributes to an exacerbation of the gender problems inherent in each of the neighboring great languages within the Fronterizo variety.

7. Conclusion As we have pointed out, Uruguayan and Brazilian Fronterizo are not simply the products of a mixture. Both are also characterized by the conservation of older forms abandoned elsewhere. In this sense, Uruguayan Fronterizo is more conservative than its Brazilian counterpart. There are only weak indices of independent innovation in the Fronterizos, while influence from neighboring varieties is frequent indeed. It turns out to be

721

722

VII. Exemplary studies

Map 39.12: The gender of ordem ‘order’ (normal state), from the Atlas lingu¨istico Diato´pico y Diastra´tico del Uruguay (Norte)

important to differentiate the donor varieties. The influence of the Brazilian Standard on Brazilian Fronterizo seems to be greater than that of Uruguayan Spanish on Uruguayan Fronterizo. Many of these Standard Portuguese elements stop at the political boundary which is also an educational border. It is typical of Uruguayan Fronterizo, but not of Brazilian Fronterizo, that the internal territory is subdivided into two zones. This bipartition is the result of various influences. The remote zone is sometimes more conservative than the rest of the Fronterizo domain. On the other hand, it is more susceptible to Spanish influence. The proximate zone is subject to greater influence from Brazil. An explanation for this may be found in the particularly intense degree of contact with Brazil. Many localities within Uruguayan Fronterizo zone are twinned with their Brazilian neighbors (cidades geˆmeas ‘twin towns’). In addition, Brazilian radio and television programs are available in the border region.

8. Reerences Bartoli, Matteo 1945 Saggi di linguistica spaziale. Turin: Rosenberg & Sellier. Blaser, Jutta 1995 Das Spanische in Südbrasilien: Die Zerstörung einer Legende durch mikrodialektologische Feldforschung. Frankfurt am Main: Lang.

39. Variety complexes in contact

723

Blaser, Jutta 2003 “Carimba˜o” oder “Corrupio”: Sprachmischung und Bewusstsein im Norden Uruguays. In: Dieter Messner and Matthias Perl (eds.), Portugiesisch in der Diaspora: Vorträge vom 4. Deutschen Lusitanistentag an der Universität Mainz (2001), 103⫺129. Germersheim: Centro de Estudios Latinoamericanos / Institut für Romanistik, Johannes GutenbergUniversität Mainz. Cos¸eriu, Eugenio 1967 [1966] Structure lexicale et enseignement du vocabulaire. In: Council of Europe, Les the´ories linguistiques et leurs applications, 9⫺51. Strasbourg: Council of Europe. [First published in: Actes du premier Colloque International de Linguistique applique´e (Nancy 1964), Annales de l’Est, Me´moire no 31, 175⫺252.] Elizaincı´n, Adolfo, Luis Behares and Graciela Barrios 1987 Nos falemo brasilero. Dialectos portugueses en Uruguay. Montevideo: Editorial Amesur. Elizaincı´n, Adolfo 1992 Dialectos en contacto. Montevideo: Editorial Amesur. Frank Kersch, Dorotea 2006 A Construc¸a˜o relativa na lı´ngua falada. Enfoque na fronteira bilı´ngüe do Brasil com o Uruguai, comparado ao espanhol e ao portugueˆs riopratense e europeu. Kiel: Westensee. Hensey, Frederik G. 1972 The Sociolinguistics of the Brazilian-Uruguayan border. Paris: Mouton. Labov, William 1994 Principles of Linguistic Change. Oxford: Blackwell. Rona, Jose´ Pedro 1965 El Dialecto del Norte del Uruguay. Montevideo: Universidad de la Repu´blica. Thun, Harald 2000 O portugueˆs americano fora do Brasil. In: Eberhard Gärtner, Christiane Hundt and Axel Schönberger (eds.), Estudos de geolingüistica do Portugueˆs Americano, 185⫺227. Frankfurt am Main: Teo Ferrer de Mesquita. Thun, Harald and Adolfo Elizaincı´n 2000a Atlas lingüistico Diato´pico y Diastra´tico del Uruguay (ADDU), vol. 1: Consonantismo y vocalismo del espan˜ol. Kiel: Westensee. Thun, Harald and Adolfo Elizaincı´n 2000b Atlas lingüistico Diato´pico y Diastra´tico del Uruguay-Norte (ADDU-Norte), vol. 1: Consonantismo y vocalismo del portugue´s. Kiel: Westensee. Weinreich, Uriel 1970 Languages in Contact: Findings and Problems. 7th ed. The Hague: Mouton.

Harald Thun, Kiel (Germany)

724

VII. Exemplary studies

40. Language as a process: A study on transnational spaces 1. 2. 3. 4. 5. 6. 7.

Operationalizing space Linguistic minority nationalism in francophone Canada The commodification and marketing of Canadian francite´ Francophone Canada goes transnational Webs and links, spaces and trajectories, resources and actors Acknowledgements References

1. Operationalizing space In the history of linguistics, language has usually been thought of as tied to some form or other of geographical space. I have argued elsewhere (Heller 2007) that this linkage is a product of the ways in which linguistics as a discipline emerged in the context of the development of the nation-state, and in particular in the ideological complex linking bounded territories to bounded and uniform languages and cultures (Hobsbawm 1990; Hutton 1999; Bauman and Briggs 2003). In many ways, it is possible to see the work of linguistics as having been devoted to a combination of creating order, boundedness and homogeneity out of linguistic diversity and flux and, simultaneously, of exploring exactly the diversity and flux which emerged as salient out of the nation-building project. Dialectology, for example, sought explicitly to identify and describe the historically and geographically most stable and most isolated forms; descriptive linguistics sought to identify and describe regularity and systematicity; and even historical linguistics looked for laws of linguistic change more attributable to the workings of the linguistic system than to the practices of speakers. So we have inherited a set of disciplinary approaches which foreground stability (or at least predictable, rule-governed change) and which construct movement and variability as problems to be explained when they cannot be swept under the rug. Lately, however, this perspective has begun to shift; movement and diversity have begun to be understood as the norm, as the phenomena to be foregrounded, and things like standardization or stability as phenomena requiring explanation. What this means, in practical terms, is that we have to re-think how we understand the relationship between language and space, imagine new ways of constructing what counts as data and figure out how to go about collecting that data. In this article, I will describe some ways in which the research team I have been working with has tried to take up that challenge in the context of the particular research questions we have been asking. This is not meant as a prescriptive or programmatic presentation, but rather as an illustration of some of the challenges before us and an exploration of some strategies, in the context of a particular theoretical approach to conceptualizing the problem. In what follows, I will set out the research question and its ties to the problem of language, space and nationalism, the conceptual underpinnings of our approach to it and some of the methods we have used to construct a set of data

40. A study on transnational spaces that would allow us to answer our question. The ways in which transnationalism has become salient for linguistic theory and methodology is embedded, I argue, in the challenges it presents for the nation-state and its accompanying ideologies of identity, community, culture and language. Ethnographic modes of tracking the circulation of actors and resources through discursive spaces, where resources are produced and distributed and given meaning, is one way of exploring how the competing values of locality and authenticity versus universality and transportability play out in late modernity.

2. Linguistic minority nationalism in rancophone Canada The major research question that I have been following for some years has been how to describe, understand and explain the changing shape of Canadian francophone nationalism (Heller 2002). This is, of course, classically a problem in which linguistic forms (in this case usually understood as varieties or a variety of French) are tied to territories with some kind of political boundary, whether municipal, regional or state-like. In this vein, much linguistic work has been devoted to describing the French found in what are understood in political terms to be francophone spaces (see for example Sankoff 1980; Pe´ronnet 1989; Thibault and Vincent 1990; Mougeon and Beniak 1991; Ostiguy 1993; Deshaies and Ouellon 1998), to detailing what has been understood as the effect of contact with English (e.g., Mougeon 1989; Mougeon and Beniak 1991), and to working in a language-policy mode towards increasing the homogeneity of language and to maximizing the language-and-territory fit (e.g., Corbeil 1980; Martel and Cajolet-Laganie`re 1996; Forest 1998; Georgeault and Page´ 2006). I understand this body of work to be ideologically informed (whether consciously or not) within a discourse of nationalism. My task is to understand the relationship between language ideologies, discourses of language and nation and language practices, whether my own or anyone else’s (and whether from the perspective of an academic or government agency researcher or one occupying any other social position). From such a perspective, it has become clear that so much has had to be done in these directions precisely because the fit is not neat and things are not staying still; that is, in the particular context of francophone Canada in the last forty years or so, the conditions for establishing a normalized, hegemonic discourse uniting language, nation, territory and individual identity have not been as propitious as its proponents might have liked. In addition, those conditions, in francophone Canada and elsewhere, have only gotten worse, not better, as the global and increasingly privatized expansion of capital forces nation-states to recast the bases of their legitimacy and to revamp their modes of regulation (Harvey 1989; Featherstone 1990; Castells 2000). The result, in the social sciences generally, has been to try to come to terms with the rapid movement and intensified linkages across time and space that characterize this period of late modernity (Harvey 1989; Giddens 1990). In anthropology and sociology especially, recent developments have been aimed both at conceptualizing challenges to the heretofore dominant sociopolitical order and to working out ways of understanding what new processes may be at work and how they function. They also aim at identifying their consequences both for the ways in which we can, may or must re-imagine the social order and its discursive legitimation, and for

725

726

VII. Exemplary studies what that means concretely for the life conditions of differently-positioned social groups. Much of this work can be understood under the rubric of studies of transnationalism, which can be understood in a number of ways, ranging from a strictly delimited concern with social processes which cross nation-state boundaries to a looser concern for multidirectional flows across social boundaries, some of which are related to ethnonational or nation-state-related identity categories, and some of which are not (Basch, Glick Schiller and Szanton Blanc 1994; Appadurai 1996; Beck and Sznaider 2006). Appadurai (Appadurai 1986, 1991), among others (cf. Burawoy et al. 2000), has suggested that what this means in practical terms for social scientists is a shift away from a focus on communities, identities, or linguistic and cultural forms understood as whole, bounded systems and towards following the flows of people, things and discourses. In Appadurai’s view, what we need to get a handle on are the trajectories of resources and of social actors across time and space. This has been operationalized in anthropology through the concept of multi-sited ethnography, understood in a variety of ways. Some of these are close to what I understand Appadurai to be getting at, that is, following things and people; some examples might include Hannerz’ studies of international journalists (Hannerz 1996, 2003), or Bestor’s study of the flow of tuna from the fishery off the northeastern coast of the United States to the sushi-bound markets of Tokyo (Bestor 2001). Many studies of migration (including many of those cited above) are of course central to this approach, and have led to an emerging argument that migration must no longer be understood as the recruitment of citizens of country A to work and reside in country B under clear rules established by that nation-state (or by A and B in collaboration with each other), but rather as a multi-faceted and multidirectional set of flows, communications and relationships in which people and resources (including communicative ones) circulate in many different ways at once. One can also, however, understand multi-sited ethnography to mean investigating the manifestations of some process in many sites at once, in order to better grasp the broader process in which local manifestations are situated; this is the version closest to the position of George Marcus, whose name is in fact most closely associated with the term (Marcus 1995). In either case, there is a shared methodological problem, which has to do with identifying the local places and moments where processes occur and their linkages across time and space with other heres and other nows. In the next section, I will discuss how one research project has tried to put some of these ideas into operation in an attempt to understand what flows mean when we are talking about things like language, identity and nation or nation-state, that is, about things which were supposed to be fixed, but which are moving around and mutating in interesting ways.

3. The commodiication and marketing o Canadian rancite´ As I mentioned earlier, francophone Canadian nationalism has, in the last forty years or so, been constructed around a modernist ideology linking language, culture, territory, nation and state. This has taken its most straightforward State-territorial form in Quebec, and a kind of quasi-territorial form outside Quebec, manifested in efforts on the part of leaders of the linguistic minority movement to establish and control homogeneous institutional spaces, especially as concerns state-controlled ones (Heller and La-

40. A study on transnational spaces brie 2003). As is the case with most nationalisms historically, francophone Canadian national identity has been linked to a Romantic vision of authenticity as rooted in the country, that is, in nature, and in continuity of occupation of the land and of cultural and linguistic practice (Williams 1973; Hobsbawm and Ranger 1983; Hobsbawm 1990). This discursive formation was sustained by a political economy in which, initially, ethnolinguistic categorization overlapped class stratification (Porter 1965; Clift and Arnopoulos 1979) and spatial segregation was the norm (Levine 1990; Lasserre 1998; Gilbert and Langlois n. d.), within an economy based on primary resource extraction and secondary sector industrial transformation of those resources. That was difficult enough to sustain as the population became increasingly urbanized, educated and both socially and geographically mobile through the 1960s and 1970s. The political economic shifts of the 1980s and 1990s however, have made it close to impossible. Those shifts include: (i) a neoliberalization of the formerly liberal welfare state; (ii) globalization; and, perhaps most importantly in this instance (iii) the collapse or restructuring of the primary and secondary sector economies on which the reproduction of nationalist francophone ethnolinguistic identity depended, in the wake of the emergence of a globalized tertiary sector economy in which cities play a prominent role. The traditional bastions of francophone Canada, that is, the areas that have long been understood as the keepers of the francophone flame and the guarantors of its authenticity, are precisely the areas which have suffered most from this shift. These are either rural areas with economies based on fishing, lumber, agriculture or mining (or some combination of these), or urban industrial working-class neighborhoods. To provide just a few key examples among many: ⫺ in Atlantic Canada, the major fishing industry, that of cod, was closed in 1992 due to overfishing, and a second iconic domain, the lobster fishery, is currently under threat, possibly because of climate change; ⫺ in northern Ontario, a majority of mines were closed and the paper industry encountered major difficulties because of a trade disagreement with the United States, a major market as well as a competitor; ⫺ the steel industry of southern Ontario suffered downsizing as industrial production increasingly went overseas, and as the US automobile industry, to which it is closely tied, suffered from severe competition from Asia; ⫺ in western Canada, the large wheat farms that were the basis of the Prairie economy have been proving unsustainable. While francophones are not alone in suffering the consequences of these changes they do feel them in two specific ways: they are over-represented demographically in these sectors, and their symbolic appropriation of them has been central to the production of legitimizing nationalist francophone discourses. (As the post-colonial dominant majority, English-speaking Canada has been concerned with nationalism in different ways; but that discussion is beyond the scope of this paper.) Indeed, the current changes have triggered a panic among institutionalized agents of production of francophone modernist nationalist discourse, both inside and outside state structures, all of whom agree on the vital political importance of what I might call “saving the space”. This is a version of Bourdieu’s notion that the first reaction when conditions change in ways that threaten the viability of symbolic markets, and therefore the value of the symbolic and material capital that people have invested there, is not to invest in new

727

728

VII. Exemplary studies capital more adequate to the new conditions and to new emerging markets, but rather to do everything possible to save the existing market (Bourdieu 1982). The connotation that I wish to emphasize here is that, for francophone Canada, saving the market means saving the idea of a homogeneous community fixed in rural locations you can point to on a map, locations which in our interviews with active participants in that market (mainly civil servants involved in official language minority funding programs, agents of those programs, or members of francophone associations) are usually referred to as nos communaute´s ‘our communities’, thereby eliminating the urban francophone population from consideration. This discursive perspective is largely shared by academic producers of knowledge about those communities, who have concerns about what is usually called the “ethnolinguistic vitality” of those communities and hence of francophone Canada, especially outside of Quebec (Landry and Allard 1996; Johnson and Doucet 2006), although the discursive perspective was first developed and remains shared by Que´be´cois nationalists. The major efforts aimed at saving the market target tourism and the knowledge economy as means of keeping young people in the rural areas that constitute francophone Canada’s traditional bastions. Secondary efforts aim at recruiting francophone immigrants into those areas (Quell 2002; CIC 2003; CIC 2005). In both cases it is assumed that urbanization is tantamount to “assimilation”, that is, to the loss of adherence to discourses and practices of collective identity (Heller 2005). But at the same time, the efforts aimed at preserving rural homogeneity by their very nature undo it: immigration, the knowledge economy and tourism all introduce ethnocultural, linguistic and racial diversity into the traditional bastions, by importing new actors and by putting existing actors into new, diverse and widespread social networks. One result is that in order to understand what is happening to francophone Canada, and to the discourses which constitute it as a relevant and meaningful social category, it is no longer adequate to carry out either community or institutional ethnographies which take for granted the existence of homogeneous and bounded social spaces. I found this out in a direct way through a school ethnography I carried out in the early 1990s in the Toronto area (Heller et al. 1999), in the course of which it became clear that the geographical mobility of students, parents and teachers was a key element of the construction of the school as a francophone space, or more properly of the contradictions between that construction and the complex and moving reality which had to be managed on a daily basis in the most quotidian of school-based practices. I have therefore had to find new ways of exploring how francophone Canada is being constructed, or possibly re-imagined, under the new conditions, both in the traditional bastions encountering late modernity and in the new spaces which francophones increasingly find themselves traversing. Most importantly, this has required moving away from understanding sociolinguistic ethnography as focused on sites, and towards understanding it as focused on spaces of discursive construction and on the trajectories of resources and actors through them. Furthermore, it has meant moving away from the quasinationalist distribution of research areas, in which each province of Canada becomes a research target in its own right, incommensurable (although comparable) with other regions, and back towards an older view of francophone Canada as stretching across political boundaries. Now, of course, it can be easily demonstrated that geographical mobility and transnationalism are scarcely new features of the francophone Canadian landscape; after all,

40. A study on transnational spaces what is francophone Canada but the product of European imperialist expansion? Ties to both England and France (and their colonial and neocolonial spheres of influence) remained active in one form or another from the time of New France on and always involved the circulation of people, of goods and of ideas. We can add to that the strong links to the United States. But in the past we had different lenses for understanding those forms of transnationalism and reasons for pushing them into the background of scholarly enquiry. Moreover, while a revisionist transnational history is almost certainly in order, the conditions of the past are not the conditions of today, and those remain at the forefront of my area of concern. In what follows, I will describe some specific attempts to carry this out empirically.

4. Francophone Canada goes transnational In this section I will begin by describing a case which represents one important dimension of the new ways in which francophone Canada is having to understand space in the globalized new economy: the circulation of identity goods (including both material products and performances) around Canada and between Canada and other parts of the francophone world, notably francophone Europe, and especially France; it involves goods and performances principally from Quebec and Acadia. I will focus mainly on the circulation of authentic products (most of them food-related) around commercial fairs and Christmas markets in France and Belgium, although much of what I have to say here applies to a variety of cultural and artistic artefacts and performances as well. What we find here is the by now well-known process whereby the current market for authentic goods and experiences provides a basis for reproduction of traditional ideologies of identity, but in ways which turn those identities into commodities (Le Menestrel 1999; Heller 2003; Coupland, Garrett and Bishop 2005; Jaworski and Pritchard 2005). This is a double-edged sword for the guarantors of authenticity of those commodities, since, while they have new value, that value is largely determined by a market constrained by consumer interest in fixed objects. Being someone else’s object of consumption has its delights, but also its disadvantages. The element I wish to explore here is how a traditional (nationalist) idea of linguistic minoritization goes hand in hand with the modes of authentication which link identity to the land and to fixed points in space, but precisely at the moment where those identities either circulate through distant markets, or are available to mobile consumers (who attribute greater value to products which have either traveled far to them or which they have traveled far to obtain). I will also keep an eye on the part of the story that is about sociolinguistic methodology, that is, what we have to do to see how things unfold. Indeed, the emphasis on this first case allows me to explore in greater depth what it means to investigate ethnographically the circulation of language, identities and resources than would be possible otherwise. However, in order not to leave the reader with a skewed idea of what transnationalism means for ideologies of language and nation such as those we find in francophone Canada, I will then briefly contrast this case with other ways in which transnationalism has come to francophone Canada, in particular: (i) the ways in which other ways of being francophone (and, of course, speaking French, or being bilingual in French and English) have entered the putatively homogeneous and fixed territories of francophone

729

730

VII. Exemplary studies Canada through the geographic and social mobility of actors (immigration is a large part of the story, though there is more to it than that), resulting in claims for a plurilocalization of francophone Canada, and (ii) the ways in which the commodification of language itself in such fields as translation sets up a market in which language is completely delocalized.

4.1. The circulation o authentic identity goods In 2005, Claudine Moı¨se, a member of our team based in France, noticed that the annual commercial fair in her hometown, Montpellier, was announcing Canada as the special guest for that year. It is not entirely clear to me when or how the practice of using states as a means of marketing such events came about, in France or elsewhere, although it is somewhat reminiscent of international competitions in sport or the arts. One person we interviewed said that in France this probably dates from an initiative in the 1990s undertaken by the then Minister of Culture, in which every year the Ministry would promote ties to a specific country and allocate funds to be made available to the organizers of regional or local events who wished to participate in that effort. It is possible that Canada was popular in 2005 as a result of programs set up in 2004 by both Canada and France to celebrate the 400th anniversary of the founding of the first French settlement in Canada. In any case, this was clearly an opportunity for us to explore what Canada as a brand might mean in the context of a French commercial fair. We used the opportunity to carry out fieldwork conceptualized as identifying and following threads. The fieldwork in Montpellier was carried out in October 2005 by team members based in France, Germany and Canada. This was largely the result of a conscious strategy used by our team in doing multi-sited ethnography within Canada since the 1990s, using teams composed of members with varying degrees of familiarity with the spaces, the networks and the resources involved, so as to maximize analytical perspectives and communicative contexts. Subsequent fieldwork at the Christmas fair of Montbe´liard was carried out in December 2005 by a similarly mixed group, including some of the same researchers, in December 2006 at the Christmas fair in Lie`ge (Belgium) and in August 2006 in the Saguenay region of Quebec on the production and circulation of the Nativity scenes exhibited in Montbe´liard. One dimension of the fieldwork can be understood as fairly conventional site ethnography: multiple visits to the site (like any visitor), checking out the kiosks and the restaurant, picking up promotional material, traveling to the site via public transport to see how it is presented in that space. Here it became clear that the image of Canada that was presented was a fairly traditional one, with kiosks made of “logs” and employees wearing iconic elements of what could be called male French Canadian national costume, which index the figure of the voyageur at the time of the fur trade or the nineteenth/twentieth-century buˆcheron (lumberjack): ceintures fle´che´es (woven multi-colored sashes) and red-and-black checked jackets. The hall was also decorated with both the Canadian and the Quebec flags, as is to be expected when state sponsors are involved, and the site also included a Plains Indian tipi that served syncretically as a site for performances of indigeneity by a member of the Maleseet First Nation from eastern Canada.

40. A study on transnational spaces Goods on display included a wide range of traditional items such as maple syrup, and more modern ones, such as hot tubs and frying pans (whose connection to the Canada brand was established via tropes such as the outdoors, healthy attention to the body, or technical competence). The restaurant served a menu designed to index authenticity, including items such as buffalo, beer, cranberries, and, of course, maple syrup (all also part of the commercial fair display). Interviews with organizers and employees allowed our team to establish what networks accounted for the particular selection of goods, people and images featured in Montpellier, and in particular to uncover the resource base for the growing industry of marketing (usually, but not only, francophone) Canadian authenticity in Europe. It was possible to establish the means with which to follow the actors and the resources. It quickly became clear that the core of the network involved entrepreneurs based in Canada (usually in Quebec) connected to entrepreneurs based in France. The Quebecbased entrepreneurs followed the goods to their markets in Europe, while the Francebased entrepreneurs were significantly less geographically mobile. The Quebec-based entrepreneurs were responsible for managing both the network of producers in Canada (some but not all of whom accompanied their goods to sell them in Europe) and a pool of temporary workers (vendors for the most part). While some of the resources (the frying pans, for example) might have been a bit difficult to authenticate, the vast majority were authenticated through a combination of tradition and rootedness in a fixed point in space. That is, their value came from being able to identify their origin in a space that was not only easily locatable, but also tied to values of the rural (nature, rootedness, the wild) which are core elements of traditional francophone Canadian-ness. They are also tied to indigenous knowledge. The following extracts are from an interview conducted on site in Montpellier by Claudine Moı¨se (C) and Emmanuel Kahn (E) with A, the Canadian partner managing the importation and sales of this set of goods at a variety of venues in francophone Europe. A explains the importance of guaranteeing the provenance of the goods on sale, both in terms of what part of the planet they are from and in terms of the knowledge which produces them. First A discusses how to make maple syrup, tying that process, interestingly, to what it means to present an indigenous person as part of the image of Canada (note that the linking of national or ethnonational identity to nature is also tied to aboriginal identity in ways which would likely be contested by First Nations within Canada). He then goes on to discuss salmon, cheese and buffalo meat (I do not provide the entirety of the discussion here for considerations of space). Extract 1 A:

ben/ il y a plusieurs personnes qui font leur petite cabane et vendent du sirop d’e´rable/ cabane qui (x) ici c’est une cabane (x) on a une bouilleuse/ on/ il y a de la (x)/ c¸a chauffe/ le monde l’avait vu plusieurs fois/ la cabane (x) le sirop preˆt a` apporter/ mais le faire/ ils avaient jamais vu c¸a/ et c’e´tait une premie`re puis c¸a nous (x) de´ja` beaucoup de/ on a au moins/ on a au moins quatre personnes la` qui veulent faire le meˆme truc l’an prochain/ comme il y a (x) il y a d’autres villes qui veulent avoir la cabane a` sucre/ c’est suˆr qu’on se prome`nera comme un charlatan de pilules avec la cabane/ il faut que ce soit bien structure´/ comme ici c’est tre`s bien structure´/ c’est de l’ouvrage/ c¸a prend un sucrier/ c¸a prend du monde compe´tent/ comme le tipi/ combien de fois qu’on voit un tipi et puis on voit quelqu’un du Que´bec qui repre´sente un Ame´rindien et qui en est pas un (…)

731

732

VII. Exemplary studies on en a un/ l’Ame´rindien qui est la` c’est un vrai/ c’est un (x) c’est une re´serve/ c’est un vrai/ le matin il part il purifie les stalles/ la` c’est un vrai/ et puis je pense les Franc¸ais maintenant ils ont accroche´ e´norme´ment (x) hey/ la` c’est pas/ c’est pas galvaude´ la`/ c’est du vrai/ A:

well/there are lots of people who do their little shack and sell maple syrup/ shack which (x) here it’s a shack (x) we have a boiler/we/there is (x)/it heats up/ people had often seen that/the shack (x) the syrup ready to go/ but making it/they had never seen that/it was a first and that we (x) already lots of/we have at least/we have at least four people who want to do the same thing next year/like there are (x) there are other cities which want to have the sugar shack/of course we could wander around like a charlatan pill-pusher with the shack/it has to be wellstructured/like here it’s very well-structured/it’s a lot of work/it takes a sugar-maker/it takes competent people/like the teepee/how often do you see a teepee and then you see someone from Quebec who represents an Amerindian and who isn’t one (…) we have one/the Amerindian who is there is a real one/he’s a (x) it’s a reserve/he’s a real one/ in the morning he goes out and purifies the stands/he’s a real one/and I think the French now they really got interested (x) hey/that’s not/that’s not fake/it’s for real

Extract 2 E: A: C: A:

C: A: E: A: C: A:

C: A:

le saumon fume´ d’ou` il vient? c’est du saumon de l’Atlantique et du saumon du Pacifique les deux oui c’est les deux ouais et sauvage/ parce que vous savez nous on peut pas (x) et c¸a c’est mon prochain de´fi/ c’est de faire comprendre/ je dis c¸a puis c’est pas pe´joratif/ je veux bien qu’on se comprenne/ parce que ici en France/ ils passent 40 000 tonnes de saumon fume´ par anne´e oui c’est tout du saumon d’e´levage where does the smoked salmon come from? it’s Atlantic salmon and Pacific salmon both yes both yeah and wild/because you know we can’t (x) and that’s my next challenge/is to get people to understand/I say this and it isn’t pejorative/I want to make sure we understand each other/because here in France/they go through 40,000 tons of smoked salmon a year yes it’s all farmed salmon

But language was also a key component of authentication. While at times members of the labor pool were recruited locally in France, the key managers all agreed that being able to provide a (largely linguistic) performance of francophone Canadian-ness was a major dimension of the symbolic value added of the products. Extract 3 C: A:

E:

et comment vous choisissez vos vendeurs la` sur le/ sur la foire? ben comme/ mes vendeurs/ a` l’exception de deux personnes c’est tout des/ des Que´be´cois/ des Que´be´coises/ euh/ les deux Franc¸ais sont excellents/ c’est pour c¸a que je les ai utilise´s/ ils connaissent le produit tre`s bien et puis euh/ comme le marche´ de Noe¨l bien/ je travaille avec le Webpage souvent avec l’Association France-Que´bec/ les vendeurs (x) les Que´be´cois ou Que´be´coises ont besoin d’un emploi/ ils appellent la`/ ah/ oui Promotion canadienne a besoin marche´ de Noe¨l/ ils envoient leur CV/ les jeunes passent tous par Internet/ ils envoient c¸a/ deux pages/ pas plus/ je regarde/ ouais/ ok/ si c’est bon on va l’e´tape plus loin et puis c’est tout des francophones?

40. A study on transnational spaces A: E: A: E: A: C: A:

pardon? des francophones ou il y a des anglo-que´be´cois ou? non/ c’est Que´be´cois/ Que´be´coises/ c’est (x) il faut l’admettre/ l’accent comme ils disent ouais souvent on va arriver/ non/ non/ on ache`te/ parlez-nous/ parlez-nous (rire) on capitalise sur c’t accent-la` aussi (x) c’est la business/ hein/ c’est

C: A:

and how do you choose your vendors for the/for the fair? well like/my vendors/except for two people they’re all Que´be´cois (masc.)/Que´be´coises (fem.)/ uh/the two French people are excellent/ that’s why I used them/ they know the product very well and uh/ like the Christmas market well/ I work with the webpage often with the Association France-Que´bec/the vendors x the Que´be´cois or Que´be´coises need a job/they call there/ah/ or Promotion canadienne needs (someone for a) Christmas market/they send their CV/young people all go through the internet/they send it/ two pages/ no more/ I take a look/yeah/ ok/ if it’s good we go to the ne(x)t stage and then they’re all francophones? pardon? francophones or are there anglo-Que´be´cois or? no/they’re Que´be´cois/Que´be´coises/it’s (x) I have to say/the accent as they say yeah often we get there/no/no/we buy/speak to us/speak to us (laughs) we capitalize on that accent also (x) it’s (the) business/eh/ it’s

E: A: E: A: E: A: C: A:

The result was that in some cases Canadians were recruited to travel to Europe, and in others, Canadians residing in Europe (for example, students) were hired. For the managers, this required a cost-benefit analysis balancing the accent’s ability to attract customers and justify a price against a calculation that includes the value of authenticity, knowledge of Canada, knowledge of the product, knowledge of the clientele and costs related to travel and subsistence for workers. Thus who circulates where is directly constrained by the conditions of a market in which authenticity in general, and linguistic authenticity in particular, are major features. That authenticity is built on the displacement across space of people and goods anchored in a traditional idea of francophone Canada as located in fixed places (and a time, roughly 1850 to 1930) into the here and now of the French market. The experience in Montpellier opened up paths to follow, both in terms of identifying trajectories of goods and people, and in terms of understanding the ideologies of transnationalism which were part of that movement. We followed many of the same actors and goods two months later to a Christmas market in Montbe´liard (France) and over a year later to one in Lie`ge (Belgium). We also followed up links to brokers and organizers of various kinds based in France and to the producers back in Canada, exploring for example the source of artisanal Nativity scenes exhibited at Montbe´liard. We traced them back to their source in a small town in Quebec, whose economic crisis led to the choice of Nativity scenes as a form of branding, in the service of developing community solidarity, of attracting tourists and of linking the remote town to the wider world. Finally, we examined the political and economic conditions, especially the work of ever-present agents of the state, in providing the funds and in facilitating the flow of information, that were important contributions to the commodification of Canadian francophone language and identity on the Euro-

733

734

VII. Exemplary studies pean francophone market. (We could, of course, have done much more; for example, we could have examined the recruitment process, or followed actors or goods physically from Canada to Europe and around European sites.) On a methodological level, this component of our research can be likened to using an observed phenomenon (in this case, the Foire de Montpellier) as a window onto a web of discursive construction. The event itself provides the kind of discursive material which has long been the object of sociolinguistic analysis. But an understanding of it is enriched by grounding that material ethnographically, by examining the social processes out of which it emerges and which it influences in turn, that is, by understanding where they lie across time and space, and what broader social processes help us understand why they occur where and when they do and why they take the shape they have. These processes, it turns out, are deeply and complexly transnational in that their very meaning and value stems from the transnational circulation of goods and people, but one in which the national categories at play are extremely traditional. This is, however, only one way in which francophone Canada is turning transnational. In the next two sections I will briefly contrast this picture with two other sets of phenomena: the incursion of the transnational into local, putatively fixed and homogeneous spaces; and the emerging importance of a language industry in which the local and authentic slugs it out with the decontextualized and universal for pride of place.

4.2. The world comes to rancophone Canada The observation which put me on the trail of the transnational in the first place, as I mentioned earlier, was the appearance, in territorial and institutional spaces set up as strategically essentialized francophone Canadian, of participants from a wide variety of backgrounds, bringing all kinds of life trajectories, linguistic repertoires and senses of belonging to spaces set up to build uniformity. There are a variety of reasons for this. One is the use of the democratic state by linguistic minority activists to create spaces of privileged control; doing so also requires legitimization of those spaces as inclusive. A second is the social mobility of Canadian francophones and their entry into a political economy in which the immigration of educated and skilled workers is of vital importance. Finally, in recent years, the world political economy has rendered emigration (or flight) necessary or at least attractive to increasing numbers of educated people from the francophone world. The result, very simply, has been work, largely on the part of immigrants and refugees, aimed at creating a symbolic place for themselves in francophone Canadian spaces. This is an ongoing, and increasingly well-documented, struggle (Jedwab 2002; Quell 2002; Laperrie`re and Beaule´ 2004; Gallant and Belkhodja 2005; Magnan and Pilote 2007), in academia and in popular culture (Bourbonnais 2005⫺2006; Sarkar and Winer 2006). Fundamentally, this a version of a broad struggle over the reproduction of ethnonational identity as we know it, through incorporation of new members into otherwise unchanging social, political and cultural categories, versus the loosening of boundaries and the redefinition of criteria of inclusion and exclusion towards more complex, dynamic and fluid forms of belonging and appropriation. This discursive struggle pits monocentric local authenticity against the polycentric; and opposes the local authentic (or local authenticities) with the fluid and universal.

40. A study on transnational spaces Methodologically, our work on this aspect of change has taken institutionalized francophone spaces such as schools and community centers as a point of departure, in order to examine the discursive struggles within them as a function of the interests of participants who can muster different kinds of linguistic and cultural capital through a variety of social networks, more and more of them transnational. Here, the transnational is embedded in the daily experience of participants in the social and institutional spaces where their trajectories converge, at least temporarily.

4.3. The language industry A final element to consider is the emerging so-called “language industry”, a feature of the globalized new economy aimed at managing the problems of globalized economic networks requiring some means of handling multilingualism. The language industry is in fact frequently associated with the growing localization industry, that is, with the development of techniques and technologies for helping globalized businesses adapt to local markets and labor pools. The Canadian State recognized several years ago that it was sitting on expertise in the management of bilingualism developed out of the necessity of running a country that was committed to an ideology of official bilingualism, but was also working (for similar ideological reasons) with aboriginal and immigrant languages (Heller 1999). It was also sitting on not one, but two, languages connected to global markets, as well as a variety of immigration-related languages of international economic importance (such as Mandarin, Cantonese, Hindi and Urdu). It has been since working out a policy aimed at capitalizing on that expertise as an investment in global capitalism (Gouvernement du Canada 2003). Current policy identifies translation, interpretation, language teaching (mainly of English and French to citizens of other countries) and speech technology as the key fields. The issue here really has to do with who gets to define what counts as English and as French (and as English-French bilingualism) in these areas. The struggle tends to be implicit, because the ideological underpinning of the field takes language as an autonomous skill that can be detached from any considerations of authenticity. Globalized capital also tends to still prefer Fordist, that is uniformized, scripts (Gee, Hull and Lankshear 1996; Cameron 2001) while simultaneously struggling with the idea of the value of niche marketing and hence the primacy of local forms of diversity, as well as with the value of authenticity itself in many domains (as we saw above). For many francophone Canadians, the language industry presents potential opportunities for capitalizing not only on the resource of their first language, but also on the bilingualism that for many years was the hallmark of their subordination. The problem is that, particularly for lower middle class or working class speakers, it is not necessarily their vernacular French, or their particular (mixed) bilingual practices which are valued there; they face stiff competition from others (anglophone graduates of French immersion programs, francophone immigrants from Africa or Europe) whose linguistic repertoire is largely school-derived (or in any case derived from spaces better dominated by prestige varieties).

735

736

VII. Exemplary studies This manifestation of transnationalism risks reproducing class distinctions, albeit on a somewhat different basis than in the past. Grasping it requires identifying the spaces where the language industry is at work, the criteria for the selection of legitimate speakers there and the nature of the language produced as a marketable product.

5. Webs and links, spaces and trajectories, resources and actors I have identified three ways (there are probably more) in which francophone Canada is encountering transnationalism, and I have tried to lay out the challenges of studying these processes as well as some of the strategies we have adopted to try to meet them. As with other linguistic nationalisms, minority or otherwise, the complexity and mobility indexed by the concept of transnationalism can no longer be swept under the rug. The question is, rather, how one copes. We have seen here the contradictory ways in which the globalized new economy values both traditional authenticity and universalized and decontextualized practices, resulting both in the development of separate niches (people writing programs for machine translation do not often work with people selling maple syrup at Christmas markets) and in some forms of struggle over access and legitimacy in others (education, for example). We see formerly dominant notions tying authenticity to fixed locality actively valued, as long as they are commodified as part of the transnational circulation of goods and consumers. At the same time, we see those notions actively challenged. The first order of challenge comes from attempts to replace single-locale authenticity with ideas about polycentric sources of authenticity and the accompanying syncretic forms of hybrid linguistic and cultural practices, such as those we see for example in multilingual rapping or bilingual and standard-vernacular theatre performances. The second order of challenge comes from ideas about linguistic and cultural forms which are held to be universal and hence easily transportable, and which remove from consideration all ideas about rootedness as essential to belonging. These discursive struggles emerge in spaces where the key resources involved are produced, distributed and given value and meaning. The challenge for linguistics is to find ways of describing how resources and actors circulate, constrain what is possible and creatively navigate those constraints. The methods I have described here require loosening up our ideas of ethnography and recognizing in particular that the idea of the bounded whole meets its limits in such conditions.

6. Acknowledgements The data on which this discussion is based are drawn from a research project, La francite´ transnationale: Pour une sociolinguistique de la mouvance, funded by the Social Sciences and Humanities Research Council of Canada (2004⫺2007). I would like to thank the many members of the team, especially those involved in the fieldwork reported on here: Gabriele Budach, Alexandre Ducheˆne, Philippe Hambye, Emmanuel Kahn, Me´lanie LeBlanc, Claudine Moı¨se, Mireille McLaughlin and Mary Richards.

40. A study on transnational spaces

7. Reerences Appadurai, Arjun, (ed.) 1986 The Social Life of Things: Commodities in Cultural Perspective. Cambridge: Cambridge University Press. Appadurai, Arjun 1991 Global ethnoscapes: Notes on queries for a transnational anthropology. In: R. G. Fox (ed.), Recapturing Anthropology, 191⫺210. Santa Fe, NM: School of American Research. Appadurai, Arjun 1996 Modernity at Large: Cultural Dimensions of Globalization. Minneapolis: University of Minnesota Press. Basch, Linda, Nina Glick Schiller and Cristina Szanton Blanc (eds.) 1994 Nations Unbound: Transnational Projects, Postcolonial Predicaments and Deterritorialized Nation-States. London: Routledge. Bauman, Richard and Charles L. Briggs 2003 Voices of Modernity: Language Ideologies and the Politics of Inequality. Cambridge: Cambridge University Press. Beck, Ulrich and Natan Sznaider 2006 Unpacking cosmopolitanism for the social sciences: A research agenda. The British Journal of Sociology 57(1): 1⫺23. Bestor, Theodore C. 2001 Supply-side sushi: Commodity, market, and the global city. American Anthropologist 103(1): 76⫺95. Bourbonnais, Jean 2005⫺2006 Pure laine. Television series, 16 eps., in French; written by Martin Forget. Canada: Christal Films Distribution/Te´le´-Que´bec. Bourdieu, Pierre 1982 Ce que parler veut dire. Paris: Fayard. Burawoy, Michael, Sean O. Riain, Joseph A. Blum, Sheba George, Zsuzsa Gille, Teresa Gowan, Lynne Haney, Maren Klawiter, Steven H. Lopez and Millie Thayer 2000 Global Ethnography: Forces, Connections and Imaginations in a Postmodern World. Berkeley, LA: University of California Press. Cameron, Deborah 2001 Good to Talk? London: Sage. Castells, Manuel 2000 The Information Age: Economy, Society and Culture. 3 vols. Oxford: Blackwell. CIC ⫽ Citoyennete´ et immigration Canada 2003 Cadre strate´gique pour favoriser l’immigration au sein des communaute´s francophones en situation minoritaire. Ottawa: Citoyennete´ et immigration Canada. CIC ⫽ Citoyennete´ et immigration Canada 2005 Vers la francophonie canadienne de demain. Sommaire des initiatives 2002⫺2006 pour favoriser l’immigration au sein des communaute´s francophones en situation minoritaire. Ottawa: Citoyennete´ et immigration Canada. Clift, Dominique and Sheila McLeod Arnopoulos 1979 Le Fait franc¸ais au Que´bec. Montreal: Libre Expression. Corbeil, Jean-Claude 1980 L’Ame´nagement linguistique du Que´bec. Montreal: Gue´rin. Coupland, Nikolas, Peter Garrett and Hywel Bishop 2005 Wales underground: Discursive frames and authenticities in Welsh mining heritage tourism events. In: Adam Jaworski and Annette Pritchard, Discourse, Communication and Tourism, 199⫺221. Clevedon: Channel View.

737

738

VII. Exemplary studies Deshaies, Denise and Conrad Ouellon (eds.) 1998 Les Linguistes et les questions de langue au Que´bec: Points de vue. Quebec: Centre international de recherche en ame´nagement linguistique. Featherstone, Mike (ed.) 1990 Global Culture: Nationalism, Globalization and Modernity. London: Sage. Forest, Jean 1998 Chronologie du que´be´cois. Montreal: Les E´ditions Tryptique. Gallant, Nicole and Chedly Belkhodja 2005 Production d’u´n discours sur l’immigration et la diversite´ par les organismes francophones et acadiens du Canada. E´tudes ethniques au Canada 37(3): 35⫺58. Gee, James P., Glynda Hull and Colin Lankshear 1996 The New Work Order: Behind the Language of the New Capitalism. Boulder, CO: Westview Press. Georgeault, Pierre and Michel Page´ (eds.) 2006 Le Franc¸ais, langue de la diversite´ que´be´coise: Une Re´flexion pluridisciplinaire. Quebec: Conseil supe´rieur de la langue franc¸aise. Giddens, Anthony 1990 The Consequences of Modernity. Berkeley, LA: University of California Press. Gilbert, Anne and Andre´ Langlois undated L’Atlas du de´veloppement des communaute´s francophones et acadiennes du Canada. Online publication accessible at Gouvernement du Canada 2003 Le Prochain Acte: Un Nouvel E´lan pour la dualite´ linguistique canadienne. Le plan d’action pour les langues officielles. Canada: Bureau du Conseil prive´/Privy Council Office. [Also published in English as The Next Act: New Momentum for Canada’s Linguistic Duality. The Action Plan for Official Languages.] Hannerz, Ulf 1996 Transnational Connections: Culture, People, Places. London: Routledge. Hannerz, Ulf 2003 On being there … and there … and there! Reflections on multi-site ethnography. Ethnography 4(2): 201⫺216. Harvey, David 1989 The Condition of Postmodernity. Oxford: Blackwell. Heller, Monica 1999 Heated language in a cold climate. In: Jan Blommaert (ed.), Language Ideological Debates, 143⫺170. Berlin: Mouton de Gruyter. Heller, Monica 2002 E´le´ments d’une sociolinguistique critique. Paris: Didier. Heller, Monica 2003 Globalization, the new economy and the commodification of language and identity. Journal of Sociolinguistics 7(4): 473⫺492. Heller, Monica 2005 Une Approche sociolinguistique a` l’urbanite´. Revue de l’Universite´ de Moncton 36(1): 321⫺346. Heller, Monica (ed.) 2007 Bilingualism: A Social Approach. London: Palgrave Macmillan. Heller, Monica, with the collaboration of Mark Campbell, Phyllis Dalley and Donna Patrick 1999 Linguistic Minorities and Modernity: A Sociolinguistic Ethnography. London: Longman. Heller, Monica and Normand Labrie (eds.) 2003 Discours et identite´s. La Francite´ canadienne entre modernite´ et mondialisation. Fernelmont: E´ditions modulaires europe´ennes.

40. A study on transnational spaces Hobsbawm, Eric 1990 Nations and Nationalism since 1760. Cambridge: Cambridge University Press. Hobsbawm, Eric and Terence Ranger (eds.) 1983 The Invention of Tradition. Cambridge: Cambridge University Press. Hutton, Christopher 1999 Linguistics and the Third Reich: Mother-tongue Fascism, Race and the Science of Language. London: Routledge. Jaworski, Adam and Annette Pritchard (eds.) 2005 Discourse, Communication and Tourism. Clevedon/Buffalo/Toronto: Channel View. Jedwab, Jack 2002 L’Immigration et l’e´panouissement des communaute´s de langue officielle au Canada: Politiques, de´mographie et identite´. Ottawa: Commissariat aux langues officielles. Johnson, Marc L. and Paule Doucet 2006 Une Vue plus claire: E´valuer la vitalite´ des communaute´s de langues officielles en situation minoritaire. Ottawa: Commissariat aux langues officielles. Also available online from . Landry, Rodrigue and Re´al Allard 1996 Vitalite´ ethnolinguistique: Une perspective dans l’e´tude de la francophonie canadienne. In: Jürgen Erfurt (ed.), De la polyphonie a` la symphonie. Me´thodes, the´ories et faits de la recherche pluridisciplinaire sur le franc¸ais au Canada, 61⫺88 Leipzig: Leipziger Universitätsverlag. Laperrie`re, Anne and Denis Beaule´ 2004 Franco-Que´be´cois et Que´be´cois des minorite´ ethniques: lignes de fracture et de suture. In: Lucille Guilbault (ed.), Me´diations et francophonie interculturelle, 55⫺81. Quebec: Presses de l’Universite´ Laval. Lasserre, Fre´de´ric 1998 Le Canada d’un mythe a` l’autre. Montreal: HMH Hurtubise. Le Menestrel, Sara 1999 La Voie des Cadiens. Paris: Belin. Levine, Marc V. 1990 The Reconquest of Montreal: Language Policy and Social Change in a Bilingual City. Philadelphia: Temple University Press. Magnan, Marie-Odile and Annie Pilote 2007 Multiculturalisme et francophonie(s): Enjeux pour l’e´cole de la minorite´ linguistique. Glottopol 9: 80⫺92. Marcus, George 1995 Ethnography in/of the world system: The emergence of multi-sited ethnography. Annual Review of Anthropology 24: 95⫺117. Martel, Pierre and He´le`ne Cajolet-Laganie`re 1996 Le Franc¸ais que´be´cois: Usages, standard et ame´nagement. Montreal: Institut que´be´cois de recherche sur la culture. Mougeon, Raymond (ed.) 1989 Le Franc¸ais parle´ hors Que´bec: Aperc¸u sociolinguistique. Quebec, Presses de l’Universite´ Laval. Mougeon, Raymond and E´douard Beniak 1991 Linguistic Consequences of Language Contact and Restriction: The Case of French in Ontario, Canada. Oxford: Oxford University Press. Ostiguy, Luc and Claude Tousignant 1993 Le Franc¸ais que´be´cois: Normes et usages. Montreal: Gue´rin universitaire. Pe´ronnet, Louise 1989 Le Parler acadien du sud-est du Nouveau-Brunswick: E´le´ments lexicaux et grammaticaux. Bern: Peter Lang.

739

740

VII. Exemplary studies Porter, John 1965 The Vertical Mosaic: An Analysis of Class and Power. Toronto: University of Toronto Press. Quell, Carsten 2002 L’Immigration et les langues officielles: Obstacles et possibilite´s qui se pre´sentent aux immigrants et aux communaute´s. Ottawa, Commissariat aux langues officielles. Also available online from . Sankoff, Gillian 1980 The Social Life of Language. Philadelphia: University of Pennsylvania Press. Sarkar, Mela and Lise Winer 2006 Multilingual code-switching in Quebec Rap: Poetry, Pragmatics and Performativity. International Journal of Multilingualism 3(3): 173⫺192. Thibault, Pierette and Diane Vincent 1990 Un Corpus de franc¸ais parle´: Montre´al 1984: Historique, me´thodes et perspectives de recherche. Quebec: De´partement des langues et linguistiques, Universite´ Laval. Williams, Raymond 1973 The Country and the City. London: Chatto and Windus.

Monica Heller, Toronto (Canada)

41. The study o language and space in media discourse 1. 2. 3. 4. 5. 6. 7. 8.

Introduction Shifting values of localness in heteroglossic mediascapes Spaces, media, language: Reviewing the main concepts Mediated local speech: The “reflection fallacy” Shifting approaches: From language variation to “techniques of localism” Contrast and voice: Issues of analysis Conclusion References

1. Introduction This article explores the relationship between language and space in the context of mass and new media, drawing on a critical review and synthesis of research literature. My starting point is the observation that media discourse may draw on linguistic variability ⫺ e.g., phonological variation, lexical or code choice ⫺ to index how social actors, institutions or reported events are related to particular spaces or places. We encounter such constructions of “linguistic locality” (Johnstone, Andrus and Danielson 2006: 79) on any ordinary day of media consumption. They include advertisements which use a regional dialect to authenticate a product’s origin; television shows which blend in the local languages of tourist destinations; pop stars who self-consciously use their urban

41. Language and space in media discourse vernacular; or films using a vernacular to underscore the origin of a character. In these and other cases, linguistic form ties in with propositional content and other semiotic means to construct “what it means linguistically to be ‘here’ or ‘from here’ and how places and ways of speaking are thought to be related” (Johnstone, Andrus and Danielson 2006: 79). Such linguistic heterogeneity is sometimes felt to be the rule rather than the exception in today’s mediascapes (Appadurai 1996), but language in the media looks back to a rather standardized or centripetal tradition (Fowler 2001; Bakhtin 1981; Androutsopoulos 2007). As a result of the nexus between mass media and the emerging nationstates of the nineteenth century, the media used standard language to reach national audiences, thereby strengthening the bond between language and nation and contributing to the standardization process (Anderson 2006; Giesecke 1992). This did not exclude vernaculars from public discourse, but restricted their status, functions and reach. The use of vernaculars to index origin or social class against the backdrop of standard language predominance has a long tradition in theatre and fiction (Blake 1981; Goetsch 1987; Lippi-Green 1997; Culperer 2001; Schröder 2005). In film, the contrastive deployment of standard and non-standard dialects as indicators of social class was established by the mid⫺World War II era (Marriott 1997). However, there is a widespread impression that the visibility of vernacular speech in the media has increased in the last few decades (Crystal 2002). In Europe, regional speech is embraced in broadcast and celebrated in advertising (Burger 2005: 364; Coupland 2007: 171). Representations of dialect in newspapers have become more common in the last decade (Betz 2006: 78, 173), and minority languages or mixed codes are surfacing in local or minority media (Jaffe 2000, 2007; Busch 2004). These processes have received relatively little attention in sociolinguistics, where, despite important forerunners such as Bell (1984), language in the media has been viewed as distinct from “authentic” vernacular speech. However, a number of current developments seem to have built up momentum for a focus on media discourse in sociolinguistics: foundational concepts of authenticity and identity are being critically questioned (Coupland 2001; Bucholtz 2003; Bucholtz and Hall 2008); the role of media in the agenda of fields such as language ideology and critical discourse analysis has been acknowledged, and “the reach and impact of media language in contemporary social life” (Coupland 2007: 28) is becoming increasingly apparent to researchers of language and society. The turn to “post-variationist”, style-centered approaches offers a pivotal point from which to engage with media sociolinguistics, and concepts such as stylization and performance are in part fleshed out with media discourse data (Coupland 2001, 2007; Schilling-Estes 2002; Auer 2007). This article is both motivated by and aims to contribute to this turn by developing a post-structuralist view of local speech as a resource that is deployed in media contexts in ways that go well beyond its classic understanding as marker of local belonging.

2. Shiting values o localness in heteroglossic mediascapes The proliferation of vernacular speech in the media reflects wider processes of social and institutional change, and I draw in this section on Fairclough’s (1995) distinction between micro, meso and macro dimensions of discourse in order to sketch out these

741

742

VII. Exemplary studies processes and how they affect the mediatization of local speech. On a macro level, the deregulation of media systems over the last decades has led to a tremendous diversification of target audiences; more recently, the digital revolution has increased grassroots access to media production technology and blurred the boundary between producers and audiences, creating new chances for vernacular voices to be heard publicly. Tendencies toward regionalization or localism, which emerge as a response to globalization pressures, may lead to a rediscovery of local linguistic values. Two other relevant aspects of the globalization process are commodification, i. e., the use of local linguistic resources for promotional purposes, and disembedding, i. e., the detachment of social relations from their original spatial context (Coupland 2003; Coupland, Bishop and Garrett 2003). Such macro-level processes can be seen as affecting meso-level changes in institutional practices of media planning and production. For instance, media diversification leads to a proliferation of local media and narrowcasting programs (see section 3.2), in which the celebration of local identity and culture gains priority. Formats of audience engagement and participation have become widely popular in the last couple of decades, resulting in a dramatic increase in the presence of lay speakers in broadcast content (Straßner 1983; Hausendorf 2003; Burger 2005). These institutional changes in turn affect the micro level of linguistic and textual processes which we focus on here. The key tendency to conversationalization (Fairclough 1995), i. e., the use of informal speech and conversational features in public discourse, is no doubt joined by vernacularization, i. e., an increase in the currency of non-standard speech. Language in the media now regularly includes “exposed dialects” (Coupland 2007: 171), and linguistic variability that is associated with specific target audiences is being transformed into linguistic capital (Richardson and Meinhof 1998). In media linguistics literature, this is reflected in, e.g., the view of media language as a “variety mix” (Burger 2005) and the suggestion that nowadays any linguistic variety has good chances of being mediatized (Schmitz 2004: 33). However, it seems important not to lose track of the ideologies and power relations surrounding the mediatization of vernacular speech. Traditional ideologies of non-standardness are still widely reproduced, and we can clearly see their workings in the example of local dialects. For instance, when regional television channels reserve use of the regional dialect exclusively for “local culture” shows, they draw on and reinforce the link of dialect to local rural tradition (Richardson and Meinhof 1998). In other cases, linguistic localness is ideologically transformed into an index of social stratification, with film, comedy or newspaper reports often drawing on vernacular speech as a signifier of low status (Marriott 1997; Sand 1999; Lippi-Green 1997). Such patterns are certainly not without exception. Local speech styles may also be represented in ways that challenge these stereotypes, and some vernaculars have ambivalent and contested meanings: for instance, London English (“Cockney”), traditionally cast as a working-class marker, is repositioned as marker of subcultural, street-smart styles in contemporary pop music culture. Indeed, the current proliferation of local speech in the media makes it increasingly difficult to allocate local speech a single, fixed meaning. Rather, it bears a range of potential meanings, whose activation depends on the specific contextual embedding of local speech (see Eckert 2008). On the other hand, as media contexts are complex and ambivalent, resorting to stereotypes of localness may offer producers and audiences a relieving reduction of complexity. An implication of this discussion is the need to steer away from fixed meanings of local speech as a marker of local identity. Rather than subscribing to an “anything goes”

41. Language and space in media discourse view of linguistic variability in the media, I suggest an understanding of media discourse as a site of heteroglossia (Bakhtin 1981). Defined as “(a) the simultaneous use of different kinds of forms or signs, and (b) the tension and conflicts among those signs” (Bailey 2007: 257), the notion of heteroglossia has been adopted by researchers interested in the intricate relationship between linguistic diversity, social difference and power (Baxter 2003: 67; Cook 2001: 187; Busch 2004, 2006). Viewing media as fundamentally heteroglossic means that our close analytical attention to the diversity of media language is complemented by a focus on the power relationships that shape the meanings of this diversity in context. The following sections attempt to flesh out such an approach.

3. Spaces, media, language: Reviewing the main concepts The nexus of language, media and space has received one dominant reading in the literature, i. e., the use of regional dialects in broadcast. While this remains a centerpiece in the following discussion, this article aims at bringing further aspects of space, media and language together in a comprehensive framework.

3.1. Spaces - rom local to virtual To begin with, it seems useful to think of spaces and places (see Johnstone in this handbook) in terms of a cline or scale, reaching from e.g., a country to a single neighborhood, with a different set of sociolinguistic contrasts becoming relevant on each level (Richardson and Meinhof 1998: 88). For example, a regional dialect in a standard-language environment has the potential to index some sort of engagement with, or proximity to that region and with what the region conventionally stands for. On a larger scale, a language as a whole may be juxtaposed with another language in order to index the national affiliation of a speaker. On a smaller scale, an urban dialect may index belonging to a city (as opposed to other cities or the surrounding region), and a slangy speech style or an ethnic or mixed variety may be mobilized to index proximity to a particular district or neighborhood. The relations between spaces and associated linguistic resources are not essentially fixed, but evoked and reproduced in discourse by means of ideological links between vernacular varieties, social types and places. But spaces indexed by language may transcend a single geographical location or lie beyond physical space altogether. In media for migrant and diasporic communities, both the community’s current location and their place of origin are made relevant in discourse and indexed by code-switching or other patterns of bilingual speech. And as digital communications technologies enable people to interact via computer networks, virtual spaces emerge on platforms such as chat channels, discussion boards or role-playing environments, which lie in a tradition dating back to the earliest days of computermediated communication and extend to currently popular formats such as Second Life.

3.2. Media - rom mass to narrowcast As these observations suggest, an open, inclusive understanding of media may increase awareness of the highly diverse ways in which language may signal local affiliations in

743

744

VII. Exemplary studies public discourse. Different delimitations of the media field will constrain the phenomena under consideration in different ways. Linguists interested in media discourse often restrict their scope to printed media and broadcast, excluding domains such as film and the new media, where linguistic constructions of localness are common and potentially influential. One point of entry for a consideration of the role of the media is to compare the use of vernacular speech across different media types. Straßner (1983) suggests that dialect is most common in Germany’s press, followed by radio and then television, whose nationwide reach impedes the use of dialects. However, the sociolinguistics of television have dramatically changed in the last two decades, and as the research focus is shifting to the discursive work done with “dialect” in particular contexts (sections 4 and 5), such a broad hierarchy seems questionable. Other researchers compare patterns of vernacular use in different media within a speech community, e.g., Jamaican Creole in newspapers and radio phone-ins (Sand 1999), Cypriot Greek dialect in newspapers, radio phone-ins and advertising (Pavlou 2004), and dialects across all media types in German-speaking Switzerland (Burger 2005; see section 5). Their findings clearly suggest moving beyond a comparison of media types as a whole. Another macro-level approach is to consider the potential impact of media institutions. Based on the tripartite system of public service media, private commercial media and alternative media (Busch 2004; Busch 2006; Androutsopoulos 2007), one might ask how the communication economy of each media sector shapes their constructions of locality and uses of local linguistic resources. Private commercial media have been shown to drive forward the processes of conversationalization, and this seems to hold true for vernacularization as well; for example, private broadcasters have been crucial in increasing the share of dialect in the Swiss media (Burger 2005). It remains to be examined whether alternative media such as community radio deal with localness in distinct ways, perhaps driven by their aim of offering a forum to social groups traditionally excluded from the public sphere (Busch 2004). Yet a different angle is to examine how the reach of a media institution or product to specific target audiences may impact on constructions of localness. This allows a connection to the study of local media and narrowcasting, i. e., “channels dedicated to particular types of programming … as well as channels designed to reach a local population rather than the country as a whole” (Richardson and Meinhof 1998: 87). Local media have been around for a long time (e.g., local newspapers), but narrowcasting refers specifically to private or public broadcasting which emerged after the deregulation of media systems. The question here is how the targeting of a spatially delimited audience may draw on language that is indexical of that area, and how this complies with relevant media aims and policies (Löffler 1998). A final perspective is to ask which media genres and formats favor the proliferation of local speech. My literature survey and observations indicate six key areas: 1. Advertising offers textbook examples of how local linguistic forms are turned into symbolic capital in the authentication of products or geographical regions (see Bell 1992, 1999; Coupland 2007; Johnstone 2004; Piller 2001). Dialect advertising comes with different degrees of audience engagement, sometimes inviting recipients to become part of an imaginary local community. Within an advertisement, dialect is often confined to the narratee, i. e., the person using or promoting the product (Piller 2001), while the voiceover or concluding slogan, which feature voices of institutional author-

41. Language and space in media discourse ity, are in standard language. Straßner (1986) finds that in southern German newspapers dialect advertising is restricted to particular product categories, especially food and beverage, and capitalizes on romantic, rural and traditional connotations. Such conjunctures of dialect, local produce and local identity indeed seem quite widespread (see e.g., Pavlou 2004 on Cyprus), but dialect is nowadays also used in ways which challenge traditional dialect values, as in examples discussed by Coupland (2007: 171⫺176) and Bell (1999) or in a series of commercials by the German federal state of Baden-Württemberg, in which dialect-speaking celebrities and professionals are framed by the slogan “We can do anything. Except speak High German”. The meanings of localness in advertising, then, are complex and multilayered, and despite enduring traditional stereotypes, local speech is sometimes coupled with high social status and recontextualized as a symbol of collective or corporate identity. 2. Linguistic landscape (Gorter 2006) seems an important site for the commodification and emblematization of vernacular speech. Local codes may be selected for use on shop signs, in company logos or tourist memorabilia. This may fulfill the informational function of the linguistic landscape, i. e., to index the territorial presence of a language and its speakers, or its symbolic function, i. e., to declare ethnolinguistic identity. These purposes may coexist with the wider commercial function of language display (Eastman and Stein 1993), i. e., to authenticate services by indexing a relation of the service to the places where that language is actually spoken. In that sense, any language may be used to index the place of origin of a product or service, as the domain of gastronomy makes clear. 3. Film and other types of fiction (comedy, drama, novel, music, television series) are important sites for the public staging of localness and local speech (see Balhorn 1998; Blake 1981; Crystal 2002; Goetsch 1987; Marriott 1997; Schröder 2005; Taavitsainen, Melchers and Pahta 1999). When local varieties are used in fiction, the contrast with other linguistic varieties is typically tied to contrasts between the characters of the plot. In film, speech styles become part of wider, multimodal “mechanisms of juxtaposition and contrast” (Marriott 1997: 184), which involve clothing, outfit, scenery, etc. As Lippi-Green suggests, In traditions passed down over hundreds of years from the stage and theatre, film uses language variation and accent to draw character quickly, building on established preconceived notions associated with specific regional loyalties, ethnic, racial or economic alliances. (1997: 81)

When dialect is used as “shortcut to characterization” (Lippi-Green 1997: 81), it will often index social stratification by being allocated to working rather than middleclass characters, people from the “street” rather than representatives of an institution (see also Johnstone, Andrus and Danielson 2006: 94). Local speech can also be used to enhance a film’s setting and to provide local flavor, and different functions of local speech in fiction may be intertwined, acting both as class and mood marker as well as reminder of the narrative setting. 4. Audience participation formats such as talk, game, quiz or reality shows contribute to the visibility and diversity of lay people’s voices. Formerly restricted to direct quotations and sound bites in news shows or documentaries, lay speech is now spread across media programs, and the drive to engage the audience can even lead to lay

745

746

VII. Exemplary studies speakers assuming small institutional roles in narrowcasting stations, such as weather presenting (Richardson and Meinhof 1998). Lay speakers are not professionally trained speakers, and in some cases are presumably selected because they display “a legitimate connection to a place” (Johnstone, Bhasin and Wittkofski 2002: 160; Jaworski et al. 2003). Their speech will often be positioned as authenticating the places, activities and experiences being reported on. In turn, by accepting and embracing lay vernacular voices into its program, a media institution may authenticate itself as representative of local concerns. 5. In media for transnational and diasporic audiences (e.g., community newspapers or ethnic websites), discourse is shaped by the tension between “here” (place or residence and everyday life) and “there” (homeland, place of ethnic identification), and their linguistic practices may relate to both relevant spaces. If the community language is still the audience’s base language, traces of contact to the majority language will index the community’s current location and cultural setting. If the community has shifted to the majority language, residual elements of home language, often intricately bound to heritage culture, remain in place and are transformed into “rich points” of diasporic culture (see section 3.4). Code-switching may be used in accentuating contrasts between home and diaspora, thereby drawing on associations of ethnic “wecode” and dominant “they-code” (Gumperz 1982). On web discussion forums popular with Jamaican expatriates, most Creole occurs in contributions from abroad, and code-switching tends to accompany deictic shifts between the writer’s place of residence and the Jamaican homeland (Hinrichs 2006). 6. Virtual spaces: in computer-mediated environments, local deixis (i. e., “being here”) becomes fundamentally ambiguous as it may index both the virtual discourse ground and a physical space that is important to the participants, and language may index orientation to both physical and virtual space. In online communities which are oriented to familiar geographical space, such as city-based chat channels (section 5), markers of local affiliation are ubiquitous, including dialect features, local references and allusions, and accompanying visual elements. In multi-user role-playing environments such as Second Life or Active Worlds (Axelsson, Abelin and Schroeder 2003), physical space may provide the backdrop for novel virtual territories, in which local language policies are established and corresponding linguistic performances emerge, authenticating the virtual space (e.g., use of Spanish expected in “mundo hispanico”, Dutch being spoken in Second Life’s “Amsterdam”). In online communities which are not defined in relation to geographical space, references to virtual space may take various forms, e.g., labels for the space and its regular visitors, a specific jargon for online activities relevant to the group. Metalinguistic conventions regulating content, style or language choice emerge, and norm-enforcing practices may occur, in which what is considered appropriate “here” gives reason to exclude participants from the virtual ground. As this overview suggests, the media environments which host local speech are those which favor heteroglossia generally: talk and interaction, fiction, audience participation, less informing or educating discourses than entertaining ones. Some of these facets of media discourse have received relatively little attention in (critical) linguistic literature, which has often tackled politically “important”, yet stylistically standardized, monoglot domains of the media (Fowler 2001).

41. Language and space in media discourse

3.3. Local speech - as language, code or symbol As suggested by the discussion so far, the range of linguistic variability that can be deployed to index space in the media is, in principle, limitless. Local dialects, varieties of a pluricentric language, creoles, mixed vernaculars or minority languages ⫺ they all may evoke, depending on context, an association to particular spaces or places, even though they are per se primarily associated with other social dimensions such as class or ethnicity. In examining how local linguistic resources are positioned in relation to other languages in a media context, I find it useful to draw on tripartite distinctions familiar from social dialectology and creole studies. Mattheier (1980) suggests that dialect may function as the “main variety” of a speech community, a “social symbol” or a “relict language”. Mair (2003) classifies contemporary uses of Jamaican Creole as “language”, “code” or “symbol” (see also Hinrichs 2006). In what follows, this latter distinction shall be used as a window onto the functions of vernacular varieties in media contexts. On the first level of this triadic scheme, a local vernacular serves as a base language for given media across various genres and topics. Radio hosts with a regional accent would be a case in point, as reported from the UK or (southern) Germany, albeit presumably restricted to spoken language media of regional reach. Media using a local dialect as a base language seem rare, Switzerland being a noticeable exception (see section 5). We might expect that the willingness to accept a regional variety as an institutional voice presupposes highly positive attitudes towards that variety. There is evidence that local media may celebrate dialects which are stigmatized in the wider speech community (see Richardson and Meinhof [1998] on Saxon dialect in regional television), but the dialect voices in this case are lay speakers, not professional journalists. The notion of code foregrounds relations of contrast, in which a code of localness is juxtaposed with other codes with different, non-local associations. Drawing on concepts from bilingualism research, their distribution may be conceived of as situational or metaphorical code-switching (Gumperz 1982). The former implies a conventionalized distribution of a code to activities or situations. When for instance local media regularly cast a small part of their content in dialect (a column on local affairs, a political cartoon, advertisements etc.), the local code is made visible and granted default status within a niche of the media product, but at the same time kept at bay from fully fledged institutional usage. The situational dimension also captures occasional uses of local codes which are perceived as obvious and “natural” in terms of topic or life-world. When for example a television documentary on the life of fishermen features the fishermen speaking exclusively their local dialect, these local voices tie in with local activities in sharp contrast to the language of the narrator or the surrounding program. On the other hand, metaphorical code-switching captures linguistic evocations of localness that are typically short and often unique in the course of a text or program: a show host switching to a local vernacular in order to caricature a social type or to address a caller; a politician in a talk show switching into his dialect to claim an understanding of local affairs; a newspaper report casting direct speech by elderly locals in the dialect (Selting 1983; Schlobinski 1988; Birkner and Gilles 2008). This is a wide, largely uncharted territory, in which language often indexes spaces and associated social types by means of double-voicing and stylization (section 5).

747

748

VII. Exemplary studies Finally, the notion of symbol implies that local speech is quantitatively and functionally restricted, but nonetheless made relevant to the participants’ identity claims. Symbols of localness are often single lexical items and proper names, which are positioned in discourse in such a way as to flag individual or institutional identities. Titles, headings, mastheads, product and company names, slogans, web forum signatures, website navigation buttons are typical textual slots in which local codes may take on symbolic functions (Coupland, Bishop and Garrett 2003; Androutsopoulos 2003, 2007; Johnstone, Bhasin and Wittkofski 2002). Here dialect is both reduced to an accessory and elevated to an icon of identity, collaborating to this end with other local indexicals of self-presentation. We return to such usage of dialect with Coupland’s (1995) notion of “rich points of culture” in the following.

3.4. Rich points: Where local orms meet local reerences This tripartite distinction is also related to Johnstone’s (2004) distinction between two types of regional speech in the media, i. e., a strategically symbolic and a habitual type. The latter corresponds in my classification to “language”, albeit a language with restricted and controlled distribution in terms of the voices and participation formats it is allocated to. Johnstone’s strategically symbolic type, corresponding to both “code” and “symbol”, entails the planned use of markers of localness in the design (Kress and van Leeuwen 2001) of a media product, thereby relying on vernacular stereotypes and anticipating audience reactions. A striking instance of this symbolically strategic type is the fit between form and content of local indexicals, by virtue of which markedly local forms come to designate a local place or category. Consider Johnstone’s example of Dahntahn, an item representing the feature of monophthongized /aw/ in media reports on “Pittsburghese” dialect. When this feature is represented in media reports, it tends to be illustrated by precisely that lexical item. This co-variance of local form and local place name is also encountered in city names whose pronunciation or spelling emphasizes their vernacular form. When they occur embedded in a standard-language environment, these variants foreground specific relations of the speaker, the addressee or a referent to the urban location. For instance, writing Saaf London ⫺ which marks two key features of London English, i. e., th-fronting and a monophthongal variant of the MOUTH vowel (Wells 1986: 309, 328) ⫺ may index either attributes of lower social class stereotypically associated with that area, or the familiarity of the speaker with local vernacular culture, or possibly both. Likewise, departing from the base language of discourse to label a social group or a key cultural concept in its own code serves as a powerful emphasis of group membership, which can be deployed in order to identify with addressees or referents or to render then as “other” (Jaworski 2007; Hinrichs 2006). This coupling of linguistic form and local reference is captured in Coupland’s (1995) notion of “rich points of culture”. For Coupland, what makes dialect a culturally powerful marker of local identity is, precisely, “the conjunction of pronunciation and ideational reference”, his example being a Cardiff place name pronounced by broadcasters in Welsh English (rather than standard English). Coupland views such conjunctions of dialect phonology and place names (or culturally relevant common nouns) as “producing

41. Language and space in media discourse expressions that are cultural mini-icons” of regional or ethnolinguistic identity (Coupland 1995: 316⫺317). Importantly, such “cultural mini-icons” are not fixed once and for all. The co-variance of local indexicals is a resource available for deployment in different socio-cultural contexts. A telling example is a slogan used by an Austrian far-right politician during the 2008 election campaign: Daham statt Islam ‘At home instead of Islam’. This uses a dialect form of the adverb, i. e., daham with long [a:] instead of the diphthong /ai/ of standard daheim, with all other campaign materials in standard German. More than just the choice of a dialect variant for rhyme, dialect-as-symbol is mobilized here to bolster the notion of local identity, which is juxtaposed to an “other” religion and thus contextualized as a national Christian identity. Interestingly, during the same period an Austrian cultural association released a CD of religious music called Daham ist Islam ‘At home is Islam’, thus mobilizing the same item of dialect-as-symbol for the opposite social meaning.

4. Mediated local speech: The relection allacy Determining the “authenticity” of mediated local speech seems a central concern in parts of the literature, and my aim in this section is to question this. By “reflection fallacy” I refer to the often tacit expectation that local speech in the media be an (accurate) reflection of non-mediated language, and to the resulting tendency to foreground this relation in the analysis. To be sure, linguistic constructions of localness in the media owe their indexical power to their tacit interpretation as representations of the informal, everyday language of a community (see Balhorn 1998; Jaffe 2007). As vernacular varieties are reproduced outside educational institutions, their mediatization contrasts with the nonlocal codes and registers required for professional media discourse practices. However, this fundamental contrast does not mean that mediated local speech will always correspond to an “authentic” counterpart. Indeed, the problematic nature of such an expectation becomes obvious in the frequently mentioned gaps and inconsistencies between the two. According to some authors, advertising features “inauthentic, synthetic, simulated” (Straßner 1986: 323) or “softened up” (Janich 1999: 171) dialects, in which deep local features are ironed out in order to avoid comprehension problems or evocations of stigma. By other accounts, based on a range of genres, the media exaggerates a few stereotypical vernacular features; combines “authentic” dialect features with less realistic ones; or even presents authorial creations as “authentically local” linguistic forms (Crystal 2002). More specifically, Bell (1992: 334) points out how advertising relies on a few “typical features to mark target dialects”. Marriott (1997: 182⫺183) shows how a British film draws together both more and less realistic dialect features in the representation of characters in terms of social class. Birkner and Gilles (2008) find that stylizations of Cologne dialect in television shows rely on a few dialect features, which are grafted onto colloquial standard German. Expecting mediated local speech to live up to an “authentic” model seems fraught with theoretical and methodological problems. First, it may easily lead to an idealized view of local speech (Bucholtz 2003) and a “fetishization” of supposedly authentic linguistic form. Second, the expectation of accuracy does not always correspond to the

749

750

VII. Exemplary studies views and motivations of media practitioners whose language use is shaped by a complex web of institutional, technological and generic constraints. A “realistic” rendering of language varieties is often not intended at all (Bell 1992). Third, in some cases an authentic model is not obvious (e.g., in advertisements featuring a mock foreign accent) while in others the mediation of local speech relies on traditions of public vernacular use (Marriott 1997; Birkner and Gilles 2008). Thus filmic representations of English dialects owe their success not to their documentary precision but their popular availability, “permitting [a] ready identification” (Marriott 1997: 183). Overall, it is perhaps realistic to assume that much media usage of local speech works no differently to literature, in that authors and speakers are “not aiming at the systematic and accurate representation of real-life sociolinguistic facts, but at supplying some markers of particular varieties, leaving readers to fill in the gaps with background knowledge” (Culperer 2001: 209). This does not mean that accurate representations of nonstandard speech do not exist; it means, however, that we should not expect them to be the default case. More generally, the search for an authentic model is rooted in the belief that language “naturally” belongs to geographical spaces. Taking inspiration from human geography and Foucault-inspired discourse theory, Johnstone (this handbook) proposes rethinking this relationship: no bond between language and space pre-dates discourse; rather, it is in discourse that the link between language and spaces is constituted. This fits well with the (equally Foucauldian) tenet of critical discourse analysis, that media discourse does not represent a reality beyond itself, but rather creates social reality in its own right (e.g., Fairclough 1995; Thornborrow 2002). Johnstone shows how, in media reports and web forum discussions, a new urban dialect, “Pittsburghese”, is literally talked into being, regardless of whether or not this status is dialectologically accurate (Johnstone 2004, Johnstone and Baumgardt 2004). We may also think in this context of the Mundartwelle ‘dialect wave’ in Germany in the 1970s, i. e., the increased use of dialect in advertising and other mass media, which was less a symptom of dialect resurgence in society at large than a trend within the media, neither reflecting nor initiating an increased use of local dialects “out there” (Löffler 1997). Indeed, a post-structuralist perspective on language and space challenges the primacy of “dialect out there”, instead drawing attention to how discourse shapes the understanding of dialect.

5. Shiting approaches: From language variation to techniques o localism The relation between media, language and space has been examined within various research frameworks so far, and I sketch here a transition from descriptive and variationist to currently influential post-variationist and interactional approaches. Parts of the literature focus on the frequency and distribution of regional dialects across media types and genres (Straßner 1983; Schmitz 2004: 31). The strengths of this approach are illustrated by the work of Burger (2005: 362⫺388) on dialect use in southern German, Austrian and Swiss German media. The diglossic situation of German-speaking Switzerland entails a split between written and spoken language media. In the former, written dialect is restricted to “occasional exceptions” (2005: 368) such as adverts for local shops and services. In spoken language media too, the choice is organized by local reach, with local

41. Language and space in media discourse radios using almost exclusively dialects, and local topics and participants favoring the use of dialect in nationwide programs, where standard German otherwise predominates. If both varieties are co-present in a program, their distribution is organized along the boundary of spoken/written or spontaneous/edited discourse (2005: 371⫺373). In research within the audience design framework (Bell 1984), the relationship between language style and target audience is examined based on variationist methodology. Bell’s original study found that colloquial phonological variables in the language of New Zealand radio stations occurred more frequently on local community stations, followed by commercial regional channels and the national public program. Even though no dialect variants were involved, localness is played out in Bell’s interpretation, which suggests that the more a station orients to a local (and younger, lower-status) audience, the more it will draw on vernacular variants as a resource to shape its relationship to that audience; by contrast, the more detached a station is from local contexts, the higher its share of standard features. This correlation of colloquial speech with restricted reach is confirmed by other studies. Reynolds and Cascio (1999) found that a British local newspaper uses more contracted forms than do national dailies, and Burger’s work, even though not variationist in methodology, offers ample evidence for the systematic association of regional speech with regional reach. Research on computer-mediated communication shows that German chat channels named after a city or region display features of the respective dialect, albeit restricted in terms of frequency and environments (Androutsopoulos and Ziegler 2004; Franke 2006). A study on Swiss-German chat communication (Siebenhaar 2006) shows that local chat channels (with a younger audience) use local dialect features almost categorically, whereas supra-regional channels (with an older audience) vary between dialect and standard forms. The attention paid in these studies to the distribution of local features is not always accompanied by an equal amount of attention to localness in discourse. Some earlier descriptive work tends to view dialect in the media in a matter-of-fact way, i. e., as a variety that happens to be used in certain genres or by certain speakers. Dialect is thus viewed as “behavior” rather than “performance” (Coupland 2001), and its function as an index of local origin is assumed a priori rather than examined in context. What I term the “reflection fallacy” (section 4) also forms part of this angle. Characteristically, Burger’s “fundamental linguistic question” is whether the distribution of varieties in broadcasting “mirrors the conditions of everyday linguistic reality” or whether the media develop their own usage, which might in turn impact upon “linguistic reality” (Burger 2005: 369; my translation). Methodologically, a strictly distributional approach implies that the occurrence of local features will be interpreted as an index of the strength of local orientation of a station or show. However, regional features may be non-deliberate, and their interactional treatment may suggest that they are less a resource for claiming localness than something participants try to avoid (Pavlou 2004). While on a more abstract level even such unintended uses of dialect are, of course, indexical to particular places, claims about the identity relevance of local speech in the media may be risky without a close look at what is actually being done with that speech within the discourse context. In work located within the perspective of interactional and style-oriented sociolinguistics (Schilling-Estes 2002; Coupland 2007) local speech is viewed as a resource in the construction of individual or institutional style. The focus shifts from linguistic features as such to the social types and activities styled or stylized through local speech, and

751

752

VII. Exemplary studies from authenticity to processes of authentication, by which local speech marks people or activities as “genuine”, and is itself constituted as genuine (Bucholtz 2003; Eastman and Stein 1993). A prime example of this line of research is from Coupland (2001), who illustrates the contrast between variationist and interactional views of dialect ⫺ “behavior” or “performance” ⫺ using the example of a Welsh radio disc jockey. In one of his examples, the radio host uses Welsh English in utterances that constitute a “local” worldview, but subsequent utterances representing an authoritative opinion are voiced in Standard English. Stereotypical conjunctions of language and social identity are still in place here, but these are more-or-less strategically selected and reflexively performed, i. e., they are an outcome of stylization. Dialect stylization emphasizes the relationship between dialect as a resource, the communicative activity at hand and the identity projected by the speaker. “To style dialect is to construct a social image or persona that interconnects with other facets of a speaker’s communicative design … in a particular event or act.” (Coupland 2001: 348) In this spirit, Birkner and Gilles (2008) examine dialect stylizations in reality shows and television comedy, pointing out that the default assumption in social dialectology, i. e., that dialect is a marker of regional belonging, is entirely off target in their data, in which practices of styling local origin, stylizing local social types or performing dialect prevail. Androutsopoulos and Ziegler (2004) show that local dialect on a German chatchannel often serves as a contextualization cue, marking shifts in topic or key, quotes and asides, or evoking the voice of local proletarian types. Franke (2006) finds similar patterns in a Berlin chat-channel. Again, far from being a reflection of the channel’s collective identity, dialect is tailored to specific interactional purposes, including the jocular staging of local stereotypes. A logical progression from this point would be to move towards a more holistic view of media discourse, extending the scope from individual speech styles, texts or shows to larger units. Local media and narrowcasting invite us to examine how local linguistic resources are strategically positioned in wider processes of media planning. As narrowcasting audiences are typically delimited by region, constructing a sense of locality can be considered a core aspect of local media design. Richardson and Meinhof find that British and German narrowcast programs deploy a broad range of “techniques of localism” (1998: 99), which involve media content (locally significant stories and topics), visual content (both contributing to a story and as an independently meaningful code), intertextuality (allowing participants to draw on and demonstrate local knowledge), and local jargon (designations of urban locations, often abbreviations, acronyms or metaphors). Narrowcasters will constantly seek occasions to produce local talk, even though these occasions may be asymmetrically distributed across participant formats and voices. In this process, bonds formed by music, topics and linguistic forms signify the importance of place, and non-standard varieties may gain niche prestige in the grassroots mediation of local communities.

6. Contrast and voice: Issues o analysis The preceding discussion hopefully makes clear the importance of the specific discourse context in interpreting the meaning of linguistic localness. Speaking simply, it makes a difference if local speech comes from a newsreader or a fictional character, the show

41. Language and space in media discourse host or an interviewee, in prime time or late at night. In order to identify and interpret these differences, I suggest that the analysis of linguistic forms needs to be complemented by the analysis of contrasts between local speech and other linguistic resources in their discourse context, the genres in which local speech occurs, the voices local speech is allocated to or presented as owned by, and the identities, i. e., social categories, which are represented or evoked through local speech in discourse. Elaborating on the interplay between contrast and voice, the importance of contrast follows from the principle that sociolinguistic styles are relational, i. e., gain their social meaning through situated contrasts to other styles (Coupland 2007). The notion of voice, on the other hand, foregrounds issues of ownership, participation format and perspective: to whom does regional speech belong, to whom is it allocated, whose perspective does it cue? (Bakhtin 1981; Baxter 2003; Blommaert 2005; Coupland 2007: 111). Local speech may be cast as the voice of institutions or of lay speakers, and it may appear as a speaker’s own, habitual voice or as an alien voice that is temporarily “put on” (in Bakhtinian double voicing). It is convenient to initially assume such simple dichotomies, even though this will not always suffice to capture the multiple layers of voices in broadcast. Local forms are in most cases not the default language in media discourse. Their choice is a matter of voice (Blommaert 2005), and a local voice in the media is always implicated in contrasts to other voices present in the media output. The analysis of this interplay operates on an ideological as well as a linguistic level. It asks how local voices are framed and contextualized by other voices (e.g., those of show hosts or reporters) or other semiotic material (music, sound, moving images). This includes an assessment of the relative frequency of local speech vis-a`-vis other resources, bearing in mind that indexing localness may be “more a matter of individual occurrences of salient variants than of quantitative sums and relative frequencies” (Bell 1992: 337; see also Bell 1999). Voice contrasts may be organized sequentially, as in code-switching or double-voicing, or in a more complex, multimodal way involving juxtapositions of speech, sound and visual event. The nexus of genre, topic and code choice may give local speech visibility and legitimacy in the media, but it also confines it to niches and excludes it from wider institutional uses. In German local newspapers, such niches are typically found in the local pages and the arts and culture section (Feuilleton; Straßner 1983), where dialects are used in articles on local everyday life or local politics, albeit often in a satirical or polemical manner (e.g., political cartoons with dialect-speaking characters who critique or mock local affairs). Straßner argues that dialect offers more drastic illustrations than standard language and enables statements that taboo or lack of evidence would not permit in standard language. In other words, dialect legitimizes certain statements by contextualizing them as originating from a non-institutional voice, perhaps a “man in the street” perspective. These observations fit in well with the written representation of “Pittsburghese” in connection with “topics that have a strong local link and about which people are not being completely serious”, such as “cartoons about the city transit system” (Johnstone, Bhasin and Wittkofski 2002: 159). Judging from the research literature, the relation of local speech to voices in media discourse follows a fairly consistent pattern. If local speech is only allocated to some participants, these will be lay speakers rather than presenters or hosts (Richardson and Meinhof 1998: 99). In fiction, it will be given to villains, low-status or minor characters

753

754

VII. Exemplary studies rather than protagonists or “goodies”. If a show host uses local speech (Coupland 2001), their delivery will be double-voiced and comic rather than thoroughly serious. If local speech is represented in newspaper reports, it will be contained within reported voices (i. e., as direct speech) rather than in the reporter’s own voice; if a broadcast feature has both on-screen voices and voiceover, vernacular will probably feature as on-screen voice (e.g., in live reporting), while the voiceover will be cast in standard language (e.g., Burger 2005: 365); and so on. In all such tendencies, the local code constructs an orientation toward everyday, extramural life, and its exclusion from professional voices eventually sustains the preclusion of accent/dialect from institutional use. While one effect of the proliferation of lay speakers is that institutional language loses its monopoly and mediascapes become more heteroglossic, this does not necessarily empower vernaculars and their speakers, as their mediated exposition is still regimented by relations of asymmetry. Many lay-speaker appearances are one-off and of short duration, and non-standard speech is thereby positioned as the speakers’ only voice, i. e., as behavior rather than performance, as habitual and thus unavoidable rather than strategically selected. Thus an analysis of voices and contrasts may show how the dominance of standard language persists even within highly heteroglossic mediascapes, but it can also be used to reveal where any exceptions and challenges to this pattern lie.

7. Conclusion Local speech in the media is employed in highly versatile ways which defy easy generalizations, and this article is only a first step towards synthesizing research insights into a coherent picture. I argued for the need to move beyond the “reflection fallacy” and to rethink the relationship between language and space in the media as inventive and creative, with the media actively constituting links between speech and spaces rather than simply reflecting existing links. Representations of localness are often done with limited linguistic material, and may echo traditions of public discourse. The heteroglossic perspective advocated here assumes that discourse does not just reflect but also creates social reality in its own right. As a consequence, the focus shifts to contrastive relations between codes and voices within media discourse, and to the workings of mediated local speech within its generic and institutional contexts. In terms of methodology, the study of language and space in the media is bound to take us beyond news discourse. Engaging with fiction and entertainment culture acknowledges that these genres may also be crucial in shaping the values of linguistic heterogeneity. The perspective developed in this article reinforces the idea, entailed in a post-variationist type of approach, that constructing localness is not simply a matter of using a local code, but of doing so in ways that evoke places and foreground “cultural icons” associated with these places. As a consequence, the attention shifts from varieties to styles and stylizations, and from linguistic forms as such to their conjunctions with other dimensions of discourse and multimodality. A question for further research is therefore how local speech in the media becomes meaningful not just by virtue of belonging to a distinct linguistic subsystem, but through its fit with particular topics, voices, genres, and visual sceneries ⫺ in short, with wider homologies of localness. We may think of the ways in which a sense of locality is constructed as working in a puzzle-like

41. Language and space in media discourse manner, complementing each other across different semiotic levels. Reconstructing this puzzle requires multi-level analyses, which would benefit from combining the methodological frameworks reviewed in this article.

8. Reerences Anderson, Benedict 2006 Imagined Communities: Reflections on the Origin and Spread of Nationalism. Revised edition. London/New York: Verso. Androutsopoulos, Jannis 2003 HipHop im Web: Zur Stilanalyse jugendkultureller Websites. In: Ulla Fix and Stephan Habscheid (eds.), Gruppenstile: Zur sprachlichen Inszenierung sozialer Zugehörigkeit, 271⫺292. Tübingen: Stauffenburg. Androutsopoulos, Jannis 2007 Bilingualism in the mass media and on the Internet. In: Monica Heller (ed.), Bilingualism: A Social Approach, 207⫺230. New York: Palgrave Macmillan. Androutsopoulos, Jannis and Evelyn Ziegler 2004 Exploring language variation on the Internet: Regional speech in a chat community. In: Britt-Louise Gunnarsson, Lena Bergström, Gerd Eklund, Staffan Fridell, Lise H. Hansen, Angela Karstadt, Bengt Nordberg, Eva Sundgren and Mats Thelander (eds.), Language Variation in Europe, 99⫺111. Uppsala: Department of Scandinavian Languages, Uppsala University. Appadurai, Arjun 1996 Modernity at Large: Cultural Dimensions of Globalization. Minneapolis: University of Minnesota Press. Auer, Peter (ed.) 2007 Style and Social Identities. Berlin/New York: de Gruyter. ˚ sa Abelin and Ralph Schroeder Axelsson, Ann-Sofie, A 2003 “Anyone speak Spanish?” Language encounters in multi-user virtual environments and the influence of technology. New Media and Society 5(4): 475⫺498. Bailey, Benjamin 2007 Heteroglossia and boundaries. In: Monica Heller (ed.), Bilingualism: A Social Approach, 257⫺274. New York: Palgrave. Bakhtin, Mikhail 1981 Discourse in the novel. In: Mikhail Bakhtin (translated by Caryl Emerson and Michael Holquist), The Dialogic Imagination, 259⫺422. Austin: University of Texas Press. Balhorn, Mark 1998 Paper representations of the non-standard voice. Visible Language 32: 56⫺74. Baxter, Judith 2003 Positioning Gender in Discourse: A Feminist Methodology. Basingstoke: Palgrave Macmillan. Bell, Allan 1984 Language style as audience design. Language in Society 13: 145⫺204. Bell, Allan 1992 Hit and miss: Referee design in the dialects of New Zealand television advertisements. Language and Communication 12(3/4): 327⫺340. Bell, Allan 1999 Styling the other to define the self: A study in New Zealand identity making. Journal of Sociolinguistics 3(4): 523⫺541.

755

756

VII. Exemplary studies Betz, Ruth 2006 Gesprochensprachliche Elemente in deutschen Zeitungen. Radolfzell: Verlag für Gesprächsforschung. Birkner, Karin and Peter Gilles 2008 Dialektstilisierung im Reality-Fernsehen. In: Helen Christen and Evelyn Ziegler (eds.), Sprechen, Schreiben, Hören, 101⫺130. Vienna: Praesens. Blake, N. F. 1981 Non-standard Language in English Literature. London: Andre Deutsch. Blommaert, Jan 2005 Discourse: A Critical Introduction. Cambridge: Cambridge University Press. Bucholtz, Mary 2003 Sociolinguistic nostalgia and the authentication of identity. Journal of Sociolinguistics 7(3): 398⫺416. Bucholtz, Mary and Kira Hall 2008 All of the above: New coalitions in sociocultural linguistics. Journal of Sociolinguistics 12(4): 401⫺564. Burger, Harald 2005 Mediensprache. 3rd ed. Berlin: de Gruyter. Busch, Brigitta 2004 Sprachen im Disput. Klagenfurt: Drava. Busch, Brigitta 2006 Changing media spaces: The transformative power of heteroglossic practices. In: Clare Mar-Molinero and Patrick Stevension (eds.), Language Ideologies, Policies and Practices, 206⫺219. Basingstoke: Palgrave Macmillan. Cook, Guy 2001 The Discourse of Advertising. 2nd ed. London: Routledge. Coupland, Nikolas 1995 Pronunciation and the rich points of culture. In: J. Windsor Lewis (ed.), Studies in English and General Phonetics, 310⫺319. London: Routledge. Coupland, Nikolas 2001 Dialect stylization in radio talk. Language in Society 30: 345⫺375. Coupland, Nikolas 2003 Introduction: Sociolinguistics and globalisation. Journal of Sociolinguistics 7(4): 465⫺ 472. Coupland, Nikolas 2007 Style. Cambridge: Cambridge University Press. Coupland, Nikolas, Hywel Bishop and Peter Garrett 2003 Home truths: Globalization and the iconising of Welsh in a Welsh-American newspaper. Journal of Multilingual and Multicultural Development 24(3): 153⫺177. Crystal, David 2002 Broadcasting the nonstandard message. In: Richard Watts and Peter Trudgill (eds.), Alternative Histories of English, 233⫺244. London: Routledge. Culperer, Jonathan 2001 Language and Characterisation. Harlow: Longman. Eastman, Carol M. and Roberta F. Stein 1993 Language display: Authenticating claims to social identity. Journal of Multilingual and Multicultural Development 14(3): 187⫺202. Eckert, Penelope 2008 Variation and the indexical field. Journal of Sociolinguistics 12(4): 453⫺476. Fairclough, Norman 1995 Media Discourse. London: Edward Arnold.

41. Language and space in media discourse Fowler, Roger 2001 The language of newspapers: Communicative and aesthetic aspects. In: Joachim-Felix Leonhard, Hans-Werner Ludwig, Dietrich Schwarze and Erich Straßner (eds.), Medienwissenschaft, vol. 2, 1693⫺1702. (Handbücher zur Sprach- und Kommunikationswissenschaft 15.2.) Berlin: de Gruyter. Franke, Katharina 2006 Language variation in #berlin. (Networx 48.) Available at ; last accessed 27 October 2008. Giesecke, Michael 1992 Sinnenwandel, Sprachwandel, Kulturwandel. Frankfurt am Main: Suhrkamp. Goetsch, Paul (ed.) 1987 Dialekte und Fremdsprachen in der Literatur. Tübingen: Narr. Gorter, Durk (ed.) 2006 Linguistic Landscape: A New Approach to Multilingualism. Clevedon: Multilingual Matters. Gumperz, John J. 1982 Discourse Strategies. Cambridge: University Press. Hausendorf, Heiko 2003 Nähe, Vertrautheit und Spontaneität: Eine Beispielanalyse zu linguistischen Aspekten der Einbeziehung des Zuschauers in die Sendung. In: Anne Betten, Richard Schrodt and Andreas Weiss (eds.), Neue Sprachmoral? Medien. Politik. Schule, 42⫺61. Salzburg: Praesens. Hinrichs, Lars 2006 Codeswitching on the Web. Amsterdam/Philadelphia: Benjamins. Jaffe, Alexandra 2000 Comic performance and the articulation of hybrid identity. Pragmatics 10(1): 39⫺59. Jaffe, Alexandra 2007 Corsican on the airways: Media discourse in a context of minority language shift. In: Sally Johnson and Astrid Ensslin (eds.), Language in the Media, 149⫺172. London: Continuum. Janich, Nina 2003 Werbung. 3rd ed. Tübingen: Narr. Jaworski, Adam 2007 Language in the media. Authenticity and othering. In: Sally Johnson and Astrid Ensslin (eds.), Language in the Media, 271⫺280. London: Continuum. Jaworski, Adam, Crispin Thurlow, Sarah Lawson and Virpi Ylänne-McEwen 2003 The uses and representations of local languages in tourist destinations: A view from British TV holiday programmes. Language Awareness 12(1): 5⫺29. Johnstone, Barbara 2004 Place, globalization, and linguistic variation. In: Carmen Fought (ed.), Sociolinguistic Variation: Critical Reflections, 65⫺83. Oxford: Oxford University Press. Johnstone, Barbara, Jennifer Andrus and Andrew E. Danielson 2006 Mobility, indexicality, and the enregisterment of “Pittsburghese”. Journal of English Linguistics 34: 77⫺104. Johnstone, Barbara and Dan Baumgardt 2004 “Pittsburghese” online: Vernacular norming in conversation. American Speech 79: 115⫺ 145. Johnstone, Barbara, Neeta Bhasin and Denise Wittkofski 2002 “Dahntahn” Pittsburgh: Monopthongal /aw/ and Representations of Localness in Southwestern Pennsylvania. American Speech 77: 148⫺166. Kress, Gunther and Theo van Leeuwen 2001 Multimodal Discourse. London: Arnold

757

758

VII. Exemplary studies Lippi-Green, Rosita 1997 English with an Accent. London: Routledge Löffler, Heinrich 1997 Dialektverlust oder Mundartrenaissance? ⫺ aus Schweizer Sicht. In: Gerhard Stickel (ed.), Varietäten des Deutschen. Regional- und Umgangssprachen, 386⫺388. Berlin/New York: de Gruyter. Löffler, Heinrich 1998 Lokalfernsehen aus der Macherperspektive. In: Werner Holly and Berndt U. Biere (eds.), Medien im Wandel, 157⫺172. Wiesbaden: Westdeutscher Verlag. Mair, Christian 2003 Language, code and symbol: The changing roles of Jamaican Creole in diaspora communities. Arbeiten aus Anglistik und Amerikanistik 28(2): 231⫺248. Marriott, Stephanie 1997 Dialect and dialectic in a British war film. Journal of Sociolinguistics 1(2): 173⫺193. Mattheier, Klaus J. 1980 Pragmatik und Soziologie der Dialekte. Heidelberg: Quelle & Meyer. Pavlou, Pavlos 2004 Greek dialect use in the mass media in Cyprus. International Journal of Sociology of Language 168: 101⫺118. Piller, Ingrid 2001 Identity constructions in multilingual advertising. Language in Society 30(2): 153⫺186. Reynolds, Mike, and Giovanna Cascio 1999 It’s short and it’s spreading: The use of Contracted Forms in British Newspapers: A change under way. In: Hans-Jürgen Diller (ed.), English via Various Media, 179⫺200. Heidelberg: Winter. Richardson, Kay and Ulrike H. Meinhof 1998 Worlds in Common? Television Discourse in a Changing Europe. London: Routledge. Sand, Andrea 1999 Linguistic Variation in Jamaica. Tübingen: Narr. Schilling-Estes, Natalie 2002 Investigating stylistic variation. In: Jack K. Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.), The Handbook of Language Variation and Change, 375⫺401. Oxford: Blackwell. Schlobinski, Peter 1988 Code-switching im Berlinischen. In: Norbert Dittmar and Peter Schlobinski (eds.), Wandlungen einer Stadtsprache. Berlinisch in Vergangenheit und Gegenwart, 83⫺102. Berlin: Colloquium. Schmitz, Ulrich 2004 Sprache in modernen Medien. Berlin: Schmidt. Schröder, Martin 2005 Zwischen Aisthesis und Mimesis. Zur Funktion des Dialekts in der standardsprachlichen Literatur. In: Eckhard Eggers, Jürgen Erich Schmidt and Dieter Stellmacher (eds.), Moderne Dialekte ⫺ Neue Dialektologie, 667⫺678. Stuttgart: Steiner. Selting, Margret 1983 Institutionelle Kommunikation: Stilwechsel als Mittel strategischer Interaktion. Linguistische Berichte 86: 29⫺48. Siebenhaar, Beat 2006 Code choice and code-switching in Swiss-German Internet Relay Chat rooms. Journal of Sociolinguistics 10(4): 481⫺509. Straßner, Erich 1983 Rolle und Ausmaß dialektalen Sprachgebrauchs in den Massenmedien und in der Werbung. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke and Herbert E. Wiegand

41. Language and space in media discourse (eds.), Dialektologie, vol. 2, 1509⫺1525. (Handbücher zur Sprach- und Kommunikationswissenschaft, 1.2.) Berlin/New York: de Gruyter. Straßner, Erich 1986 Dialekt als Ware. Zeitschrift für Dialektologie und Linguistik 53: 310⫺342. Taavitsainen, Irma, Gunnel Melchers and Päivi Pahta (eds.) 1999 Writing in Nonstandard English. Philadelphia: John Benjamins. Thornborrow, Joanna 2002 Power Talk: Language and Interaction in Institutional Discourse. Harlow: Pearson. Wells, John C. 1986 Accents of English, vol. 2: The British Isles. Cambridge: Cambridge University Press.

Jannis Androutsopoulos, Hamburg (Germany)

759

VIII. Structural domains: Methodological problems 42. Areal variation in segmental phonetics and phonology 1. 2. 3. 4. 5. 6. 7.

Introduction Traditional phonetic approaches and methods Instrumental phonetic approaches and phonological interpretation Taxonomic approaches Phonology in structural dialectology Conclusion References

1. Introduction As disciplines at the very core of linguistics, both phonetics and phonology provide general means of describing, analyzing and explaining the sound structure of language, represented as utterances, sentences, words, syllables, sounds and features. Thus, the subject of analysis is any language or language variety, regardless of its status as a standard language, a regional variety or even a specific speaking style of an individual. In this article, we will only briefly touch on general concepts of phonetic and phonological analysis. We will instead concentrate on the phonetic and phonological aspects that may prove relevant to the study of areal variation, i. e., how the notion of space, and variation in space, is reflected in research on phonetics and phonology. For general information about key concepts and methods the reader is referred to up-to-date works by, e.g., on phonetics, Reetz and Jongman (2008), Ladefoged (2004), Ladefoged and Maddieson (1996), or Johnson (2003) and, on phonology, Goldsmith (1996), Gussenhoven and Jacobs (2005) or de Lacy (2007). Ladefoged (2003) provides a thorough introduction to phonetic fieldwork.

2. Traditional phonetic approaches and methods Following some preliminary dialectological work in the first half of the nineteenth century, the debate proper about accurate phonetic reality started with Eduard Sievers’ Grundzüge der Lautphysiologie ‘Fundamentals of the physiology of sounds’, and with its dialectological implementation in Jost Winteler’s account of the Kerenzer vernacular, both published in 1876. At the same time, Georg Wenker was collecting data for his linguistic atlases (between 1876 and 1887). But Wenker gathered his data via the medium of the Latin alphabet (in kurrent script) transcribed by non-experts. Hence, a phonetic

42. Segmental phonetics and phonology interpretation of these data is only possible via a reconstruction of the grapheme⫺phoneme relation and is thus only viable for parts of the sound system. Notwithstanding these concerns, these data can still be successfully used for phonetic and phonological analysis (e.g., Kehrein, Lameli and Nickel 2006 on h-dropping). In contrast to the indirect surveys of Georg Wenker, exact phonetic representation grew increasingly important in the neogrammarian dialect monographs. The data for these monographs were collected in a direct survey by either the researcher himself or a phonetically trained fieldworker and the sounds were recorded phonetically accurately. On the basis of these exact phonetic descriptions, the history of the sounds, i. e., the sound laws for the specific location, were presented. Although a strict interpretation of the hypothesis of the exceptionlessness of such sound laws could not be maintained, the notion of a regular correspondence between the historical and current sounds is still one of the pillars of the investigation of sound and space. All modern Germanic linguistic atlases (except the first volume of the ALA [Beyer and Matzen 1969], which lies more in the Romance tradition) ground their representation in the sounds of the historical sound system. Using example words they show how the sounds in these words are currently realized. To correct and verify the phonetic development, several words containing the respective sound are usually mapped. Thus, Hugo Schuchardt’s emphasis on word history is taken at least partially into consideration. In the linguistic geography of the Romance tradition (Gillie´ron and Edmont 1902⫺1910, ALF; Jaberg and Jud 1928⫺1940, AIS), which opposes the neogrammarian adherence to sound laws, transcriptions of complete words are mapped not individual sounds. Nevertheless, an exact phonetic representation is considered equally important by the Romance and German linguists. This recognition of the importance of the sound values of spoken language first gave rise to the development of phonetic alphabets with which to capture phonetic reality. The different transcription conventions are not addressed here (cf. Richter 2005); common to all phonetic transcription systems is the fact that they symbolically represent the articulation of sounds (cf. Almeida and Braun 1986). The investigator tries to capture the auditory impression exactly. This auditory impression, however, cannot be rendered directly, and the investigator must articulatorily emulate it. Only the articulatory emulation can be represented symbolically in the transcription. Despite investigators’ intensive phonetic training, this representation of a speaker’s articulation, with a detour via the transcriber’s acoustic impression and subsequent rearticulation, can hardly be described as objective (cf. Almeida and Braun 1982; Vieregge 1985). For this reason, the same interviews are often transcribed by multiple investigators. These control transcriptions guarantee intersubjectivity, not objectivity, and indicate differences between investigators that can give rise to false distinctions and isoglosses on the maps of linguistic atlases. For the ALA (Beyer and Matzen 1965; Bothorel-Witz, Philipp and Spindler 1984) and the SDS (Hotzenköcherle 1962⫺1997), which recorded their data at the same time, several interviews were transcribed in parallel by the fieldworkers from both projects. These differences, as well as differences between individual SDS investigators, are described in Hotzenköcherle (1962: 60⫺73), and such potential pitfalls are addressed in the introduction to every linguistic atlas. Despite their shortcomings, transcriptions remain ⫺ up until the most recent surveys ⫺ the basic data for phonetic and phonological analysis in linguistics and especially for the representation of variation across space in linguistic atlases. In the Romance

761

762

VIII. Structural domains: Methodological problems dialect atlas tradition, these transcripts are written directly onto maps, where they are immediately accessible as raw data. In the German tradition, the transcripts are processed so that the sound maps depict only the relevant sounds as symbols. This abstraction from various transcription symbols to a single graphic symbol on the map is itself a reduction and interpretation of data. The transcripts and their conversion into linguistic maps are the basis for broader interpretation and analysis. The distribution of the sound representations on the map is the basis for the traditional historical and extralinguistic interpretation of the distribution of sound in space (cf. Case Study 1: which follows). Precise phonetic data are a prerequisite for phonological analyses (cf. the discussion of Case Studies 2 and 3). They are also the prerequisite for interpretations of phonetic distributions in space by means of dialect distance measurements (e.g., Herrgen and Schmidt 1985, 1989) and for dialectometric analyses (cf. Nerbonne and Heeringa in this volume).

Case study 1: Traditional dialect geography The point of origin for all sound maps in traditional dialectology is a sound from a historical phoneme system. As a tertium comparationis the historical system represents the putative common ancestor of a diachronic development with all the forms shown on the map as its end. In German dialectology, the Middle High German or Old High German vowel system is usually selected as such a common ancestor for vocalic developments; for consonants it is usually a West Germanic consonant system. In Romance linguistics it is normally the Vulgar Latin system. During data collection, an informant is asked to produce a carrier word containing this sound; the exact phonetic realization is transcribed auditorily ⫺ or measured acoustically, as in the Atlas of North American English (Labov, Ash and Boberg 2005). Despite the fact that the historical reference point is a sound system, the current forms of the historical sounds are interpreted phonetically, not phonologically. Only in exceptional cases are phonological questions raised, e.g., when multiple sounds are counterposed in a combination map. However, in this kind of map, only the contrast within part of the system can be shown, not all of the systemic contrasts (cf. section 5.1 on structuralist approaches). Map 42.1 is a typical example of a phonetic map from a modern linguistic atlas. The Sprachatlas der deutschen Schweiz (SDS; Hotzenköcherle 1962⫺1997) has played a formative role for all subsequent linguistic atlases of the German area⫺VALTS (Gabriel 1985⫺), SSA (Steger, Gabriel and Schupp 1989⫺), BSA (Hinderling 1996⫺), MRhSA (Bellmann, Herrgen and Schmidt 1994⫺2002). Map 42.1 depicts the realization of MHG aˆ in German-speaking Switzerland (Hotzenköcherle 1962⫺1997: vol. I, 61). MHG aˆ in Abend ‘evening’ is regarded as a good representative for the development of the aˆ of the Middle High German sound system. However, phonological contrasts such as regionally differentiated mergers with a lengthened MHG a or MHG oˆ cannot be elicited from this map, nor can the realization of MHG aˆ in other words (e.g., Haar ‘hair’, Schaf ‘sheep’, gebracht ‘brought’, gaa ‘go’, laa ‘let’, or staa ‘stand’) whose isoglosses follow slightly different paths be seen. Dialectological maps thus always imply a diachronic view, and the example shows the present-day endpoint of the development of MHG aˆ. There are three possible ways to find an explanation for the distribution presented on this map. First, the present-day realizations can be interpreted phonetically, as in the neogrammarian tradition. Secondly, extralinguistic models can be utilized (cf. articles 13⫺21 in this volume). Finally,

42. Segmental phonetics and phonology

Map 42.1: Symbol map of MHG aˆ Abend ‘evening’ in German-speaking Switzerland (Hotzenköcherle 1962⫺1997: vol. I, 61); see text for an explanation of the symbols

the current forms can be explained phonologically. A “sound physiological” (lautphysiologisch) explanation in the neogrammarian tradition, as developed by Bachmann (1908), would be as follows. Middle High German aˆ is raised to [o/c] in most of the southern parts of the German area; only the most southern corners of Alemannia do not take part in this sound change. This raising of MHG aˆ is often seen as one of the core elements of an areal structuring of Swiss German. The south of German-speaking Switzerland retains MHG aˆ as [a:/a:] (symbols , , ). In the northern parts, represented by open circles (o), MHG aˆ was raised to [c:]. In the northwest and in a northern strip on the border to Germany, in the Zürich uplands, in Thurgau, St. Gallen, Appenzell and Schaffhausen, this rise continued on to [o:], depicted by filled circles (쐌). Finally, the crossed circles (✦) represent a diphthongization ([o:] > [ow4]). This diachronic and physiological development [a: > a: > c: > o: > ow4] is thus implicatively reflected in space. However, the region around Zürich builds an exception to this spatiotemporal sequence. Here we find [A:] (depicted as ). In this physiological/phonetic interpretation, [A:] would be inserted between [a:] and [c:] in the series above. Hence, [A:] can be regarded

763

764

VIII. Structural domains: Methodological problems as a relict form of the Zürich region (Bachmann 1908). Hotzenköcherle (1961), however, has convincingly shown that this interpretation has shortcomings. Bringing other maps to this interpretation has unearthed an additional motivation for this [A:]. Normally, the umlaut of a preserved MHG aˆ is [æ:], for a raised [c:] it is [œ:] and for an [o:] it is [ø:]. So we have a systematic relation between a back vowel and a fronted variant. But in the Zürich region (and only there), the umlaut of [A:] is [œ:], not [˜:]. This opposition /A: ~ œ:/ can only be explained by hypothesizing an originally raised [c:] that has since been lowered again to [A:]. Lexical hypercorrections of MHG oˆ that are lowered to [A:] provide additional evidence (e.g., Trog > [trA:g˚] ‘trough’ and Tor > [tA:r] ‘door’). Coupled with these observations, Map 42.1 makes clear what a phonetic interpretation can achieve, and also that a purely phonetic interpretation has its drawbacks. Based on this and similar misinterpretations and corrections of the phonetic interpretation increasing numbers of structural phonological approaches have been pursued since the 1960s.

3. Instrumental phonetic approaches and phonological interpretation Compared to the surveys of traditional dialectology, recent developments in instrumental phonetics offer a range of new possibilities for the analysis and representation of sounds which have consequences for dialect geography and phonological interpretation. These instrumental phonetic approaches rely not on articulation, which was the basis for auditory transcription, but on the acoustic features of sounds and therefore on more objective data. There is no space here to provide an introduction into instrumental phonetics (available in, e.g., Johnson 2003; Ladefoged 2003; Reetz and Jongman 2008), but because formant analysis has been used several times in the representation of spatial variation (Adank, van Hout and van de Velde 2007; Iivonen 1987, 1994; Jacewicz, Fox and Salmons 2007; Labov, Ash and Boberg 2005), the principle behind this measurement shall be briefly described. During voiced phonation, each sound is characterized by a fundamental frequency and its harmonics, which are integer multiples of this fundamental frequency. While speaking, we rapidly change the size and shape of the vocal tract by altering the shape of the tongue, the rounding or spreading of the lips and opening or closing of the nasal tract. These changes in the vocal tract and the chambers of resonance filter out most of the harmonics and amplify other frequencies, giving each sound its own specific quality. The remnants of this filtering process are frequency bands of higher intensity ⫺ the socalled formants ⫺ which show characteristic values for each sound quality. In order to classify the different vowel qualities, it is sufficient to take just the first two lower formants into account (for an elaborate account of the interpretation of vowel formants, see Moosmüller 2006). Thus, the first formant (F1) roughly corresponds to the articulatory dimension of vowel height, while the second formant (F2) corresponds to the back⫺front dimension of vowel articulation. The first formant of an [i] spoken by male speakers, for example, is around 300 Hertz, the second formant has a value of around 2300 Hertz. Figure 42.1 is a spectrogram of three vowels illustrating the different values for the

42. Segmental phonetics and phonology

Fig. 42.1: Spectrogram with highlighted first and second formant for the vowels [i:], [a:] and [u:]

Fig. 42.2: Spectra with highlighted first and second formant of the vowels [i:], [a:] and [u:] from Figure 42.1

formants; the frequencies of the first two formants are highlighted. Figure 42.2 shows the corresponding spectra, which allow more exact measurements than the spectrogram. Differences between languages/dialects arise because of different formant structures in the vowels. Measuring formants in the sound spectrum therefore allows the sound quality of different varieties to be surveyed and compared on the basis of the acoustic features. Early applications of this technique to a comparison of regional variation in the Germanic dialects area include Göschel (1973), Barry (1986) (a comparison of two

765

766

VIII. Structural domains: Methodological problems

Fig. 42.3: Depiction of formants by Iivonen (1994: 222, 226)

Low German dialects), Herrgen and Schmidt (1986) (vowel centralization in West Middle German dialects), Lauf (1993) (diphthongization), Schouten and Peters (1996) (diphthongization) and Iivonen (1983, 1987, 1994). Present-day research takes different distance measurements and gender normalizations into account. Fabricius (2007) discusses sound change in Received Pronunciation, Jacewicz, Fox and Salmons (2007) analyze regional varieties of American English and Adank, van Hout and van de Velde (2007) examine regional variation in standard Dutch. Figure 42.3 shows the comparison of acoustic vowel systems for two speakers of Standard German. In this figure, F1 is depicted on the y-axis with the higher values at the bottom and F2 on the x-axis with the higher values to the left. The advantage of this mirrored coordinate system is that it corresponds roughly to the representation of the articulatory vowel space. The differences between the two systems are due to regional variation in the realizations of Standard German vowels in Hamburg and Zürich, respectively. Notice, e.g., the much more obvious tense⫺lax relationship between the vowel pairs [i:] ~ [=] and [u:] ~ [w] for a speaker from Hamburg as compared to the Zürich speaker. For the axes, the perceptually based bark scale is used, which allows conclusions

42. Segmental phonetics and phonology to be drawn about the perceptual distance between vowel realizations (Iivonen 1994). Iivonen has published several such comparisons of Standard German vowel systems, allowing him to sketch a rough spatial structure of Standard German pronunciation. For a comprehensive analysis of a whole linguistic area, however, the costs remain too high. This analysis only became a reality with The Atlas of North American English (Labov, Ash and Boberg 2005), which consistently takes acoustic measurements into account. Formant measurements play an important role in analyzing sub-phonemic sound changes in progress. Labov, Ash and Boberg (2005: 49⫺53), for instance, give a detailed acoustic account of the merger in progress between the word classes /ohr/ (porch) and /chr/ (sort, cort) in American English.

Case Study 2: The Atlas o North American English With its phonetic and phonological interpretation of acoustic data in space, the seminal Atlas of North American English (ANAE; Labov, Ash and Boberg 2005) is groundbreaking in many ways. By utilizing acoustic measurements on a large scale it becomes possible to describe patterns of regional variation in phonetics and phonology for the dialects of North America. For 297 cities in the United States and Canada, language data was recorded in telephone interviews to gather information on all of the phonemes and allophones of English. The innovative maps of this atlas reveal not only aspects of the phonological system but also the phonetic dynamics of sound changes in progress. Data analysis was done both auditorily and acoustically. Using acoustic measurements of formant structures makes it possible to trace vocalic sound changes in progress, which are barely, if at all, analyzable using auditory transcriptions alone (but see Künzel 2001 on the disadvantages of analyzing telephone recordings acoustically). This then allows a reconstruction of the progress of a sound change across geographical space. Labov, Ash and Boberg apply very exact methods of measurement (well-defined placement of the measurements points and LPC [“linear predicted coding”] to obtain the formant values) plus the normalization of the frequency values (using Nearey’s 1978 algorithm) in order to make speech data from informants of different ages and sexes comparable. The quantification of variability is symbolized using various shades of the same color. The maps detail the regional distribution of phonetic/phonological processes, which affect the realization of phonemes and their allophones (e.g., /r/-vocalization, fronting of /uw/ [e.g., in boot], lowering of /o/ [hot, fox], raising of /æ/ [bad ], as well as relationships between phonemes). The latter include the various kinds of phoneme mergers, distinguishing between “nearly completed mergers” (e.g., the merger of /ohr/ and /chr/ in words like hoarse and horse) and “mergers in progress”. The status of a merger is determined both auditorily and on the basis of acoustic measurements of the corresponding vowels in the informant’s vowel system. Different symbols on the map describe the status of a merger and reveal core areas and transition areas. In most cases, phonetic/phonological processes form part of more far-reaching changes in the overall system by virtue of being triggers for, or reactions to, other processes. Such chain shifts often occur in a vowel system, and were first described in Bavarian dialects by Pfalz (1918) and among German dialects by Wiesinger (1970). With regard to chain shifts, regional variation arises when related dialects exhibit different stages of the same chain shift. The mapping of the different stages of a chain shift also allows us to capture the progression of this sound change in time. One of the best investigated

767

768

VIII. Structural domains: Methodological problems

Map 42.2: Phonetic processes involved in chain shifts in the dialects of North America (from Labov, Ash and Boberg 2005: 123)

chain shifts is the so-called “Northern Cities Shift” taking place in the northern inland region of the United States around the major cities of Chicago, Detroit, Rochester, Syracuse and Buffalo (see Labov, Yaeger and Steiner 1972; Labov 1991). In Labov, Ash and Boberg (2005), the stages of this chain shift are analyzed acoustically and mapped. It becomes clear that the temporally consecutive stages of the shift have their own regional spread. In stage (1), the phoneme /æ/ in words like bad, laugh is raised to a more closed front vowel, which can lead in part to diphthongization (“breaking”). For an instance of /æ/ to be classified as raised, the normalized value for F1 has to be lower than 700 Hertz. This initial process in the chain shift is mostly found in the area designated “Inland North” (red isogloss on Map 42.2). In stage (2) of the shift, the now empty position of /æ/ attracts the neighboring vowel position /o ~ ah/ and words like top, bottle are being fronted to occupy the now empty position of the laugh-vowel. In step (3), the vowel /oh/ in words like caught, dog is lowered and moves into the empty position of /o ~ ah/. In stage (4), due to the pressure of the raised /æ/, short /e/ in words like flesh, ten undergoes backing. This, in turn, makes the /v/ phoneme (lunch, tough) move to the now free position of /oh/ (former slot for dog, caught). The consistent application of acoustic measurements allows the regional spread of the Northern Cities Shift to be systematically captured. Labov, Ash and Boberg (2005: 121⫺124) use the formula ED ⫽ F2 (e) ⫺ F2 (o), which utilizes the second formant to measure the relation between backing of /e/ and fronting of /o/. Since backing is shown by a decrease and fronting by an increase in F2, and since these two processes characterize the Northern Cities Shift decisively, this chain shift is most advanced in areas where the difference

42. Segmental phonetics and phonology between F2 (e) and F2 (o) is the lowest. If, on the other hand, ED is high, the chain shift has not advanced very far. Map 42.2 details several aspects of the Northern Cities Shift, as well as other chain shifts in the dialects of the United States. With regard to the former, the round dark blue symbols and the dark blue isogloss mark out an area around the cities in the north where the chain shift has proceed the farthest, based upon the acoustic formula F2 (e) ⫺ F2 (o) < 375 Hertz; this frequency value has proven useful in calculating the advancement of the chain shift. This map shows also the isogloss for the raising of /æ/ (red isogloss), i. e., the initial stage of the Northern Cities Shift. It is noticeable that this encompasses a larger area than the blue isogloss. This is a typical instance of the nesting of isoglosses, which can be utilized to show the temporal dynamics of a chain shift within an area. All the processes described and shown on the map are clearly sub-phonemic and have effects on the phonological system. In contrast to purely phonemic analyzes, like Moulton’s (1961) study of the “Ostschweizerische Vokalspaltung”, these maps draw upon a phonetically and process-based mapping. Instead of comparing (parts of) phoneme systems, this kind of mapping makes the gradual spread of phonetic processes visible in geographical space and allows subsequent interpretations about phonological restructuring and the progress of sound change. While the Atlas of North American English relies mainly on acoustic measurements to detect regional patterns in the phonetics of dialects, sound perception is also moving more and more into focus in dialectology. Plichta and Preston (2005) present impressive results from a perception experiment on the ongoing process of monophthongization of /ay/ (as in tide) in the dialects of the United States, which is most prominent in the dialects of the South. They prepared a seven-step continuum of synthetic stimuli (F1 and F2 synthesis) from clearly diphthongal [a= ] to clearly monophthongal [a:]. Listeners then had to match the randomly presented stimuli to a location on a map of a North⫺ South continuum. They found that listeners were able to locate the stimuli correctly by transferring the diphthong⫺monophthong continuum perfectly to the North⫺South continuum on the map of the United States.

4. Taxonomic approaches Besides traditional phonetic and phonological analysis, the dialectometric approach should also be addressed. A dialectometric analysis employs numerical classification methods to detect basic patterns in large data sets, such as those used in data mining for market analysis. The data for these linguistic analyses are those drawn from the linguistic atlases. While phonetic analyses focus on individual or closely related phenomena, phonology analyzes the relations within the system. However, the selection of phenomena is often restricted to a subsystem. A taxonomic approach, on the other hand, tries to capture differences and similarities between linguistic systems quantitatively, normally without relating them to a phonological or structural analysis. The point of departure for the dialectometric approach is the insight that traditional linguistic maps, with their many different isoglosses, can hardly be overlooked, so that it is difficult to structure the space (cf. the linguistic, chiefly phonological maps Wiesinger [1983a] drew in his overview of the German-speaking area). Moreover, structuring the space via a selec-

769

770

VIII. Structural domains: Methodological problems tion of isoglosses often seems arbitrary, and individual divergent isoglosses can hardly be integrated into a synopsis. Besides, a structuring on the basis of individual isoglosses gives the impression that dialect area and dialect continuum are divorced form one another. One either concentrates on the area, and therefore looks at the isoglosses that border the area, or one concentrates on the continuum, which means the dialect area disappears. An alternative is to look at the differences in space that are not prestructured by linguistic premises underlying the selection of isoglosses, but which are found in a huge set of data. To attain such an overview, the similarities between all the dialects of an area are calculated. The measurement of similarity/dissimilarity between all local dialects then results in an area structure that is not biased by singular events. Indeed, it is the models of numeric classification, the taxonomy and the measure of similarity which influence the results. In the following we examine projects that focus on phonetic aspects. However, studies with a morphological or syntactic (or even lexical) focus have been put forward (cf. e.g., Goebl 1982, 1984 and 1993 for atlases of different Roman languages), and it is also possible to combine different phenomena. Many algorithms (e.g., Goebl 1982 passim; Herrgen and Schmidt 1985 and 1989; Kessler 1995) have been proposed to define the phonetic similarity of dialects. Based on Herrgen and Schmidt’s (1989) algorithm, early maps were drawn to show the degree of similarity between numerous localities in the West Middle German dialect area on the one hand and their similarity to the German standard language on the other (see Bellmann, Herrgen and Schmidt 1994⫺2002: Map 314/1⫺2). A comparison of different measurement tools is given in Heeringa (2004; see also Nerbonne and Heeringa in this volume). He concludes (2004: 184⫺186) that binary segment differences outperform more refined feature-based measures of segment difference. As a consequence, he uses the Levenshtein distance algorithm for analyses of different atlases that are presented here in summary. Levenshtein distance was originally developed to compare mathematical strings and has been transformed to suit acoustic data by calculating the costs of transforming one sequence of sounds into another. It comprises the following operations: (a) insertion of a single sound; (b) deletion of a single sound; and (c) substitution of one sound for another. For every change necessary to transform a word into another the Levenshtein distance rises. When the differences between all the dialects in an area are calculated, the distances are grouped into classes using multidimensional scaling and cluster analysis, thus generating classes of greater similarity. Representing these classes in space produces clear dialect borders (Map 42.3). Similar techniques have been used for different regions, based on the appropriate linguistic atlases ⫺ cf. Heeringa (2004) for Norway; Nerbonne and Siedle (2006) for German dialects; Nerbonne (to appear) for the Middle and South Atlantic States of the USA; Shackleton (2007) for the traditional English Dialects; Wieling, Heeringa and Nerbonne (2007) for Belgium and Holland. Map 42.3, from Heeringa (2004: 231), suggests that we have clear categorical dialect borders, similar to those generated within the traditional approaches on the basis of the interpretation of a handful of isoglosses, but resting on a broader and therefore more reliable database. These borders are, however, not per se more genuine. They are a product of the data selection, the statistical techniques and their interpretation. Another clustering technique, another selection of the clusters, another measurement of similarity, another numerical classification or another scaling technique ⫺ all these factor can result

42. Segmental phonetics and phonology

Map 42.3: Dialectometric structuring of Dutch by means of a cluster analysis (Heeringa 2004: 231)

in different regional distributions. So, these structures are based on more data than the classical approaches and in many cases they support the structures of the classical analyses, however, they can recognize additional or more deeply nested patterns.

5.

Phonology in structural dialectology

5.1. Structuralist phonology Structuralist approaches in dialectology differentiate themselves from purely historical and phonetic perspectives such as those put forward by the Neogrammarians or the Marburg School. According to Moulton (1968: 456), the splitting up of a phonetic continuum into regions with clear-cut borders, as presented on dialect maps for instance, only becomes possible with the application of a phonological perspective. In developing his axiomatic structuralist phonology, Trubetzkoy (1931) formulated early thoughts on the application of phonological principles to both language and dialect geography. By distancing himself from the neogrammarian diachronic sound-change theory, he transferred his principles of synchronic phonology to the areal dimension, developing phonological descriptions for individual lects in geographical space, which are subsequently compared with each other in a separate step. Three areas of difference in the sound system can be identified. Phonetic differences concern the phonetic realization of a phoneme ⫺ either context-free or context-sensitive (“kombinatorisch”). Phonological differ-

771

772

VIII. Structural domains: Methodological problems ences refer either to the sound inventory or to functional differences in the phonemic systems. While these general phonological principles can be applied to the comparison of any languages or dialects, Trubetzkoy’s third type of difference relates explicitly to the comparison of dialects ⫺ for so-called etymological differences, the “sound history” of words with a common origin is taken into account. Uriel Weinreich (1954) introduced an influential formalization to capture differences between phonological inventories and phonological distributions with his notion of the “diasystem”. This basic structuralist principle was taken up rather late in German dialectology. The concept of the diasystem has been and still is (implicitly or explicitly) used in various dialectological investigations. Under the label of strukturelle Sprachgeographie ‘structural language geography’, the comparative/diatopical method of Goossens (1969) has also been influential. Here, for various (ideally all) localities of a dialect region, the phonemic systems and their distributional characteristics are determined and, in a subsequent step, compared with each other. To explain the phonological differences thus analyzed, diachronic processes relating to older language stages have been deployed (e.g., split or merger of a phoneme, etc.). A thorough example of this method is the description of the vowel system of the Swiss German dialects by Haas (1978; cf. Haas in this volume). Within this perspective, dialect geography is a branch of diachronic linguistics which employs synchronic/structuralist methods. This becomes evident in the fact that a historical system has often been used as a methodological reference point for the description of a present-day dialect, as in neogrammarian phonetic analyses. As already mentioned, in German dialectology, “classical” Middle High German and West Germanic (Keller 1961), and for Romance scholars, late-Latin forms, have been chosen as reference points; in the dialectology of English, word classes have been set up based on presentday English (Wells 1982). After preliminary work (e.g., by Pfalz 1918, 1936) on the systematic nature of vowel systems and their systematic changes in related dialects, groundbreaking research in this paradigm was done in the 1960s and 1970s (e.g., Moulton 1961; Goossens 1969; Wiesinger 1970; Werlen 1977; Haas 1978; for a systematic comparison of the phonological vowel systems of German dialects, see Wiesinger 1983b). Since that time, basic structuralist concepts like minimal pairs, phonological oppositions, distributions, neutralizations and allophony have been widely applied to obtain phonological descriptions of vowel and consonant inventories as closed systems. With this method, areal variation becomes visible when the phonemic systems of several dialects are presented in the same manner. For the vowel system, then, these phonological differences may apply to the number of phonemes, the structure of the inventory (e.g., degrees of openness, presence or absence of rounded front vowels), phonemic vowel length versus a tense⫺lax system, the implementation of phoneme split or mergers, etc. A classic example of the application of structuralist principles and methods is a vowel split in Swiss German dialects (the “nordostweizerische Vokalspaltung”, see Moulton 1961), which demonstrates how the internal dynamics of a phonological vowel system result in a sound change (“Lautwandel durch innere Kausalität”). The Middle High German system, the putative historical predecessor, is asymmetrical with regard to the phonologically relevant degrees of opening of the short vowels: while there are four degrees of openness for the unrounded front vowels, there are only three degrees for the rounded front vowels and the back vowels (see Figure 42.4a). According to Moulton, this asymmetry creates pressure within the system, which leads to subsequent restructuring in the later stages of phonological development. It is

42. Segmental phonetics and phonology

773

Original MHG system i e ε æ

y ø

u o a

Fig. 42.4a: MHG short vowel system Solution 1 i e æ

y ø

Solution 2 u o a

i e ε æ

y ø œ

u o c a

Fig. 42.4b: Short vowel systems in Swiss German dialects

assumed that phonological systems strive for symmetry and, to regain this, two remedies are conceivable in this case, both of which are attested in neighboring northeastern Swiss German dialects. In the first solution, the degrees of openness are reduced from four to three when MHG /e/ merges with /æ/ (cf. Figure 42.4b). In the remaining dialects, the phoneme split of /ø/ and /o/ leads to the emergence of the new phonemes /œ/ und /c/ respectively, which then occupy the empty slots of the halfopen rounded vowels (see the right-hand side of Figure 42.4b). In terms of symmetry, the number of rounded and unrounded vowels is thus in balance. This sound change has been initiated by the drag of the empty slots in an asymmetrical inventory and preceded by the emergence of contextual allophones. This example demonstrates how the notions of phoneme merger and phoneme split are used to capture structural differences in the phonological inventories of neighboring dialects. Aside from its temporal dimension, sound change also implies the dimension of gradual progression in space (i. e., speech community), which may lead to a situation where all instances of a sound in one region are affected by a change, with no exceptions, whereas in other regions the same sound is only affected in specific contextual or prosodic contexts. This contextually bound progression of a sound change in geographical space then leads to a so-called “graded landscape” (Staffellandschaft), the systematic structure of which can also be captured in phonological terms. Map 42.4 shows the distribution of the phoneme /s / in a dialect cluster in the Limburg area (Netherlands/ Belgium; following Goossens 1969), where the consonant is permitted in differing contexts (e.g., initially before a vowel, initially before certain consonants, in a word-final position, etc.). Whereas in the westernmost region, (1), /s / is only permitted in foreign words, the consonant is allowed in an increasing number of contexts, the further east one goes. In region (6), then, /s / is possible in nearly all positions within a word. Thus, the regional distribution of the phoneme /s / can be described in terms of phonological contexts, and each combination of contexts represents a specific region in this dialect cluster. In the dialectological literature one finds several instances of a regional gradation of sound changes, which are dependent upon the phonological context. For example, there is the regionally graded application of the Second Germanic consonant shift (Szczepa-

774

VIII. Structural domains: Methodological problems

Map 42.4: Distribution of /s/ in a dialect cluster in Limburg (Goossens 1969)

niak 2007) or the palatalization of /s/ before plosives (/st/ >/st/) in Luxembourgish dialects (Gilles 1999). In recent dialect atlases, systematic maps which illustrate the regional patterns of truly phonological aspects rather than simply the regional variation of a single specific word can also be found. The Südwestdeutscher Sprachatlas (Steger, Gabriel and Schupp 1989⫺) for instance, contains several maps to show the structure of the subsystem of the plosives in southwestern German dialects, in such a way that it becomes possible to distinguish core areas and transition zones.

5.2. Recent developments in phonology With the advent of generative phonology and publication of the influential Sound Patterns of English (Chomsky and Halle 1968), the distinction between underlying form, surface form and phonetic implementation became manifest in phonological thinking. As described in Barbiers (in this volume), phonological rules were introduced to express the derivation of the different forms at these phonological levels. With regard to differences between the phonological systems of related dialects, one way to describe these differences is “rule ordering”: differences between these dialects may be the result of different orderings of phonological rules. In the following example, the difference between the Standard German (and southern) surface form [lan] ‘long’ and the northern German variant [lank] is explained in terms of different rule orderings (following Ramers 2001: 72⫺73; see Figure 42.5). In Standard German, the rules “final g-deletion” and

42. Segmental phonetics and phonology

775

Standard German

Northern German

/lang/ B /lang/ B /lan/ B [lan]

/lang/ B /lang/ B /lank/ B [lank]

nasal assimilation final g-deletion final devoicing (not applicable)

nasal assimilation final devoicing final g-deletion (not applicable)

Fig. 42.5: Derivation of an underlying form in two varieties of German and application of different rule orderings (following Ramers 2001: 72⫺73)

“final devoicing” stand in “bleeding order”, and hence the latter phonological rule cannot be applied and this results in the correct form [lan]. In northern German, however, “final devoicing” is ordered before “final g-deletion” and this leads to the correct northern variant [lank]. Thus, in this model regional differences are expressed in terms of the different orderings of rules operating on one common underlying form. A similar analysis has been proposed by Rein (1974), who describes l-vocalization in Bavarian dialects with the help of several phonological rules. The characteristics of the different Bavarian regions, then, are formulated on the bases of various rule orderings (for a general criticism of the generative concept of rule ordering, see Hooper 1976). In the last decades, phonological research has increasingly taken dialect data into account (see, e.g., Repetti 2000 on the dialects of Italy or Hinskens, van Hout and Wetzels 1997). However, as in structuralist phonology, from this perspective dialects are seen as clearly delimitable, self-contained varieties; this fails to systematically take into account the fact that dialects are virtually always characterized by internal variation, the origins of which lie in language contact with neighboring dialects (horizontal dimension) and ⫺ these days perhaps more importantly ⫺ with overarching regional varieties and the standard language (vertical dimension). To circumvent the traditional method of comparing the phonological systems of idealized self-contained dialects, variation has to be systematically taken into account by developing more complex, multi-layered models for the phonological interrelationship between regional varieties (see Auer 1990, 1995; Auer, Hinskens and Kerswill 2005: 19 for a general critique). Ongoing developments in phonology ⫺ Autosegmental Phonology (Goldsmith 1990), Lexical Phonology (Mohanan 1986), Natural Phonology (Donegan and Stampe 1979; Dressler 1985), Optimality Theory (Prince and Smolensky 2004) and Stochastic Optimality Theory (Boersma 1998) ⫺ tend to strengthen a dynamic perspective in phonology by bringing internal variation more and more into focus. Certain elements of these phonological theories have been applied in so called “standard/dialect phonology” to model the dynamic relationship between small-scale varieties like dialects and large-scale varieties like the standard languages. In this perspective the whole linguistic repertoire of the individual or the speech community is at the center of attention, not discrete, selfcontained varieties. One radical proposal in this context is for a “bi-dialectal phonology” (Rennison 1981), in which a common underlying form for both dialect and standard language is posited. On the basis of the standard/dialect repertoire found in the city of Salzburg (Austria), Rennison shows how filtering rules, standard processes and dialect

776

VIII. Structural domains: Methodological problems processes ensure that either dialect or standard forms are generated out of a single common deep structure. However, to be able to generate strongly diverging dialect and standard language surface forms, rather abstract underlying forms are required. This drawback is avoided if one stipulates a repertoire with two separate phonological systems, for dialect and standard language respectively, interconnected by a system of rules and processes (Auer 1990, 1995; Dressler and Wodak 1982; Gilles 1999; Moosmüller 1987, 1988). This, then, allows a more adequate description of language contact between small-scale and large-scale varieties. One constitutive factor is the modular and hierarchical composition of phonology, a concept borrowed from Lexical Phonology. Here, surface forms are derived from an underlying form via various steps. At a prelexical level, the possible phonemic contrasts and permitted phonotactic structures are specified. The lexical level is organized by phonological rules, which are sensitive to morphological operations. They model the phonological consequences of morphological operations like affixation and stem alternation (e.g., final devoicing, velar softening, resyllabification, umlaut, etc.). This level in itself is also organized hierarchically by assuming different sublevels (“strata”), so as to be able to formulate certain rule orderings.

Fig. 42.6: Two-dimensional model of phonological variation between a dialect and a standard variety (from Auer 1995: 25)

42. Segmental phonetics and phonology Finally, on the postlexical level, the phonetic implementation of the surface forms takes place. For postlexical rules, morphological boundaries are no longer “visible”. Here we find speech-style related and sandhi rules, which in most cases follow a lenition teleology, i. e., they are production oriented and strive for articulatory simplification (e.g., certain kinds of assimilation, monophthongization, “flapping”, segment reduction and deletion; for an extensive list of prelexical/redundancy, lexical and postlexical rules, see Dressler 1985; Mohanan 1986; Auer 1990). This hierarchical organization of phonology is utilized by Auer (1990, 1995) to model the dynamic nature of the relationship between dialect, unmarked intermediate varieties and a standard variety and to capture variability within a language repertoire (Figure 42.6). Contact between the varieties (i. e., along the horizontal dimension of focusing) is possible on the prelexical (via redundancy rules) and lexical level (via lexical rules) of the phonological hierarchy (i. e., along the vertical dimension of lexicalization/grammaticalization) by virtue of different rule types. Postlexical rules, on the other hand, are not sensitive to focus on a specific variety; they are language universal in the sense that they reflect general natural/phonological processes. Hence, a dialect may differ from other varieties (either the standard language or neighboring varieties) on the prelexical and/or lexical level. The fact that the varieties share certain phonological aspects is accounted for in terms of common rules (the “unmarked” region in the center of the model). Unique traits of dialect or standard language are expressed in terms of dialect or standard rules, respectively. Rules of correspondence are a special type of rule used to describe certain “trends” in the relationships between lexical structures that have to be learned item by item. This complex model of phonological contact between adjacent varieties thus describes, in a theoretically grounded way, how a certain pronunciation is oriented to either the dialect, the unmarked, or the standard pole of a given repertoire. This model should also be seen as an attempt to combine or reconcile traditional dialectology with modern phonological theory.

Case Study 3: Optimality Theory in dialectology In the realm of recent advancements in phonological theory, Optimality Theory (OT; Prince and Smolensky 2004) offers a fresh view on the representation of the phonological system and phonological variation. In OT, underlying representations are transferred to surface forms without any intermediate levels or strata, solely using the evaluation of various possible candidates through ranked constraints only. The core concepts of OT, i. e., the violability and re-ranking of constraints, make it possible to capture the differences between languages and/or dialects and to derivate variability from within the grammar itself (Anttila 1997, 2002, 2007). With regard to dialectology, several studies have been published in the context of OT: Auer (2005) gives an OT account of vowel epenthesis in German dialects, Alber 2001 tackles glottal stop epenthesis and dissimilation in two varieties of German, van Oostendorp (2005) the interaction of empty vowels in Dutch dialects, Noske (2005) variation between Northern and Southern Dutch and Smith (2003) syllabification differences between two related dialects of Campidanian Sardinian. To exemplify the application of OT to dialectology, Herrgen’s (2005) analysis on final st-clusters in German dialects will be presented in more detail. These clusters can be found in simple words (like Standard German Brust ‘breast’) and inflectional suffixes (hast ‘you have [2sg]’) and are, depending on the dialect, subject to obstruent dissimila-

777

778

VIII. Structural domains: Methodological problems Tab. 42.1: OT tableaux to demonstrate the interaction of obstruent dissimilation and deletion in the Alsatian dialect Alsatian Input: /hast/ 2sg.pres ‘you have’

*ComplexIS

DissObstr

*!

*

hast has 췽 hast

*ComplexCod *

*! *!

* *

has Input: /brust/ ‘Breast’

Ident-IO

*

**

brust

*!

brus

*!

췽 brust

* * *

brus

*

**!

tion (st > st: hast > hast) and/or t-deletion (Brust > Brus; hast > has ). The combination of these two processes lead to four input forms for the evaluation by the OT constraint mechanism and four constraints are needed to evaluate the candidates: Ident-IO (“Input and output are identical”; the faithfulness constraint), *ComplexCod (“Syllable codas are simple”), *Complex IS (“Inflectional suffixes are simple”) and DissObstr (“Obstruent clusters are dissimilated”). As an example, the constraint ranking for the Alsatian dialect in Table 42.1 shows that the (correct) candidates has and Brust are evaluated best. In this dialect, among other aspects, the constraint prohibiting complex suffixes (i. e., *ComplexIS) favors the inflected word forms has and has over hast and hast, whereas the constraint DissObstr favors the dissimilated clusters st in the word form Brust over Brust. Other dialects exhibit different constraint rankings and thus reveal their different phonologies (Table 42.2). For instance, while on the one hand the constraint DissObstr is ranked very highly in the neighboring dialects of Alsatian and Swabian (leading to forms like hast and Brust), it is, on the other hand, ranked lowly in the remaining three dialects, which captures the fact that underlying st clusters remain undissimilated (leading to the forms hast and Brust). Thus, in OT, phonological differences between dialects are expressed in terms of different constraint rankings. With regard to the structural difference between a standard language and a dialect, Butskhrikidze and van de Weijer (2001: 49) make an important and far-reaching claim. While standard languages tend to rank faithfulness constraints highly, i. e., constraints which require that input forms and output forms be identical (e.g., Ident-IO), and markedness constraints (constraints modifying the output form)

Tab. 42.2: OT constraint rankings for five Germanic dialects Alsatian Swabian East Franconian Moselle Franconian Ripuarian

* ComplexIS DissObstr Ident-IO * ComplexIS * ComplexCod

>> DissObstr >> Ident-IO >> * ComplexIS >> Ident-IO >> * ComplexIS

>> Ident-IO >> * ComplexIS >> * ComplexCod >> * ComplexCod >> Ident-IO

>> * ComplexCod >> * ComplexCod >> DissObstr >> DissObstr >> DissObstr

42. Segmental phonetics and phonology lower, it is the other way round for dialects. In our Case Study, it becomes obvious that for most dialects the only relevant faithfulness constraint, Ident-IO, is ranked lowly, while the remaining markedness constraints are ranked highly. Interestingly, in the EastFranconian dialect the faithfulness constraint Ident-IO achieves a top ranking, and it is precisely this dialect that shows the least divergence from the standard language, while the other dialects, with a lower ranked Ident-IO faithfulness constraint, are characterized by greater structural distance from the standard language.

6. Conclusion This overview has pointed out the core role phonetics and phonology have played in the presentation and interpretation of linguistic variation in space since the last quarter of the nineteenth century. For concrete interpretations of variation in space, phonetics and phonology go hand in hand and stimulate each other. The traditional method of relating present-day regional variation to a (presumably) common historical reference system, which has been the guiding principle of many dialect atlases, is increasingly being challenged by purely synchronically oriented phonetic or phonological approaches. Nevertheless, for nearly all of the studies cited here, it has become clear that sound changes are still an important means of explaining certain patterns of regional variation. The increased availability of easy-to-use and reliable instrumental phonetic techniques is not only helping to overcome or correct for the shortcomings of auditory transcriptions, it also allows for the analysis of large amounts of acoustic data. Furthermore, it brings phonetic detail back to the center of analysis, viz., by the use of fine-grained phonetic measurements to disclose the pattern of regional spread of a sound change in progress. For dialectology, then, a strict distinction between (substance-based) phonetics and (form-oriented) phonology is not an issue. Quite the contrary, analysis shows the smooth transition from phonetic convergence to a switch in the phonological system. The huge amount of data that has been collected for linguistic atlases is also a prerequisite for statistical, taxometric approaches, which complement traditional phonetic and phonological explanations. With regard to phonology, though, it still needs to be pointed out that to date phonological research in the strict sense still shows only limited interest in dialectological issues. Data from regional varieties are often described only for the sake of applying or improving a specific theory or formal account. To overcome this shortfall, increased consideration of the social and cognitive aspects of phonological description and explanation seems a promising avenue. To our knowledge, so far it is only in so-called standard/ dialect phonology that aspects of the different sociolinguistic status of varieties are systematically incorporated into the model of phonological variation. A further reconciliation of dialectology and phonology under the umbrella of sociophonetic/sociophonological approaches is therefore desirable.

779

780

VIII. Structural domains: Methodological problems

7. Reerences Adank, Patti, Roeland van Hout and Hans van de Velde 2007 An acoustic description of the vowels of northern and southern standard Dutch II: Regional varieties. Journal of the Acoustic Society of America 121(2): 1130⫺1141. Alber, Birgit 2001 Regional variation at edges: Glottal stop epenthesis and dissimilation in Standard and southern varieties of German. Zeitschrift für Sprachwissenschaft 20(1): 3⫺41. Almeida, Antonio and Angelika Braun 1982 Probleme der phonetischen Transkription. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke and Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 1, 597⫺615. (Handbücher zur Sprach- und Kommunikationswissenschaft, 1.1.) Berlin/New York: de Gruyter. Almeida, Antonio and Angelika Braun 1986 “Richtig” und “Falsch” in phonetischer Transkription. Vorschläge zum Vergleich von Transkriptionen mit Beispielen aus deutschen Dialekten. Zeitschrift für Dialektologie und Linguistik 53: 158⫺171. Anttila, Arto 1997 Deriving variation from grammar. In: Hinskens, van Hout and Wetzels (eds.), 35⫺68. Anttila, Arto 2002 Variation and phonological theory. In: Jack Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.), Handbook of Language Variation and Change, 206⫺243. Oxford: Blackwell. Anttila, Arto 2007 Variation and optionality. In: de Lacy (ed.): 519⫺536. Auer, Peter 1990 Phonologie der Alltagssprache. Eine Untersuchung zur Standard/Dialekt-Variation am Beispiel der Konstanzer Stadtsprache. (Studia Linguistica Germanica 28.) Berlin/New York: de Gruyter. Auer, Peter 1995 Modelling phonological variation in German. In: Iwar Werlen (ed.), Verbale Kommunikation in der Stadt, 9⫺37. Tübingen: Narr. Auer, Peter 2005 Areale Variation und phonologische Theorie: Überlegungen am Beispiel der mitteldeutschen “Vokal-Epenthese”. In: Gerhard Stickel (ed.), Varietäten des Deutschen, 46⫺ 87. (Akten der IdS-Jahresstagung 1996.) Berlin: de Gruyter. Auer, Peter, Frans Hinskens and Paul Kerswill 2005 Dialect Change: The Convergence and Divergence of Dialects in Contemporary Europe. Cambridge: Cambridge University Press. Bachmann, Albert 1908 Sprachen und Mundarten in der Schweiz: I. Deutsch. Sprachgrenze und geschichtliche Entwicklung, Gebrauch des Deutschen im Innern, Mundart und Schriftsprache, Charakter und Gliederung der Mundart, Bibliographie. In: Geographisches Lexikon der Schweiz, vol. 5, 58⫺76. Neuenbug: Gebrüder Attinger. Barry, William 1986 Vokalqualitätsunterschiede in Dithmarschen und Angeln. Eine kontrastive Vokaluntersuchung zweier niederdeutscher Dialekte, mit instrumentalphonetischer Unterstützung. Zeitschrift für Dialektologie und Linguistik 53: 145⫺157. Bellmann, Günter, Joachim Herrgen and Jürgen Erich Schmidt 1994⫺2002 Mittelrheinischer Sprachatlas. 5 vols. Tübingen: Niemeyer.

42. Segmental phonetics and phonology Beyer, Ernest and Raymond Matzen 1965 Atlas linguistique et ethnographique de l’Alsace, vol. 1. (Atlas linguistique de la France par re´gion 22.) Paris: CNRS. Boersma, Paul 1998 Functional phonology: Formalizing the interactions between articulatory and perceptual drives. Unpublished Ph.D. dissertation, University of Amsterdam. Bothorel-Witz, Arlette, Marthe Philipp and Sylviane Spindler 1984 Atlas linguistique et ethnographique de l’Alsace, vol. 2. (Atlas linguistique de la France par re´gion 22.) Paris: CNRS. Butskhrikidze, Marike and Jeroen M. van de Weijer 2001 On de-affrication in Modern Georgian. In: A. van der Wouden and H. Broekhuis (eds.), Linguistics in the Netherlands, 41⫺51. Amsterdam/Philadelphia: Benjamins. Chomsky, Noam and Morris Halle 1968 The Sound Pattern of English. New York: Harper Row. de Lacy, Paul 2007 The Cambridge Handbook of Phonology. Cambridge: Cambridge University Press. Donegan, Patricia and David Stampe 1979 The study of natural phonology. In: D. A. Dinnsen (ed.), Approaches to Phonological Theory, 126⫺176. Bloomington: Indiana University Press. Dressler, Wolfgang 1985 Morphonology. Ann Arbor: Karoma. Dressler, WoIfgang and Ruth Wodak 1982 Socio-phonological methods in the study of sociolinguistic variation in Viennese German. Language in Society 11: 339⫺370. Fabricius, Anne 2007 Vowel formants and angle measurements in diachronic sociophonetic studies: Foot-fronting in RP. In: Jürgen Trouvain and William J. Barry (eds.), Proceedings of the ICPhS XVI, 1477⫺1480. Saarbrücken: Universität des Saarlandes. Gabriel, Eugen (ed.) 1985⫺ Vorarlberger Sprachatlas mit Einschluß des Fürstentums Liechtenstein, Wettirols und des Allgäus. Bregenz: Vorarlberger Landesbibliothek. Gilles, Peter 1999 Dialektausgleich im Le¨tzebuergeschen. Zur phonetisch-phonologischen Fokussierung einer Nationalsprache. (Phonai 44.) Tübingen: Niemeyer. Gillie´ron, Jules and Edmond Edmont 1902⫺1910 Atlas linguistique de la France (ALF). 13 vols. Paris: Champion Goebl, Hans 1982 Dialektometrie. Prinzipien und Methoden des Einsatzes der Numerischen Taxonomie im Bereich der Dialektgeographie. (Denkschriften der Österreichischen Akademie der Wissenschaften, Philosophisch-historische Klasse 157.) Vienna: Verlag der österreichischen Akademie der Wissenschaften. Goebl, Hans 1984 Dialektometrische Studien anhand italoromanischer, rätoromanischer und galloromanischer Sprachmaterialien aus AIS und ALF. Tübingen: Niemeyer. Goebl, Hans 1993 Probleme und Methoden der Dialektometrie: Geolinguistik in globaler Perspektive. In: Wolfgang Viereck, (ed.), Proceedings of the International Congress of Dialectologists, vol. 1, 31⫺81. Stuttgart: Steiner. Göschel, Joachim 1973 Strukturelle und intrumentalphonetische Untersuchungen zur gesprochenen Sprache. Berlin/ New York: de Gruyter.

781

782

VIII. Structural domains: Methodological problems Goldsmith, John A. 1990 Autosegmental and Metrical Phonology. Oxford: Blackwell. Goldsmith, John A. (ed.) 1996 The Handbook of Phonological Theory. (Blackwell Handbooks in Linguistics.) Oxford: Blackwell. Goossens, Jan 1969 Strukturelle Sprachgeographie. Eine Einführung in Methodik und Ergebnisse. (Sprachwissenschaftliche Studienbücher ⫺ Zweite Abteilung.) Heidelberg: Winter. Gussenhoven, Carlos and Haike Jacobs 2005 Understanding Phonology. London: Hodder Arnold. Haas, Walter 1978 Sprachwandel und Sprachgeographie. Untersuchung zur Struktur der Dialektverschiedenheit am Beispiel der schweizerdeutschen Vokalsysteme. (Zeitschrift für Dialektologie und Linguistik. Beiheft 30.) Wiesbaden: Steiner. Heeringa, Wilbert 2004 Measuring dialect pronunciation differences using Levenshtein distance. (Groningen Dissertations in Linguistics 46.) Available at . Herrgen, Joachim 2005 Sprachgeographie und Optimalitätstheorie. Zeitschrift für Dialektologie und Linguistik 72(3): 278⫺317. Herrgen, Joachim and Jürgen Erich Schmidt 1985 Systemkontrast und Hörerurteil. Zwei Dialektalitätsbegriffe und die ihnen entsprechenden Meßverfahren. Zeitschrift für Dialektologie und Linguistik 52: 20⫺42. Herrgen, Joachim and Jürgen Erich Schmidt 1986 Zentralisierung. Eine phonetisch-phonologische Untersuchung zu Konstanz und Wandel vokalischer Systeme. In: Günter Bellmann (ed.), Beiträge zur Dialektologie am Mittelrhein, 56⫺100 (Mainzer Studien zur Sprach- und Volksforschung 10.) Stuttgart: Steiner. Herrgen, Joachim and Jürgen Erich Schmidt 1989 Dialektalitätsareale und Dialektabbau. In: Wolfgang Putschke, Werner Veith and Peter Wiesinger (eds.), Dialektgeographie und Dialektologie. Günter Bellmann zum 60. Geburtstag von seinen Schülern und Freunden, 304⫺346 (Deutsche Dialektgeographie 90.) Marburg: Elwert. Hinderling, Robert (ed.) 1996⫺ Bayerischer Sprachatlas. Heidelberg: Winter. Hinskens, Frans, Roeland van Hout and W. Leo Wetzels (eds.) 1997 Variation, Change and Phonological Theory. Amsterdam: Benjamins. Hooper, Joan B. 1976 An Introduction to Natural Generative Phonology. New York: Academic Press. Hotzenköcherle, Rudolf 1961 Zur Raumstruktur des Schweizerdeutschen. Zeitschrift für Mundartforschung 28: 207⫺ 227. Hotzenköcherle, Rudolf 1962 Einführung in den Sprachatlas der Deutschen Schweiz. A: Zur Methodologie der Kleinraumatlanten. Berne: Francke. Hotzenköcherle, Rudolf 1962⫺1997 Sprachatlas der deutschen Schweiz. Founded by Heinrich Baumgartner and Rudolf Hotzenköcherle. Edited by Rudolf Hotzenköcherle in cooperation with Konrad Lobeck, Robert Schläpfer and Rudolf Trüb and with the assistance of Paul Zinsli. Basel: Francke. Iivonen, Antti 1983 Zur Frage der regionalen Variation der hochdeutschen Vokale. Neuphilologische Mitteilungen 84: 45⫺52.

42. Segmental phonetics and phonology Iivonen, Antti 1987 Zur regionalen Variation der betonten Vokale im gehobenen Deutsch. Kontrastive Evidenz unter besonderer Berücksichtigung des Ostmitteldeutschen und Wienerdeutschen. In: Leena Kahlas-Tarkka (ed.), Neophilologica Fennica, 87⫺119. (Me´moires de la Socie´te´ Ne´ophilologique de Helsinki XLV.) Helsinki: Socie´te´ Ne´ophilologique. Iivonen, Antti 1994 Zur gehobenen regionalen phonetischen Realisierung des Deutschen. In: Wolfgang Viereck (ed.), Verhandlungen des Internationalen Dialektologenkongresses Bamberg 29. 7.⫺ 4. 8. 1990, vol. 3: Regional Variation, Colloquial and Standard Languages, 311⫺330. (Zeitschrift für Dialektologie und Linguistik. Beiheft 76.) Stuttgart: Steiner. Jaberg, Karl and Jakob Jud 1928⫺1940 Sprach- und Sachatlas Italiens und der Südschweiz. 8 vols. Zofingen: Ringier. Jacewicz, Ewa, Robert Allen Fox and Joseph Salmons 2007 Vowel space areas across dialects and gender. In: Jürgen Trouvain and William J. Barry (eds.), Proceedings of the ICPhS XVI, 1465⫺1468. Saarbrücken: Universität des Saarlandes. Johnson, Keith 2003 Acoustic and Auditory Phonetics. 2nd ed. Oxford: Blackwell. Kessler, Brett 1995 Computational dialectology in Irish Gaelic. In: European Chapter of the ACL (ed.), Proceedings of the European ACL, 60⫺67. Dublin: ACL. [Available from ; last accessed 14 September 2008.] Kehrein, Roland, Alfred Lameli and Jost Nickel 2006 Möglichkeiten der computergestützten Regionalsprachenforschung am Beispiel des Digitalen Wenker-Atlas (DiWA). In: Georg Braungart, Peter Gendolla and Fotis Jannidis (eds.), Jahrbuch für Computerphilologie 7: 149⫺170. Paderborn: mentis. Keller, Rudolf Ernst 1961 German Dialects: Phonology and Morphology, with Selected Texts. Manchester: Manchester University Press. Künzel, Hermann J. 2001 Beware of the “telephone effect”: The influence of telephone transmission on the measurement of formant frequencies. Forensic Linguistics 8(1): 80⫺99. Labov, William 1991 The three dialects of English. In: Penelope Eckert (ed.), New Ways of Analyzing Sound Change, 1⫺44. New York: Academic Press. Labov, William, Malcah Yaeger and Richard Steiner 1972 A Quantitative Study of Sound Change in Progress. Philadelphia: US Regional Survey. Labov, William, Sharon Ash and Charles Boberg 2005 The Atlas of North American English: Phonetics, Phonology, and Sound Change. Berlin/ New York: Mouton de Gruyter. Ladefoged, Peter 2003 Phonetic Data Analysis. An Introduction to Instrumental Phonetic Fieldwork. Oxford: Blackwell. Ladefoged 2004 Vowels and Consonants: An Introduction to the Sounds of Languages. 2nd ed. Oxford: Blackwell. Ladefoged, Peter and Ian Maddieson 1996 The Sounds of the World’s Languages. Oxford: Blackwell. Lauf, Raphaela 1993 Phonetische Aspekte des Vokalismus münsterlandischer Dialekte. Eine auditive und akustische Analyse vokalischer Merkmale mit einem Ansatz zu ihrem Vergleich mit der Standardsprache. Stuttgart: Steiner.

783

784

VIII. Structural domains: Methodological problems Mohanan, Karuvannur P. 1986 The Theory of Lexical Phonology. Dordrecht: Reidel. Moosmüller, Sylvia 1987 Soziophonologische Variation im gegenwärtigen Wiener Deutsch. Eine empirische Untersuchung. Stuttgart: Steiner. Moosmüller, Sylvia 1988 Sociophonology. In: Peter Auer and Aldo di Luzio (eds.), Variation and Convergence. Studies in Social Dialectology, 76⫺93. Berlin/New York: de Gruyter. Moosmüller, Sylvia 2006 Some relevant aspects of vowel formant interpretation. Grazer Linguistische Studien 65: 1⫺21. Moulton, William G. 1961 Lautwandel durch innere Kausalität: Die ostschweizerische Vokalspaltung. Zeitschrift für Mundartforschung 28: 227⫺251. Moulton, William G. 1968 Structural dialectology. Language 44/3: 451⫺466. Nearey, Terrance M. 1978 Phonetic feature systems for vowels. Bloomington: Indiana University Linguistics Club. [Ph.D. dissertation, University of Alberta. Also available at .] Nerbonne, John to appear Various variation aggregates in the LAMSAS South. Accepted to appear in: Catherine Davis and Michael Picone (eds.), Language Variety in the South III. Tuscaloosa: University of Alabama Press. [Available from ; last accessed 22 September 2008.] Nerbonne, John and Christine Siedle 2006 Dialektklassifikation auf der Grundlage Aggregierter Ausspracheunterschiede. Zeitschrift für Dialektologie und Linguistik 72(2): 129⫺147. Noske, Roland 2005 A prosodic contrast between Northern and Southern Dutch: A result of a Flemish-French sprachbund. In: Hans Broekhuis, Riny Huybrechts, Ursula Kleinhenz and Jan Koster (eds.), Organizing Grammar. Linguistic Studies in Honor of Henk van Riemsdijk, 474⫺ 482. Berlin: Mouton de Gruyter. Oostendorp, Marc van 2005 The first person singular in dialects of Dutch. In: Leah Bateman and Cherlon Ussery (eds.), Proceedings of the North Eastern Linguistic Society, 134⫺166. Amherst: GLSA Pfalz, Anton 1918 Reihenschritte im Vokalismus. In: Beiträge zur Kunde der bayerisch-österreichischen Mundarten, 22⫺42. (Kaiserliche Akademie der Wissenschaften in Wien. Philosophischhistorische Klasse. Sitzungsberichte 190/2.) Vienna: Rohrer. Pfalz, Anton 1936 Zur Phonologie der baierisch-österreichischen Mundarten. Vienna: Hölder. Plichta, Bartłomiej and Dennis R. Preston 2005 The /ay/s have it: The perception of /ay/ as a North⫺South stereotype in US English. Acta Linguistica Hafniensia 37: 107⫺130. [Theme issue: Tore Kristiansen, Nikolas Coupland and Peter Garrett (eds.), Subjective Processes in Language Variation and Change.] Prince, Alan and Paul Smolensky 2004 Optimality Theory: Constraint Interaction in Generative Grammar. Oxford: Blackwell. Ramers, Karlheinz 2001 Einführung in die Phonologie. Munich: Fink. Reetz, Henning and Allard Jongman 2008 Transcription, Production, Acoustics, and Perception. (Blackwell Textbooks in Linguistics.) Malden, MA: Blackwell.

42. Segmental phonetics and phonology Rein, Kurt 1974 Die mittelbairische Liquiden-Vokalisierung. Zeitschrift für Dialektologie und Linguistk 41: 21⫺37. Repetti, Lori (ed.) 2000 Phonological Theory and the Dialects of Italy. (Current Issues in Linguistic Theory 212.) Amsterdam: Benjamins. Rennison, John R. 1981 Bi-dialektale Phonologie. Die Kompetenz zweier Salzburger Sprecher. (Zeitschrift für Dialektologie und Linguistik. Beiheft 34.) Wiesbaden: Steiner. Richter, Helmut 2005 Transkriptionssysteme. In: Ulrich Ammon, Norbert Dittmar and Klaus J. Mattheier (eds.), Sociolinguistics: An International Handbook of the Science of Language and Society, 1126⫺1133. (Handbücher zur Sprach- und Kommunikationswissenschaft, 3.2.) Berlin/New York: de Gruyter. Schouten, Bert and Wim Peters 1996 The Middle High German Vowel Shift, measured acoustically in Dutch and Belgian Limburg: Diphthongization of short vowels. Zeitschrift für Dialektologie und Linguistik 63(1): 30⫺48. Shackleton, Robert G. 2007 Phonetic variation in the traditional English dialects: A computational analysis. Journal of Englisch Linguistics 35(1): 30⫺102. Sievers, Eduard 1876 Grundzüge der Lautphysiologie ⫺ zur Einführung in das Studium der Lautlehre der indogermanischen Sprachen. (Bibliothek indogermanischer Grammatiken 1.) Leipzig: Breitkopf & Härtel. Smith, Jennifer L. 2003 Onset sonority constraints and subsyllabic structure. Presented at the Ninth International Phonology Meeting, University of Vienna, 3 November 2002. Available from; last accessed 21 October 2008. (Rutgers Optimality Archive 608.) Steger, Hugo, Eugen Gabriel and Volker Schupp (eds.) 1989 Südwestdeutscher Sprachatlas. Marburg: Elwert. Szczepaniak, Renata 2007 Der phonologisch-typologische Wandel des Deutschen von einer Silben- zu einer Wortsprache. (Studia Linguistica Germanica 85.) Berlin/New York: de Gruyter. Trubetzkoy, Nikolai S. 1931 Phonologie und Sprachgeographie. Travaux du Cercle Linguistique de Prague 4: 228⫺234. Vieregge, Wilhelm H. 1985 Ein Maß zur Reliabilitätsbestimmung phonetisch-segmenteller Transkriptionen. Zeitschrift für Dialektologie und Linguistik 52: 167⫺180. Weinreich, Uriel 1954 Is a structural dialectology possible? Word 10: 388⫺400. Wells, John C. 1982 Accents of English. 3 vols. and cassette. Cambridge: Cambridge University Press. Werlen, Iwar 1977 Lautstrukturen des Dialekts von Brig im schweizerischen Kanton Wallis. Ein Versuch zur Integration strukturaler und generativer Beschreibungsverfahren in der Dialektologie. Wiesbaden: Steiner. Wieling, Martijn, Wilbert Heeringa and John Nerbonne 2007 An aggregate analysis of pronunciation in the Goeman⫺Taeldeman⫺van Reenen-Project data. Taal en Tongval 59(1): 84⫺116.

785

786

VIII. Structural domains: Methodological problems Wiesinger, Peter 1970 Phonetisch-phonologische Untersuchungen zur Vokalentwicklung in den deutschen Dialekten. Berlin: de Gruyter. Wiesinger, Peter 1983a Die Einteilung der deutschen Dialekte. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke, Herbert E. Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 2: 807⫺900. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.2.) Berlin/New York: de Gruyter. Wiesinger, Peter 1983b Phonologische Vokalsysteme deutscher Dialekte. Ein synchronischer und diachronischer Überblick. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke, Herbert E. Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 2: 1042⫺1076. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.2.) Berlin/ New York: de Gruyter. Winteler, Jost 1876 Die Kerenzer Mundart des Kantons Glarus in ihren Grundzügen dargestellt. Leipzig/Heidelberg: Winter.

Peter Gilles, Walferdange (Luxembourg) Beat Siebenhaar, Leipzig (Germany)

43. Areal variation in prosody 1. 2. 3. 4. 5. 6. 7. 8. 9.

Introduction Basic concepts of prosody The temporal domain Intensity Voice quality Intonation Prosodic structure of word and syllable Conclusion References

1. Introduction Research into regional variation in prosody has increased tremendously since the 1990s. Although regarded as one of the most pertinent features since the beginnings of dialectology, prosody was long neglected due to the predominance of purely segmental issues. Research into prosody has not yet achieved full coverage of a larger geographical area, e.g., in the form of a language atlas. Rather, several studies of various prosodic features in single dialects are available, which can be used to gain initial insights into areal variation in prosody. For descriptions of general aspects of prosody, the reader is referred to Botinis, Granström and Möbius (2001), Fox (2000), Gussenhoven (2004, 2007), Gussen-

43. Areal variation in prosody hoven and Riad (2007), Ladd (2008), Riad and Gussenhoven (2007) and Sudhoff et al. (2006), which together provide an up-to-date overview of all of the relevant theoretical, descriptive and methodological issues.

2. Basic concepts o prosody Prosodic or suprasegmental features of speech are attributed to aspects covering more than one segmental sound or to single sound properties that are governed by larger units. As on the segmental level, the analysis of prosody has a phonetic and a phonological dimension. Phonetics is concerned with the physical aspects of prosody, while phonology studies the systematic use of prosody to encode meaning in a specific language. In phonology, a hierarchy of prosodic domains is widely accepted, ranging from the syllable through foot, phonological word, clitic group, phonological phrase, intonational phrase to the phonological utterance (see Nespor and Vogel 2007). With regard to phonetics, prosody manifests itself in the properties of length, pitch, loudness, and their respective acoustic correlates: duration, fundamental frequency (F0) and intensity. Current linguistic research on prosody is focused on intonation, pitch and fundamental frequency, while temporal/timing aspects have mainly been the province of speech technology research, predominantly research into speech synthesis. Intensity, however prominent it was at the beginning of the twentieth century, with the distinction between tone accent and dynamic accent in historical linguistics, tends to be sidelined in current research. Voice quality is still only peripherally recognized as a prosodic feature. Current linguistic research into prosody still concentrates on methodological questions, on variation in standard varieties, or on pragmatic and psycholinguistic questions. Nevertheless, all aspects can be analyzed in relation to their distribution in space, and the more these questions are raised, the more obvious their spatial distribution becomes.

Fig. 43.1: Display of prosodic features in the acoustic phonetics software package Praat (Boersma and Weenink 2008)

787

788

VIII. Structural domains: Methodological problems Contemporary research on prosody only became possible when affordable and easy to use computer tools for phonetic analysis became available. Figure 43.1, generated with the Praat software package (Boersma and Weenink 2008), displays several acoustic features relevant to prosody, with an oscillogram on top, followed by a spectrogram, a pitch curve displayed as a blue line and intensity as a yellow line. In the bottom window there are manually attributed labels for each segment and tone. The highlighted section is part of a diphthong realized with creaky voice.

3. The temporal domain Temporal aspects at the level of the phrase are investigated by analyzing the position ˇ apkova´ and Megyand duration of pauses in different speaking styles (e.g., Gustafson-C esi 2001; Peters, Kohler and Wesener 2005; Zellner 1994). However, a comparison of regional varieties is barely addressed and differences are rarely found. Initial attempts are presented in Hove (2004). However, individual and stylistic variation are very high in this regard, so that generalizations about the influence of dialects or regional varieties cannot yet be made. The phonologically relevant distribution of segment duration is a prominent aspect within the temporal domain. If a variety distinguishes phonemes on the basis of quantity, it is usually between long and short segments. A ternary distinction is rare (Fox 2000: 42⫺46). Globally, quantity distinctions are more common for vowels than for consonants (Laver 1994: 436). Within a language, different dialects can feature different oppositions. While most German dialects only differ with respect to vowel quantity, Alemannic and Bavarian dialects also show distinctions of consonant quantity, as documented in the BSA and SDS linguistic atlases (Hinderling 1996⫺; Hotzenköcherle 1962⫺1997). Moreover, the Standard and northern German opposition between voiced and voiceless obstruents is expressed as a temporal distinction between long and short consonants in the southern dialects (Willi 1996). This distinction is even observed in initial position, as Kraehenmann (2003) documents for the Thurgovian dialect (in Switzerland). Since the distinction was drawn by Pike (1945), typological differences between syllable-timed, stress-timed and mora-timed languages have been considered. Reliable phonetic correlates to this distinction were only found in the late 1990s, with different algorithms that take into account the relation between the durations of consecutive consonantal and vocalic parts in an utterance. These typological differences are mainly evident between languages (Ramus, Nespor and Mehler 1999); however, the algorithms also reveal differences between British English and Singapore English (Low, Grabe and Nolan 2000) or between Arabic dialects (Ghazali, Hamdi and Barkat 2002); the latter conclude that differences in vowel duration and syllabic complexity seem to be the main factors responsible for differences in stress patterns and rhythmic structure. Comparing four Swiss German dialects, Siebenhaar (2004) and Leemann and Siebenhaar (2007) show that the timing relation between accented and unaccented syllables distinguishes these dialects; in effect they exhibit a different way of producing accents. Further, different patterns are found in the marking of phrase boundaries in these four Alemannic dialects. Final lengthening is documented in all of the dialects; however, the first and the penultimate syllable of a phrase also appear lengthened, not just the last

43. Areal variation in prosody

Fig. 43.2: Duration of nuclei by position in the phrase (f: first, m: medial, p: penultimate, u: ultimate syllable) in Berne (left) and Wallis (right) dialect; diamonds indicate the mean and the 95 %-confidence interval; from Leemann and Siebenhaar (2007: 958)

syllable. The temporal differences between the positions of the syllables in a phrase reveal distinctive patterns in midland and alpine dialects. Figure 43.2 compares the midland Bernese dialect and the alpine Wallis dialect. In the Bernese dialect, the first syllable is significantly lengthened compared to the syllables in the middle of a phrase while the Wallis dialect has no significant phrase-initial lengthening. Moreover, the final lengthening in the Bernese dialect is much more pronounced than in the Wallis dialect, especially in the last syllable. The alpine dialects show much less lengthening around phrase boundaries than do the midland dialects. Hence, the phrase boundaries are clearly temporally marked in the midland dialects, with a slowing down and a speeding up, while this temporal markedness is less obvious in the alpine dialects. Moreover, general speech rate can also be used to distinguish dialects (Ruoff 1973), as is also clearly visible in Figure 43.2. These phonetic differences between the dialects of one dialect group are to an extent minimal; nevertheless they are perceptually relevant (Leemann and Siebenhaar 2008a).

4. Intensity Due to the technical difficulties in obtaining reliable measurements results, systematic research into intensity is rare in dialectology (for a review of older research, see Heike 1983). With regard to its influence on syllable structure, Spiekermann (2000) presents findings on syllable cut prosody in German dialects. In this study, syllable cut is expressed in terms of the intensity contour of the vowel. Syllables with a smooth syllable cut show a flat intensity contour at a high level until the end of the vowel, while an abrupt cut is characterized by a steeply falling intensity contour. In German regional varieties, we find syllable cut prosodies mainly in northern German dialects. Southern German dialects, in contrast, do not show these differences in intensity distribution, but rather a system with short and long vowels. As for the central Franconian tone accents (see section 7), Heike (1964) and Peters (2006c) show that the so-called accent 1 is characterized by a sharp drop in intensity, while accent 2 does not show any such drop.

789

790

VIII. Structural domains: Methodological problems

5. Voice quality To date, voice quality has not usually been considered in prosodic studies. However, an increasing number of studies examine voice quality as an integral dimension of prosody (Nı´ Chasaide and Gobl 2004). It has been shown that changes in voice quality distinguish prominent and non-prominent words and phrase-initial and phrase-final words in English (Epstein 2002). Sound quality changes are also used for marking phrase boundaries (Peters, Kohler and Wesener 2005). Nevertheless, a spatial dimension to voice quality has not yet been explored and therefore remains an open field for research.

6. Intonation Regional differences in intonation have been repeatedly reported since the beginnings of dialectology, with Sievers’ (1912: 62⫺63) idea of a northern and a southern German intonation system (“zwei konträre Generalsysteme der Melodisierung”) being probably the most prominent proposal (for a review of older intonation research, see Gilles 2005 and Peters 2006a). However, it was not until the end of the twentieth century that substantial intonation research in dialectology began to gain ground. It was not only the improvement in and easy availability of acoustic phonetic techniques mentioned above, but also the conceptual and theoretical work on the structure of intonation itself, which contributed to this substantial progress (see Ladd 2008). Today, intonation is seen as a linguistic feature that is phonetically constituted by the movements of the fundamental frequency (F0) over a specific stretch of speech. The function of intonation lies on the one hand in signalizing certain semanto-syntactic aspects like broad or narrow focus or sentence type (Fe´ry 1993). On the other hand, intonation serves several conversational functions in the context of turn taking or turn projection (Couper-Kuhlen and Selting 1996). In the influential “British School” of intonation research (Cruttenden 1997; Crystal 1969), basic concepts like “head”, “nucleus” and “tail” have been introduced to capture the structural constituents of a “tone unit”. In present-day autosegmental-metrical accounts of intonation (Gussenhoven 2004; Ladd 2008), the intonation contour in an “intonation phrase” is constituted by target tones, which are anchored to certain landmarks in the intonation phrase ⫺ syllables with focus accents function as anchor points for pitch accents. Falling pitch accents are said to consist of a prominent high tone (H*) and a less prominent low tone (L); rising pitch accents are conceptualized in a similar fashion. The beginning and end of an intonation phrase can be marked by boundary tones (i. e., %H, %L, H%, L%). The actual phonetic shape of the intonation contour in a given utterance then results from the interpolation of the melody between the tones. This structural decomposition into well-defined tonal elements also provides the descriptive framework for the transcription and phonological interpretation of intonation contours for dialectology. According to Ladd (2008: 119), four types of differences between intonation systems, which can also be adopted to regional variation, can be distinguished: (1) semantic ⫺ differences in the meaning of identical intonational tunes; (2) systemic ⫺ differences in the inventory of tunes; (3) realizational ⫺ differences in the phonetic make-up of tunes; (4) phonotactic ⫺ differences in the text⫺tune association and in the permitted tone

43. Areal variation in prosody structure. Phonotactic differences in the intonational systems of German regional varieties are discussed extensively in Peters (2006a) and Kügler (2007); an overview is given in Peters (2006b). A collection of comparative studies can be found in Gilles and Peters (2004). In the following, only realizational and systemic differences will be addressed in more detail.

6.1. Realizational dierences in intonation Realizational differences may be evident in the phonetic implementation of a particular phonological category. Several phonetic aspects can be subject to regional variation, including but not limited to alignment of F0 movement to an accented syllable, truncation of F0 movements because of lack of sonorant material, the F0 range used or the steepness of the F0 slope over an utterance. Tonal alignment has attracted much research activity in recent years. “Peak delay”, for example, is described in Scottish English by Aufterbeck (2004), in Irish dialects by Dalton and Nı´ Chasaide (2005) and in the Swabian dialect by Kügler (2004). For falling accents of the H*L type, Gilles (2005) describes the regional pattern for alignment of the F0 maximum in the accented syllable in German regional varieties. In order to control for interfering factors, only phrase-final structures consisting of exactly two syllables were analyzed (nucleus and tail). Several acoustic measurements ensure that the characteristics of the F0 movement were captured. The raw data from several instantiations were then converted into relative measurements to allow pooling of different speakers. The resulting schematic contours, shown in Figure 43.3, exhibit clear differences. Common to the regional varieties of Berlin, Dresden and Munich (on the left) is that the F0 maximum is reached in the second half of the nucleus, whereas in Hamburg and Mannheim (on the right), it is reached quite early (after just 22 and 13 percent of the duration, respectively). One can also observe a steep falling movement in the nucleus in the latter dialects, due to the early F0 maximum, while the former group is characterized by a falling movement predominantly after the nucleus, i. e., on the unaccented syllable of the tail. Hamburg and Mannheim, then, are said to feature an early alignment of the F0

Fig. 43.3: Mean intonation contours for bi-syllabic falling contours in five regional varieties of German (based on Gilles 2005: 342)

791

792

VIII. Structural domains: Methodological problems

Fig. 43.4: Typical rise patterns for statements and questions conflated over two utterances, and broken down by speaker dialect (Shetland or Orkney) and by sex (from van Leyden 2004: 65)

maximum, while Berlin, Dresden and Munich show a middle alignment. Further studies on alignment in German regional varieties were conducted by Atterer and Ladd (2004) and Kleber and Rathcke (2008). They show that alignment is very early for prenuclear accents in East Middle German, while it is late in southern German. Leemann and Siebenhaar (2008b) show that this pattern is continued in the Swiss German dialects, where an accent begins to rise 20 milliseconds before the accented syllable in eastern dialects and 100 milliseconds after the onset of the accented syllable in the western dialects. Using a similar technique, van Leyden (2004) presents a thorough discussion of alignment for the closely related Shetland and Orkney dialects. Figure 43.4 shows the alignment of accent-lending rises in statements and questions. (Test sentences of the type There are many gardens in Bergen and Are there many houses in Bergen? are used.) While the onset of the rising movement is located right at the beginning of the accented syllable in the Shetland dialect, the rise begins considerably later on the Orkneys. A second aspect of realizational differences in intonation should be mentioned: truncation and compression have been suggested as typological parameters in intonation (Grønnum 1989; Ladd 2008). For example, analyzing the IViE data, Grabe (2004) compares Leeds and Newcastle dialects. At a phonological level, the two dialects both have

43. Areal variation in prosody

Fig. 43.5: Example of compression in Newcastle English and truncation in Leeds English (from Grabe 2004: 18)

falling accents (H*L). The dialects show no difference when the accent is realized over a long word with a large scope for voicing (see the top row in Figure 43.5). But if the accent is produced at the end of an intonation phrase, on a short word with little scope for voicing, the two dialects show distinct realizations (bottom row in Figure 43.5): Newcastle English speakers compress the accent, i. e., the pattern is produced more quickly, while Leeds English speakers truncate the accent: the pitch movement remains incomplete, the fall of the H*L-accent is not realized. It should become clear that all of the regional differences described so far, although clearly discernible, are purely phonetic processes, in that they do not affect the phonological system. The phonetic processes underlying intonational differences nonetheless represent a central area of research, because they are to a wide extent responsible for differences between (closely related) dialects.

6.2. Systemic dierences in intonation The intonation system of a language/dialect consists of an inventory of intonation contours or tunes. Closely related dialects can have a high degree of system overlap, differing in only a few inventory related aspects. Thus, numerous studies are found in the research literature, which concentrate on single specific intonational differences typical of the intonation system of a given variety. The focus of these studies lies in the detailed (formal and functional) description of this salient intonation contour, and they fail to take into account the intonational inventory as a whole. To cite only a few works that apply this method: Auer (2001a) discusses the so-called “high onset”, which sets Hamburg German off from other regional varieties (and the standard language); Bergmann (2008) describes a rising-falling contour typical for Cologne German; Selting (2000) analyzes a typical high rise contour in Berlin German; and Fletcher and Harrington (2001) explore the characteristic high-rising terminals in Australian English. Peters (2006a: 384⫺390), elaborating on the pioneering work of Guentherodt (1973), analyzes the nuclear contours

793

794

VIII. Structural domains: Methodological problems

Map 43.1: Intonation isogloss in the Palatinate dialect cluster (Peters 2006a: 388)

used in wh-questions in a cluster of Palatinate dialects (“Pfälzisch”, southwestern Germany). Because data was available for several locations in the area, it proved possible to discover an intonational isogloss that separates an area in the southwest with a risingfalling nuclear contour from a northern territory characterized by (high-)falling contours (see Map 43.1). These kinds of selective studies on isolated intonation contours in regional varieties should be considered as an important step towards obtaining an increasingly exhaustive picture of the possible variation of intonation in dialects. As for the specification of a whole inventory of intonation contours in a given dialect and the subsequent comparison with inventories from other, related dialects, research findings are still limited. In order to establish an inventory system, extensive formal and functional analysis of all intonation patterns in a variety has to be done. By applying, e.g., the autosegmental-metrical approach, it is necessary to determine all pitch accent types, the boundary tones and the allowed combinations of these elements. Barker (2005) uses this procedure for his account of the intonation patterns in Tyrolean German. Grabe (2004), Grabe and Post (2002) and Grabe et al. (2001) present extensive findings about inventory differences and overlap for the dialects of the British Isles. Peters (2006a) develops finite state grammars of the intonation systems for six regional varieties of German (as an example, see Figure 43.6 for the intonation grammar of Cologne German).

Fig. 43.6: Autosegmental-metrical finite state grammar for the intonation system of Cologne German (based on Peters 2006a: 459)

43. Areal variation in prosody

Fig. 43.7: Schematic nuclear contours and autosegmental-metrical symbolization for northern Standard German and Freiburg German (based on Peters 2006a: 455)

By identifying these intonation grammars for several regional varieties, differences and similarities become evident. Figure 43.7 shows a subset of the nuclear contours for northern Standard German and Freiburg German exhibiting the phonological differences between the two varieties. One striking differences concerns the type of pitch accent used in the nucleus: Standard German (like many other German varieties) has a falling (H*L) accent, while in Freiburg a rising accent (L*H) is observed. As for the end of the intonation phrase, Standard German has simple boundary tones (Li, Hi, ⵰i), whereas Freiburg has more complex structures involving phrase accents of the type Hi⫹L.

7. Prosodic structure o word and syllable This last section is devoted to aspects that are most adequately described in terms of the (phonological) word and the syllable. Auer (1998) provides a detailed account of the variability of the intervocalic position in disyllabic words in several German dialects. His starting point is a clearly defined prosodic structure of a word that takes syllabic positions into account and thus makes it possible to describe the regional variation of a specific prosodic structure rather than just a single sound. It turns out that considerable weakening processes are at work in the intervocalic (in many cases ambisyllabic) position in German dialects. Applying a similar methodology, Auer (1997) analyzes the epenthesis of schwa in Central German varieties by making crucial reference to syllable structure. With regard to more global prosodic features, Nübling and Schrambke (2004) elaborate on the distinction between “word languages” and “syllable languages”. This distinction, established by Donegan and Stampe (1983), is related to the distinction between “stress-timed” and “syllable-timed” languages. These two prosodic language types are characterized by the features listed in Table 43.1. Nübling and Schrambke (2004), then, analyze prosodic properties in several Alemannic dialects (southwestern Germany, Switzerland and Alsace). They are able to show that in the north of this region the (phonological) word is the most important unit for

795

796

VIII. Structural domains: Methodological problems Tab. 43.1: Characteristics of syllable languages and word languages (cf. Auer 2001b) Syllable languages

Word languages

simple syllable structure, often CV

complex syllable structure, differences between accented and unaccented syllables

clear-cut syllable boundaries

partly unclear syllable boundaries, speed related variability

hardly any signals for word boundaries

phonological signals of word boundaries

syllable structured according to sonority hierarchy

frequent violations of the sonority hierarchy

geminates possible

no geminates

accentuation has no influence on the syllable structure

distinction between accented and unaccented syllable

tone systems possible

hardly any tone system (if so, restricted to accented syllables)

vowel harmony possible

vowel harmony impossible

external sandhi possible

only internal sandhi

liaison possible

no liaison

phonological processes, while syllable related processes play the more important role in the southern part. These sandhi processes had already been described by Moulton (1986); however, the general prosodic difference had not been recognized. Map 43.2 (left) displays the regional distribution of cross-word assimilation processes. The Standard German phrase sollte man ‘should one’ shows up as [zodme(r)] in the north and [zøpme] in the south. Whereas the first form conforms to a prohibition on phonological processes crossing the word boundary, the latter form is characterized by regressive assimilation across this boundary: the initial nasal in the pronoun spreads its labiality to the final consonant of the verbal form ([zød] > [zøp]). These two prosodically defined areas are separated by a thin transition zone, where both non-assimilated and assimilated forms are observed. Similar cases involving place assimilation of a clitic article or negative are shown on the right-hand map of Map 43.2. Here, assimilation across word boundaries is blocked in the north ([dfensder] ‘the window’, [(n)idfolge] ‘not following’, [dmil(i)c¸] ‘the milk’) and is possible in the south ([pfensder], [(n)ipfolge], [pmil(x)]). By presenting several other prosodic isoglosses with a similar shape (e.g., no geminates in the north versus geminates in the south; appearance of the glottal stop in the north versus its absence in the south because resyllabification processes can cross word boundaries), Nübling and Schrambke (2004) provide arguments for the first bundle of prosodic isoglosses detected in German dialects, a bundle which divides varieties with word-language aspects from those with syllable-language aspects. Research in this field has only recently begun, and a prosodic reanalysis of regional patterns reported in older dialectological descriptions and atlases should offer new insights into the phonological structure of regional varieties. A long-standing research topic concerns the tone accent system found in a few Germanic languages/dialects. See de Vaan (2006) for an overview; Gussenhoven and Bruce

43. Areal variation in prosody

Map 43.2: Prosody-induced isoglosses in Alemannic dialects (from Nübling and Schrambke 2004: 312)

(1999), Kristoffersen (2000) and Riad (2003) on the tone accents of Swedish and Norwegian; Nilsen (2001) for a discussion of regional patterns in Norwegian dialects; Gilles (2002), Gussenhoven and Peters (2004, 2008), Künzel and Schmidt (2001), Peters (2006c), Schmidt (1986) and Schmidt and Künzel (2006) on the “Tonakzente” (“Schärfung” versus “Trägheitsakzent”) of Central Franconian; Gussenhoven and Aarts (1999) and Gussenhoven and van der Vliet (1999) on stoottoon ‘push tone’ and sleeptoon ‘dragging tone’ in Limburgish. The origin of this tonal contrast is explored in de Vaan (1999), Gussenhoven (2000) and Schmidt (2002), drawing on voicing of the following consonant and apocope of an unstressed syllable as the key aspects of tonogenesis. The regional distribution of the tone accents in Limburg and Central Franconian is documented in Schmidt and Künzel (2006: 139) and on the corresponding maps of the MRhSA (Bellmann, Herrgen and Schmidt 1994⫺2002). In these “tonal” languages/dialects, the prosodic features of the accented, bi-moraic syllable in two segmentally otherwise identical words establish a phonological opposition, which manifests itself in two tone accents, often called accent 1 and accent 2. Phonetically, specific tonal shapes are the most pertinent prosodic features distinguishing the two accents, but duration and intensity also appear to play a (minor) role; accent 2 is often longer than accent 1 (see Heijmans 2003 for a discussion of Limburg dialects). While it is generally accepted that the tone accents constitute a phonological opposition, e.g., in minimal pairs like /ra=1f/ ‘grater’ versus /ra=2f/ ‘frost’ in Central Franconian, the realization of the two tones varies considerably from dialect to dialect and is also influenced by the global intonation contour of the phrase or sentence. Figure 43.8 exemplifies the tonal contrast for the minimal pair [mø1l] ‘mill’ versus [mø2l] ‘garbage’ embedded in the middle of the global intonation contour of a statement (top) and a question (bottom). In the statement, tone accent 1 is characterized by a sharp F0 drop within the rhyme of the syllable (shaded area), while accent 2 falls only slightly. In

797

798

VIII. Structural domains: Methodological problems

Fig. 43.8: Examples of the two Germanic tone accents in statements and questions for the words [mø1l] ‘mill’ and [mø2l] ‘garbage’; speaker from the Cologne region (Central Franconian, Germany)

questions, the reverse picture emerges, i. e., accent 1 shows a steep rise, while the tone contour of accent 2 remains low throughout the accented syllable. The main question for a phonological account of the tone accents is how to describe their different phonetic forms in a unified manner. In the autosegmental-metrical approach put forward by Gussenhoven (e.g., Gussenhoven and van der Vliet 1999), tone accent 2 is characterized by a so-called “lexical tone” absent from accent 1. The specific surface-phonetic forms of the two tone accents are due to their implementation and interaction with the tonal structure of the intonation of the utterance as a whole. Dialects, then, may differ in both the quality of the lexical tone and its interaction with utterance intonation. For example, in the Limburgish dialects of Roermond, Venlo and Maastricht, the lexical tone of accent 2 is a high one (Hlex), while the closely related dialects of Tongeren and Hasselt exhibit a low tone (Llex) here. In both areas, the lexical tone follows the focal tone of the accented syllable. Peters (2006c) assumes that the dialect of Cologne also has a lexical tone, but it is not further specified whether it is high or low (Tlex). Rather, its quality is derived directly from the following utterance intonation and can thus always be predicted from the tonal context. Furthermore, and in contrast to the Limburg dialects, this lexical tone precedes the focal tone of the accented syllable. This Tlex-tone, then, shows up as a high tone (Hlex) when it precedes a high tone (H*) and is low when it precedes a low tone (L*). As can be seen in Figure 43.8, the level high tone of accent 2 in statements has the phonological form Hlex H*, while the corresponding accent 2 realization in the question context has the form Llex L*.

43. Areal variation in prosody

8. Conclusion Since the end of the twentieth century, research in prosody has also been gaining momentum with regard to regional variation through the presentation of numerous findings about various prosodic features ⫺ from both a phonetic and a phonological perspective. These advances are due to tremendous improvements in and the easy availability of acoustic phonetic techniques as well as the theoretical grounding of phonetic and phonological conceptual categories (e.g., [pitch] accent, boundary, intonation phrase, etc.). For future research it would seem desirable to obtain yet more empirically based prosodic descriptions of dialects, so as to gather information about not only regional differences but also the domains of possible prosodic variation. In our view, not all of the prosodic features responsible for dialect differences are known yet. Experimental work that also takes speech perception into account could be promising in this respect (see, e.g., Peters et al. 2002). Furthermore, it is to be hoped that the results of phonetic science experiments are complemented by research on spontaneous speech.

9. Reerences Atterer, Michaela and D. Robert Ladd 2004 On the phonetics and phonology of “segmental anchoring” of F0: Evidence from German. Journal of Phonetics 32(2): 177⫺197. Auer, Peter 1997 Areale Variation und phonologische Theorie: Überlegungen am Beispiel der mitteldeutschen “Vokal-Epenthese”. In: Gerhard Stickel (ed.), Varietäten des Deutschen, 46⫺ 87. (Akten der IdS-Jahresstagung 1996.) Berlin: de Gruyter. Auer, Peter 1998 Variabilität der intervokalischen Position in deutschen Trochäen. Germanistische Linguistik 141⫺142: 304⫺333 [Special issue: Matthias Butt and Nanna Fuhrhoop (eds.), Variation und Stabilität in der Wortstruktur.] Auer, Peter 2001 “Hoch ansetzende” Intonationskonturen in der Hamburger Regionalvarietät. Germanistische Linguistik 157⫺158: 125⫺165 [Special issue: Jürgen Erich Schmidt (ed.), Neue Wege der Intonationsforschung.] Auer, Peter 2001b Silben- und akzentzählende Sprachen. In: Martin Haspelmath, Ekkehard König, Wulf Oesterreicher and Wolfgang Raible (eds.), Language Typology and Language Universals: An International Handbook, 1391⫺1399. (Handbücher zur Sprach- und Kommunikationswissenschaft, 20.) Berlin: de Gruyter. Aufterbeck, Margit 2004 Identifying sociolinguistic variables in intonation: The onset onglide in Anstruther Scottish English. In: Gilles and Peters (eds.), 33⫺48. Barker, Geoffrey 2005 Intonation Patterns in Tyrolean German. An Autosegmental-Metrical Analysis. New York: Peter Lang. Bellmann, Günter, Joachim Herrgen and Jürgen Erich Schmidt 1994⫺2002 Mittelrheinischer Sprachatlas. 5 vols. Tübingen: Niemeyer. Bergmann, Pia 2008 Regionalspezifische Intonationsverläufe im Kölnischen. Formale und funktionale Analysen steigend-fallender Konturen. (Linguistische Arbeiten 525.) Tübingen: Niemeyer.

799

800

VIII. Structural domains: Methodological problems Boersma, Paul and Weenink, David 2008 Praat: Doing Phonetics by Computer [computer program]. Available from ; last accessed 20 November 2008. Botinis, Antonis, Bjorn Granström and Bernd Möbius 2001 Developments and paradigms in intonation research. Speech Communication 33: 263⫺ 296. Couper-Kuhlen, Elizabeth and Margret Selting (eds.) 1996 Prosody in Conversation: Interactional Studies. Cambridge: Cambridge University Press. Cruttenden, Alan 1997 Intonation. Cambridge: Cambridge University Press. Crystal, David 1969 Prosodic Systems and Intonation in English. Cambridge: Cambridge University Press. Dalton, Martha and Ailbhe Nı´ Chasaide 2005 Tonal alignment in Irish Dialects. Language and Speech 48(4): 441⫺464. de Vaan, Michiel 1999 Towards an explanation of the Franconian tone accents. Amsterdamer Beiträge zur älteren Germanistik 51: 23⫺44. de Vaan, Michiel (ed.) 2006 Germanic Tone Accents. Proceedings of the First International Workshop on Franconian Tone Accents, Leiden, 13⫺14 June 2003. (Zeitschrift für Dialektologie und Linguistik. Beiheft 131.) Stuttgart: Steiner. Donegan, Patricia and David Stampe 1983 Rhythm and the holistic organization of language structure. In: John F. Richardson, Mitchell Marks and Amy Chukerman (eds.), Papers from the Parasession on the Interplay of Phonology, Morphology and Syntax, 337⫺353. Chicago: Chicago Linguistic Society. Epstein, Melissa Ann 2002 Voice quality and prosody in English. Ph.D. dissertation, University of California, Los Angeles. Available from ; last accessed 19 November 2008. Fe´ry, Caroline 1993 German Intonational Patterns. Tübingen: Niemeyer. Fletcher, Janet and Jonathan Harrington 2001 High-rising terminals and fall-rise tunes in Australian English. Phonetica 58: 215⫺229. Fox, Anthony 2000 Prosodic Features and Prosodic Structure. The Phonology of Suprasegmentals. Oxford: Oxford University Press. Ghazali, Salem, Rym Hamdi and Melissa Barkat 2002 Speech rhythm variation in Arabic dialects. In: Bernard Bel and Isabelle Marlien (eds.), Proceedings of Speech Prosody 2002: 331⫺334. Aix-en-Provence: Laboratoire Parole et Langage. Available from ; last accessed 16 November 2008. Gilles, Peter 2002 Einflüsse der Rheinischen Akzentuierung auf die segmentelle Ebene. Evidenz aus dem Luxemburgischen. In: Peter Auer, Peter Gilles and Helmut Spiekermann (eds.), Silbenschnitt und Tonakzente, 265⫺283. (Linguistische Arbeiten 463.) Tübingen: Niemeyer. Gilles, Peter 2005 Regionale Prosodie im Deutschen. Variabilität in der Intonation von Abschluss und Weiterweisung. (Linguistik ⫺ Impulse & Tendenzen 6.) Berlin/New York: de Gruyter. Gilles, Peter and Jörg Peters (eds.) 2004 Regional Variation in Intonation. Tübingen: Niemeyer. Grabe, Esther 2004 Intonational variation in urban dialects of English spoken in the British Isles. In: Gilles and Peters (eds.), 9⫺32.

43. Areal variation in prosody Grabe, Esther and Brechtje Post 2002 Intonational variation in English. In: Bernard Bel and Isabelle Marlien (eds.), Proceedings of Speech Prosody 2002: 343⫺346. Aix-en-Provence: Laboratoire Parole et Langage. Available from ; last accessed 16 November 2008. Grabe, Esther, Brechtje Post, Francis Nolan and Kimberly Farrar 2001 Modelling intonational variation in English: The IViE system. In: Stanislaw Puppel and Grazyna Demenko (eds.), Proceedings of Prosody 2000, 51⫺58. Poznan: Adam Mickiewicz University. Grønnum, Nina 1989 Stress group patterns, sentence accents and sentence intonation in Southern Jutland (Sønderborg and Tønder) ⫺ with a view to German. Annual Report of the Institute of Phonetics, University of Copenhagen (ARIPUC) 23: 1⫺85. Guentherodt, Ingrid 1973 A prosodic isogloss in German dialects. Zeitschrift für Dialektologie und Linguistik 40: 29⫺35. Gussenhoven, Carlos 2000 On the origin and development of the Central Franconian tone contrast. In: Aditi Lahiri (ed.), Analogy, Levelling, Markedness. Principles of Change in Phonology and Morphology, 215⫺260. Berlin/New York: de Gruyter. Gussenhoven, Carlos 2004 The Phonology of Tone and Intonation. Cambridge: Cambridge University Press. Gussenhoven, Carlos 2007 Intonation. In: Paul de Lacy (ed.), The Cambridge Handbook of Phonology, 253⫺280. Cambridge: Cambridge University Press. Gussenhoven, Carlos and Flor Aarts 1999 The dialect of Maastricht. Journal of the International Phonetic Association 29: 155⫺166. Gussenhoven, Carlos and Gösta Bruce 1999 Word prosody and intonation. In: Harry van der Hulst (ed.), Word Prosodic Systems in the Languages of Europe, 233⫺271. Berlin/New York: de Gruyter. Gussenhoven, Carlos and Jörg Peters 2004 A tonal analysis of Cologne Schärfung. Phonology 21: 251⫺285. Gussenhoven, Carlos and Jörg Peters 2008 De tonen van het Limburgs. Tijdschrift voor Nederlandse Taalkunde 13: 87⫺114. Gussenhoven, Carlos and Tomas Riad (eds.) 2007 Tones and Tunes. Experimental Studies in Word and Sentence Prosody. Berlin/New York: de Gruyter. Gussenhoven, Carlos and Peter van der Vliet 1999 The phonology of tone and intonation in the Dutch dialect of Venlo. Journal of Linguistics 35: 99⫺135. ˇ apkova´, Sofia and Bea´ta Megyesi Gustafson-C 2001 A Comparative Study of Pauses in Dialogues and Read Speech. In: Proceedings of Eurospeech 2001, 931⫺935. Available from ; last accessed 16 November 2008.) Heijmans, Linda 2003 The relationship between tone and vowel length in two neighboring Dutch Limburgian dialects. In: Paula Fikkert and Haike Jacobs (eds.), Development in Prosodic Systems, 7⫺ 45. Berlin/New York: Mouton de Gruyter. Heike, Georg 1964 Zur Phonologie der Stadtkölner Mundart: Eine experimentelle Untersuchung der akustischen Unterscheidungsmerkmale. Marburg: Elwert. Heike, Georg 1983 Suprasegmentale dialektspezifische Eigenschaften. Überblick und Forschungsbericht. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke and Herbert E. Wiegand (eds.), Dialek-

801

802

VIII. Structural domains: Methodological problems tologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, 1154⫺1169. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.2.) Berlin/New York: de Gruyter. Hinderling, Robert (ed.) 1996⫺ Bayerischer Sprachatlas. Heidelberg: Winter. Hotzenköcherle, Rudolf 1962⫺1997 Sprachatlas der deutschen Schweiz. Founded by Heinrich Baumgartner and Rudolf Hotzenköcherle. Edited by Rudolf Hotzenköcherle in cooperation with Konrad Lobeck, Robert Schläpfer and Rudolf Trüb and with the assistance of Paul Zinsli. Basel: Francke. Hove, Ingrid 2004 Pausen in spontan gesprochenem Schweizerdeutsch. Deutsche Sprache 32: 97⫺116. Kleber, Felicitas and Tamara Rathcke 2008 More on the “segmental anchoring” of prenuclear rises: Evidence from East Middle German. In: Plinio A. Barbosa (ed.), Proceedings of Speech Prosody 2008, 583⫺586. Aix-enProvence: Laboratoire Parole et Langage. Available from ; last accessed 16 November 2008. Kraehenmann, Astrid 2003 Quantity and Prosodic Asymmetries in Alemannic. Synchronic and Diachronic Perspectives. (Phonology & Phonetics 5.) Berlin/New York: Mouton de Gruyter. Kristoffersen, Gjert 2000 The Phonology of Norwegian. Oxford: Oxford University Press. Kügler, Frank 2004 The phonology and phonetics of nuclear rises in Swabian German. In: Gilles and Peters (eds.), 75⫺98. Kügler, Frank 2007 The Intonational Phonology of Swabian and Upper Saxon. Tübingen: Niemeyer. Künzel, Hermann J. and Jürgen Erich Schmidt 2001 Phonetische Probleme bei Tonakzent 1. In: Angelika Braun (ed.), Beiträge zu Linguistik und Phonetik. Festschrift für Joachim Göschel zum 70. Geburtstag, 421⫺439. (Zeitschrift für Dialektologie und Linguistik. Beiheft 118.) Stuttgart: Steiner. Ladd, D. Robert 2008 Intonational Phonology. 2nd, revised edition. Cambridge: Cambridge University Press. Laver, John 1994 Principles of Phonetics. Cambridge: Cambridge University Press. Leemann, Adrian and Beat Siebenhaar 2007 Intonational and temporal features of Swiss German. In: Jürgen Trouvain and William J. Barry (eds.), Proceedings of the 16th International Congress of Phonetic Sciences (ICPhS XVI), Saarbrücken 6⫺10 August 2007, 957⫺960. Saarbrücken: Universität des Saarlandes. Available from . Leemann, Adrian and Beat Siebenhaar 2008a Perception of dialectal prosody. In: J. Fletcher, D. Loakes, R. Gocke, D. Burnham and M. Wagner (eds.), Proceedings of Interspeech 2008, 524⫺527. Brisbane: ISCA. Leemann, Adrian and Beat Siebenhaar 2008b Swiss Alpine and Midland intonation. In: Plinio A. Barbosa (ed.), Proceedings of Speech Prosody 2008, 289⫺292. Aix-en-Provence: Laboratoire Parole et Langage. Available from ; last accessed 16 November 2008. Low, Ling E. E., Esther Grabe and Francis Nolan 2000 Quantitative characterizations of speech rhythm: Syllable-timing in Singapore English. Language and Speech 43(4): 377⫺401. Moulton, William G. 1986 Sandhi in Swiss German dialects. In: Henning Andersen (ed.), Sandhi Phenomena in the Languages of Europe, 385⫺392. Berlin/New York: de Gruyter.

43. Areal variation in prosody Nespor, Marina and Irene Vogel 2007 Prosodic Phonology. 2nd ed. Berlin/New York: de Gruyter. Nı´ Chasaide, Ailbhe and Christer Gobl 2004 Voice Quality and f0 in Prosody: A Holistic Account. In: Rüdiger Hoffmann and Hansjörg Mixdorff (eds.), Speech Prosody 2004, 189⫺196. Dresden: TUDpress/Verlag der Wissenschaften. [Also available from ; last accessed 16 November 2008.] Nilsen, Randi Alice 2001 “Borderline cases”. Tonal characteristics of some varieties of spoken south Norwegian. In: Wim A. van Dommelen and Thorstein Fretheim (eds.), Nordic Prosody. Proceedings of the VIIIth conference, Trondheim 2000: 173⫺186. Frankfurt: Peter Lang. Nübling, Damaris and Renate Schrambke 2004 Silben- versus akzentsprachliche Züge in germanischen Sprachen und im Alemannischen. In: Elvira Glaser, Peter Ott and Rudolf Schwarzenbach (eds.), Alemannisch im Sprachvergleich. Beiträge zur 14. Arbeitstagung für alemannische Dialektologie in Männedorf (Zürich) vom 16.⫺18. 9. 2002, 281⫺320. Stuttgart: Steiner. Peters, Benno, Klaus J. Kohler and Thomas Wesener 2005 Phonetische Merkmale prosodischer Phrasierung in deutscher Spontansprache. In: Klaus J. Kohler, Felicitas Kleber and Benno Peters (eds.), Prosodic Structures in German Spontaneous Speech, 143⫺185. (AIPUK 35a.) Available from ; last accessed 16 November 2008. Peters, Jörg 2006a Intonation deutscher Regionalsprachen. (Linguistik ⫺ Impulse & Tendenzen 21.) Berlin/ New York: de Gruyter. Peters, Jörg 2006b Dialektintonation. Osnabrücker Beiträge zur Sprachtheorie 71: 179⫺203. Peters, Jörg 2006c The Cologne word accent revisited. In: de Vaan (ed.), 107⫺134. Peters, Jörg, Peter Gilles, Peter Auer and Margret Selting 2002 Identification of regional varieties by intonational cues. An experimental study on Hamburg and Berlin German. In: Language & Speech 45(2): 115⫺139. Pike, Kenneth L. 1945 The Intonation of American English. Ann Arbor: University of Michigan Publications. Ramus, Franck, Marina Nespor and Jacques Mehler 1999 Correlates of linguistic rhythm in the speech signal. Cognition 72: 1⫺28. Riad, Thomas 2003 The origin of Scandinavian tone accents. Diachronica 15: 63⫺98. Riad, Tomas and Carlos Gussenhoven (eds.) 2007 Tones and Tunes. Typological Studies in Word and Sentence Prosody. Berlin/New York: de Gruyter. Ruoff, Arno 1973 Grundlagen und Methoden der Untersuchung gesprochener Sprache. Tübingen: Niemeyer. Schmidt, Jürgen Erich 1986 Die Mittelfränkischen Tonakzente. Rheinische Akzentuierung. Stuttgart: Steiner. Schmidt, Jürgen Erich 2002 Die sprachhistorische Genese der mittelfränkischen Tonakzente. In: Peter Auer, Peter Gilles and Helmut Spiekermann (eds.), Silbenschnitt und Tonakzente, 201⫺233. (Linguistische Arbeiten 463.) Tübingen: Niemeyer. Schmidt, Jürgen Erich and Hermann J. Künzel 2006 Das Rätsel löst sich: Phonetik und sprachhistorische Genese der Tonakzente im Regelumkehrgebiet (Regel B). In: de Vaan (ed.): 135⫺163.

803

804

VIII. Structural domains: Methodological problems Selting, Margret 2000 Berlinische Intonationskonturen: “Die Treppe aufwärts”. Zeitschrift für Sprachwissenschaft 20: 66⫺116. Siebenhaar, Beat 2004 Berner und Zürcher Prosodie. Ansätze zu einem Vergleich. In: Elvira Glaser, Peter Ott and Rudolf Schwarzenbach (eds.), Alemannisch im Sprachvergleich. Beiträge zur 14. Arbeitstagung für alemannische Dialektologie in Männedorf (Zürich) vom 16.⫺18.9.2002, 419⫺437. (Zeitschrift für Dialektologie und Linguistik. Beiheft 129.) Stuttgart: Steiner. Sievers, Eduard 1912 Über Sprachmelodisches in der deutschen Dichtung. In: Eduard Sievers (ed.), Rhythmisch-melodische Studien, 56⫺77. Heidelberg: Winter. Spiekermann, Helmut 2000 Silbenschnitt in deutschen Dialekten. (Linguistische Arbeiten 425.) Tübingen: Niemeyer. Sudhoff, Stefan, Denisa Lenertova´, Roland Meyer, Sandra Pappert, Petra Augurzky, Ina Mleinek, Nicole Richter and Johannes Schließer (eds.) 2006 Methods in Empirical Prosody Research. (Language, Context and Cognition 3.) Berlin/ New York: de Gruyter. van Leyden, Klaske 2004 Prosodic Characteristics of Orkney and Shetland Dialects. An Experimental Approach. Utrecht: LOT. Willi, Urs 1996 Die segmentale Dauer als phonetischer Parameter von “fortis” und “lenis” bei Plosiven im Zürichdeutschen. Eine akustische und perzeptorische Untersuchung. (Zeitschrift für Dialektologie und Linguistik. Beiheft 92.) Stuttgart: Steiner. Zellner, Brigitte 1994 Pauses and the temporal structure of speech. In: Eric Keller (ed.), Fundamentals of Speech Synthesis and Speech Recognition, 41⫺62. Chichester: Wiley.

Peter Gilles, Walferdange (Luxembourg) Beat Siebenhaar, Leipzig (Germany)

44. Areal variation in morphology 1. 2. 3. 4. 5. 6. 7.

Introduction Morphological variation and its areal manifestation Morphology and the mental construction of space Effects of geographic space on morphology Methodological issues Explanatory power of the study of areal variation References

1. Introduction Morphology is the study of word structure. Morphology deals with the linguistic rendering of meaning in word formation and inflection. Morphological variation involves morphological variants which encompass formal units (words, affixes, etc.) and paradigmatic

44. Areal variation in morphology distinctions. For example, the 1sg pronouns mi in Czech and mi in Nigerian Pidgin English (Faraclas 1996: 179) are phonologically identical and of the same Indo-European origin. Yet, the paradigms to which these words belong are completely different: while mi in Nigerian Pidgin English performs all the syntactic functions of the 1sg pronoun, mi in Czech is only one form (dative, alongside mneˇ) in a rich case paradigm that also contains ja´ (nom), mne and meˇ (gen/acc), (o) mneˇ (prep) and mnou (instr). There is a broad transition zone between morphology and neighboring disciplines. In the Italian 1sg future tense of the verb, all inflectional features are coded as affixes in a single word, e.g., leggero` ‘I will read’, whereas in the corresponding English form I will read, a periphrastic construction composed of three words is needed. The English expression is thus morphosyntactic. In Russian the vowels [e] and [o] in the last syllable of the stem are sometimes deleted if a suffix is added, e.g., аок ‘castle’ (nom.sg), аки ‘castles’ (nom.pl) (cf. Haspelmath 2002: 27). The alternation of [o] and zero has both morphological and phonological properties and is, therefore, morphophonological (see section 5 for more on the delimitation problem). Areal variation refers to spatial variants of morphological variables within and between languages. Variation within languages is the topic of dialectology. As dialects of the same language share most of their morphological features, this article evidences, first of all, the differences displayed by dialect variants of a variable. For instance, in most varieties of English the subject pronoun they is opposed to the object pronoun them, but in Tyneside English them realizes both functions and is of special interest here (see section 3). With respect to the variation between languages, in this article, similarities between the languages compared are pointed out. Such similarities may be a consequence of language universals, of common historical origins (e.g., as in the above-mentioned Indo-European 1sg subject pronouns), of chance convergences, or they may be geographically conditioned: consider the sprachbund characteristics of linguistic areas such as the Balkans, Meso-America or the Indian subcontinent (cf. Campbell 2006). An additional aspect in this context is the center⫺periphery pattern (cf. the notions “aree centrali vs. laterali” of the Italian neolinguistic school, Bartoli 1925), in the sense that structural similarities may be found at the margins of a linguistic area, regardless of whether or not the languages in question are closely genetically related (see section 4.2). Typically, areal variation is a feature of the horizontal dimension, i. e., with one variant close to another in space. This variation can be dynamic (diffusion of features within a geographical area) or static (geographic boundaries inhibit the diffusion of features). However, horizontal contrasts also reappear in the vertical dimension, i. e., two originally geographically separated variants are now present within the same area, but with different pragmatic or social values. Nowadays this is often the case for dialect variants which are in competition with variants from standard languages that originated from other dialects (e.g., modern Standard French from Paris French, cf. Holtus 1990: 583). Vertical reinterpretations of horizontal contrasts increase the effect of language contact, originally restricted to spatially contiguous varieties. With respect to its speakers’ competence, variation within a language can be seen either as an inherent property of the language’s grammar or as the result of competition between different grammars. The first viewpoint is that of variationist sociolinguistics, the second is preferred by formal linguistics (cf. Meyerhoff 2006: 403; Henry 2002). Following the second assumption, variation is, essentially, removed from the grammar, in that variants are regarded as the result of code-switching between different grammars.

805

806

VIII. Structural domains: Methodological problems In the following sections, areal variation in morphology is discussed from various perspectives. However, this description cannot be complete. Hence, as mentioned above, two types of variation are singled out, and related examples will be presented in some detail: (i) similarities in the morphology of non-cognate languages and (ii) differences in the morphology of cognate languages/dialects whose relationships to geographic space is clear enough to be explicated.

2. Morphological variation and its areal maniestation Areal variation may be found in both aspects of morphological units, meaning and form. Meaning aspects encompass inflectional, derivational and compound meaning. However, derivational and compound meaning is much more lexeme or lexeme-group specific than inflectional meaning and can be much less easily generalized and classified (there are inflectional categories, but “derivational categories” are not usually assumed, cf. Haspelmath 2002: 60⫺61). Since the precondition for the analysis of the distribution of morphological variants is the stipulation of the variable, which must be identical for all languages under investigation, areal variation in morphology is predominantly studied with respect to inflectional features ⫺ as is attested, for instance, by the choice of features mapped in The World Atlas of Language Structure (WALS, Haspelmath et al. 2005). (Contrary to some other usages, inflectional category in this article refers to the inflectional dimension, e.g., “number”, whereas inflectional feature refers to the specific value, e.g., “singular”.)

2.1. Variation o morphologically coded meaning In the inflectional domain, the most general distinction between languages is the presence or absence of an inflectional category, i. e., whether or not a grammatical category is expressed solely by morphological means (i. e., within the boundaries of a single word, see also section 5). Think about distinctions of voice. Such distinctions change the relation that the surface subject has to the verb, and voice features (e.g., active, passive) are selected according to the perspective from which the action or event is viewed: in the active it is usually the doer, in the passive usually the object affected by the action, that is accentuated by being made the grammatical subject of the sentence. While such perspectivizations are possible in every language, only some languages allow the voicefeature expression to be inflectionally bound to the verb. Among the Germanic languages, Swedish and Norwegian do so: Norwegian uses the suffix -r for the active, e.g., han spiser ‘he eats’, and the suffix -s for the passive, e.g., maten spises ‘the food is eaten’. On the contrary, the cognate languages of German and English use periphrastic expressions composed of an auxiliary and a past participle for the passive voice. Hence, voice is an inflectional category in Swedish and Norwegian but not in German or English. Within the same inflectional category, different languages and dialects may display different sets of feature values. In Standard Italian and most Italian dialects there are two gender features, namely masculine and feminine. However, in some Umbrian dialects there is a kind of “neuter”, used mainly for mass nouns, that gives rise to a three-

44. Areal variation in morphology

807

Tab. 44.1: Paradigm distinctions for plural verbs in German dialects Cell

Type 1

Type 2

Type 3

Type 4

Type 5

1pl 2pl 3pl

A B C

A B B

A B A

A A B

A A A

way gender distinction in the article system: la fija ‘the daughter’ (fem); lu coccu ‘the egg’ (masc); lo granu ‘the grain’ (ntr) (Haase 1999: 89⫺91). Most Slavic languages have two number features (sg, pl) but in the Sorbian language island in eastern Germany the old three-way distinction of singular, dual and plural has been preserved in both the verbal and the nominal system, e.g., meˇsto ‘city’ (nom.sg), meˇsc´e (nom.dual), meˇsta (nom.pl) (Stone 1998: 180⫺183). The tense category of Standard German and northern German dialects features both present and preterite, e.g., ich komme ‘I come’ vs. ich kam ‘I came’. In most German dialects of the south, as well as in many other Germanic languages, the inflectional preterite forms have ⫺ with the exception of a few very frequent forms like ich war ‘I was’ ⫺ disappeared and have been substituted by a periphrastic perfect tense, ich bin gekommen ‘I have come’ (cf. Abraham and Conradie 2001). Features can be expressed by monoexponential or by cumulative markers. In IndoEuropean languages the expression is typically cumulative, e.g., in the Russian word uроков ‘lessons’ (gen.pl) the suffix -ов expresses both plural and genitive case. The cumulative expression is attested in many European languages but is very rare outside Europe: monoexponential formatives dominate everywhere (cf. Bickel and Nichols 2005: 92⫺93). A good example is Turkish. In the word yılların ‘years’ (gen.pl) all exponents are monofunctional: yıl expresses the semantic content ‘year’, -lar the plural and -ın the genitive case. However, monoexponential expression is not restricted to agglutinative languages. A particularly high degree of areal variation is observed in the field of paradigmatic distinctions, which lies halfway between the form and meaning aspects of morphological units. Despite the cases quoted above, the dialects of a language usually share the same inflectional categories with the same feature sets. But they may have very different paradigmatic distinctions. Whereas a single inflected word form may correspond to a single feature, this is not the only possibility. A single word form may also correspond to multiple feature values, i. e., syncretisms occur. Consider the syncretisms in the present plural of the verbs in Upper German dialects (cf. Rabanus 2004, 2006, 2008): Table 44.1 shows all possible distinctions between the extremes of “a distinct word form in each cell” (Type 1) and “just a single word form for all cells”, i. e., complete syncretism of person (Type 5). In the Upper German dialects all types except Type 4 are well attested. Type 1 is typical for a part of the Central Bavarian dialects, especially in the region of Lower Bavaria (Niederbayern, cf. Rabanus 2008: 233⫺238), and for the Highest Alemannic and Walser dialects in the Swiss Alps. Type 2 appears across a considerable area in northern Swabia and is spreading in the northeast Swabian dialects of Bavarian Swabia (Bayerisch Schwaben), around the city of Donauwörth (cf. Rabanus 2004: 345⫺349, 2008: 170⫺ 188). Type 3 characterizes most Bavarian and East Franconian dialects as well as the western High Alemannic ones (Rabanus 2008: 140⫺144), and it is the paradigm type of Standard German. Type 5 is the configuration of most Low Alemannic, Swabian and

808

VIII. Structural domains: Methodological problems eastern High Alemannic dialects. Type 4 appears only in rare and doubtful cases of present-day Upper German dialects (Rabanus 2008: 139). However, cross-linguistically, Type 4 is the most common plural syncretism on the global scale (cf. Cysouw 2003: 302). To sum up: there is a high degree of variation in paradigmatic distinctions, and paradigmatic distinctions contribute notably to the constitution of linguistic areas.

2.2. Variation in the orm o morphological units Meaning features, obviously, have different formal exponents across space. This holds for both the marker type (additive, modificatory, suppletive markers) and the phonological structure of the marker (different words, affixes). Not surprisingly, morphological features are often formally rendered with different words or affixes even in closely related dialects (see the example of the 2pl verb affixes in Bavarian and Standard German in section 3). An example of marker-type variation is found at the dialect border between Moselle Franconian and Rhine Franconian in western Germany. The noun plural in the Rhine-Franconian dialects is usually marked by an affix, such as the plural suffix -e: Stan ‘stone’ (sg) vs. Stane ‘stones’ (pl). In the Moselle Franconian dialects the number difference is often expressed by an intonational modificatory marker only: singular and plural forms have the same segmental structure but use different tone accents, e.g., Sta 2n ‘stone’ (sg) vs. Sta 1n ‘stones’ (pl) (Schmidt 1986: 136; the superscript numbers indicate two different tone accents). There is also variation in the word-internal order of morphological units. Leaving aside the problematic status of infixation or circumfixation, affixes generally occur before or after the lexical stem of the word. On a global scale, suffixes are preferred to prefixes (Dryer 2005: 112⫺113). The most consistent area with predominantly prefixing languages is found in a long stretch in the southeast of Africa. The area contains mainly Bantu languages, but there are also Niger-Congo and Nilo-Saharan languages. Where there is more than one affix, their order seems to universally follow the “Inflectionoutside-Derivation Principle” (cf. Stump 2001: 712⫺713), i. e., derivational affixes are closer to the base than inflectional affixes, as in English sweetens where -en is derivational and -s inflectional. Thus, despite some arguable counterexamples (Stump 2001: 712 draws attention to the multifunctional Russian suffix - я which follows the person and number marking, e.g., uчиш я ‘you study’ where -иш is the 2sg.pres suffix and - я is a verb-class marker), there is no areal variation with respect to the derivational versus inflectional affix order. In contrast, the order of head and modifier in nominal compounds can vary even in cognate languages or dialects of the same language. The nominal compounds in the South Bavarian language island of Deutsch-Fersentalerisch in northern Italy (Rowley 2003) are an example. Older speakers in particular apply the modifier⫺head order in compounding, which is common to almost all German dialects: in the noun⫺noun compound hennensto`ll ‘hen house’ the modifier hennen precedes the head sto`ll. However, younger speakers prefer the head⫺modifier order, i. e., the modifier (preceded by preposition and article, if necessary) follows the head. Thus, younger speakers use sto`ll va den hennen ‘hen house’ instead of hennensto`ll (Rowley 2003: 241). This variation (some might say that sto`ll va den hennen is not a compound any more but an NP) is due to the intense present-day language contact between Deutsch-Fersentalerisch and Italian which is the majority language of the area. In Standard Italian and in

44. Areal variation in morphology the Italian dialects of the area the order in compounds is usually head⫺modifier, e.g., pesce spada ‘swordfish’, with the head pesce ‘fish’ preceding the modifier spada ‘sword’. Finally, the same marker may have different degrees of productivity in different dialects or cognate languages. Consider the noun plural suffix -er, which exists in German as well as in Danish. In German the plural suffix -er occurs only in a small class of native nouns and is no longer productive in new nouns (loan words, neologisms, derivations from native words of other word classes). On the contrary, in Danish -er is highly productive and used with new words, e.g., sushi ‘sushi’ (sg) vs. sushier ‘sushis’ (pl). The plural suffix -er is so productive that in colloquial Danish speakers even tend to reanalyze stem-final -er as in kilometer ‘kilometer’ (sg) as a plural suffix and form the new singular kilomet (Pia Quist, personal communication).

3. Morphology and the mental construction o space Perceived geographic space is a mental construct. Physical geography may play a role (see section 4) but only insofar as it conditions human behavior. Members of social communities establish the extent and the boundaries of their geographic space principally according to cultural and political criteria. Linguistic factors also matter. However, if people are asked to indicate the structural differences between their own language or dialect and the speech of others, they predominantly refer to phonological and lexical characteristics: they say that the others have a “different way of saying” the “same words” or that they use “different words” for the “same things”. Grammatical contrasts (e.g., different paradigmatic distinctions) are less accessible to the observations of naı¨ve speakers, and even segmentally different markers are less noted than slight phonological differences. Nevertheless, sometimes morphological marking is relevant for the mental construction of cultural geographic space. This is illustrated here with examples from dialects of German, English and Chinese. Consider first the Bavarian 2pl verb suffix -ts. Unlike Standard German in which the 2pl suffix -t is homonymous with the 3sg.pres suffix -t ⫺ resulting in numerous syncretisms (e.g., er/ihr geht ‘he/you go[es]’) ⫺ in the Bavarian dialects the suffix -ts exclusively codes the 2pl (er geht vs. ihr gehts). In contrast to other morphological features that have been replaced by Standard German forms in the last decades, -ts is a very resistant marker. It not only remains unchanged in dialect speech but is even attested above the dialect level, in colloquial regional German in Bavaria and Austria (cf. Scheuringer 1990: 343). Particularly instructive is the situation in Bavarian Swabia. Bavarian Swabia is a region which, linguistically, lies mostly in the Alemannic rather than the Bavarian dialect area, but politically belongs to Bavaria. Language-change studies (Renn 1994: 106⫺107; Kleiner 2003: 142⫺143) report that in recent times -ts has spread in the Alemannic dialects of Bavarian Swabia, replacing the Alemannic verb plural suffixes. The reason for this language change is the prestige of Bavarian dialect features compared to Alemannic ones. What matters here is that a morphological feature has become a geographical marker in the speakers’ perception. Speakers who are not linguistically “Bavarians” use this marker to identify themselves as politically and culturally belonging to Bavaria and to distinguish their area from the adjacent Alemannicspeaking regions. In this way, morphology is used to construct and delimit cultural space.

809

810

VIII. Structural domains: Methodological problems A similar case of marking geographical space morphologically is found in China. In Mandarin Chinese there is a class of semantically empty elements that attach to monosyllabic, nominal, countable bound roots (cf. Packard 2000). One of these empty elements is (-r). This “vacuous derivation” (cf. Pirani 2007: 41⫺42) is not dialect specific. However, has a special status in the city dialect of Beijing. In this dialect attaches not only to the above class of roots but to the vast majority of roots, regardless of their syllabic, word-class, or semantic features (cf. Pirani 2007: 145⫺147). The socalled “Beˇijı¯ng e´rhua`” (cf. Peng 2006) or “-erization” (cf. Barnes 1977) is an even stronger example of a morphological geographic marker than the Bavarian -ts because it bears no other meaning than the geographic one: it symbolizes the speaker’s Beijing origin. Since Beijing is China’s capital and the Beijing dialect is a very prestigious variety of Mandarin Chinese, also has a social value, i. e., it is a social and geographic marker. The third case is slightly different. The pronoun system of Tyneside English (and many other English varieties) is characterized by different instances of “pronoun exchange”: e.g., in the 3pl the object pronoun is also used in subject position, i. e., instead of a distinction between the subject pronoun they and the object pronoun them there is just one generalized form, them (cf. Beal 2004: 118⫺119). In the 1pl the functions of we and us are reversed in Tynside when compared to Standard English, i. e., us functions as subject and we as object pronoun, see (1a, b) (Beal 1993: 205): (1) a. Us’ll do it. b. Give it we. The feature is so salient and well known even outside northern England that it has become a shibboleth for Tyneside English. From a theoretical point of view it is instructive to note that here it is not the word form that serves as a geographical marker: the pronouns are the same words as in Standard English. It is the paradigmatic distinction that makes the difference, i. e., the syncretism of (subjective/objective) case or the exchange of subject and object function. To sum up: morphology is less important for the perception and mental construction of space than phonology and the lexicon. But there are examples of morphological space marking, including cases in which abstract paradigm distinctions are recognized as relevant for spatial structures by naı¨ve speakers, as exemplified by Tyneside English.

4. Eects o geographic space on morphology The point of departure for this section is the observation that linguistic phenomena usually exhibit well-formed spatial patterns. However, physical geographic space as such has no effect on morphology or any other linguistic feature. Correlations between the topographic and the linguistic structure of an area exist insofar as topographic space inhibits or supports communicative contacts between populations. In ancient times, mountains, swamps and large forests presented serious traffic barriers; consequently they became language and dialect boundaries. On the other hand, precisely because of the absence of good roads, rivers were more important traffic routes than they are today.

44. Areal variation in morphology Hence, rivers are hardly ever major linguistic boundaries. If they delimit linguistic spaces, this is due to additional border functions that they have adopted, e.g., political borders. Nation-state borders constrain communication. In today’s Europe with its “open frontiers” there are almost no restrictions on cross-border movements anymore, at least within the Schengen area. However, long-term or regularly repeated cross-border movements for purposes like education, work, or marriage are restrained by national economies and national cultures as well as by infrastructure development that is not usually internationally oriented. Thus, the majority of movements that, unlike holiday traffic, could seriously contribute to cultural and language contacts remain within the borders of the nation state. Religious boundaries have similar effects, even in Europe, as not only the recent history of the Balkan region has taught us. In sum: geographical space has effects on language only insofar as it conditions whether and what kind of language or dialect contact occurs. Language contact presupposes some kind of geographical adjacency. There are three possible types of adjacency which are addressed in the following subsections: isolation, i. e., no adjacency (section 4.1); physical, i. e., topographic adjacency (section 4.2); social adjacency (section 4.3).

4.1. Eects o isolation If a language is completely isolated, no language contact occurs. This is, historically, the situation for language islands in topographically remote locations. The effect of the isolation on (the rare cases of almost) homogeneous languages or dialects is often the preservation of conservative morphological features that have died out in related languages or dialects on the “mainland”. Good examples are the Alemannic and Bavarian language islands in the Italian Alps. Consider the Walser German dialect of Issime, a small village in the Aosta Valley region in northern Italy, which for centuries remained isolated both from the German language area and the surrounding Italian and FrancoProvenc¸al dialects. In the inflection of some nouns distinct genitive case forms still exist both in the singular and the plural, e.g., wegsch ‘way’ (gen.sg), we egu ‘ways’ (gen.pl) (Zürrer 1999: 156⫺157), whereas in almost all other German dialects the (Old High German) genitive case forms have completely disappeared from the noun paradigms (cf. Koß 1983: 1242) while the surrounding Romance varieties do not have morphological case on nouns at all. However, there are examples in which isolation has the opposite effect, i. e., language islands in which case reduction goes beyond the level reached by the dialects of the core territory. Here the development is also independent of the structure of the surrounding languages. In Pennsylvania German simplifying effects of language contact (English has fewer cases than German) cannot be excluded (cf. Louden 1994: 90). However, Rosenberg (2005: 229) reports that “case reduction is stronger in ‘sectarian’ Mennonite and Amish groups who consistently use German in everyday conversation than in ‘non-sectarian’ groups with intense language contact”. Case reduction is also attested in German language islands in Russia, despite the rich case morphology of Russian. What both types (case reduction and case preservation) have in common is that topographically or culturally induced spatial isolation has effects on morphology: the morphology of the language-island varieties is different to the morphology of the related “mainland” varieties.

811

812

VIII. Structural domains: Methodological problems

4.2 Eects o topographic adjacency Topographic adjacency is the typical situation for language contact in Western Europe with its dialect and language continuums based on rural, non-mobile societies. Often, one variety influences the variety next to it in space. In this situation, linguistic innovations might spread out horizontally from the margins of a linguistic area, and represent a pattern which is sometimes referred to as “contagious diffusion” (Bailey et al. 1994: 366). The magnitude of the influence is a function of geographical distance between two places: the greater the distance the less similar the varieties. Topographic adjacency is also the necessary precondition for the occurrence of sprachbund phenomena. Probably the best-studied sprachbund is the Balkan Sprachbund that features, among others, morphological similarities. Bulgarian and Macedonian are the only Slavic languages which have a (definite) article (cf. Haspelmath 2001: 1494; see also section 5). This is the result of their geographical position on the Balkan peninsula between Albanian, Greek and Romanian, i. e., in between languages with article systems. The effect of topographic adjacency is evidenced as well by the postposition of the definite article common to all of these topographically contiguous but genetically only distantly related languages (cf. Feuillet 2001: 1518; Trudgill 1975: 239). The article system is, additionally, a good example of the center⫺periphery pattern. Both a definite and an indefinite article exist in the center of the so-called Standard Average European (SAE) linguistic area whose nucleus is formed by French and German (cf. Haspelmath 2001: 1505). Icelandic, although a Germanic language, has only a definite article. Thus, it resembles the nearby Celtic languages but is also similar to Bulgarian and Macedonian, languages which are located on the opposite side of the SAE center. Another well known sprachbund with a common morphological property is that of East-Central Africa. African languages do not usually display morphological case on nominals. Exceptions are Nilo-Saharan and Cushitic languages which have up to seven case forms. These languages cluster in East-Central Africa (mainly Sudan, Ethiopia and some Central African areas, cf. Iggesen 2005: 204⫺205) and form the so-called “ChadEthiopia” sprachbund (Heine 1975: 41⫺43). A special type of topographic adjacency involves the contact between two languages simultaneously present in the same geographical space. Consider the linguistic situation in Kazakhstan, in which Kazakh, the local Turkic language, has been in intense contact with Russian for centuries (similar situations exist in other countries which were part of the former Soviet Union and the Russian Empire). Until the end of the Soviet Union, there were an almost equal percentage of ethnic Russians and Kazakhs in Kazakhstan, and Russian today is still an official language that plays an important role, especially in business, with many bilingual speakers. In a recent study (Muhamedowa 2005), the effect of this intense language contact on Russian adjective inflection is discussed. Russian adjectives, like nouns, display a rich inflectional paradigm; in attributive function they agree in case, number and gender with the head of the NP. In contrast, Kazakh adjectives do not agree with the head in any of these categories, their syntactic function as attributive adjectives is marked exclusively by their position in the sentence (cf. Muhamedowa 2005: 269⫺271). The study shows that in code-mixing contexts, when Russian NPs are embedded in Kazakh sentences, the morphology of many Russian adjectives is simplified, i. e., they partially lose their formal agreement features. However, even in the purely Russian utterances of even Russian-dominant bilingual speakers, a small but consider-

44. Areal variation in morphology able proportion of attributive adjectives (seven percent of the non-masculine NPs, cf. Muhamedowa 2005: 304) do not formally display gender agreement. The adjective is “neutralized” to the masculine form, as illustrated in (2) (Muhamedowa 2005: 294; glossed according to the Leipzig Glossing Rules): (2) кажд- every-masc.nom.sg

ациоало т nationality (fem)[nom.sg]

While the noun ациоало т ‘nationality’ is intrinsically feminine, the adjective takes the “neutral” default suffix - (masc.nom.sg). The resulting NP is, obviously, not well-formed in Standard Russian. Hence, the spatially mediated language contact clearly has effects on morphology. The reduction of gender and number agreement under language-contact conditions can also be found in Paraguayan Spanish. In Paraguay, the regional Spanish variety is in competition with Guaranı´ in a situation of general bilingualism. In contrast to the situation in the neighboring Spanish- and Portuguese-speaking countries, Guaranı´ is an official language in Paraguay and still has a relatively strong position: in rural areas there are even speakers who know very little Spanish (cf. Dietrich 2001). Hence, Guaranı´ is influenced by Spanish, but regional Spanish is influenced by Guaranı´, too. In Guaranı´, number and gender are not marked as morphological categories. As a consequence, in the Paraguayan Spanish sentence la banana son buenas ‘the bananas are good’ (Dietrich 2001: 69) neither the article la nor the noun banana bear any plural marker (the sentence would be considered ungrammatical by Spanish speakers from other regions). Another very interesting example is the use of perfective aspect markers in Paraguayan Spanish (cf. Dietrich 2001: 68). Guaranı´ has aspect suffixes, Standard Spanish does not. Regional Spanish adopts perfective aspect markers in two different ways, which are illustrated in (3): (3) a. su novia-kue his girlfriend-perf ‘his ex-girlfriend’ b. voy a comprar para mi vestido I go to buy perf for me a dress ‘I will buy me a dress’ In (3a) the Guaranı´ perfective suffix -kue is transferred into Spanish preserving both its phonological form and its grammatical meaning (“state or activity which was ended”). In (3b) the Spanish preposition para adopts the meaning of the Guaranı´ perfective suffix -ra˜ (‘state or activity which will be ended’) assigning to (3b) a perfective meaning without transferring the original Guaranı´ suffix itself ⫺ no speaker of Spanish from Madrid or Buenos Aires could assign any meaning to para in (3b). Example (3b) is especially instructive for language-contact theory as it shows the establishment of an aspect suffix in a language in which aspect suffixes were formerly unknown by reanalyzing a grammatical word from the target language without transferring the affix from the source language itself.

813

814

VIII. Structural domains: Methodological problems

4.3. Eects o social adjacency The idea of “social adjacency” is that areas sharing social features can be considered adjacent although they are not spatially contiguous. This holds, for instance, for cities. Many communicative contacts do not affect the countryside but are transmitted directly from city to city. The spread of features from larger cities to smaller cities and thence to rural areas is usually referred to as “hierarchical diffusion” or “cascade diffusion” (Bailey et al 1994: 361). Outside linguistics, a good example is the spread of the Beatles’ musical style from Liverpool to London and then to other major cities in the world. Only later did it move on downwards to smaller cities and towns (cf. Haggett 2004: 319). In linguistics, since Trudgill’s classic studies (1974, 1975), hierarchical diffusion has mainly been demonstrated in the spread of phonological features. For the opposite pattern, contrahierarchical diffusion, a morphological example exists. In the south of the United States the quasi-modal fixin to (e.g., I feel like I’m fixin to die ‘I feel like I am about to die’) has been reported to have spread rapidly throughout Oklahoma in the twentieth century (cf. Bailey et al. 1994: 371⫺378). The spread occured in a contra-hierarchical fashion from the rural areas to small towns and then to larger cities. Hence, it is not primarily topographic adjacency that drives the spread of the feature. The reason for the diffusion is rather the social value of the feature as a marker of local identity, in this case, a rural, explicitly non-urban identity.

5. Methodological issues There are two preconditions for the analysis of areal variation in morphology. The first is the determination of a variable which is valid for the whole area under investigation. The second is the availability of data on the areal distribution of morphological variants. These preconditions, in principle, apply to all structural domains, not just morphology. However, the meanings that frame morphological variables are particularly heterogeneous, and their definitions are very theory dependent. Additionally, the available data sources (linguistic atlases, lists of structural features with explicit areal reference points) usually do not contain complete descriptions of morphological systems but only a few selected features (whereas for phonology complete systems may be available). The descriptive situation of morphology has improved notably in recent years: while traditional dialect atlas projects have focused primarily on phonology and the lexicon and paid little to no attention to morphology (or syntax), nowadays things are more balanced. Morphological variation both within and between languages has been surveyed in several projects that have already produced good descriptive results. The Morfologische Atlas van de Nederlandse Dialecten ‘Morphological Atlas of the Dutch Dialects’ (MAND, De Schutter et al. 2005; Van Oostendorp et al. 2008) and The World Atlas of Language Structure (WALS) are just two examples of geolinguistic projects that concentrate solely on morphology (MAND) or in which morphological and syntactic maps far outnumber the phonological and lexical ones (WALS). However, a map that depicts the areal distribution of the variants of a morphological feature will still contain large amounts of white space.

44. Areal variation in morphology

815

The data problem is common to all structural domains to varying degrees. Specific to morphological analysis is the variable-determination problem. It is relatively simple to define a phonological variable (e.g., “vowel nasalization”) or a lexical variable (e.g., expressions for “hand” and “arm”). It is also possible to fix an inflectional variable (e.g., “plural”). But meanings regarding derivation and, to an even greater extent, compounding cannot be defined and classified in this way, since the number of relational concepts that can underlie a compound is potentially infinite (cf. Downing 1977). A successful survey of, for instance, the areal distribution of the possible interpretations of English noun⫺noun compounds like plate length seems an almost impossible task. For this reason, derivational and compound features are usually not studied with respect to areal variation. Another specific problem, already addressed in section 1, is the delimitation of the structural domains: it is not easy to decide where morphology ends and phonology or syntax begins. Turning first to the morphology⫺phonology transition zone (cf. Wiese 1996), in section 1 the example of vowel deletion in Russian noun inflection was mentioned. A similar morphophonological phenomenon is German umlaut. Originally, umlaut or i-mutation referred to the fronting of back vowels caused by assimilation to an original [i] in the following syllable, as illustrated in (4a): (4) a. lamb Lamm lamb[sg] b. muoter Mutter mother[sg]

lemb-ir Lämm-er lambs\pl-pl

Old High German New High German

muoter Mütter mothers\pl

Old High German New High German

The Old High German umlaut in (4a) could be regarded as an automatic phonetic process. But since it was primarily caused by [i] in morphological markers (here the plural suffix -ir) the umlaut began to participate in the symbolization of the morphological feature. In the New High German plural word Lämmer one could still argue that the plural suffix -er is simply an additive exponent and that the umlaut of the stem vowel is the fossilized result of the Old High German i-mutation, rather than a second exponent of the plural feature. However, the umlaut was later extended to noun plurals that never carried -ir or any other plural suffix, as in Mütter in (4b) from Old High German muoter ‘mothers’ (pl). Hence, the fronting of the stem vowel is morphologically conditioned; the vowel is fronted in order to express the plural feature. The morphology⫺syntax relationship is no easier. In section 4.2, reference was made to the definite article in Bulgarian. In fact, the situation is more complicated there, since what is labeled an “article” is not an autonomous word but an element added to the first word of the NP, as illustrated in (5): (5) a. кига book b. ова кига new book

кига-та book-def ова-та кига new-def book

(Hill 1991: 314)

816

VIII. Structural domains: Methodological problems The element -та in (5) could also be regarded as a suffix expressing definiteness. Perhaps it is best classified as a clitic since it exhibits a certain degree of freedom of selection with respect to the host ⫺ the noun in (5a) and the adjective in (5b). Clitics are in many ways intermediate between affixes, which belong to the domain of morphology, and free word forms, whose behavior in the sentence is governed by syntactic rules. In addition to the determination of variables discussed above, compounding is also problematic with respect to the domain delimitation problem. Spelling differences such as English flower pot, flower-pot, flowerpot are not simply issues of orthographic codification but rather indicators of serious word delimitation problems. These problems are more obvious in languages like Chinese which in written form never use blank spaces between characters (cf. Haspelmath 2002: 148).

6. Explanatory power o the study o areal variation In the foregoing sections, the effects of morphological features on (the construction of cultural) geographic space (section 3) and the effects of geographic space on morphological features (section 4) have been discussed. It has been argued that morphological differences create spatial delimitations and vice versa. Taking the issue one step further, we can now ask: what does the geographic reach of a morphological feature tell us about the properties of that feature? In the first instance, areal variability reveals that the phenomenon under analysis is not a linguistic universal. Geographic factors do not affect linguistic universals like the “Inflection-outside-Derivation Principle” (see section 2.2). If areal variation exists, then the question of its explanatory power can be framed with regard to its actual extent and its diachronic development. The actual extent of the morphological expression of a grammatical category, then, may tell us something about the relevance of this variable in the speaker’s mind. Bybee’s (1985) cross-linguistic survey of morphological categories of verbs revealed that in ninety percent of the languages examined, morphological expression of valence was attested whereas person and number were morphologically coded in only little more than half of the sample (cf. Bybee 1985: 29⫺33). Bybee did not depict the spatial extensions of the coding patterns. But since she included a remarkable tally of fifty languages, each chosen from a different language group and cultural area so as to create a representative sample, a map of the world’s languages can be drawn on the basis of her results. The almost complete coverage of the world with valence-marking languages ⫺ the ten percent which comprise the exceptions are scattered between Haitian, Serbo-Croatian and Vietnamese ⫺ proves the high degree of mental relevance of this category. It seems that speakers consider valence features (transitivity, intransitivity, causation, etc.) so important for the semantic content of the verb that they universally tend to express them morphologically. Compared with that, number and person features are less relevant for verbs; their morphological marking on verbs can be more readily omitted. Hence, a picture of areal distribution reveals a property which may otherwise have escaped notice. The conclusions that can be drawn from the actual spatial extent of the realization of a morphological variable depend on the representativity of the language sample. Even well-composed samples are never free of biases: it is almost inevitable that some areas are

44. Areal variation in morphology

817

overrepresented and others are missing (cf. Bybee 1985: 25). That is why the diachronic perspective should not be forgotten. The diachronic view is more independent of the samples’ representativity, since regardless of whether the phenomenon under investigation is representative or not for a language group or area, its diffusion in time from a point X to a point Y offers reliable information on the factors governing the organization and reorganization of morphological systems. Consider the plural-verb paradigm of Type 3 in the western High Alemannic (mostly Swiss German) dialects, depicted in Table 44.1 in section 2.1. A two-way distinction with homonymy of the 1pl and 3pl but a distinct form for the 2pl is cross-linguistically a very rare phenomenon (cf. Cysouw 2003: 134, 302). However, in southwest Germany its area is found to be expanding. During the twentieth century Type 3 spread from a very small area at the Swiss border towards the northeast in the former Type 5 area (cf. Rabanus 2004: 342⫺345). Since, as discussed, person marking is not highly relevant for verbs but is nonetheless a feature of Standard German, the change in the dialect implies convergence towards the standard language. Hence, the geographic mapping of diachronic changes confirms changes in the vertical dimension, i. e., an increasing importance of the standard language. Finally, the study of variation in small areas, i. e., dialect variation, can contribute to the development of linguistic theory in that it evidences aspects of “microvariation”. Table 44.2 shows the distribution of the 3sg present suffix -t in four preterite-present verbs in the Walser German villages in the Italian and Swiss Alps (data presented by Silvia Dal Negro in her contribution to the Verona workshop on German dialect morphology and syntax, 14 December 2008: cf. Rabanus, Alber and Tomaselli 2008; also cf. Dal Negro 2004: 205). Preterite-present verbs in Standard German and most German dialects are a small group of modal verbs which have acquired a present meaning from a preterite form but which preserve some formal properties of the preterite, e.g., they do not take the suffix -t in the 3sg present. In the Walser German dialects -t is added in a process of regularization. However, the process is not complete. A look at the areal distribution pattern reveals an implicational scale: dialect variants of mögen are the first that take the suffix -t while the occurrence of -t in the 3sg present forms of können implies that the suffix -t is already also used in wissen, dürfen and mögen (cf. the syntactic implicational scales in the same dialects in Dal Negro 2000: 47). Once again, areal distribution patterns tell us something about morphology ⫺ here about the readiness of preterite-present verbs to give up their formal irregularity.

Tab. 44.2: 3sg present suffix -t in the paradigms of preterite-present verbs Location Formazza Salecchio Bosco Gressoney Alagna Macugnaga Issime Rima Rimella

mögen ‘be willing’

dürfen ‘be allowed’

wissen ‘know’

können ‘can’

-t -t -t -t -t -t

-t -t -t -t

-t -t -t

-t -t -t

818

VIII. Structural domains: Methodological problems In section 5, the data situation was listed among the problematic aspects in the study of areal variation in morphology. In order to conclude with a positive outlook, this lack of information needs to be seen as an incentive for further research: the “white spaces” on a map of variants of a morphological variable are exactly the areas whose morphology awaits exploration!

7. Reerences Abraham, Werner and Jac Conradie 2001 Präteritumschwund und Diskursgrammatik. Amsterdam: Benjamins. Bailey, Guy, Tom Wikle, Jan Tillery and Lori Sand 1994 Some patterns of linguistic diffusion. Language Variation and Change 5: 359⫺390. Barnes, Dayle 1977 To er or not to er. Journal of Chinese Linguistics 5: 211⫺236. Bartoli, Matteo 1925 Introduzione alla neolinguistica. Principi, scopi, metodi. Geneva: Olshki. Beal, Joan 1993 The grammar of Tyneside and Northumbrian English. In: James and Lesley Milroy (eds.), Real English: The Grammar of English Dialects in the British Isles, 187⫺213. London: Longman. Beal, Joan 2004 English dialects in the north of England: Morphology and syntax. In: Bernd Kortmann and Edgar W. Schneider (eds.), A Handbook of Varieties of English, vol. 2: Morphology and Syntax, 114⫺141. Berlin: Mouton de Gruyter. Bickel, Balthasar and Johanna Nichols 2005 Exponence of selected inflectional formatives. In: Haspelmath et al. (eds.), 90⫺93. Brown, Keith (ed.) 2006 Encyclopedia of Language and Linguistics. 2nd ed. Amsterdam: Elsevier. Bybee, Joan L. 1985 Morphology. A Study of the Relation between Meaning and Form. Amsterdam: Benjamins. Campbell, Lyle 2006 Areal linguistics. In: Keith Brown (ed.), vol. 1: 454⫺460. Cysouw, Mıˆchael 2003 The Paradigmatic Structure of Person Marking. Oxford: Oxford University Press. Dal Negro, Silvia 2000 Altertümlichkeit, Sprachwandel und Sprachtod. Das Gleichnis vom “Verlorenen Sohn” in zwei piemontesischen Walserdialekten. Zeitschrift für Dialektologie und Linguistik 67(1): 28⫺52. Dal Negro, Silvia 2004 The Decay of a Language. The Case of a German dialect in the Italian Alps. Bern: Lang. De Schutter, Georges, Boudewijn van den Berg, Ton Goeman and Thera de Jong 2005 Morfologische Atlas van de Nederlandse Dialecten (MAND), vol. 1. Amsterdam: Amsterdam University Press. Dietrich, Wolf 2001 Zum historischen Sprachkontakt in Paraguay: Spanische Einflüsse im Guaranı´, Guaranı´Einflüsse im regionalen Spanisch. In: Gerda Haßler (ed.): Sprachkontakt und Sprachvergleich, 53⫺73. Münster: Nodus. Downing, Pamela 1977 On the creation and use of English compound nouns. Language 53(4): 810⫺842.

44. Areal variation in morphology Dryer, Matthew S. 2005 Prefixing versus suffixing in inflectional morphology. In: Haspelmath et al. (eds.), 110⫺ 113. Faraclas, Nicholas G. 1996 Nigerian Pidgin. London: Routledge Chapman & Hall. Feuillet, Jack 2001 Aire linguistique balkanique. In: Haspelmath et al. (eds.), 1510⫺1528. Haase, Martin 1999 Dialektdynamik in Mittelitalien. Sprachveränderungsprozesse im umbrischen Apenninenraum. Tübingen: Stauffenburg. Haggett, Peter 2004 Geografia, vol. 1: Geografia Umana. Translated by Alfredo Suvero. Bologna: Zanichelli. Haspelmath, Martin 2001 The European linguistic area: Standard Average European. In: Haspelmath et al. (eds.), vol. 2: 1492⫺1510. Haspelmath, Martin 2002 Understanding Morphology. London: Arnold. Haspelmath, Martin, Matthew S. Dryer, David Gil and Bernard Comrie (eds.) 2005 The World Atlas of Language Structure. Oxford: Oxford University Press. Haspelmath, Martin, Ekkehart König, Wulf Oesterreicher and Wolfgang Raible (eds.) 2001 Language Typology and Language Universals: An International Handbook. (Handbücher zur Sprach- und Kommunikationswissenschaft 20.) Berlin: de Gruyter Heine, Bernd 1975 Language typology and convergence areas in Africa. Linguistics 144: 27⫺47. Henry, Alison 2002 Variation and syntactic theory. In: Jack C. Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.), The Handbook of Language Variation and Change, 267⫺282. Malden, MA: Blackwell. Hill, Peter 1991 Das Bulgarische. In: Rehder (ed.), 310⫺325. Iggesen, Oliver A. 2005 Number of cases. In: Haspelmath et al. (eds.), 202⫺205. Holtus, Günter 1990 Französisch: Gliederung der Sprachräume. In: Günter Holtus, Michael Metzeltin and Christian Schmitt (eds.), Lexikon der Romanistischen Linguistik, vol. 5.1: Französisch, 571⫺595. Tübingen: Niemeyer. Kleiner, Stefan 2003 Bairisches in der Regionalsprache von Bayerisch-Schwaben: Die Übernahme des Flexionssuffixes {-ts} für die 2. Person Plural. In: Edith Funk, Stefan Kleiner, Manfred Renn and Bernadette Wecker (eds.), Sprachgeschichten. Ein Lesebuch für Werner König zum 60. Geburtstag, 137⫺152. Heidelberg: Winter. Koß, Gerhard 1983 Realisierung von Kasusrelationen in den deutschen Dialekten. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke, Herbert Ernst Wiegand (eds.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung, vol. 2, 1242⫺1250. (Handbücher zur Sprach- und Kommunikationswissenschaft 1.2.) Berlin: de Gruyter. Louden, Mark L. 1994 Syntactic change in multilingual language islands. In: Nina Berend and Klaus J. Mattheier (eds.), Sprachinselforschung. Eine Gedenkschrift für Hugo Jedig, 73⫺91. Frankfurt: Lang. Meyerhoff, Miriam 2006 Syntactic variation. In: Brown (ed.), vol. 12, 402⫺404.

819

820

VIII. Structural domains: Methodological problems Muhamedowa, Raihan 2005 Kasachisch-russisches Code-Mixing: Ein Fall von morphologischer Vereinfachung. Zeitschrift für Sprachwissenschaft 24(2): 263⫺309. Packard, Jerome L. 2000 The Morphology of Chinese: A Linguistic and Cognitive Approach. Cambridge: Cambridge University Press. Peng Zongping 2006 [Beijinghua Erhuaci Yanjiu ‘Study of the words with -er in the language of Beijing’]. Beijing: Beijing Guangbo Xueyuan Chubanshe. Pirani, Laura 2007 Morphological processes in the Mandarin nominal domain. Unpublished Ph.D. thesis, Universita` degli Studi di Verona. Rabanus, Stefan 2004 Morphological change in German dialects. Two cases of plural verbs in Alemannic. In: Britt-Louise Gunnarsson et al. (eds.), Language Variation in Europe. Papers from the Second International Conference on Language Variation in Europe, ICLaVE 2: Uppsala University, Sweden, June 12⫺14, 2003, 339⫺352. Uppsala: Universitetstryckeriet. Rabanus, Stefan 2006 An der Schnittstelle von Morphologie und Syntax: Einheitsformen der Personalpronomen der 1. und 2. Person Plural im Nordbairischen. Zeitschrift für Dialektologie und Linguistik 73(3): 301⫺317. Rabanus, Stefan 2008 Morphologisches Minimum. Distinktionen und Synkretismen im Minimalsatz hochdeutscher Dialekte. Stuttgart: Steiner. Rabanus, Stefan, Birgit Alber and Alessandra Tomaselli 2008 Erster Veroneser Workshop “Neue Tendenzen in der deutschen Dialektologie: Morphologie und Syntax”, 13.⫺14. Dezember 2007. Vorschläge für die Ausrichtung zukünftiger Dialektsyntaxprojekte. Zeitschrift für Dialektologie und Linguistik 75(1): 72⫺82. Rehder, Peter (ed.) 1998 Einführung in die slavischen Sprachen. Darmstadt: Wissenschaftliche Buchgesellschaft. Renn, Manfred 1994 Die Mundart im Raum Augsburg. Untersuchungen zum Dialekt und zum Dialektwandel im Spannungsfeld großstädtisch-ländlicher und alemannisch-bairischer Gegensätze. Heidelberg: Winter. Rosenberg, Peter 2005 Dialect convergence in the German language islands (Sprachinseln). In: Peter Auer, Frans Hinskens and Paul Kerswill (eds.), Dialect Change: Convergence and Divergence in European Languages, 221⫺235. Cambridge: Cambridge University Press. Rowley, Anthony R. 2003 Liacht as de sproch. Grammatik des Deutsch-Fersentalerischen. Palu` del Fe`rsina (Trento): Istituto culturale mo`cheno-cimbro. Scheuringer, Hermann 1990 Sprachentwicklung in Bayern und Österreich. Eine Analyse des Substandardverhaltens der Städte Braunau am Inn (Österreich) und Simbach am Inn (Bayern) und ihres Umlandes. Hamburg: Buske. Schmidt, Jürgen Erich 1986 Die mittelfränkischen Tonakzente (Rheinische Akzentuierung). Stuttgart: Steiner. Stone, Gerald 1998 Das Obersorbische. In: Rehder (ed.), 178⫺187. Stump, Gregory 2001 Affix position. In: Haspelmath et al. (eds.), vol. 1, 708⫺714.

45. Lexical variation in space

821

Trudgill, Peter 1974 Linguistic change and diffusion: Description and explanation in sociolinguistic dialect geography. Language in Society 3: 215⫺246. Trudgill, Peter 1975 Linguistic geography and geographical linguistics. In: Christopher Board, Richard J. Chorley, Peter Haggett and David R. Stoddart (eds.), Progress in Geography. International Reviews of Current Research, vol. 7, 227⫺252. London: Arnold. Van Oostendorp, Marc, Piet van Reenen, Oele Koornwinder, Boudewijn van den Berg and Ton Goeman 2008 Morfologische Atlas van de Nederlandse Dialecten (MAND), vol. 2. Amsterdam: Amsterdam University Press. Wiese, Richard 1996 Phonological versus morphological rules: On German Umlaut and Ablaut. Journal of Linguistics 32: 113⫺135. Zürrer, Peter 1999 Sprachinseldialekte. Walserdeutsch im Aostatal (Italien). Aarau: Sauerländer.

Stefan Rabanus, Verona (Italy)

45. Lexical variation in space 1. 2. 3. 4. 5. 6.

Lexical variation as a sociolinguistic variable Methodological requirements for sociolexicology Alternative strategies for dealing with the semantic problem Case studies Overview and prospects References

1. Lexical variation as a sociolinguistic variable Let us suppose that we try to treat lexical variation as a sociolinguistic variable, in the Labovian sense. What methodology would we have to use? In this article, we will explain why lexical variation presents a specific challenge for the quantitative study of linguistic variation, we will suggest that the challenge has not been adequately dealt with yet, and we will present an overview of various partial attempts to live up to the requirements that follow from the specific nature of lexical variation.

1.1. Spatial variation and lectal variation The emphasis of this article will be on lexical variation as a generic sociolinguistic phenomenon rather than on spatial lexical variation per se. Specifically, the concrete details of the traditional methods of dialect lexicology will hardly be touched upon. There are two reasons for this generic focus.

822

VIII. Structural domains: Methodological problems In the first place, diatopic variation is an integral part of the larger canvas of social variation in language use, systematically interacting with the other dimensions of sociolinguistic variation. As is well known, dialect switching is sensitive to situation; see, for instance, Werlen (2005) for an example involving rural dialects. More generally, the traditional dialect is only “the local language” because it happens to be the language of a locally dominant or well-established group. Move the so-called local language to a different locale and it reveals its essential nature as a marker of group identity, as studies on immigrant language show. In this sense, spatial dimensions are no different to other variational dimensions, and need to be treated on a par. In the second place, diatopic variation itself is a generic phenomenon that extends beyond traditional dialect variation ⫺ in two directions, in fact: the scale of the geographic variation may be broader than the dialect, as when we consider regiolects or the national varieties of standard languages, or it may be more restricted, as when unexpectedly local forms of spatial variation are detected. A conspicuous example of the latter situation is provided by Tway (1975), who shows that topology may influence language even on such a small scale as a factory. Distinguishing between the front and the back sections of a pottery factory in Pennsylvania (a functional distinction, since most of the activities associated with forming and firing take place at the back of the plant, while decorating and warehousing take place at the front), she illustrates how different objects receive a different conventional name according to the functional division of the factory. One ware shape may be called platter in the back section but dish in the front section; similarly, oatmeal versus nappy, museum versus coupe. The phenomenon is not restricted to ware shapes, as witnessed by the distribution of ware buggy versus upright, high lift versus truck, or inspector versus selector.

1.2. The speciicity o lexical variation as a sociolinguistic variable Given, then, the necessity of a generic perspective on lexical variation, why is lexical variation specific if we want to treat it as a sociolinguistic variable? Let us first recall what we mean by a sociolinguistic variable. Put simply, a sociolinguistic variable in the sense of contemporary sociolinguistics (see Labov 1966) is a set of alternative ways of expressing the same linguistic function or realizing the same linguistic element, where each of the alternatives has social significance: “Social and stylistic variation presuppose the option of saying ‘the same thing’ in several different ways: that is, the variants are identical in reference or truth value, but opposed in their social and/or stylistic significance” (Labov 1972: 271). As such, a sociolinguistic variable is a linguistic element that is sensitive to a number of extralinguistic independent variables like social class, age, sex, geographical group location, ethnic group, or contextual style and register. Classical cases of sociolinguistic variables involve pronunciation. Pronouncing the t in butter as a glottal stop is indicative of a Cockney accent, just like a full pronunciation of the n in chemin is typical of southern French in contrast to Standard French. Examples like these have long been studied in traditional dialectology, but modern sociolinguistics enlarged the scope of investigation beyond the traditional dialects to encompass other “lects” (as a cover term for dialects, sociolects, regiolects and the like, i. e., socially defined language varieties, like class-related speech or contextual styles).

45. Lexical variation in space Now, to see why lexical variation is specific as a sociolinguistic variable, let us consider how we choose words. Lexical choices in discourse may be determined by different factors: next to (obviously) the topic of the text, there is variation of a sociolinguistic or stylistic nature: a speaker of British English will, for instance, make different lexical choices than a speaker of American English, at least for a number of concepts. Choices of this kind often involve differences between language varieties. Lexical choices of this type are not choices of specific concepts (like the topic-related choices would be); they are choices of one word over another that expresses the same concept: we can recognize American English when we come across the word subway in contrast to British English underground, but the type of public transport referred to is the same. We will call this type of variation (i. e., the subway/underground type) formal onomasiological variation (FOV), in contrast to conceptual onomasiological variation (COV), which involves thematic choices, like talking about public transport rather than beer, biology, or Bach. For the third type of variation, we will use the term speaker and situation related variation (SSV). It covers all relevant features of the speech situation: not just the lectal variation that comes with more or less permanent speaker characteristics (like being American or British), but also the more transient, interactional characteristics of the communicative context, like whether the speech event is a dialogue or a monologue. The interaction between these various dimensions has not yet been systematically investigated. Such interactions may surely be expected, and they are likely to work in different directions. The underground/subway case is an example of an SSV-FOV interaction, as would be choosing an informal term rather than a more formal one according to the formality of the speech situation. But at the same time, the choice of an informal expression might correlate with thematic factors: there might well be more dirty words for dirty topics than colloquial words for scientific topics. While such a case would constitute a COV-FOV interaction, COV-SSV interactions may occur just as well. Conceptual choices, in fact, are not just determined by the topic of a text: rather, for a number of specific concepts, they derive from the situational, interactional characteristics of the communicative context. Second-person pronouns constitute an obvious example: they are typically likely to occur in dialogues rather than monologues. Similarly, persuasive texts contain different modal verbs than do informative texts. Let us consider a further example. In naming a specific garment worn on the lower part of the body and covering both legs separately, speakers using the general category name may choose between British English trousers and pants as used in the US, Canada and South Africa; but at the same time, they may make a conceptual choice for a more specific name, like jeans, rather than the general category name of trousers or pants. The lexical variation between trousers, pants and jeans does not just depend on sociostylistic factors (like whether you are talking to Americans or to Brits) but also on conceptual factors (like whether you opt for a more general or a more specific conceptual category). An SSV-FOV interaction, then, involves the classical type of sociolinguistic variable: the choice of pants or trousers correlates with specific sociostylistic features. An SSV-COV interaction occurs when the choice of specific concepts correlates with sociostylistic factors, such as when a professional or academic context invites more specialized language than an everyday context does. A COV-FOV interaction may be found when the choice of a particular expression correlates with conceptual factors, such as when the taboo-laden nature of a concept influences lexical choices. Given such interactions, the basic research question for sociolexicology can be defined as follows: What is the overall structure of lexical variation in terms of the relation-

823

824

VIII. Structural domains: Methodological problems ship between FOV, COV and SSV? The specific situation of sociolexicological research in the context of sociolinguistics follows straightforwardly from this question: while the paragon cases of sociolinguistic research involve formal variables and a binary relationship between formal variation and lectal context, sociolexicological research has to come to terms with a ternary relationship between form, meaning and context. In the next section, we will spell out in more detail the methodological consequences of this observation.

2. Methodological requirements or sociolexicology While sociolinguistic variables have readily been studied in the phonology and the grammar of languages, lexical variation has only rarely been studied as a sociolinguistic variable, even though lexical differences are conspicuous variety markers and, on a very practical level, often the source of misunderstandings. There are, we think, two methodological reasons for that neglect. These reasons are easy to identify when we recognize that treating lexical variation as a sociolinguistic variable requires a systematically onomasiological point of view. To explain: according to the traditional terminology of structuralist lexicology, a semasiological point of view is one in which the lexical item is taken as a point of departure, and in which the meaning(s) of that item is/are described. An onomasiological point of view is one in which a concept to be expressed is taken as the point of departure, and in which the various ways of naming that concept are described. (For the distinction between semasiology and onomasiology, see, e.g., Quadri [1952]; for a contemporary view on onomasiology, see Blank [2003].) If, then, the variation between subway and underground is a form of sociolinguistic variability, an onomasiological approach (charting the choices between the two items) clearly provides the most natural perspective. In general, however, an onomasiological perspective of this kind is much less common in lexicological research than a semasiological perspective, and this may have hindered the development of sociolexicology. Now, given the ideal of an onomasiological perspective in sociolexicology, there are two methodological hurdles for quantitative sociolexicology to take: the need for a rich documentation of lexical usage; and the need to control for meaning. Let us look more closely at each of these problems.

2.1. Usage-based data in sociolexicology The onomasiological perspective, like the sociolinguistic perspective in general, implies a usage-based approach to language: we are not dealing with structural variation among different languages, but we are interested in variation in language use within the confines of a single language. The examples we have quoted so far may suggest that the differences in lexical choices between languages varieties, like American English and British English, are a matter of black and white choices: subway is the American term and underground is the British term. In actual practice, however, the choices are seldom of a categorical nature. Rather, the selection from among the various lexical alternatives that occur within a language is a gradual matter. If we compare, for instance, the different terms for the concept of ‘jeans’ in Dutch, we notice that between the two national vari-

45. Lexical variation in space eties of Dutch (Netherlandic Dutch and Belgian Dutch), the choice for either jeans or jeansbroek or spijkerbroek is a matter of relative preferences, not of absolute choices. In that sense, the onomasiological enquiry will have to proceed in the quantitative way that is customary in sociolinguistics in general: what are the relative preferences for the competing forms? The kind of lexical documentation found in traditional dialectology, then, will not suffice. In most cases, traditional dialect dictionaries present us with a single name for a given concept, coming from just a single informant for a given place: the assumption would seem to be that (except maybe for diachronic changes in the vocabulary of the dialect in question) the dialect is internally homogeneous and lexical usage among the speakers of the dialect is basically uniform. That assumption runs counter to the sociolinguistic perspective but, at the same time, we need to recognize that the restrictions on the dialectological data are to a large extent practical ones: if the study of lexical variation is to be pursued in a quantitative way, the data we need are much more difficult to obtain than in the case of formal variation. You need only a relatively limited amount of spontaneous speech to gather a number of attestations of the majority of the phonemes in a language, but you will need a large amount of text before you come across a sufficient number of instances of, say, either jeans or jeansbroek or spijkerbroek. Certainly, there are ways of circumventing the problem by special elicitation techniques, but in general, there is a methodological hurdle for lexical research that is less prominent in the study of phonetic variation as the core of sociolinguistic research. It is in fact only in our current era, with the growing availability of large corpora of spontaneous text, that the first methodological handicap of sociolexicology is gradually losing some of its importance. The internet, for instance, provides a wealth of materials that may be subjected to quantitative analysis and that also (with the availability of blogs, forums and chat channels) provide a stylistically diversified set of materials. At the same time, classical sociolinguistic speaker-related information (like age, class and educational level) or even geographic origin is often difficult to identify when using internet sources. (See Kilgarriff and Grefenstette 2003 for some of the issues involved.) Even apart from the internet, however, a host of new text materials is gradually becoming available for lexical research. The first methodological hurdle for sociolexicology, in other words, seems to be becoming less of a problem than it used to be.

2.2. Controlling or meaning Treating onomasiological variation as a sociolinguistic variable means coming to terms with the meaning of words: the selection of a word is also the selection of a conceptual category, so if we are interested in the contextual choice between synonyms as an expression of sociolinguistic factors, we first need to control for meaning ⫺ and that, needless to say, is not an obvious matter. The problem is methodological in a high-level sense: in what ways can we push semantic description beyond intuitive interpretation? But the problem is also methodological in a very practical sense: in what ways can we achieve a method of semantic interpretation ⫺ and more specifically, synonym identification ⫺ that works efficiently enough to allow for an easy demarcation of a large set of concepts? Getting a good grip on the interrelations between FOV, COV and SSV cannot be achieved unless we can study a sufficiently high number of concepts, but that ideally requires a method of semantic analysis that is as fast as it is trustworthy.

825

826

VIII. Structural domains: Methodological problems In the broad context of sociolinguistics, the problem of semantics was recognized early on, in an important article by Beatriz Lavandera. She argued that it is inadequate at the current state of sociolinguistic research to extend to other levels of analysis of variation the notion of sociolinguistic variable originally developed on the basis of phonological data. The quantitative studies of variation which deal with morphological, syntactic, and lexical alternation suffer from the lack of an articulated theory of meanings. (Lavandera 1978: 171)

In the context of this article, we may forgo the question of whether the problem is as striking in morphology and syntax as it is in lexicology, but the lexical problem is definitely real: how exactly do we establish whether a number of potential synonyms actually express “the same thing” or whether they are merely referential near-synonyms? Neither sociolinguistic nor lexical-semantic theory has a ready-made answer to that, but sociolinguistic research features a number of basic strategies for dealing with the problem which we will review in the next section. As an intermediate summary, let us recall that the proper treatment of lexical variation as a sociolinguistic variable has so far been slowed down primarily by two methodological hurdles: the difficulties involved in constructing a sufficiently large dataset, and the difficulties involved in identifying concepts in a way that is minimally subjective and maximally efficient. The first problem is greatly alleviated by the advent of ever larger corpora, but the second question is still moot.

3. Alternative strategies or dealing with the semantic problem Given the question of whether a number of potential synonyms actually do express “the same thing” in referential terms, there are in essence two approaches that sociolexicological research may follow: to restrict the domain of investigation to the referential overlap between the near-synonyms (and specifically, to establish the referential synonymy as objectively and systematically as possible); or to include the semantic differences between the near-synonyms in the investigation. Given two words with an overlapping referential range of application, the first strategy focuses on the area of overlap without including the non-overlapping areas. As a borderline case, this applies specifically if it can be established that the non-overlapping areas are semantically insignificant, i. e., if it can be established that the two items are indeed pure referential synonyms. The second strategy looks at the entire ranges, including subtle semantic differences. The alternative approaches will be made more tangible and concrete in the following sections, but a general comparison is due at this point already.

3.1. The restrictive approach The first approach is obviously the one that corresponds most strictly to the concept of a sociolinguistic variable: we can only talk about sociolinguistic variables if there are alternative ways of expressing the same thing. However, we need to be aware that a

45. Lexical variation in space

b jeans

a

827

c spijkerbroek

Fig. 45.1: A graphical representation of lexical overlap jeans a blue denim trousers, with visible stitches b brown denim trousers, with visible stitches spijkerbroek a blue denim trousers, with visible stitches c blue denim trousers, without visible stitches

situation such as that in Figure 45.1 may represent a number of semantic configurations, which do not all allow for a simple reduction of the area of investigation to the area of overlap. For a schematic overview of the various configurations, let us consider the following fictitious example. If we are looking at the items jeans and spijkerbroek, the areas in Figure 45.1 might be filled with the types of referents listed. (The “visible stitches” feature refers to the fact that jeans often have prominent stitching on the seams.) The three configurations we need to consider are the following: complete synonymy, partial synonymy, and mere referential overlap. To begin with, the clearest situation from a sociolinguistic point of view occurs when the differences in the range of application of both items (the fact that [b] occurs with jeans but not with spijkerbroek, and that [c] occurs with spijkerbroek but not with jeans) are superficial. If we can argue that the meaning of both jeans and spijkerbroek is ‘denim trousers’, the observed differences with regard to (b) and (c) could in principle be disregarded, because both would in principle fall within the range of the two words ⫺ which might then obviously be treated as pure synonyms. Clearly, the superficial nature of (b) and (c) could also follow from a statistical analysis, if their absence in the observational range of the other term would be due to chance. Suppose, however, as a second interpretation, that jeans and spijkerbroek are each considered to be polysemous, with jeans having the meanings (a) and (b), and spijkerbroek having the meanings (a) and (c). The two words are then partially synonymous. They could still be treated as manifestations of a sociolinguistic variable, but only if the investigation is restricted to the meaning they have in common. A further possibility obtains when jeans and spijkerbroek receive unitary definitions that imply a certain degree of referential overlap, but no conceptual equivalence, either complete or partial. This is what happens if jeans is defined as ‘denim trousers with visible stitches’ and spijkerbroek as ‘blue denim trousers’. (These definitions are neither realistic nor plausible; they merely the serve purposes of illustration.) In this case, we can still build up our investigation starting from the area of overlap (a), but we should be aware that the choice between the two words is a matter of conceptual construal: selecting jeans or spijkerbroek implies different ways of conceptualizing the referents in the overlapping area. In other words, whereas the cases in which we are confronted with full or partial synonymy lead to an investigation of FOV, the situation in which we are

828

VIII. Structural domains: Methodological problems dealing with referential overlap between items that are semantically non-equivalent leads to an investigation of COV. It follows from this analysis of the various semantic configurations potentially corresponding with Figure 45.1 that the possibility of treating lexical variation as a sociolinguistic variable hinges on the feasibility of identifying synonymy and polysemy. That is not, in the current state of methodology in semantics, a straightforward matter. For instance, the distinction between polysemy and vagueness (to mention one of the crucial steps) is known to be methodologically unstable (Geeraerts 1993; Taylor 2003). The semantic analysis required for the implementation of the restrictive strategy, in other words, will tend to be a painstaking and tentative process. At the same time, such a semantic analysis is absolutely necessary to avoid confusion of FOV and COV. Simply starting from the items that may be found in the area of overlap ⫺ i. e., starting from a purely referential rather than a conceptual level ⫺ will tend to confuse FOV and COV: if there is an unheeded conceptual difference between jeans ‘denim trousers with visible stitches’ and spijkerbroek ‘blue denim trousers’, simply asking people what they call a pair of blue denim trousers with visible stitches would not be sufficient to establish the difference between FOV and COV. This is basically what the traditional Wörter und Sachen method in dialectology did, which was appropriately criticized by lexical field theory. (Conversely, traditionally collected dialectological data may be reanalyzed from the point of view of contemporary semantic models: see Swanenberg 2002; Szelid and Geeraerts 2008.)

3.2. The comprehensive approach Given the difficulties involved in establishing semantic equivalences, a less restrictive approach may be envisaged: what if we investigated the entire area covered by the two items? This is the approach that was advocated by Lavandera in her critical appraisal of the concept of sociolinguistic variable: she argued that “unless we examine the entire distribution of the apparent synonyms, the possibility of an explanation of the variation is ruled out” (1978: 179), and accordingly proposed “to relax the condition that the referential meaning must be the same for all the alternants and substitute for it a condition of functional comparability” (1978: 181). As Labov remarks in his reaction to Lavandera, taking this suggestion at face value and simply pretending that all the instances of use in the combined range of the nearsynonyms are semantically equivalent would result in the loss of the precision that sociolinguistic research tries to achieve (Labov 1978: 6). The suggestion may, however, also be taken more constructively, as an injunction to study the actual equivalence of the potential synonyms in detail: how do differences of meaning and contextual features simultaneously determine the choice of one item rather than another, and do these factors work in the same way for both items? To see the difference between this and the restrictive approach, it is instructive to distinguish clearly between the dependent and the independent variables in the investigation. In the restrictive approach, FOV (lexical selection) is the dependent variable, and SSV (variation of context) is the independent variable. (But when the near-synonyms are not partially or fully synonymous, the approach mentioned under 3.1 takes COV as the dependent variable and SSV as the independent variable.) In the comprehensive ap-

45. Lexical variation in space

Fig. 45.2: A schematic representation of the different approaches

proach suggested here, semantic factors are included in the set of independent variables, with FOV as the dependent variable. In the example we have used in 3.1, the comprehensive approach could take the form of a regression analysis in which the selection of either jeans or spijkerbroek is modeled in terms of lectal features (like the difference between Netherlandic Dutch and Belgian Dutch) and semantic features (like whether the item to be named is made of denim, what color it has, whether the stitches are prominent). The different approaches that we have been able to identify, then, may be summarized in Figure 45.2. (The arrows within the triangles point from independent to dependent variables.) The comprehensive approach has the disadvantage of not leading in a straightforward way to a quantification of the SSV effect, which in this type of research emerges with other factors in a combined analysis, and is thus less easily assessed on its own. On the other hand, the comprehensive approach has the advantage of providing a detailed description of the relationship between the near-synonyms, and may thus help to answer the questions about semantic equivalence that lie at the basis of the restrictive approach. If the approach is implemented in the form of a regression analysis, for instance, two items would be fully synonymous if there is no interaction (in the statistical sense) between the lectal factors and the semantic factors included in the regression model. In this respect, it would seem that we could use the comprehensive approach as a heuristic step towards the restrictive approach.

4. Case studies Now that we have gained an insight into the two basic strategies that are currently being followed to tackle the specificity of lexical variation as a sociolinguistic variable, two case studies may help to make the methodological strategies more tangible. We will present them in a different order than the one in which we introduced the two approaches: because the comprehensive approach is, as we saw, less narrowly linked to the idea of a sociolinguistic variable, we will start with a brief example of the comprehensive approach.

4.1. An example o the comprehensive approach In a comprehensive approach, lectal variation and semantic variation among near-synonyms are studied in combination. Differences exist as to the type of semantic description that is taken as the point of departure. In Geeraerts, Grondelaers and Bakema

829

830

VIII. Structural domains: Methodological problems (1994), the referential type of description illustrated by Labov (1973) in the field of cups and mugs is applied to the lexical field of clothing terms: on the basis of pictures and photographs in magazines, the referents of a number of Dutch clothing terms receive a componential description which is then used to judge identity of meaning among nearsynonyms (like, indeed, jeans and spijkerbroek). In contrast to the original Labov study, lectal variation is included in the analysis, in the form of register variation, and in the form of a spatial factor, viz. the distinction between Netherlandic Dutch and Belgian Dutch. In other cases, the semantic description does not take a referential form, but rather consists of a featural coding of textual corpus data. In Glynn (2007), a corpus-based analysis of the verbs bother, hassle and annoy is conducted with specific emphasis on the combination of lectal, semantic and constructional variation. Some 2000 examples taken from internet sources are coded for a variety of features: first, the distinction between US attestations and UK attestations; second, a number of formal and constructional variables (whether the verb is used in a transitive or intransitive construction, whether or not it is used predicatively, whether or not it is used as a gerund, etc.); and third, a variety of (manually attributed) semantic features. Among others, these semantic features include a distinction between humorous and non-humorous contexts and an indication of the thematic meaning of the utterances: whether the verb expresses anger, pain of an emotional or physical nature, concern, a request, an attempt to impose, etc. These coded data are subjected to a number of statistical analyses that help to reveal the underlying pattern of polysemy and lectal differentiation. In Figure 45.3, for instance, a multiple correspondence analysis shows how the different verbs cluster with regard to the semantic characterization. In the top right corner, for instance, we notice that us_annoy_trans (i. e., the verb annoy in a transitive construction in US sources) and uk_annoy_trans are situated close to each other, i. e., there is not a lot of difference in the way they are used. In general, lectal differences between the UK sources and the US sources are not striking. The UK attestations and the US attestations cluster together, except for bother in predicative use: us_bother_pred patterns with the expression of anger (ang in the plot) while uk_bother_pred ties in with themes like “energy”, “request”, “agitation”. This example demonstrates how the contrastive analysis of polysemy in near-synonyms may profit from the inclusion of lectal features ⫺ which is basically what the comprehensive approach amounts to. At the same time, it makes clear how an analysis of this type may help to establish synonym sets. As mentioned, a lectal difference correlates with a semantic difference only in the predicative use of bother; in the other readings of bother and in the senses expressed by annoy and hassle the lectal difference plays no role. The semantic grouping of the items in the plot reveals the polysemy of the expressions. In the top right corner of the plot, for instance, we note how transitive annoy, transitive bother and predicative bother (at least in the US) are all associated with the “anger” reading. In the left hand side of the plot, we see how an “imposition” and “teasing” reading is associated with a transitive use of hassle. In the lower right corner, intransitive uses of bother, nominal uses of hassle, and predicative uses of bother (restricted to the UK) are associated with the expression of agitation. A correspondence analysis of this kind is an exploratory technique only, and one would like to take a further step by introducing inferential statistics. In Speelman and Geeraerts (to appear), for instance, the selection of Dutch doen ‘make’ or laten ‘let’ as a

45. Lexical variation in space

Fig. 45.3: Correspondence analysis of hassle, bother, annoy

causative auxiliary is subjected to a logistic regression analysis including lectal factors together with semantic ones. The lectal factors (viz. the distinction between Netherlandic Dutch and Belgian Dutch and differences of stylistic register) clearly emerge as relevant for the distribution of the two verbs. While such statistical analyses of near-synonyms are relatively scarce in lexical studies, they do form an emerging trend in the study of syntax. In the context of construction-oriented grammar research, a number of younger scholars are using advanced quantitative techniques to distinguish between near-synonymous constructions. Including lectal factors in such analyses is not yet standard procedure, but a number of publications do incorporate lectal variation as a potentially distinguishing factor for constructional near-synonyms: see for instance Wulff, Stefanowitsch and Gries (2007), Grondelaers and Speelman (2007).

4.2. An example o the restrictive approach As an example of a restrictive approach, we may have a closer look at Geeraerts, Grondelaers and Speelman (1999). Following up on the comprehensive approach developed in Geeraerts, Grondelaers and Bakema (1994), this study investigates the relationship between the vocabulary of the two national varieties of Dutch, Netherlandic Dutch and Belgian Dutch.

4.2.1. Materials and method The empirical basis of the investigation consists of over 40,000 observations of actual language use, involving two lexical fields (soccer terms and clothing terms). The concepts in the lexical fields are chosen in such a way that potential polysemy is minimal, and can be easily controlled for by a manual inspection of the data.

831

832

VIII. Structural domains: Methodological problems Tab. 45.1: Onomasiological profiles for ‘shirt’ in the Belgian and Netherlandic data (1990)

hemd overhemd shirt

B90

N90

31 % 69 % 0%

17 % 46 % 37 %

The database allows, for instance, for a calculation of the relative proportions in Belgian and Netherlandic sources of the term buitenspel ‘offside’ and the loanword offside as names for the “offside” concept. In the case of the concept of “dress”, it can be determined whether the lexical choices involve a preference for one of the items jurk, japon or kleed. The core of the observed material consists of magazine and newspaper material recorded in 1990. This core is extended in two ways. In the first place, similar material from 1950 and 1970 was collected, which allows for a “real-time” investigation of lexical convergence or divergence processes. In addition, the stratification of language use was taken into account. Between standard and dialect, there are a number of “strata” on which register differences may co-occur with an increasing geographical specialization. For an investigation of the relationship between Belgian and Netherlandic Dutch, these strata ⫺ viz. the regionally colored informal variants of the standard language ⫺ are extremely relevant: it can be expected that the linguistic differences between Belgium and The Netherlands will increase on this regiolectal level. This intermediate level between dialect and written standard language was represented by the clothing terms collected from labels and price tags in shop windows in two Belgian (Leuven and Kortrijk) and two Netherlandic towns (Leiden and Maastricht). The intended audience of this form of communication is more restricted than the national or binational audience which is the target of the magazines from which the core material was selected. Using written language in a semi-formal situation, on the other hand, implies that the purely dialectal pole of the stratificational continuum is not represented in the data. The collected data are analyzed from two perspectives: a diachronic one (involving the convergence or divergence between Belgian Dutch and Netherlandic Dutch in the period between 1950 and 1990), and a synchronic one (involving the stratification of stylistic registers in the two national varieties of Dutch). To see how the analysis the treatment of lexical variation as a sociolinguistic variable may happen in practice, we first need to specify two methodological concepts. First, the onomasiological profile of a concept in a particular source is the set of synonymous names for that concept in that particular source, differentiated by relative frequency. Table 45.1 contains the onomasiological profiles for overhemd ‘shirt’ in the Belgian and the Netherland 1990 databases. Second, uniformity is a measure of the correspondence between two onomasiological profiles. The computation of uniformity has its starting point in the idea that a common language norm triggers uniform linguistic behavior. In its most extreme form, lexical uniformity in the naming of a concept obtains when two language varieties have an identical name for that concept, or several names with identical frequencies in each of the two varieties. Much more frequent than these examples of “ideal” uniformity, however, are such partial correspondences as illustrated in Table 45.1. Let us, for the sake of illustration, assume that the relative frequencies in Table 45.1 represent 100 actual

45. Lexical variation in space naming instances in each of both profiles, rather than percentages. The partial overlap between the profiles in Table 45.1 is quantified by counting the naming instances for which there is a counterpart in the other profile. In the ideal scenario outlined above, each of the 100 naming events in each of both profiles has its counterpart in the other profile, yielding a maximal uniformity of 100 percent. In Table 45.1, however, 14 instances of hemd in B90 have no counterpart in N90, 23 Belgian overhemden have no Netherlandic counterpart, and there are no Belgian counterparts for the 37 Netherlandic shirts. Of the grand total of 200 naming events in the two profiles, only 126 (i. e., 200 ⫺ 14 ⫺ 23 ⫺ 37) instances have counterparts in the other profile, yielding a uniformity index of 126/2, or 63 percent. For the sake of quantitative convenience, it should be noticed that this percentage equates the sum of the smallest relative frequency for each alternative term, i. e., 17 ⫹ 46 ⫹ 0 ⫽ 63 %. If more than one concept is investigated, uniformity index U is defined as the average of the uniformity indexes of the separate concepts, whereas uniformity index U1 is defined as a weighted average, in which the relative frequency of each concept in the investigated samples is taken into account. In this presentation of the results obtained by Geeraerts, Grondelaers and Speelman (1999), we will focus exclusively on the weighted uniformity index U1, in which high-frequency concepts have a more outspoken impact on the overall uniformity.

4.2.2. Hypotheses Given the database and the methodology, what can we expect to find with regard to the relationship between the various language varieties included in the investigation? Some historical background is necessary at this point. In Flanders ⫺ the Dutch-speaking part of Belgium⫺the standardization process that began (as it did in many European countries) in the Early Modern Period was slowed down as a result of Flanders’ political separation from The Netherlands in the Eighty Years’ War (1568⫺1648). Standard Dutch developed in The Netherlands over the course of the seventeenth century, but as Flanders was politically separated from The Netherlands, remaining under foreign ⫺ Spanish, Austrian, French ⫺ rule, it did not participate in this process of standardization. Rather, French was used more and more as the language of government and high culture, a practice that received an important impulse after the birth of the Belgian state in 1830. Dutch then survived basically in the form of a variety of Flemish dialects. However, as a result of a social and political struggle for the emancipation of Flanders and the Flemish-speaking part of the Belgian population, Dutch regained ground as the standard language in Flanders, i. e., as the language of learning, government and high culture. This process started somewhat hesitantly in the late nineteenth century, gained momentum during the first half of the twentieth century, and finally made a major leap after World War II and during the booming 1960s. Still, most linguists agree that the standardization process has not progressed as far as in The Netherlands. The traditional local dialects, for instance, have a much stronger position in Flanders than in The Netherlands. With respect to the status and the development of Belgian Dutch, two hypotheses may now be formulated. First, there is an expectation of diachronic convergence between

833

834

VIII. Structural domains: Methodological problems Belgian and Netherlandic Dutch. The standardization process in Flanders is characterized by an explicit normative orientation towards Netherlandic Dutch: the standardization of Belgian Dutch over the course of the twentieth century took the form of an adoption of the Dutch standard language that already existed in The Netherlands. In addition, the less advanced character of the standardization of Belgian Dutch is believed to manifest itself in a larger synchronic distance between local and national language in Belgium than in The Netherlands; even to the untrained observer, it is obvious that the differences between regional and supraregional registers are much larger in Dutch-speaking Belgium than in The Netherlands.

4.2.3. Empirical results The diachronic and synchronic hypotheses may now be made operational in terms of uniformity values as defined above. Diachronically, convergence and divergence can be quantified as increasing or decreasing uniformity. Synchronically, the larger distance between national and local language we expect in Belgian Dutch will manifest itself in less uniformity between magazine and shop-window material in Belgian Dutch than in Netherlandic Dutch. Table 45.2 contains the relevant results. B50 stands for “Belgian data from 1950”, N50 stands for “data from the Netherlands from 1950”; Bsw90 refers to the shop-window materials from Belgium, in contrast to B90, which stands for the data taken from magazines and newspapers. The data in Table 45.2 unambiguously confirm both the diachronic and the synchronic hypotheses. Diachronically, the increase in uniformity between Belgian and Netherlandic Dutch suggests an evident lexical convergence between both varieties: U1 (B50, N50) < U1 (B70, N70) < U1 (B90, N90) 69.84 < 74.59 < 81.70 Synchronically, the delayed or unfinished standardization of Belgian Dutch manifests itself in a distinctly lower uniformity between the Belgian magazine and shop-window data than between the Netherlands magazine and shop-window material: U1 (B90, Bsw90) < U1 (N90, Nsw90) 45.90 < 67.75

Tab. 45.2: U1 values comparing Belgium and The Netherlands (1950⫺1970⫺1990) and comparing written data from magazines and newspapers with local shop window data (1990) B50 / N50: B70 / N70: B90 / N90: B90 / Bsw90: N90 / Nsw90:

69.84 74.59 81.70 45.90 67.75

45. Lexical variation in space Further extensions of this type of approach include taking into account the effect of specific factors, like the level of foreign influence or the degree of linguistic purism that shows up in the data.

5. Overview and prospects Starting from the assumption that the methodological problems involved in studying the relationship between lexicon and space boil down to the problems involved in studying lexical variation as a sociolinguistic variable, we have argued that the dearth of quantitative lexical variation studies compared to other types of sociolinguistic studies results from two features: first, the difficulties involved in constructing a sufficiently large dataset for studying lexical variation, and second, the specific nature of lexical variation as a sociolinguistic variable. That specificity resides in the ternary rather than binary nature of lexical variation: a full picture of lexical variation requires an insight into the multiple relations between lectal (and more generally, contextual) features, conceptual onomasiological variation and formal onomasiological variation. The first difficulty (that of constructing a sufficiently large dataset) is gradually becoming less daunting, thanks to the growing availability of large corpora. The second difficulty (that of disentangling conceptual variation and formal variation) is still a major one. We have identified two fundamental ways in which quantitative sociolexicology responds to the challenge of semantic analysis. A restrictive approach stays close to the original idea of the sociolinguistic variable, and tries to determine the synonym sets that constitute the input for the lectal analysis as accurately as possible. A more comprehensive approach relaxes the requirement of semantic identity, and conducts an analysis at short range of the differences between near-synonyms, including both lectal and semantic factors as independent variables in the analysis. The restrictive approach examines choices within sets of strict synonyms in the light of external, lectal variables. The comprehensive approach examines choices within sets of near-synonyms in the light of such contextual variables as well as conceptual ones. The case studies we have described as an illustration of both approaches show the advances that are currently being made in comparison with the older studies of lexical variation in space that may be exemplified by traditional dialectological work: in both case studies, diatopic variation is but one variable among a number of relevant factors; both approaches use extended datasets of actual usage materials, rather than sample responses from a limited number of informants, and both subject the data to advanced forms of quantitative analysis. At the same time, both approaches also have their restrictions and disadvantages. While the comprehensive approach, focusing on local differences between a small set of near-synonyms, lends itself less easily to large-scale investigations, the restrictive approach has yet to come to terms with the influence of external factors on conceptual rather than just formal onomasiological variation. Further, both approaches could profit from reliable and efficient methods of semantic analysis. In this respect, recent developments in Natural Language Processing may provide a promising avenue to methodological innovation. In the Word Sense Disambiguation paradigm in particular, methods for the automatic recognition of polysemy are being developed, and one of the major further steps to be taken could be to introduce

835

836

VIII. Structural domains: Methodological problems these algorithms into sociolexicology, i. e., to introduce statistical, distributional measures of semantic similarity in order to automate or semi-automate the semantic analysis underlying the identification of onomasiological profiles. Such distributional measures for semantic similarity (Agirre and Emonds 2006) rely on the assumption that words appearing in similar textual contexts will also have similar meanings. Given the current state of the art in distributional measures for semantic similarity, it is not yet obvious at this point which of the various measures is best for synonymy extraction, but we may be fairly confident that methodological advances may be achieved by pursuing this path.

6. Reerences Agirre, Eneko and Philip Edmonds (eds.) 2006 Word Sense Disambiguation. Algorithms and Applications. Berlin: Springer. Blank, Andreas 2003 Words and concepts in time: Towards diachronic cognitive onomasiology. In: Regine Eckardt, Klaus von Heusinger and Christoph Schwarze (eds.), Words in Time: Diachronic Semantics from Different Points of View, 37⫺65. Berlin/New York: Mouton de Gruyter. Geeraerts, Dirk 1993 Vagueness’s puzzles, polysemy’s vagaries. Cognitive Linguistics 4: 223⫺272. Geeraerts, Dirk, Stefan Grondelaers and Peter Bakema 1994 The Structure of Lexical Variation: Meaning, Naming, and Context. Berlin/New York: Mouton de Gruyter. Geeraerts, Dirk, Stefan Grondelaers and Dirk Speelman 1999 Convergentie en divergentie in de Nederlandse woordenschat. Amsterdam: Meertens Instituut. Glynn Dylan 2007 Mapping meaning. Toward a usage-based methodology in cognitive semantics. Unpublished Ph.D. thesis, Katholieke Universiteit Leuven. Stefan Grondelaers and Dirk Speelman 2007 A variationist account of constituent ordering in presentative sentences in Belgian Dutch. Corpus Linguistics and Linguistic Theory 3: 161⫺193. Kilgarriff, Adam and Gregory Grefenstette (eds.) 2003 The Web as Corpus. Computational Linguistics 29 (Special issue). Labov, William 1966 The Social Stratification of English in New York City. Washington: Center for Applied Linguistics. Labov, William 1972 Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Labov, William 1973 The boundaries of words and their meanings. In: Charles-James N. Bailey and Roger W. Shuy (eds.), New Ways of Analyzing Variation in English, 340⫺373. Washington: Georgetown University Press. Labov, William 1978 Where does the sociolinguistic variable stop? A response to Beatriz Lavandera. Working Papers in Sociolinguistics 44. Austin: Southwest Educational Development Laboratory. Lavandera, Beatriz 1978 Where does the sociolinguistic variable stop? Language in Society 7: 171⫺183. Quadri, Bruno 1952 Aufgaben und Methoden der onomasiologischen Forschung. Berne: Francke.

46. Areal variation in syntax

837

Speelman, Dirk and Dirk Geeraerts to appear Causes for causatives: The case of Dutch ’doen’ and ’laten’. In: Ted Sanders and Eve Sweetser (eds.), Linguistics of Causality. Berlin: Mouton de Gruyter. Swanenberg, Jos 2002 On ethnobiological nomenclature in Southern Dutch Dialects. In: Jan Berns and Jaap van Marle (eds.), Present-Day Dialectology: Problems and Findings, 353⫺365. Berlin/New York: Mouton de Gruyter. Szelid, Veronika and Dirk Geeraerts 2008 Usage-based dialectology. Emotion concepts in the Southern Csango dialect. Annual Review of Cognitive Linguistics 6(1): 23⫺49. Taylor, John R. 2003 Linguistic Categorization. Oxford: Oxford University Press. Tway, Patricia 1975 Workplace isoglosses: Lexical variation and change in a factory setting. Language in Society 4: 171⫺183. Werlen, Iwar 2005 Von Brig nach Bern. Dialektloyalitat und Dialektanpassung bei Oberwalliser Migrierenden in Bern. In: Eckhard Eggers, Jürgen Erich Schmidt and Dieter Stellmacher (eds.), Moderne Dialekte ⫺ Neue Dialektologie, 375⫺404. Stuttgart: Steiner. Wulff, Stefanie, Anatol Stefanowitsch and Stefan Gries 2007 Brutal Brits and persuasive Americans: Variety-specific meaning construction in the intocausative. In: Günter Radden, Klaus-Michael Köpcke, Thomas Berg and Peter Siemund (eds.), Aspects of Meaning Construction in Lexicon and Grammar, 265⫺281. Amsterdam: Benjamins.

Dirk Geeraerts, Leuven (Belgium)

46. Areal variation in syntax 1. Introduction 2. Theories, methods, research traditions 3. The nature of areal variation in dialect syntax: Small and large-scale variation in European dialect 4. Areal variation and (areal) typology 5. Conclusion and outlook 6. References

1. Introduction What we spontaneously associate with areal variation are regional dialects, language (or rather dialect) atlases and isoglosses. The examples of isoglosses and dialect maps which immediately spring to mind are phonological, lexical and, to a much lesser extent, morphological in nature. Syntax, by contrast, is a relative newcomer in this respect. It has only been in the late twentieth century that widespread interest, within both dialectology

838

VIII. Structural domains: Methodological problems and general linguistics, has developed in issues of syntactic variation, or dialect syntax or ⫺ as it is alternatively labeled in studies working within a broadly formalist research tradition ⫺ microparametric syntax, microcomparative syntax. This late awakening started very modestly in the 1980s, but has gained enormous momentum since the mid1990s. As a matter of fact, when judged on the basis of relevant major research projects and the range and number of publications, the period beginning in the second half of the 1990s may justly be called “the decade of dialect syntax”, and this is going to be a very long decade as it were ⫺ with this very handbook no doubt contributing its fair share to this development. This article strives to provide an overview of some of the major developments of this decade. Moreover, given the fact that the bulk of the relevant studies on dialect syntax of the last ten to fifteen years have been concerned with Germanic and to a lesser extent Romance (especially Italian) dialects, and given the research profile of the author, it is areal variation in Continental Western Europe, in the British Isles and the wider Englishspeaking world that will take center stage. The article will show that it is indeed especially owing to the study of dialect syntax that modern dialectology can no longer be accused of working “in isolation from the full spectrum of language variation” or scolded for its “separation from mainstream linguistics”, as Auer and Schmidt correctly point out concerning classical dialectology in their introduction to this volume. This significant turn that modern dialectology has taken is also clearly reflected in the following opening statement from the introduction to the volume by Filppula et al. (2005: vii): Until fairly recently, the word “dialect” carried the connotation of being something antiquated and having low social status. The situation is now very different: along with the general rise of language awareness in modern societies, non-standard varieties of languages have become an object of new interest, which in turn is reflected in their generally improved position even in educational systems. A similar change has taken place in scholarly research on dialects. This is partly due to the recent advances in the methods used in dialectological research. The advent of computer-assisted methods has enabled study of significantly larger databases than in traditional dialectological research. Also, methods derived from sociolinguistic and variation studies have greatly added to the general interest of dialect studies and distanced them even further from the “butterfly collecting” mentality often associated with traditional dialectology. New language-theoretical frameworks form yet another source of inspiration for dialect studies today: typology, cognitive linguistics, discourse analysis and pragmatics have provided fresh perspectives on old problems and opened up completely new lines of research such as cognitive dialectology, folk linguistics, and perceptual dialectology.

In the following, various dimensions of the “syntactic turn” in dialectology on the one hand, and the “dialect turn” in modern linguistic theorizing on the other, will be explored. Current theories, methods and research projects in the field of dialect syntax will be introduced in section 2. Selected findings from these (exclusively European) projects and other studies on the areal variation of dialect-syntactic phenomena will be presented in section 3: including a subsection which addresses the issue of historical stability of such dialect-syntactic areas/isoglosses and some general problems to be kept in mind by historical dialectologists when wanting to study dialect syntax. In section 4: the important links between the study of areal variation in dialect syntax and language typology, especially areal typology, will be pointed out. The article will be rounded off by a sketch of some of the most promising research avenues in exploring syntactic variation in dialects.

46. Areal variation in syntax Given the nature of a survey article, many issues can be addressed only briefly. To some extent, the rather long list of (largely recently published) references as well as the links to the most important research centers and projects on dialect syntax are meant to help compensate for this, guiding the reader to helpful sources of information. In general, however, the kind of kaleidoscopic approach adopted here seems best suited to providing a state-of-the-art account of this extremely dynamic and fast-developing field.

2. Theories, methods, research traditions What is typically associated with classical studies on areal variation is strength as regards methodology, but weakness, or even blissful ignorance, as regards theory, i. e., the complete lack of (interest in) entering into a dialogue with current developments in linguistic theorizing. That way, dialectology for a long time deprived itself of important sources of inspiration concerning data to be analyzed and questions to be asked of the material and, at the same time, failed to feed the dialectological expertise back into the discussions contributing to the further developments of theories in different fields of mainstream linguistics, such as formalist and functionalist syntactic theories, sociolinguistics, language typology, and historical linguistics (e.g., research on grammaticalization). In this section, it will be shown how radically this has changed ⫺ most dramatically in the realm of theory (section 2.1), less so but still noticeably in terms of methodology (section 2.2) ⫺ and which research traditions have played a key role in this recent change.

2.1. Theory The major driving force behind the “syntactic turn” in the study of dialects has clearly been the “dialect”, or rather “microparametric”, turn in generative linguistic theorizing. This turn made itself modestly felt in the 1980s and gained significant momentum in the course of the 1990s (cf., for example, Abraham and Bayer 1993; Beninca` 1989; Black and Motapanyane 1996; Kayne 1993). Triggered by the Principles and Parameters approach, there was a growing awareness among syntacticians working within this theory that alongside cross-linguistic variation (macroparametric syntax), language-internal variation (microparametric or microcomparative syntax) should also matter in the further advancement of generative theory. This microparametric turn, which has alternatively been described in terms of a marriage of “biolinguistics” (i. e., generative syntactic theory) and sociolinguistics (just take the programmatic title which Cornips and Corrigan [2005b] have chosen for their collective volume ⫺ Syntax and Variation: Reconciling the Biological and the Social), is also clearly reflected in quotations like the following from the editors’ preface (Trousdale and Adger 2007) and scene-setting introduction (Adger and Trousdale 2007) to the most recent collection of articles on English dialect syntax: The articles in this special issue present new theoretical research into grammatical variation in English, which yields a number of findings of interest to those involved in work in English dialectology, typology, and syntactic theory. (Trousdale and Adger 2007: 257)

839

840

VIII. Structural domains: Methodological problems It is clear from these articles, and from the other ongoing work on microparametric variation in English, that research based on nonstandard varieties of English can enhance and refine theoretical models of language structure. (Trousdale and Adger 2007: 259) Yet the modelling of (syntactic) variation in and across dialects is a critical issue in any theoretical framework, as variation is ubiquitous in language, and the fact that language can vary raises important questions regarding what the theory is actually modelling. (Adger and Trousdale 2007: 261)

In fact, the microparametric turn in generative syntactic theory culminates in the following statement by Adger and Trousdale (2007: 274): “The articles in this issue show how theoretical modelling can be enriched by taking variation as a core explanandum”. This is as far as one can get from the “ideal speaker’s/listener’s competence” in a (standard variety of) language which has for decades formed the central object of research in generativist theory formation. Today the whole spectrum of broadly formalist syntactic theories is brought to bear on the study of dialect syntax: still dominant are studies essentially operating within the Government and Binding (GB, P&P) model of the 1980s or the Minimalism of the 1990s (apart from the early publications on microparametric syntax mentioned at the beginning of this section, cf., for example, Adger and Trousdale [2007] for a brief survey discussion and, for a cross-section of microparametric studies on different European dialects, many of the contributions to Barbiers, Cornips and van der Kleij [2002] and Barbiers et al. [2008]). But there are also, albeit less frequently, formalist-driven studies working within other, more recent formalist frameworks like Head-driven Phrase Structure Grammar (HPSG; e.g., Bender 2007), Optimality Theory (OT; e.g., Seiler 2004) or Stochastic Optimality Theory (e.g., Bresnan and Deo 2001; Bresnan, Deo and Sharma 2007). Right from the humble beginnings in the 1980s, formalism has retained a firm grip on the study of dialect syntax and clearly dominates the field today ⫺ and this is to the benefit of anyone interested in dialect syntax, regardless of the theoretical framework he or she is working in. At the same time, it is not only the “microparametric” turn in generative syntax that has made dialect syntax attractive within linguistic theorizing. Since the late 1990s, language typologists, too, have increasingly turned their attention to cross-dialectal variation in light of what is known about cross-linguistic variation, just as dialectologists have come to realize that typology offers as a very useful source of inspiration and an additional reference frame for investigating syntactic variation and for putting their results in perspective (cf., e.g., Kortmann [2004c] for a collection of papers in this spirit; see also section 4). This is also acknowledged by Trousdale and Adger in their preface: “Equally, a study of English dialects can make significant contributions to work on language typology, and to our understanding of the place of sociolinguistics in linguistic theory” (Trousdale and Adger 2007: 259). Language typology, in turn, especially given the nature of the explanations for the observable patterns and limits of variation across the world’s languages offered in functional (i. e., Greenbergian) typology, can be seen as part of the broad research tradition(s) of functionalism and cognitivism. And just as there are a host of formalist syntactic theories which are currently used in research on dialect syntax, there are also broadly functional/cognitive theories other than functional typology which are used in modeling language-internal syntactic variation: usage-based theories (e.g., certain variants of Con-

46. Areal variation in syntax

841

Modelling syntactic variation

Formalist theories

Functional-cognitive theories

GB functional typology Minimalism usage-based theories (e.g., Construction Grammar) HPSG Word Grammar OT Stochastic OT (sharing properties with both research traditions) Fig. 46.1: Current syntactic theories aiming at modeling syntactic variation

struction Grammar, such as Haser and Kortmann 2009; Hollmann and Siewierska 2007; Pietsch 2005a) or Word Grammar (e.g., Hudson 1999: 2007). Figure 46.1 provides an overview. One of the many positive points about the current interest in dialect syntax from different theoretical angles is that there is a healthy competition between these theories and, even more importantly, that there is a lot of interaction between researchers and research units working within different theoretical frameworks. Moreover, integration is the buzzword in the current research scene. Integrated approaches to the study of language (internal) variation are called for in many publications (for forceful arguments cf. Bisang 2004; Cornips and Corrigan 2005a; or Croft 1995; cf. also the volume by Dufter, Fleischer and Seiler 2009). Three central facets of such an integration of approaches to the study of dialect syntax are the following: i.

bridging the gap between syntactic theory and dialectology/sociolinguistics (recall the motto: “reconciling the biological and the social”); above all, this involves the following two points. ii. taking variation (across the members of a speech community, but crucially also within the individual members of a speech community) as the core explanandum, it has become the major testing ground for all cognitivist theories, be they cognitive in the Chomskyan sense or cognitive in the Labovian/Langackerian sense of “Cognitive Linguistics” (e.g., usage-based models), which now have to face the challenge of developing models of the representation and organization of linguistic knowledge in light of what we know about linguistic reality, namely variation. In other words, for the purpose of developing theories of the representation of linguistic knowledge in the mind, language is now investigated in its natural habitat, as it were, and no longer in the vacuum of the competence of the “ideal” speaker/listener. How, for example, are competing rules or grammars in the individual speaker to be modeled ⫺ in terms of competing grammars (Kroch 1994: 2001) or a single variable grammar (e.g., Henry 2002; Seiler 2004), and if the latter, how would such a variable grammar need to be modeled? iii. bridging the gap between language typology and dialectology/sociolinguistics/language contact studies; the following quotation from Bisang (2004: 13⫺14) makes this point succinctly. Dialectology is of particular relevance for functional approaches subsumed under the term of “integrative functionalism” by Croft (1995). Departing from the existence of language-internal

842

VIII. Structural domains: Methodological problems variants, this approach assumes that individual speakers select the variant they are going to use in a given situation according to sociolinguistic criteria, i. e., according to grammar- external criteria. If this is true at the level of individual speakers, it is also true at the level of distributional patterns of linguistic types across languages because these patterns are ultimately the result of the successful diffusion of particular linguistic types and because successful diffusion depends on social criteria. From this perspective, typologists cannot but integrate findings on the diffusion of particular language structures from dialectological and sociolinguistic studies.

Finally, it should be noted that two developments within the research community of dialect syntacticians went hand in hand with the major turns in linguistics described above. These developments made dialect syntax a new research focus in dialectology, sociolinguistics, syntactic theory, and language typology. On the one hand, a generational change has taken place since the 1990s: young researchers with a broad and solid theoretical grounding, ideally but not necessarily paired with a solid dialectological training, entered the field of dialect syntax research. On the other hand, as a consequence of the “dialect turn” in syntactic theorizing, a substantial number of people working on dialect syntax are not trained dialectologists, but started out as, for example, syntacticians or typologists who often tackled the study of dialect syntax as dialectological novices but, thanks to their interest in syntactic variation, over time worked their way into dialectology. This explains why the community of dialect syntacticians in particular diverges quite markedly from the community of traditional dialectologists of several decades ago.

2.2. Methodology In contrast to phonological and lexical variation, syntactic variation can be characterized as follows: it is much subtler and less salient; it is less categorical, a matter of statistical frequency rather than presence or absence; it has a wider areal reach and is less restricted to very confined areas or individual dialects. All this taken together implies that significantly more material is necessary for the analysis of dialect syntax than for the study of accents and dialect vocabularies. The question is what the nature of this material should be, and how it is to be collected. Another crucial point to be taken into consideration in this respect follows from the shift in dialect syntax research away from studying isolated syntactic phenomena in isolated dialects towards large-scale comparative studies for a broad range of syntactic phenomena, with the ideal of a complete coverage of all dialects of a language. The greatest advances in the collection of large amounts of dialect data from as many dialects and dialect areas as possible have been made by research projects on Dutch and Flemish dialects, German (especially Alemannic and Bavarian) dialects, Scandinavian dialects, English dialects, and northern Italian dialects. (For a survey of [and links to] these projects, readers may turn to the website , which has been set up as “a digital meeting place for researchers in the field of micro-comparative syntax”. It aims at stimulating international cooperation and will for the foreseeable future serve as the single most important window onto European research on dialect syntax.) The methods used in these projects and other studies on dialect syntax are a combination of new and established methodology, with the indi-

46. Areal variation in syntax vidual methods best suited to different and often complementary purposes. Speaking of established versus new methodology in the context of dialect syntactic research is to be interpreted as follows: established essentially means established in dialectology, but new in its systematic application to the study of syntactic variation (and, of course, always exploiting the latest state-of-the-art wisdom as regards, for instance, technology and questionnaire design); new means new for the field of dialectology and the study of dialect syntax, but established as a means for exploring syntactic (and other types of) variation in other fields of linguistics, such as sociolinguistics and language typology. Briefly, the three most important of these established, (i), and new⫺ (ii), (iii) ⫺ methods applied in current research on dialect syntax can be characterized as follows, including a sketch of the specific advantages each of them offers: i.

ii.

The traditional dialectological method: standardized data collection via solid questionnaire-based fieldwork with native-speaker informants from a large number of measuring points evenly distributed across all dialect areas. Note that what is traditional about this is solely the approach as such. It is not that variation in dialect syntax has been investigated this way all along. Quite to the contrary: for example, in the Survey of English Dialects, compiled in the 1950s and for a long time (and in certain respects still) serving as the most important data source for dialectologists and dialect geographers working on English dialects, the relevant questionnaire was not at all geared to the systematic collection of syntactic data. Out of the more than 1300 questions in the SED questionnaire only a fraction had been designed to collect morphological, let alone syntactic information. This is radically different in most recent and current dialectological work, where the relevant questionnaires are exclusively designed for the collection of data on individual (morphological and) syntactic phenomena. Most advanced in this respect are the Syntactic Atlas of the Dutch Dialects (SAND; cf. Barbiers et al. 2005) with 266 survey locations in The Netherlands and Flanders and a minimum of two informants per location, and the Syntactic Atlas of Swiss German Dialects (SADS) with 385 survey locations and more than 2700 informants (on the SAND methodology cf. Cornips and Jongenburger 2001a, 2001b; for SADS methodology, Bucheli and Glaser 2002; Seiler 2005). The questionnaires used in both projects (and planned for use in the Newcastlebased Syntactic Atlas of Northern English ⫺ SANE) consist of a combination of translation, multiple choice, insertion and elicitation tasks for a selection of syntactic phenomena. Such questionnaires may include, in particular, syntactic phenomena which are only rarely observable in natural dialect discourse. Besides mere data collection (especially via the translation task), such a combination of task types allows the dialect syntactician to tackle the same phenomenon (e.g., pronominal reference, agreement phenomena, negation, quantification, word order) from different angles and in different contexts. Moreover, attitudes and consistency in the judgments of the individual informant can be tested. Another advantage offered by such a mix of questionnaire tasks is that it helps to capture several options that coexist for the individual informant; among other things, this allows the dialectologist to identify and characterize transition areas between clear-cut syntactic zones in which only one option is possible to the exclusion of all others (for examples of syntax-based areal variation, including transition zones, cf. section 3). Typology-style analytic questionnaires: filled in by (possibly even native-speaker) specialists for individual (areas of) dialects, such questionnaires ensure a high degree

843

844

VIII. Structural domains: Methodological problems of comparability among judgments made on a large number of dialects and varieties. Even in its crudest form such analytic questionnaires help collect information on whether individual syntactic phenomena are pervasive in individual dialects or dialect areas, whether they can at least occasionally be observed (by individual speakers and/or in individual contexts), or whether they are simply not part of the grammar of the relevant dialect at all. They also help making statements on syntactic features which are otherwise often overlooked in studies on individual dialects, dialect areas, or world regions (e.g., when looking just at dialects in the British Isles) because they are not considered distinctive (enough) since they are either not used pervasively or are also found in other individual dialects or dialect areas, or because they are considered strongly supraregional, e.g., as part of spontaneous spoken British English in general. Such questionnaires also help pin down the areal spread of the dialect features analyzed, ranging from very low degrees of spread (in the extreme case: restriction to a single dialect) to a very high degree, with individual features possibly even qualifying as candidates for “vernacular universals” in the sense of Chambers (2004). On a large scale, just such an analytic questionnaire (for 76 morphosyntactic features from 11 domains of grammar) has been made use of in the investigation of morphological and syntactic variation in 46 varieties of English around the world (cf. Kortmann and Szmrecsanyi 2004). Here is, for example, the relevant set of features from the domain of relative clauses: ⫺ relative particle what (e.g., This is the man what painted my house); ⫺ relative particle that or what in non-restrictive contexts (e.g., My daughter, that/what lives in London, …); ⫺ relative particle as (e.g., He was a chap as got a living anyhow); ⫺ relative particle at (e.g., This is the man at painted my house); ⫺ use of analytic that his/that’s, what his/what’s, at’s, as’s instead of whose (e.g., The man what’s wife has died); ⫺ gapping or zero-relativization in subject position (e.g., The man __ lives there is a nice chap); ⫺ resumptive/shadow pronouns (e.g., This is the house which I painted it yesterday). The other grammatical domains on which information on morphosyntactic variation across varieties of English around the world has been collected in this format are the following: pronouns, pronoun exchange and pronominal gender; the noun phrase; tense and aspect; modal verbs; verb morphology; adverbs; negation; agreement; complementation; discourse organization; and word order. Some major results stemming from this questionnaire concerning areal variation will be presented in section 3. iii. Computerized natural-discourse corpora based on interviews and/or free speech samples of dialect speakers (either monologues or conversations): as pointed out earlier, syntactic variation often operates on a very subtle level; it is in many cases not categorical, but instead a matter of statistical frequency, of preferences and preference patterns rather than mere presence or absence. In order to identify such subtle aspects of syntactic variation within and across dialects (even within the speech of the individual dialect speaker) much more material is necessary than for the study of accents and dialect vocabularies. According to some estimates, about

46. Areal variation in syntax forty times the amount of text is needed for a syntactic analysis as opposed to a phonetic one. Large amounts of electronically stored material allow for solid quantitative and qualitative morphosyntactic research, especially on high-frequency phenomena like pronominal usage, negation, tense and aspect, etc. From an areal perspective, one crucial lesson to be learnt from dialect corpora is that clear regional patterns typically emerge solely on the basis of regionally restricted (dis)preference patterns, i. e., the extent to which a certain morphosyntactic phenomenon is used in different dialects and dialect areas. (This will be shown for negative concord and relativization strategies in English dialects in section 3.1) Dialect corpora also allow an immediate comparison with the results from in-depth morphosyntactic studies based on (written and especially spoken) corpora of the standard variety/varieties of a given language (e.g., spoken Standard British English). Against this and other strengths of computerized dialect corpora, one major weakness of any corpus-based approach in linguistics is, of course, that rare phenomena may not be represented at all, and that even for relatively frequent syntactic phenomena not all of the contexts that the dialect syntactician would like to investigate are represented in the corpus. (Or take a very simple situation: in oral history interviews, dialect syntacticians interested in future time marking or markers of hypotheticality will most likely run into problems compiling a critical mass of examples which allows them to identify patterns of variation and draw general conclusions.) This is why the elicitation questionnaire method sketched in (i) above and the corpus method complement each other in important ways. So far, the corpus method in dialect syntactic research is most advanced in the study of English dialects. Currently, the largest relevant corpus which has been compiled for the purpose for morphosyntactic variation across dialects is the Freiburg English Dialect Corpus (FRED). The full version of the corpus consists of approximately 2.5 million words (which equates to about 300 hours of transcribed speech), with representative subsamples for all major English dialect areas including data from Scotland and Wales. The data in FRED are orthographically transcribed interviews collected all over the British Isles, for the most part during the 1970s and 1980s in the course of oral history projects. The majority of the informants were born between 1890 and 1920: i. e., they are roughly a generation younger than the generation of informants who were interviewed for the Survey of English Dialects (SED). For details on FRED compare Kortmann and Wagner (2005), and Anderwald and Wagner (2007). Many of the results concerning areal variation in the British Isles sketched in the following will draw on FRED-based research (for book-length accounts cf. Kortmann et al. 2005; Anderwald 2002 and 2008). Two other (much smaller and regionally restricted) corpora quite heavily exploited for dialect syntactic research on varieties spoken in the British Isles are the Northern Ireland Transcribed Corpus of Speech (NITCS), recorded in the mid1970s and comprising some 230,000 words, and the Newcastle Electronic Corpus of Tyneside English (NECTE) which consists of two subcorpora compiled in the 1960s and 1990s respectively and is designed for (primarily phonological) research on the dialect(s) spoken in the northeast of England. More information on this corpus can be found on the NECTE homepage and as part of the overview of the corpus-based approach in current dialect syntactic research edited by the makers of NECTE, namely Beal, Corrigan and Moisl (2007); for another overview on the marriage of corpus linguistics and dialectology compare Anderwald and Szmrecsanyi (2009).

845

846

VIII. Structural domains: Methodological problems Viewed together, what these three methods allow the dialectologist to do is to collect information of different degrees of granularity. Coarse-grained (macro-level) information, as it were, gleaned from typology-style analytic questionnaires primarily allows us to see the wood for the trees and develop a first large, necessarily rather crude picture of syntactic variation. This can be very helpful, however, when increasing the resolution, i. e., wanting to collect much more detailed information on more fine-grained questions. The crude picture will guide us to phenomena which are worth investigating in more detail, thus selecting syntactic features for investigation using either the more traditional elicitation questionnaires or large naturalistic corpus data. All three methods complement each other, allowing us to zoom in step by step from the global level (comparing, for example, syntactic variation in different Anglophone world regions with each other) via individual world regions (e.g., syntactic variation in the British Isles, North America, Australia), individual dialect areas (e.g., the dialects of Northern or Southwest England), and individual dialects (e.g. East Anglia, Somerset, Northumberland and Tyneside) to, ultimately, the individual dialect speaker, drawing our information on his or her dialect profile either from elicitation questionnaires or from the digitized transcript of a long stretch of recorded speech, the latter even allowing us to search for persistence (i. e., syntactic priming) effects in individual dialect speakers (cf. Szmrecsanyi 2006).

3. The nature o areal variation in dialect syntax: Small and large-scale variation in European dialects To start with, it is clearly possible to identify distinct areas of syntactic variation for dialects. This statement may sound trivial but is worth making, since traditional dialect atlases, for example, provide information almost exclusively on phonological and lexical variation. In terms of areal variation, however, dialect (morpho)syntax is not different, in principle, from what is known about areal variation in accents and dialect vocabularies. You find everything from extremely local features, documented in just one or very few dialects, via regional features (individual dialect areas or dialect zones) and supraregional features (cutting across different dialect areas) to (allegedly) universal features found in vernaculars, e.g., multiple negation, also known as negative concord. (For dialects of English, detailed accounts of the regional spread of a large set of morphosyntactic features are given in Kortmann 2002, 2004a, 2006, and in Kortmann and Szmrecsanyi 2004.) To detect and determine the areal reach of syntactic phenomena, however, you must look much harder, since (as pointed out earlier) syntactic variation is much subtler and less salient, less categorical, and in many cases a matter of statistical frequency rather than the presence or absence of a feature (both across the members of a given speech community and for the individual community member). There is also widespread agreement that grammatical variation has a wider areal reach than phonological and lexical variation. For example, there are far fewer features of dialect grammar which are restricted to just one or two local dialects (e.g., for English, the after-perfect ⫺ I am after going ‘I have just gone’, which is found only in Irish English and, exported from there, in Newfoundland English) when compared with the list of features characteristic of an entire dialect area or even several areas. In section 3.1, some examples will be given of areal variation in dialect syntax deriving from some of the research projects on

46. Areal variation in syntax European dialect syntax discussed in section 2.3. What we can learn about transitional zones in such syntax-based areal patterns will briefly be addressed in section 3.2, followed by a short account of stability and change in syntax-based dialect zones/isoglosses over time (section 3.3). For reasons of space, a map is only included for a very few of the relevant patterns

3.1. Areal patterns Among the Swiss German dialects, it is possible to identify clear north⫺south and east⫺ west contrasts with regard to individual morphosyntactic phenomena. North⫺south contrasts have emerged, for example, with regard to the passive construction (Bucheli Berger 2005) and the marking of so-called depictives, more exactly copredicative adjectives and past participles (Bucheli Berger and Glaser 2004; Bucheli Berger 2005). Depictive marking always involves the relevant adjective (e.g., heiss ‘hot’ in [1a]) or past participle (e.g., gfror- ‘frozen’ in [1b]) exhibiting agreement in gender and number with its controller, i. e., either the subject or the direct object. This is illustrated in the following examples from the Wallis dialect: (1) a. Dü müoscht d-Milch de heiss-i triich-u. you must the-milk.f.sg then hot-f.sg drink ‘You must drink the milk (while it is still) hot.’ b. Fisch-stäb-jini müoscht doch gfror-ni abraat-u. fish-finger-dim.pl must:2sg but pst.ptcp-freeze-pl fry-inf ‘But you have to fry fish fingers (while they are still) frozen.’ In Map 46.1 the north⫺south contrast is apparent in the absence of depictive marking in the northern dialects as opposed to the fully productive depictive marking in the Highest Alemannic dialects in the south and southwest of German-speaking Switzerland (see the triangles). Map 46.1 at the same time exhibits an east⫺west contrast, with the southern and southwestern area marked with triangles as opposed to the area with the dots in the northeast of Switzerland. The dialects in the latter area also exhibit depictive marking, but the relevant marker is invariable; it does not follow the regular agreement patterns. Regardless of the gender and number of the subject/object which the adjective or past participle co-predicate something of, an invariable marker -ä/nä (or -a/-na, -e/ -ne) is used, depending on the phonology of the relevant northeastern dialect (thus, for example, gfror-nä instead of the regular plural marker gfror-ni in the south). Along the east⫺west axis in Swiss German dialects there is also clear-cut areal variation in the marking of purpose clauses (cf. Seiler 2005) and verb clusters consisting of a minimum of two verbs (cf. Seiler 2004). With respect to the marking of purpose clauses, Seiler (2005) finds that the für … z marker (formally and functionally equivalent to English for … to) is the western variant, whereas z’/zum ⫹ inf or z’/zum …⫹ inf (formally equivalent to English to/to the ⫹ inf/gerund) are characteristic purpose-marking patterns among the eastern dialects. Contrast (2a) and (2b), both of which mean ‘We need that for repairing the car’:

847

848

VIII. Structural domains: Methodological problems

Map 46.1: Swiss German dialects: Depictives (or co-predicatives), as in ‘You must fry fish fingers frozen’ (Bucheli Berger 2005: 144)

(2) a. Mir bruche das für dr Wage z’repariere. we need that for the car to repair b. Mer bruuche daas zum de Waage reperiere. we need that to.the the car repair

(Berne) (Zürich)

From the study of Dutch dialect syntax, one phenomenon which yields nice areal patterns is subject pronoun doubling (cf. De Vogelaer 2008 and, for subject doubling in Dutch dialects in general, De Vogelaer and Devos 2008). Here are some examples of subject doubling in sentences with inverted (VS) word order (3a) and normal (SV) word order (3b and c): (3) a. Ga-de (gij) naar Brussel? go.2sg-youclitic youstrong to Brussels ‘Are you going to Brussels?’ b. Ge-gaat naar Brussel. (gij) youclitic-go.2sg youstrong to Brussels ‘You are going to Brussels.’ c. Gij gaat gij naar Brussel. youstrong go.2sg youstrong to Brussels ‘You are going to Brussels.’

46. Areal variation in syntax

Map 46.2: Dutch dialects: Subject doubling, inverted word order (De Vogelaer 2008)

Of interest to us is subject doubling such as in (3a), i. e., sentences with inverted word order, an obligatory subject clitic, -de, and an optional pronoun (gij) doubling the subject clitic for pragmatic or stylistic reasons. In some Dutch dialects, subject doubling is found only for a limited set of subject pronouns, e.g., first and second person singular as well as second person plural in the city dialect of Antwerp. Map 46.2 shows that three broad areas can be distinguished on the basis of whether or not a given dialect may use all personal pronouns in contexts such as (3a), and if not, which personal pronouns are allowed: just second person (singular and plural), or first person singular as well. Roughly, the three areas follow a west⫺(north)east axis. Before moving on in greater detail to dialects of England, one major finding about areal variation in northern Italy may be mentioned here (for general information on the Syntactic Atlas of Northern Italy (ASIS) cf. Beninca` and Poletto 1998⫺2002 and Poletto 2000). For the Lombard dialects, three broad groups of dialects can be distinguished when exploring subject clitics (cf. Murelli 2006: 47⫺50 based on ASIS data): alpine dialects, eastern and western dialects. This threefold distinction can be boiled down to an even more basic one, namely dialects spoken in the mountainous regions (alpine plus northeastern dialects) and those spoken in the plains region (rural western and southeastern dialects). The former are more conservative than the latter, in that they rely more strongly on subject clitics in order to distinguish between different grammatical persons. In the dialects of England there are many interesting north⫺south contrasts. First of all, there are features which are exclusively found in either the northern dialects or in the southern, more precisely southwestern, dialects ⫺ both dialect areas constituting highly traditional dialect areas of England reaching back into the Middle English period (see also section 3.3). Within England, the so-called Northern Subject Rule (NSR), for example, is found exclusively in the dialects of the North; however, it is also widely

849

850

VIII. Structural domains: Methodological problems attested in other northern parts of the British Isles, namely in Scotland, Northern Ireland and the Republic of Ireland. This rule is all about subject⫺verb agreement and can be formulated roughly as follows: every verb in the present tense can take an -s ending unless its subject is an immediately adjacent simple pronoun. (Third person singular verbs always take the -s ending, as in Standard English.) In other words, the NSR involves a type-of-subject constraint (pronoun versus common/proper noun) and a position constraint (⫹/⫺ immediate adjacency of pronominal subject to verb). Thus, in NSR varieties we get the following examples: (4) a. I sing (vs. *I sings) b. Birds sings. c. I sing and dances Zooming in on one of the NSR regions in the British Isles, Pietsch (2005a) observed an interesting areal distribution applying to the NSR in Northern Ireland. On the basis of the Northern Ireland Transcribed Corpus of Speech, Pietsch identified one central northwest-to-southeast belt in which the NSR usage is most pronounced. Surprising about this is that this core NSR area is not identical with the core Ulster Scots areas, contrary to expectations based on the widely entertained view that the NSR is historically most closely linked to Scots. Rather, its regional distribution supports independently formulated views that the NSR in Ulster has its historical roots in varieties spoken by both Scottish and English settlers. The dialects of the southwest exhibit a number of morphosyntactic features which are either found exclusively here, or which are highly distinctive of this dialect area and not found in many others (cf. Kortmann 2002; Wagner 2005). A syntactic feature of the former type is known as pronominal gender (or alternatively, gendered pronouns, gender diffusion; cf. Wagner 2004), a feature which relates to a typologically very rare semantic system of gender marking via pronouns (cf. also Siemund 2008). What we encounter in Somerset, in particular, is pronominal gender that is primarily sensitive to the mass/ count distinction and only secondarily to the animate/inanimate and human/nonhuman distinctions. It is only used for mass nouns. Count nouns take either he or she: she is used if the count noun refers to a female human, and he is used for count nouns either referring to male humans or to nonhuman entities. Thus we get a contrast as in (5a) and (5b): (5) a. Pass the bread⫺it’s over there. b. Pass the loaf⫺he’s over there.

[bread ⫽ mass noun] [loaf ⫽ count noun]

Again zooming in on individual regions of southwest England, we see that the areal variation we find in the FRED corpus and in the SED fieldworker notebooks differs from standard dialectological textbook wisdom. The established view is that most southwestern dialect features are found in the Somerset dialect, and that Cornwall, especially West Cornwall, should generally not be considered part of the southwestern dialect area. However, as Wagner (2004) found, the story told by both the relevant SED and FRED material is a quite different one, not only with respect to the phenomenon of pronominal gender, but for other morphosyntactic phenomena, too. Typical dialectal features of the southwest are quite (partly even most) pronounced in West Cornwall and, at least in a

46. Areal variation in syntax

851

Tab. 46.1: Distribution of relative markers along the north⫺south axis in percentages (Herrmann 2005: 27)

North

왖 Ô Ô 왔

South

Northern Ireland Scotland Central North Central Midlands East Anglia Central Southwest

zero

that

what

as

46.9 % 23.6 % 34.0 % 17.7 % 20.4 % 28.9 %

50.1 % 46.2 % 43.5 % 40.3 % 22.0 % 26.5 %

⫺ 0.4 % 2.4 % 5.8 % 15.9 % 22.3 %

0.5 % ⫺ 1.4 % 2.4 % ⫺ ⫺

few cases (notably including pronominal gender), considerably less pronounced in Somerset. On a larger scale, i. e., with regard to syntactic phenomena traditionally assumed not to exhibit a marked regional distribution, it is also possible to identify clear differences along the north⫺south axis in England. Most of these differences are not of a categorical type, as for the NSR or pronominal gender, but of purely quantitative nature, i. e., clear preference patterns can be identified on the basis of large-scale cross-dialectal (microcomparative) analyses of SED and, especially, corpus (FRED, NITCS) data. For relative markers, for example, Hermann (2005) found the following north⫺south distribution in her FRED-based comparative study of relative clauses in six major dialect areas (see Table 46.1). Zero relatives for the subject position, as in There’s a man __ sits in the garden, are used much more frequently in Northern Ireland than in any of the other five dialect areas investigated. As for relative particles, the established marker that (The man that came in is my brother) is used much more frequently in a large area stretching from the central Midlands to the north than in the south, whereas what (as in The man what came in is my brother), the newcomer among the relative markers in English dialects, seems to make its inroads clearly from the south (East Anglia, central Southwest) and is hardly used of yet in the central Midlands and the north. The receding traditional relative particle as, on the other hand, has clearly remained a supraregional feature, but only in the north and the central Midlands, and is very much on the way out even there. A perhaps even more surprising areal distribution concerning a syntactic dialect (or general nonstandard) feature has been identified by Anderwald (2002: 109⫺114, 2005). The phenomenon at issue is negative concord (or “multiple negation”), as in I’ve never been to market to buy no heifers. If there is one safe candidate for a supraregional feature of nonstandard syntax that probably even a nonlinguist would spontaneously point to, it is surely negative concord. The surprising finding for England, Scotland and Wales is this: in analyzing both the spoken subsample of the British National Corpus and FRED data, what emerges is a clear south⫺north cline, with rough proportions of multiple negation usage in the FRED data of 40⫺45 percent in the south of England, 30 percent in the Midlands, and around ten percent in the north of England, Scotland and Wales. Finally, as a result of the analytic questionnaire method (see [ii] in section 2.2) for capturing areal variation on a very large scale, the comparative study of the morphology and syntax of the nonstandard varieties of the British Isles yields a recurrent north⫺south pattern (cf. Kortmann 2004a). Based on the 76-feature catalogue investigated worldwide in the Handbook of Varieties of English (Kortmann and Schneider 2004), a marked north⫺south divide emerges for two subsets of morphosyntactic properties. The core of

852

VIII. Structural domains: Methodological problems the north is constituted by Scottish English, the Orkney and Shetland dialects, and the dialects of North England, whereas the south is constituted by the southwest, the southeast, and East Anglia. For most of the relevant features, Irish English (not least due to the English spoken in Northern Ireland) patterns with the varieties in the north, and Welsh English with those in the south. The features in (6) are exclusively or almost exclusively found in the varieties of the north; those in (7) form the corresponding set for the south: (6) a. special forms or phrases for the second person plural pronoun (e.g., youse, y’all, aay’, yufela, you… together, all of you, you ones/’uns, you guys, you people) b. wider range of uses of the progressive (e.g., I’m liking this, What are you wanting?) c. be as perfect auxiliary (e.g., They’re not left school yet) d. double modals (e.g., I tell you what we might should do) e. epistemic mustn’t ‘cannot; it is concluded that … not’ (e.g., This mustn’t be true) f. resumptive/shadow pronouns (e.g., This is the house which I painted it yesterday) g. lack of inversion / lack of auxiliaries in wh-questions (e.g., What you doing?) h. lack of inversion in main clause yes/no-questions (e.g., You get the point?) i. Northern Subject Rule j. the relative particle at (e.g., This is the man at painted my house) a-prefixing on ing-forms (e.g., They wasn’t a-doin’ nothin’ wrong) ain’t as the negated form of be (e.g., They’re all in there, ain’t they?) ain’t as the negated form of have (e.g., I ain’t had a look at them yet) invariant non-concord tags (e.g., innit/in’t it/isn’t in They had them in their hair, innit?) e. relative particle what (e.g., This is the man what painted my house)

(7) a. b. c. d.

From here one could even move on to areal variation on a global scale, namely identifying the areal reach and preference patterns of morphosyntactic features when comparing different anglophone world regions with each other (for L1 varieties, i. e., traditional and modern dialects, notably the British Isles and North America). It is indeed possible to identify features which are either restricted to or strongly preferred in one of these two world regions and not the other. For instance, using was sat or was stood with a progressive meaning is a British Isles feature not to be observed in US dialects. Vice versa, ain’t as the negated form of be and have is a pervasive North American feature found only in some dialects of England. For an in-depth account of areal variation of such a global scale, which so far has only been studied for nonstandard varieties of English, the reader is referred to Kortmann and Szmrecsanyi (2004). It should also be kept in mind that in such a large-scale comparison of the morphosyntax of varieties of English around the world, variety type ⫺ i. e., whether L1, L2, or contact varieties (notably pidgins and creoles) are studied ⫺ has been found to be a much better predictor of their morphosyntactic behavior than geography (cf. Szmrecsanyi and Kortmann 2009a). In other words, regardless of the region in which a given variety is spoken, each of the three broad variety types exhibits a distinctive morphosyntactic profile: for example, L2 varieties of English in Africa will be much more similar to L2 varieties of English from elsewhere in the world than to L1 varieties of English spoken in Africa.

46. Areal variation in syntax

3.2. Transitional zones In identifying such dialect areas for individual syntactic phenomena as in section 3.1, what typically also emerges are transitional zones: recall the subject pronoun doubling in Dutch dialects shown on Map 46.2, or the English Midlands as a transition zone from infrequent to fairly frequent use of negative concord among the dialects of England. In other words, the typical situation for syntactic variation, especially on a supraregional scale, is that there are no sharp isoglosses (which are largely a myth of the pre-quantitative age in the study of areal variation in traditional dialectology anyway), and that the situation can rather be likened to a cline or slope. The following properties can be observed in such transitional zones (cf. especially Seiler 2005). Each individual transition zone may exhibit at least one, but typically a combination, and upon close investigation perhaps even all, of these properties: i.

Frequency: frequency rates for the morphosyntactic phenomenon at hand (be it text frequency in corpora, frequency in terms of survey locations or of mention by informants) are somewhere in between the relevant rates for the clearly distinguishable syntactic areas (just recall section 3.1: compare the percentages for the English Midlands with the percentages of the dialects of the north and the south for relative markers [Table 46.1] and negative concord). ii. Number of variants: where two dialect areas clearly delineated on syntactic grounds differ in terms of the number of variants they allow (e.g., in Map 46.2 on subject pronoun doubling, all subject pronouns are doubled in sentences with inverted SV order in the (south)western core area, in sharp contrast to the peripheral area in the northeast, where only the second person pronouns are affected), the number of variants to be observed in the transition zone lies between the maximum and minimum in the two neighboring zones (e.g., in Map 46.2, there where three options are allowed, namely the two second person pronouns and the first person singular); but a transition zone may also allow more variants than either of its neighboring zones, as will next be shown. iii. Preference and optionality: the variants of a given syntactic phenomenon which are found exclusively or strongly preferred in clearly identifiable syntactic areas (e.g., variant A in zone 1 and variant B in zone 2) are found side by side in the transition zone (i. e., both A and B) and may even combine (AB). The latter can, for example, be very nicely observed with regard to variation in negative concord along an east⫺ west axis among the Dutch dialects. As shown by van der Auwera and Neuckermans (2004), differing types of double negation are used in West Flemish dialects (en niemand ‘not no-one’) and the dialects of Brabant and Limburg (niemand niet ‘noone not’). What separates West Flanders from the provinces of Brabant and Limburg is East Flanders. And it is precisely in the East Flemish dialects that the following can be observed: triple negation is used by many speakers (en niemand niet ‘not no-one not’), but these speakers also opt for one of the double negation variants found in the dialect areas west and east of East Flanders. What may also be observed is a clear regional skewing in the transition zone as regards additional options: if, say, variant A is the sole or strongly preferred variant in zone 1 and, correspondingly, variant B in zone 2: then what can be observed in a transition zone is, for example, that there is a shift from variant A to variant B as the variant preferred by informants as we move away from zone 1 towards zone 2.

853

854

VIII. Structural domains: Methodological problems iv.

v.

Division of tasks among competing variants: in a transition zone where both the variants of the dialect zones it separates can be found, the use of these variants is typically subject to syntactic constraints, i. e., in the transition zone each of variants A and B can be used only in a subset of the contexts in which they are used in zones 1 and 2 respectively. As a consequence of this, it may be observed, as Pietsch (2005a) has done, that it is in the zone at the southern boundary of the Northern Subject Rule that relevant constructions cannot be used as freely as in the core areas in the north, and are instead more or less frozen, approximating fixed expressions.

3.3. Continuity and change Of course there is also an important historical dimension to the areal distribution of syntactic phenomena in dialects. As far as we have reliable information on syntax-based areal patterns in historical dialectology ⫺ and reliable information meeting the standards of current dialect syntactic research we have on virtually nothing in historical dialect syntax, not even for the very well-described West Germanic dialects (cf. Kortmann and Wagner 2007, to appear) ⫺ it is possible to identify quite a number of syntactic features for which the areal reach seems to have remained relatively, or even absolutely, stable for a century or longer. An extreme example is the Northern Subject Rule, the areal distribution of which is largely identical to that which can be observed, for example, in the Late Middle English period (cf. Pietsch 2005b: 178). The southern boundary of the NSR zone in the SED data (based on informants born mainly between 1870 and 1890) is practically identical with the southern boundary in Late Middle English times. A high degree of regional stability over a period of around a century (mid-nineteenth until midtwentieth century) has also been observed for the phenomenon of pronoun exchange (subject pronoun in object position: They always called I “Willie”, see; object pronoun in subject position: We used to stook it off didn’t us?) by Kortmann and Wagner (to appear). In German dialectology, a remarkable stability has been observed for some of the syntactic isoglosses based on the famous nineteenth-century Wenker sentences (cf. Schmidt 2005; for a fabulous state-of-the-art cartographic representation of the Wenker maps the reader is strongly encouraged to consult the Digital Wenker Atlas ). For example, the isogloss for the interrogative pronoun ‘what’ (wat versus was), which is part of the Rhenish fan, has not changed at all between 1880, when Wenker collected his information, and 1980/1985, when this phenomenon was investigated for different age groups of dialect speakers by members of the team from the Deutscher Sprachatlas in Marburg, the research center for German dialects, but not (yet) for dialect syntax. Besides these continuities, there are clearly also changes in the areal reach of individual features of dialect syntax which have taken place over time. In section 3.1, we mentioned for example the use of as (and likewise at) as a relative particle has strongly receded, even in its northern core distribution area, whereas the relative particle what seems to be an innovation currently spreading from the south of England. So-called pseudo-passives with stood and sat (i. e., passive forms with progressive meaning: I was stood/sat … ‘I was standing/sitting …’) are also most likely a Late Modern English, chiefly nineteenth-century, innovation, which is currently diffusing from the northern

46. Areal variation in syntax and western parts of England to spontaneous spoken British English (nonstandard and semistandard) in general (cf. Kortmann and Wagner, to appear). Readers should note in passing that, from a global perspective, this feature is relatively rare in nonstandard varieties of English around the world and not found at all in L1 varieties of English in the US. Another interesting story of change can be told about unstressed do as a pure tense marker in declaratives, as in John did see you yesterday. Among the dialects of England it is a distinctive feature of the southwest and has been so since Middle English times. During the Early Modern English period it spread vigorously, though, and could also frequently be found in the standard (both spoken and written), not just for rhythmic purposes in poetry and drama. Then, however, it fell into disuse again in the standard, receding to that small area which had been its heartland since medieval times, the southwest. From a larger comparative perspective, it is interesting to note that precisely this use of periphrastic do ⫺ the rarest of the various uses of do as a tense and aspect marker in the dialects of England and varieties of English worldwide ⫺ is the most widespread of the uses of periphrastic do (German tun, Dutch doen) in the Continental West Germanic dialects (cf. Kortmann 2004b: 247). In general, statements on continuities and changes in the domain of dialect syntax should be made and taken only with a generous portion of caution. This even applies to statements reaching back to the best-documented historical dialects, namely those of the nineteenth century, but much more so the further back in time we go. Even where information on the syntax of nineteenth century dialects is available, the relevant studies invariably mention dialect features only for their presence or absence. Current accounts of dialect syntax, as we have tried to show in sections 3.1 and 3.2 in particular, tend to focus on the distribution of features in terms of frequencies of occurrence, and the range of contexts in which the relevant (variants of) constructions are used, etc., which makes it very difficult to compare feature catalogues from the nineteenth century or earlier with descriptions of late nineteenth or twentieth-century variation that are focused on quantitative statements.

4. Areal variation and (areal) typology Areal variation is central not only in dialectology, of course, but also in areal typology (cf. also Goebl 2001 and Bisang 2004). Above all, the (search for and) study of convergence areas across a set of (especially genetically unrelated) languages spoken in the same geographical area cannot afford to ignore dialectal (including sociolinguistic) variation, as is frankly admitted by Dahl (2001: 1460): Almost by definition, areal linguistics neglects the social dimension of language diversity. More generally, the whole notion of “areal phenomena” as well as the idea of drawing language maps, build on the convenient fiction that each language has a specific location in space, that no more than one language is spoken in each place, and that language contact takes place between adjacent languages.

Moreover, language contact does not take place between the (written) standard varieties of languages, but in day-to-day language contact between the spoken (typically nonstandard) varieties of the adjacent languages involved. Thus, on the one hand, the inclusion

855

856

VIII. Structural domains: Methodological problems of dialects is bound to offer new insights to areal typology on a small scale, i. e., concerning the areal distribution of individual syntactic variables in regionally restricted convergence areas. (This is established practice for studies on the Balkan Sprachbund, but is necessary also in the exploration of other contact superposition zones like, for example, the Circum-Baltic area and the Mediterranean.) On a larger scale, the systematic study of grammatical variation in dialects could make a significant contribution to exploring whether Europe, or a large part of continental Europe, forms a linguistic convergence area (Sprachbund Europe). It is striking, and worrying at the same time, that dialects played a rather marginal role in such a major international project as “Typology of European Languages” (EUROTYP; 1990⫺1994). The ultimate question which the inclusion of dialects in European areal typology might answer is what the “real”, i. e., spoken nonstandard, landscape of Europe looks like: What is the nature of “real” Standard Average European, which was hypothesized by Benjamin Lee Whorf about sixty years ago and largely confirmed by the findings of the EUROTYP project (cf. Haspelmath 2001)? Compare Kortmann (to appear) on the contribution that the cross-dialectal study of dialect syntax can make to the study of cross-linguistic variation in language typology, especially in an exploration of the linguistic landscape of Europe. Another aspect to be kept in mind in this respect is that in a number of cases the grammar of the standard variety of a language is strikingly different from the (vast majority of the) nonstandard varieties. Negative concord is a notorious example: it is typical only of the standard varieties of the Germanic languages, but not for the nonstandard varieties (cf. also Weiss 2004: 27⫺30). A more subtle example can be given from English: in typological studies of relative markers in the European languages, English is characterized as a language using, like almost all (standard varieties of) European languages, relative pronouns. Take Bernard Comrie, who characterizes the relative pronoun strategy as “a remarkable areal typological feature of European languages, especially the standard written languages” (quoted in Haspelmath 2001: 1495). However, in nonstandard varieties of English, as Hermann (2005) has clearly shown, there is not a single dialect that does not use the relative particle(s) (that, what, as/at) more frequently than the relative pronoun (i. e., the wh-) strategy. For other morphosyntactic features, too, Standard British English clearly turns out to be the odd man out among the varieties of English. As a consequence, findings from studies on dialect syntax may also serve as a crucial corrective for typological classifications, in general. Vice versa, language typology may serve as an important (additional) reference frame and source of inspiration for dialectology when, for example, selecting phenomena to be investigated and appropriate methodologies (recall, for example, the analytic questionnaire method in section 2.2), shedding light and interpreting the observable variation in light of what is known about patterns and limits of large-scale cross-linguistic variation. One example may suffice: interpreting dialect variation in the domain of relative clauses with the help of a typological hierarchy, more exactly the very hierarchy which made the concept of hierarchies such a successful and firmly established one in functional typology, namely Keenan and Comrie’s (1977: 1979) NP Accessibility Hierarchy: (8) Accessibility Hierarchy (AH) subject > direct object > indirect object > oblique > genitive > object of comparison

46. Areal variation in syntax

857

In various studies on English and German dialects this hierarchy has been drawn upon (notably in Herrmann 2005 and Fleischer 2004). Let us look at just one particular observation concerning relative clauses in Standard English vis-a`-vis English dialects. According to the Accessibility Hierarchy, if a language can relativize any NP position further down on the hierarchy, it can also relativize all positions higher up, i. e., to the left of it. This constraint applies to whatever relativization strategy a language employs. For the relativization strategy known as zero-relativization (or gapping) there is thus a clear prediction that the relativized NP is most likely to be gapped if it is the subject of the relative clause, next most likely if it is the direct object of the relative clause, etc. However, this is clearly not the case for Standard English: the direct object position can be gapped (9a), whereas the subject position cannot (9b): (9) a. The man I called ____ was our neighbor. b. *The man ____ called me was our neighbor.

[direct object] [subject]

English dialects, on the other hand, conform to the Accessibility Hierarchy prediction. Examples like (9b) are nothing unusual at all (cf. Table 46.1); in fact, gapping of the subject position is an extremely widespread phenomenon in nonstandard varieties of English in and outside the British Isles: (10) a. I have a friend ___ lives over there. b. It ain’t the best ones ___ finish first. So here we have a striking instance of the situation where the nonstandard varieties of English conform to a typological hierarchy whereas the standard variety does not.

5. Conclusion and outlook The very end of the twentieth century witnessed a most powerful awakening of the long dormant interest in dialect syntax in general, and in areal patterns of dialect syntax in particular. Virtually no collection of dialectological articles is published these days which does not include chapters on dialect syntax. All of a sudden, the morphology and syntax of regional varieties (but also other nonstandard, e.g., social or contact, varieties) have attracted attention from fields as diverse as syntactic theory, variationist linguistics and language typology. In conclusion, some of the most promising research avenues in exploring (morpho)syntactic variation in dialects for the next two decades will be outlined, with the mapping of areal variation in a more narrowly dialectological sense as only one research avenue, albeit a central one. To start with, given the existing research groups on dialect syntax in Europe, with a powerhouse such as the Meertens Institute Amsterdam and its European Dialect Syntax (Edisyn) project helping to establish a Europe-wide network of dialect syntacticians, the data situation for the study of dialect syntax in a number of regions in Europe is set to steadily and significantly improve in the near future. Data will be collected in as refined a way as in the SAND project in The Netherlands and Flanders, the SADS project in Switzerland, the current NORMS project in Scandinavia (Nordic Centre of Excellence

858

VIII. Structural domains: Methodological problems in Microcomparative Syntax; cf. Bentzen and Vangsnes 2007), or the envisaged Syntactic Atlas of Northern English (SANE) project in England. For a growing number of dialects, the whole spectrum of data sources needed for the analysis of syntactic phenomena will be available, notably including interview-based data, data based on both elicitation and analytical questionnaires, and electronically stored (orthographic) transcriptions of longer stretches of natural discourse (monologues or spontaneous conversation) to which the whole corpus linguistic toolkit can be applied. The crucial point, however, is not simply quantity, but the quality of the data. What is called for, and what the Edisyn project in particular tries to orchestrate, is an agreement on standards of data collection, so that ideally syntactic data are collected on whatever dialect of whatever language according to a uniform pattern, ensuring immediate cross-dialectal, cross-linguistic and cross-theoretical comparability. Moreover, one may even define a core of syntactic domains that are known to exhibit interesting regional variation in many languages, so that, for example, the same set of sample sentences is taken as a basis for translational questionnaires. Among the current European projects on dialect syntax, the following domains appear to be most promising candidates for such a core: agreement, complementation, negation, relativization, pronominal systems, clitics and, cutting across various of these domains, syntactic doubling, which has been made the research focus of Edisyn (cf. the volume on doubling phenomena in the dialects of Europe by Barbiers et al. [2008]). These significant advances concerning the data situation will allow the following issues to be addressed in future research: i.

The areal scope of dialect syntactic studies should be increased. On the one hand, this can be achieved by exploring variation concerning syntactic phenomena in dialects of genetically closely related languages. For example, for any speaker of a West Germanic language, the parallels in many domains of syntax to be observed for the regional dialects of Dutch, English and German are quite striking. On the other hand, following the model set for English by the global approach taken in the Handbook of Varieties of English (Kortmann and Schneider 2004), similar studies of syntactic variation around the globe should be carried out especially for European languages with a colonial history to rival that of English, notably Spanish, French, Portuguese and Dutch. ii. Findings from dialect syntax should be systematically integrated into language typology and areal typology, much along the lines suggested in section 4 (cf. also Kortmann 2004c, to appear). iii. Findings from dialect syntax should be systematically integrated into research on grammaticalization. This is to some extent a natural consequence of the proposition in (ii), since grammaticalization has for some time now been the central research topic in diachronic typology. What is more important, however, is that grammaticalization starts from spontaneous spoken language, and that regional and social (nonstandard) varieties are thus to be seen as goldmines for identifying further grammaticalization phenomena and shedding light on known ones. For some initial studies conducted in this spirit compare Dahl (2004), Kortmann (2004b) and De Vogelaer (2008). A suitable transition to the next research issue can be made in the knowledge that, increasingly, attempts are being made to model grammaticalization phenomena in dialects in terms of constructionist grammars.

46. Areal variation in syntax iv.

v.

Another rich area of future research on dialect syntax will be the theoretical modeling of syntactic variation: there will be formalist-driven, functionalist-driven, constructionist, integrated, variationist and possibly other approaches, all in lively and fruitful competition with each other ⫺ the crucial point in all of them being that, after decades of disregard, it is variation which is at the centre of linguistic reality and that it is thus variation that is considered to be the core explanandum in syntactic theory. More studies are needed which systematically test hypotheses such as the one put forward by Trudgill (2009) on the correlation between the degree of contact to which varieties of a language are or have been exposed and their degree of structural complexity (in grammar and phonology). Trudgill’s claim is that high-contact varieties (mainstream varieties, including the standard varieties of English, L2 varieties, contact and shift varieties) undergo simplification, whereas the grammars and phonological systems of low-contact varieties (especially the traditional rural dialects) are (and can afford to have) more complex grammatical and phonological systems. For varieties of English, a first set of studies by Szmrecsanyi and Kortmann (2009a⫺c) seems to confirm this hypothesis. But more studies on English and, in particular, on the varieties of other languages are needed. Is there, for example, evidence across languages that the still central object of study for dialectologists, namely the low-contact L1 variety, is morphologically and syntactically more complex than its high-contact L1 counterpart?

There is a historical dimension to the previous point, of course, in that extensive language contact (among adults) plays a crucial role in the shaping of the structure of a language or variety. This historical dimension is covered by points (vi) and (vii): vi. Present-day data from low-contact rural communities in Europe in particular may increasingly play a key role in reconstructing paths of language change, for example in the reconstruction (of missing links in the history) of younger, transplanted or contact varieties based on a European superstrate in other parts of the world. For English, this line of research, using refined quantitative variationist methodology, has been explored in a large number of studies by Tagliamonte and co-workers (e.g., Tagliamonte, Smith and Lawrence 2005). vii. Finally, there is still a lot of work to be done on syntactic variation in historical dialectology: every effort possible needs to be made in order to improve the data situation before, in a second step, research questions can be asked of the historical material that would allow an immediate comparison of the findings on historical dialects with findings and interpretations concerning syntactic variation, including the areal reach of syntactic phenomena, in present-day dialects.

6. Reerences Abraham, Werner and Josef Bayer (eds.) 1993 Dialektsyntax. Opladen: Westdeutscher Verlag. Adger, David and Graeme Trousdale 2007 Variation in English syntax: Theoretical implications. In: Graeme Trousdale and David Adger (eds.), 261⫺278.

859

860

VIII. Structural domains: Methodological problems Anderwald, Lieselotte 2002 Negation in Non-standard British English: Gaps, Regularizations and Asymmetries. London/New York: Routledge. Anderwald, Lieselotte 2005 Unexpected regional distributions: Multiple negation in FRED. In: Yoko Iyeiri (ed.), Aspects of Negation, 113⫺137. Kyoto: Kyoto University Press. Anderwald, Lieselotte 2008 The Morphology of English Dialects: Verb-Formation in Non-Standard English. Cambridge: Cambridge University Press. Anderwald, Lieselotte and Benedikt Szmrecsanyi 2009 Corpus linguistics and dialectology. In: Merja Kytö and Anke Lüdeling (eds.), Handbook of Corpus Linguistics, 1126⫺1139. Berlin/New York: Mouton de Gruyter. Anderwald, Lieselotte and Susanne Wagner 2007 The Freiburg English Dialect Corpus (FRED): Applying corpus-linguistic research tools to the analysis of dialect data. In: Beal, Corrigan and Moisl (eds.), 35⫺53. Barbiers, Sjef, Hans Bennis, Gunther De Vogelaer, Magda Devos and Margreet van der Ham 2005 Syntactic Atlas of Dutch Dialects, vol. 1: Pronouns, Agreement and Dependencies. Amsterdam: Amsterdam University Press. Barbiers, Sjef, Leonie Cornips and Susanne van der Kleij (eds.) 2002 Syntactic Microvariation. Amsterdam: SAND. Available from . Barbiers, Sjef, Olaf Koeneman, Marika Lekakou and Margreet van der Ham (eds.) 2008 Microvariations in Syntactic Doubling. Bingley: Emerald. Beal, Joan, Karen Corrigan and Hermann Moisl (eds.) 2007 Using Unconventional Digital Language Corpora, vol. 1: Synchronic Corpora. Basingstoke: Palgrave MacMillan. Bender, Emily M. 2007 Socially meaningful syntactic variation in sign-based grammar. In: Trousdale and Adger (eds.), 347⫺381. Beninca`, Paola (ed.) 1989 Dialect Variation and the Theory of Grammar. Dordrecht: Foris. Beninca`, Paola and Cecilia Poletto 1998⫺2002 Quaderni di lavoro dell’ASIS (Atlante Sintattico dell’Italia Settentionale). Available from . Bentzen, Kristine and Øystein A. Vangsnes (eds.) 2007 Scandinavian Dialect Syntax 2005. Nordlyd⫺Tromsø University Working Papers in Language & Linguistics [special issue]. Bisang, Walter 2004 Dialectology and typology ⫺ An integrative perspective. In: Kortmann (ed.): 11⫺45. Black, James R. and Virginia Motapanyane (eds.) 1996 Microparametric Syntax and Dialect Variation. Amsterdam/Philadelphia: Benjamins. Bresnan, Joan and Ashwini Deo 2001 Grammatical constraints on variation: ‘Be’ in the Survey of English Dialects and (Stochastic) Optimality Theory. Unpublished manuscript, Stanford University. Bresnan, Joan, Ashwini Deo and Devyani Sharma 2007 Typology in variation: A probabilistic approach to be and n’t in the Survey of English Dialects. In: Trousdale and Adger (eds.), 301⫺346. Bucheli, Claudia and Elvira Glaser 2002 The syntactic atlas of Swiss German dialects: Empirical and methodological problems. In: Barbiers, Cornips and van der Kleij (eds.), Syntactic Microvariation, 41⫺74. Bucheli Berger, Claudia and Elvira Glaser 2004 Zur Morphologie des (ko)prädikativen Adjektivs und Partizips II im Alemannischen und Bairischen. In: Franz Patocka and Peter Wiesinger (eds.), Morphologie und Syntax

46. Areal variation in syntax deutscher Dialekte und Historische Dialektologie des Deutschen. Beiträge zum 1. Kongress der IGDD, Marburg/Lahn, 5.⫺8. März 2003, 189⫺226. Vienna: Edition Praesens. Bucheli Berger, Claudia 2005 Depictive agreement and the development of a depictive marker in Swiss German dialects. In: Nikolaus Himmelmann and Eva Schultze-Berndt (eds.), Secondary Predication and Adverbial Modification. Crosslinguistic Explorations in the Syntax and Semantics of Depictives, 141⫺171. Oxford: Oxford University Press. Chambers, Jack 2004 Dynamic typology and vernacular universals. In: Bernd Kortmann (ed.): 127⫺145. Cornips, Leonie and Karen Corrigan 2005a Towards an integrated approach to syntactic variation: A retrospective and prospective synopsis. In: Cornips and Corrigan (eds.), 1⫺27. Cornips, Leonie and Karen Corrigan (eds.) 2005b Syntax and Variation: Reconciling the Biological and the Social. Amsterdam: Benjamins. Cornips, Leonie and Willy Jongenburger 2001a Het design en de methodologie van het SAND project. Nederlandse Taalkunde 16: 215⫺232. Cornips, Leonie and Willy Jongenburger 2001b Elicitation techniques in a Dutch syntactic dialect atlas project. In: Hans Broekhuis and Ton van der Wouden (eds.), Linguistics in the Netherlands 18, 57⫺69. Amsterdam/Philadelphia: Benjamins. Croft, William 1995 Autonomy and functionalist linguistics. Language 71: 490⫺532. Dahl, Osten 2001 Principles of areal typology. In: Martin Haspelmath, Ekkehard König, Wulf Oesterreicher and Wolfgang Raible (eds.), Language Typology and Language Universals: An International Handbook, 1456⫺1470. (Handbücher zur Sprach- und Kommunikationswissenschaft, 20.2.) Berlin: de Gruyter. Dahl, Osten 2004 Definite articles in Scandinavian: Competing grammaticalization processes in standard and non-standard varieties. In: Kortmann (ed.): 147⫺180. De Vogelaer, Gunther 2006 Actuation, diffusion, and universals: Change in the pronominal system in Dutch dialects. Zeitschrift für Dialektologie und Linguistik 73: 259⫺274. De Vogelaer, Gunther 2008 (De)grammaticalisation as a source for new constructions: The case of subject doubling in Dutch. In: Alexander Bergs and Gabriele Diewald (eds.), Constructions and Language Change, 229⫺257. Berlin/New York: Mouton de Gruyter. De Vogelaer, Gunther and Magda Devos 2009 On geographical adequacy, or: how many types of subject doubling in Dutch. In: Barbiers et al. (eds.), 249⫺274. Dufter, Andreas, Jürg Fleischer and Guido Seiler (eds.) 2009 Describing and Modeling Variation in Grammar. Berlin/New York: Mouton de Gruyter. Eggers, Eckhard, Jürgen Erich Schmidt and Dieter Stellmacher (eds.) 2005 Moderne Dialekte ⫺ Neue Dialektologie. Akten des 1. Kongresses der Internationalen Gesellschaft für Dialektologie des Deutschen (IGDD) am Forschungsinstitut für deutsche Sprache “Deutscher Sprachatlas” der Philipps-Universität Marburg vom 5.⫺8. März 2003. (Zeitschrift für Dialektologie und Linguistik, Beiheft 130.) Stuttgart: Steiner. Filppula, Markku, Juhani Klemola, Marjatta Palander and Esa Penttilä (eds.) 2005 Dialects across Borders: Selected Papers form the 11th International Conference on Methods in Dialectology (Methods XI), Joensuu, August 2002. Amsterdam/Philadelphia: Benjamins.

861

862

VIII. Structural domains: Methodological problems Fleischer, Jürg 2004 A typology of relative clauses in German dialects. In: Kortmann (ed.): 211⫺243. Goebl, Hans 2001 Arealtypologie und Dialektologie. In: Haspelmath et al. (eds.), 1471⫺1491. Haser, Verena and Bernd Kortmann 2009 Agreement in English dialects. In: Dufter, Fleischer and Seiler (eds.), 273⫺299. Haspelmath, Martin 2001 The European linguistic area: Standard Average European. In: Haspelmath et al. (eds.), 1492⫺1510. Haspelmath, Martin, Ekkehard König, Wulf Oesterreicher and Wolfgang Raible (eds.) 2001 Language Typology and Language Universals. 2 vols. (Handbücher zur Sprach- und Kommunikationswissenschaft, 20.1./20.2.) Berlin/New York: Mouton de Gruyter. Henry, Allison 2002 Variation and syntactic theory. In: Jack Chambers, Peter Trudgill and Natalie SchillingEstes (eds.), The Handbook of Language Variation and Change, 267⫺282. Malden: Blackwell. Herrmann, Tanja 2005 Relative clauses in English dialects of the British Isles. In: Kortmann et al.: 21⫺123. Hollmann, Willem and Anna Siewierska 2007 A construction grammar account of possessive constructions in Lancashire dialect: Some advantages and challenges. In: Trousdale and Adger (eds.), 407⫺424. Hudson, Richard A. 1999 Subject⫺verb agreement in English. English Language and Linguistics 3: 173⫺207. Hudson, Richard A. 2007 English dialect syntax in word grammar. In: Trousdale and Adger (eds.), 383⫺405. Kayne, Richard 1993 Towards a modular theory of auxiliary selection. Studia Linguistica 47: 3⫺31. Keenan, Edward and Edward Comrie 1977 Noun phrase accessibility and universal grammar. Linguistic Inquiry 8: 63⫺99. Keenan, Edward and Edward Comrie 1979 Data on the noun phrase accessibility hierarchy. Language 55: 333⫺351. Kortmann, Bernd 2002 New prospects for the study of dialect syntax: Impetus from syntactic theory and language typology. In: Barbiers, Cornips and van der Kleij (eds.), 185⫺213. Kortmann, Bernd 2004a Synopsis: Morphological and syntactic variation in the British Isles. In: Kortmann and Schneider (eds.), vol. 2: 1089⫺1103. Kortmann, Bernd 2004b Do as a tense and aspect marker in varieties of English. In: Kortmann (ed.): 245⫺275. Kortmann, Bernd (ed.) 2004c Dialectology Meets Typology. Berlin/New York: Mouton de Gruyter. Kortmann, Bernd 2006 Syntactic variation in English: A global perspective. In: Bas Aarts and April McMahon (eds.), Handbook of English Linguistics, 603⫺624. Oxford: Blackwell. Kortmann, Bernd to appear Non-standard varieties in the areal typology of Europe. In: Bernd Kortmann and Johan van der Auwera (eds.), Field of Linguistics: Europe. Berlin/New York: Mouton de Gruyter. Kortmann, Bernd, Tanja Herrmann, Lukas Pietsch and Susanne Wagner 2005 A Comparative Grammar of British English Dialects: Agreement, Gender, Relative Clauses. Berlin/New York: Mouton de Gruyter.

46. Areal variation in syntax Kortmann, Bernd and Edgar Schneider (eds.) [In collaboration with Kate Burridge, Rajend Mesthrie and Clive Upton] 2004 A Handbook of Varieties of English, vol. 2: Morphology and Syntax. Berlin/New York: Mouton de Gruyter. Kortmann, Bernd and Benedikt Szmrecsanyi 2004 Global synopsis ⫺ Morphological and syntactic variation in English. In: Kortmann and Schneider (eds.), 1142⫺1202. Kortmann, Bernd and Susanne Wagner 2005 The Freiburg English Dialect project and corpus. In: Kortmann et al.: 1⫺20. Kortmann, Bernd and Susanne Wagner 2007 A fresh look at late modern English dialect syntax. In: Javier Pe´rez-Guerra, Dolores ´ lvarez, Jorge L. Bueno-Alonso and Esperanza Rama-Martı´nez (eds.), “Of Gonza´lez-A Varying Language and Opposing Creed”: New Insights into Late Modern English, 279⫺ 300. Berne: Peter Lang. Kortmann, Bernd and Susanne Wagner to appear Changes and continuities in dialect grammar. In: Raymond Hickey (ed.), EighteenthCentury English: Ideology and Change. Cambridge: Cambridge University Press. Kroch, Anthony 1994 Morphosyntactic variation. In: Katharine Beals, Robert Knippen and Jeannette Denton (eds.), CLS 30. Papers from the 30th regional meeting of the Chicago Linguistic Society, vol. 2: The Parasession on Variation in Linguistic Theory, 180⫺201. Chicago: Chicago Linguistic Society. Kroch, Anthony 2001 Syntactic change. In: Mark Baltin and Chris Collins (eds.), The Handbook of Contemporary Syntactic Theory, 699⫺729. Malden/Oxford: Blackwell. Murelli, Antonio 2006 Subject clitics in Lombard dialects: Evidence from the ASIS (Syntactic Atlas of Northern Italy). Unpublished Master’s thesis, University of Freiburg. Pietsch, Lukas 2005a The Grammar of Variation: Verbal Agreement in Northern Dialects of English. Tübingen: Niemeyer. Pietsch, Lukas 2005b “Some do and some doesn’t”: Verbal concord variation in the north of the British Isles. In: Kortmann et al.: 125⫺209. Poletto, Cecilia 2000 The Higher Functional Field: Evidence from Northern Italian Dialects. Oxford: Oxford University Press. Sampson, Geoffrey, David Gil and Peter Trudgill (eds.) 2009 Language Complexity as an Evolving Variable. Oxford: Oxford University Press. Schmidt, Jürgen Erich 2005 Sprachdynamik. In: Eggers, Schmidt and Stellmacher (eds.), 15⫺44. Seiler, Guido 2004 On three types of dialect variation and their implications for linguistic theory. Evidence from verb clusters in Swiss German dialects. In: Kortmann (ed.): 367⫺399. Seiler, Guido 2005 Wie verlaufen syntaktische Isoglossen, und welche Konsequenzen sind daraus zu ziehen? In: Eggers, Schmidt and Stellmacher (eds.), 313⫺341. Siemund, Peter 2008 Pronominal Gender in English: A Study of English Varieties from a Cross-Linguistic Perspective. London: Routledge.

863

864

VIII. Structural domains: Methodological problems Szmrecsanyi, Benedikt 2006 Morphosyntactic Persistence in Spoken English: A Corpus Study at the Intersection of Variationist Sociolinguistics, Psycholinguistics, and Discourse Analysis. Berlin/New York: Mouton de Gruyter. Szmrecsanyi, Benedikt and Bernd Kortmann 2009a Between simplification and complexification: Non-standard varieties of English around the world. In: Sampson, Gil and Trudgill (eds.), 64⫺79. Szmrecsanyi, Benedikt and Bernd Kortmann 2009b The morphosyntax of varieties of English worldwide: A quantitative perspective. Lingua 119(11): 1643⫺1663. [Special issue: John Nerbonne (ed.), The Forests behind the Trees.] Szmrecsanyi, Benedikt and Bernd Kortmann 2009c Vernacular universals and angloversals in a typological perspective. In: Markku Filppula, Juhani Klemola and Heli Paulasto (eds.), Vernacular Universals and Language Contacts: Evidence from Varieties of English and Beyond, 33⫺35. London/New York: Routledge. Tagliamonte, Sally, Jennifer Smith and Helen Lawrence 2005 English dialects in the British Isles in cross-variety perspective: A base-line for future research. In: Filppula et al. (eds.), 87⫺117. Trudgill, Peter 2009 Sociolinguistic typology and complexification. In: Sampson, Gil and Trudgill (eds.), 98⫺109. Trousdale, Graeme and David Adger 2007 Preface. In: Trousdale and Adger (eds.), 257⫺259. Trousdale, Graeme and David Adger (eds.) 2007 Special issue on dialect syntax. English Language and Linguistics 11(2). van der Auwera, Johan and Annemie Neuckermans 2004 On the interaction of predicate and quantifier negation in Flemish. In: Kortmann (ed.): 453⫺478. Wagner, Susanne 2004 “Gendered” pronouns in English dialects ⫺ A typological perspective. In: Kortmann (ed.): 379⫺493. Wagner, Susanne 2005 Gender in English pronouns: Southwest England. In: Kortmann et al.: 211⫺367. Weiss, Helmut 2004 Vom Nutzen der Dialektsyntax. In: Franz Patocka and Peter Wiesinger (eds.), Morphologie und Syntax deutscher Dialekte und Historische Dialektologie des Deutschen: Beiträge zum 1. Kongress der Internationalen Gesellschaft für Dialektologie des Deutschen, Marburg/ Lahn, 5.⫺8. März 2003, 21⫺41. Vienna: Edition Praesens.

Bernd Kortmann, Freiburg (Germany)

47. Areal variation and discourse

47. Areal variation and discourse 1. 2. 3. 4. 5.

Problem space Areas, communication communities and discourse styles Variation of particular discourse segments Variation in discourse: Prospects and perspectives References

1. Problem space Language and communicative practices often vary between and within regions, mostly due to local and social diversity (e.g., heterogeneity in big cities). The ecological environment leaves its marks on communicative formulas and particular patterns of institutional and everyday use as well as on shifting or switching styles in urban and regional social spaces where dialect and standard, minority and majority languages coexist. Nowadays, media have a great deal of influence on the distribution of routines and interactional styles as is particularly evident in the language of adolescents. American anthropologists were the first to introduce discourse variation into the international discussion surrounding sociolinguistic methodology, surveys and case studies (Gumperz and Hymes 1964; Hymes 1964). With the emergence of pragmatics in Europe, Schlieben-Lange and Weydt (1978) suggested the investigation of regional patterns of politeness in dialectology by looking at the making and reception of compliments in different regional settings of Germany. They observed that Westfalians reacted to compliments with delight and gratitude whereas Swabians were embarrassed by them and played them down. Variation affects discourse in essentially two dimensions: (1) particular discourse segments in various communicative genres (cf. Günthner and Knoblauch 1995) are realized by pragmatic variants in different areas; (2) in the process of performing social activities, discourse and interactional conversations are subject to variation (formal and informal register variants, code shifting and code switching, particular prosodic markers). But first of all, what do we understand by the notions of variation and discourse?

1.1. Areal variation The idea of an areal dimension to variation is commensurable with the idea of maps of landscapes, cities, routes, highways, etc., and their conveyance of a visual impression at a glance (Chambers and Trudgill 1980; Labov, Ash and Boberg 2005). The old dialectal perspective of “one variant per location” (see the critical review by Weinreich 1954) has given way to the principle that the coexistence of multiple variants is the rule, sole variants the exception. Sounds and words are the easiest to present on maps. Larger units or stretches of discourse cannot be represented on a map unless they are defined by a keyword or an abstract symbol in the form of a label that “stands for something”. The goal of a map is an informative overview of a regional repertoire of variants, whose distribution is easily perceivable (“at one glance”). In looking at German as a polycentric

865

866

VIII. Structural domains: Methodological problems language, Ammon and coauthors (2004) identify some phraseologisms, formulaic expressions and routines which differ ⫺ above and beyond all lexical variation ⫺ between Austria, Germany and Switzerland. Elspass and Möller (2005⫺) present the results of questionnaires sent to many informants all over Germany in an online atlas. They obtained variants in (among others) (i) phraseological collocations for managing duties and obligations in everyday communication (das braucht’s nicht ‘this is not necessary’), (ii) existentials (es gibt versus es hat ‘there is’; cf. the detailed description of these items by Schiffrin [1994: 239, 305]), (iii) refusals (ach geh! ‘take it easy!’) and (iv) requests for continuation (ach komm! ‘go ahead!’). Some variation involved interactive items: tag questions (nicht wahr? ‘isn’t it?’), greetings when entering a store in the afternoon (Grüss Gott, Hallo, Guten Tag, Grüezi) and answers to the question, “What do you say when somebody expresses thanks and you have to respond?” Dealing with discourse in this article, it will become immediately clear that the areal limits and boundaries of conversational practices and discourse patterns cannot be as straightforwardly determined as those of lexical items or speech-act variants that can be elicited by questionnaires. Studies on variation in discourse have to focus on natural, authentic interactions and communicative genres documented in everyday life. Such studies have only recently become the subject of sociolinguistic analysis. Exploring the state of the art, this article will look at urban verbal styles, dialect shifting and code switching (section 2) as well at the variation of particular discourse segments across communicative genres (section 3).

1.2. Discourse Of the competing views on discourse, I prefer a sociolinguistic definition of this notion in this article. Global pragmatic approaches (cf. Levinson 1983) include deictic expressions, speech acts, Gricean implicatures, (semantic) presuppositions and conversational structure. The conversation analysis (CA) approach looks at very fine-grained structures of everyday language use, treating the patterns it finds as mostly general and universal. Both approaches do not take into account regional and socially conditioned variation, which are central to this article. Paying as much attention to interactional sequencing as possible, discourse can “best be thought of as … utterances as units of language production … that are inherently contextualized”. This view can be applied to several relevant goals of discourse analysis. One the one hand, there are sequential goals: Are there principles underlying the order in which one utterance, or one type of utterance, follow another? Second is what might be called semantic and pragmatic goals: how does the organization of discourse, and the meaning and use of particular expressions and constructions within certain contexts, allow people to convey and interpret the communicative content of what is said? (Schiffrin 1994: 41)

Crucial to the study of variation in discourse is the concern over the use of particular ways of speaking (cf. Hymes 2003) and the organization of turns and utterance sequences in interaction in particular social contexts. A full account of the methodology used in discourse analysis can be found in Schiffrin (1994) and Brinker et al. (2001). The latter is an 1800-page handbook which includes an overview of the methods of conversation

47. Areal variation and discourse linguistics (survey methods, transcription and analysis), the organization of conversation (prerequisites, structures, procedures and modalities) and conversation typology. We rarely find a theoretical text introducing studies on discourse and variation. Schiffrin’s chapter on “Variation analysis” in her 1994 book, Approaches to Discourse, is a discussion of possible concepts. Another text is the introduction to the “variationist approach” by Dubois and Sankoff (2001), who look at the discourse organization of lists in conversations. Amenable introductions to the analysis of ways of speaking in the process of discourse and interaction are Gumperz (1982) and Auer (1998), who identify variation in discourse with dialect shifting and code switching in conversation.

2. Areas, communication communities and discourse styles 2.1. Conversation styles in big cities Research on regional variation in dialects has often neglected the major cities because urban dialects were supposedly characterized by large social variation as a result of immigration, social mobility, commuting, etc. In fact, urban dialects are mostly a compromise between dialectal features of the surrounding areas and some kind of spoken vernacular substandard. In the case of the urban vernacular of Berlin, Dittmar and Schlobinski (1988) gained empirical evidence that the Low German substratum influenced the Berlin Urban Vernacular (BUV) in a few outstanding phonetic, lexical and morphological features, but the original way of speaking that both community insiders and outsiders associated with the way “Berliners talk” was a particular Berlin style known in the vernacular as the Berlin wit/gob (Berliner Schnauze). It is the Berliner’s famed combination of brash, impudent remarks and humor, quickness or repartee, incisiveness of verbal expression, self-assertive aggressiveness, and loud-mouth bluster! (1988: 45).

This style co-occurs with particular pragmatic and rhetorical features. The vernacular style is marked by the following key characteristics, which all match the most prominent underlying social maxim: Be a cool, tough talker and storyteller and make it clear through your discourse style that you are superior. 1. In conflict narratives (disputes), Berliners like to show that they are winners in the long run although their opponents may be in a better starting position. That they are tougher, more powerful and in the end successful is a result of their fine-grained stylistic and rhetorical techniques which correspond to their superior self-image. Narratives end up with them as winners and others as losers. 2. Core members of native groups (adolescents in the 1988 study by Dittmar and Hädrich) use the most prestigious vernacular features when they react to other group members in performing jokes, challenges and provocations, or when underlining their success in a quarrel at the culminating point of a narrative by verbally focusing on the target of their actions towards the “loser” (as opposed to taking the perspective of the agent). They prefer addressee-oriented patterns of formulation like they’ll get clobbered so hard they won’t know what hit ’em (Dittmar and Hädrich 1988: 87) which is coded in what in German is called the Adressatenpassiv (‘addressee-oriented passive

867

868

VIII. Structural domains: Methodological problems voice’) instead of active voice constructions like I’ll belt you in the kisser (1988: 87). Thus, highly scored and successful storytellers prefer particular BUV patterns of formulation at the peak of narratives so as to enhance their image in the group. These speakers are accorded the highest appreciation for sophisticated verbal skills in local patterns of formulation. 3. When Levinson (1988: 161) states that “there is more language variation than just the superficial phonological and morphosyntactic variables”, he is referring to reported quarrels in his corpus of narratives, where he finds a particular characteristic of an urban way of speaking: “[a] push-down structure, … where the ‘plates’ (turns) continue to be added to the stack, but are never lifted off again in the reverse order” (1988: 181). The “combative” way Berliners have of talking leads them to indulge in less “face-preserving” behavior. We can therefore expect Berliners (a) to routinely use dispreferred responses to adjacency pair first parts without the normal dispreferred format (e.g. to refuse request with a simple “No!”); (b) to correct other speakers instead of prompting them to correct themselves; (c) to not indulge in pre-emptive offers, or other helpful truncations of conversational sequences. (Levinson 1988: 185)

4. Conversational studies in institutions and in everyday life have shown that BUV speakers control and dominate the speech of others by leaving out the “hedges” of politeness, interrupting speakers in the course of their turn and speaking quickly. Similar stylistic features have been found in other major cities, including Zurich, London, Montreal and Paris. Metropolitan speakers have the reputation of being arrogant. “That is to say [in terms of variation in discourse] … that a whole slew of interactional features should follow as motivated consequences of a certain lack of interest in face-preservation” (Levinson 1988: 185) Other studies have identified variation in discourse in prosodic patterns (cf. Gilles and Siebenhaar in this volume). Schwitalla (1994: 519⫺527) shows that evaluative statements in narratives and argumentations are prosodically contrasted to the preceding speech flow. Particular formulaic speech is attested in the studies of the urban vernacular of Mannheim: parallel to the BUV, there are specific communicative routines and formulas which represent typical ways of expressing the “Mannheim” social identity (Kallmeyer and Keim 1994b).

2.2. Dialect shiting and code switching in discourse Inspired by the pioneer work of Auer (1986) who identified standard⫺dialect continua in the form of conversational code shifting in the everyday speech of citizens of Constance (Germany), Schlobinski (1988) looked at code shifting in BUV. He isolated three dimensions which contribute to the initiation of code shifting: the emotional (trouble), the interactive (i. e., the selection of the speaker) and the discourse dimension (i. e., quotation). In a detailed analysis of conversational interaction, he established that a prominent German politician, Richard von Weizsäcker, shifted between standard and BUV on a famous talk show in Berlin on the occasion of his retiring as Mayor of Berlin to take up his new post as the recently elected Federal President of Germany. In a verbal spar with a famous artist native to Berlin (Wolfgang Neuss), he switched from standard to dialect

47. Areal variation and discourse when he was attacked and had to defend himself. Another reason for shifting was to accommodate the BUV in which he was addressed. Schlobinski (1988) identified “quotations” as a third source of shifting. Quotations in BUV served to bring in the original speaker into the actual (“authentic”) situation or scenario. Thus Weizsäcker was speaking in the standard as an “institutional” representative and in the vernacular as an “individual” person reacting “emotionally” to particular interactional or speech events. According to Kallmeyer and Keim (1994a: 161), phonological variation mainly serves the functions of organizing the conversation (assignment of turns) and structuring the utterances (segmentation and control of focalizations). Particular devices which are the sources of code shifts and variation are quotations and addressee specifications. According to these authors, variation in discourse has social and symbolic meaning. Evaluations may be affected by them, but code shifts saliently symbolize whether somebody who is addressed using particular variants is an outsider or an insider, a local and familiar person or a stranger. Speakers constitute social meaning by shifting their speech behavior from standard (formal) to dialect/vernacular (informal) and vice versa according to the well-known parametric oppositions: (a) social proximity versus social distance, (b) wecode versus they-code, (c) formality versus informality and (d) subjectivity versus objectivity. Prosodic and phonological features reflect these parameters. The variation is defined as the gradual deviation from the informal, basic vernacular speakers use when they talk to friends, lovers, children and the members of their family. Illustrating the dynamics of code shifting over large stretches of discourse and the social construction of particular prosodic patterns at the beginning and end of utterances with many examples, Kallmeyer and Keim (1994a) find significant social meaning in variation in discourse. When others are accepted as “superior, noble, better people”, the use of the standard is positively evaluated. The same holds true for speakers who are recognized by participants as persons of “authority” for whom they have respect. Divergence is signaled prosodically and phonologically when somebody sanctions the norms of the group. Group members may also imitate the standard of somebody in order to make a fool of them and to negatively mark the distance to them (“they-code”). On the other hand, the use of dialect also has significant functions: (i) an alien, somebody external to the group, can be excluded from understanding by using the dialect; (ii) “convergence from above” (Kallmeyer and Keim 1994a: 233), a kind of solidarity function; (iii) a salient use of dialect in order to sanction socially unacceptable roles or role performances of group members (“drunkard”, “slut”). The different ways of marking the quality of vowels, consonants and syllables or of accentuating, lengthening or shortening words or discourse segments carry social-symbolic values which constitute the social meaning of variation in discourse. The ethnographic study on the language use and communicative practices in Mannheim is published in four separate volumes (Kallmeyer 1994). In a similar vein to these four volumes, Denkler describes in a recent study (2007) the functions of language variation in informal talk (conversations). The methodological background is conversation analysis as practiced by Auer in his pragmatic approach to code switching in discourse (Auer 1995), who in turn founds his work on Gumperz’ model of “contextualization cues” (Auer 1992). The informants in Denkler’s study are native speakers of a variety of Plattdeutsch ‘Low German’ from the German area of Münsterland surrounding the city of Münster (in northern Germany). Participants at an evening party switch between the Low German Münsterlandish dialect (which is treated as a separate minority language, as different from German as Catalan is from Caste-

869

870

VIII. Structural domains: Methodological problems llano) and the northern German oral substandard. In the discourse stretches under investigation (personal narratives), participants switch in order to indicate, using specific contextualization cues, how particular discourse segments need to be understood in their sequential context. Participants make conversational inferences from the way the switches between the two “languages” are performed (phonetic and prosodic quality of sounds, syllables, words and chunks). There are particular communicative routines which help to organize conversational sequences. Quotations show significant code switches which represent contextualization cues for specific pragmatic inferences. Code switching is used (among other functions) “to gain the floor and to both build a link to and distinguish an account from previous utterances” (Denkler 2007: 192). Another situation of covariation is described in Hove (2008). Analyzing 6381 emergency calls where a Swiss policeman and a native or foreign caller interact with each other in a situation of great pressure, Hove looks at external and discourse specific motivation for using Swiss German dialect or Standard Swiss German. Most obviously, discourse determines the variation of both caller and policeman here ⫺ callers initially react to the local Swiss dialect of the policemen in dialect (natives) or in some kind of (Swiss) standard (non-natives). Discourse-specific considerations which affect the choice of Standard Swiss or dialect are (i) whether the relevance of the particular utterance favors standard (e.g., in order to prevent misunderstanding of an important message from the policeman), (ii) communicative formulas at the beginning and at the end of the call are most often in dialect (they do not convey relevant information and, being in dialect, they provoke the calling person to either answer in dialect or indicate that they do not have mastery of the dialect), (iii) changes in the modality of interaction (a shift to dialect during an interaction conducted mostly in standard language can indicate a higher degree of spontaneity and emotionality), (iv) addressee specifications (addressing a verbal message to himself or to a third person present during the call can lead to code switching). Code switching, as mentioned in (iv), is a common event when policemen talk to foreigners (speakers of other languages). In the course of the interaction they switch to the standard, often in a way that is known as foreigner talk (a register in which grammatical structures are simplified to different degrees in order to be better understood). Beyond these discourse-specific factors there are participant-related factors which favor a change in variety. When Germans or Austrians answer the dialectal identification of the policeman in Standard German, policemen switch and also continue in the standard. Naturally, individual competence in Swiss dialect with respect to production and reception is a factor which has an important impact on code switching. So far I have presented the results of investigations studying variation in the course of discourse and verbal interaction affecting conversational sequences, turns and turn taking. I call this variation, which marks the process of discourse, vertical dynamic variation in discourse. The next section is devoted to another dimension of variation in discourse which I call horizontal variation of discourse segments. The idea is to compare welldefined segments (including formulas, routines, lists, discourse markers and particles) at different sequential positions in discourse. On the one hand, the verbal realization of communicative functions is compared across varieties or languages; on the other hand, pragmatic connectors or particles are compared with respect to their (different) role in discourse.

47. Areal variation and discourse

3. Variation o particular discourse segments There are a number of studies which examine the sociolinguistic implications of variation in particular discourse segments ⫺ such as formulas, lists, discourse markers and particles ⫺ and the latter at least suggest an areal dimension to such variation (albeit often in conjunction with other factors). Changes in areal distribution can be seen to go handin-hand with changes in meaning and pragmatic function, as a more detailed examination of shifts in the use of the modal particles halt and eben in section 3.2 reveals.

3.1. Formulas, lists and discourse markers Kallmeyer and Keim (1994b) show that everyday conversations are interspersed with formulaic comments, evaluations and routines which serve the pragmatic orientation of hearers. They have their own phonetic and prosodic shape and carry the local values and norms of positive and negative evaluation. Such routines can take the form of rounds, parallel routinized comments or speech acts which can be differentiated into those which exhibit thematic progression and those that do not. An example of the former would be a Berlin football team’s training discussions reported in Dittmar and Hädrich (1988), in which rounds determine a ritual order of participant turns. With regard to lists, Schiffrin (1994: 293) states that “one of the most basic differences between narratives and lists is that narratives tell what happened and lists describe a category”. She suggests a whole range of different types of lists which reveal a rich variation. But what is most important in the context of this article is her “assumption about discourse analysis: syntactic and semantic differences among linguistic items are sensitive to text structure both within and across text types” (Schiffrin 1994: 315). Narrowing down Schiffrin’s pragmatic approach to a restrictive analysis of lists in discourse, Dubois and Sankoff (2001) look at “enumeration as a discourse strategy” in interviews from the perspective of a variationist approach (Dubois and Sankoff 2001: 285). There is a clear operationalization of those utterances which belong to an “enumeration” in their sense and the authors identified and analyzed a total of 3464 enumerations in their corpus. They found variation according to (a) expressivity (“the more components there are … the more the expressive potential of enumeration” [2001: 290]), (b) processing (the more complex the components are, the greater the processing difficulty), and (c) length (“the shorter the enumeration … the more efficient is the use of enumeration in carrying out its function” [2001: 290]). Dubois and Sankoff find differences in verbal style between older and younger speakers (reference to experience, informative function, biographical relevance among other features) and between middle-class speakers (who tend to use parts of sentences, each well coordinated with the other) and speakers of working class, who use “full sentences” (2001: 293, 297). The authors conclude their pioneering work with the following statement: We propose our study of enumeration as a prototype of an approach which succeeds in operationalizing discourse concepts on many levels, so that an exhaustive study of a large corpus can reveal and characterize with some precision the deep connections among the various processes implied in the motivation, construction, use, and interpretation of this figure. (Dubois and Sankoff 2001: 299)

871

872

VIII. Structural domains: Methodological problems The authors develop an explicit methodology, but unfortunately treat variation in discourse only on a “monological” level. Their study nonetheless exemplifies how a potential areal variation might be approached. In an article on dialect divergence, Cheshire, Kerswill and Williams (2006) make an important point about variation in discourse. They argue that patterns of variation are not the same at different levels of linguistic description: they find sharp stratification for morphosyntactic variables, whereas phonological variation demonstrates gradient patterns. This pattern is reversed in France: phonological variants show sharp social stratification, in contrast to morphosyntactic variables illustrating gradient variation. They argue that in comparison to the bulk of sophisticated studies on phonological and morphosyntactic variation, discourse features are barely analyzed. Nevertheless, there are some studies on the variation of discourse particles (Schiffrin 2001), set marking tags and particular words. Cheshire, Kerswill and Williams (2006: 142) emphasize that “the analysis of discourse features […] requires a more complex analysis than a simple counting of the number of tokens”. Interview data alone are not adequate for the pragmatic analysis of variation in discourse. Discourse markers (and elements such as particles) often only show up in informal interactions. They require more qualitative analyses, e.g., the studies of Erman 1993 and Holmes (1995), who found gender differences in the use of discourse marker like I mean, you see and you know, women using them to connect arguments or signal affective meaning, men to signal referential meaning or repair work or to gain attention. The point is that solid qualitative analysis and description have to be achieved first because “important differences in the way that different social groups use discourse features in interaction may be obscured if we simply count numbers of tokens” (Cheshire, Kerswill and Williams 2006: 143). One example of a discourse marker which exhibits areal variation is like, which has been observed to function as a focus marker (and we were like rushing home) and as a marker of reported speech and thought (and she was like “Where are you off to?”) throughout the urban centers of the anglophone world (Cheshire, Kerswill and Williams 2006: 154). For three different British cities, Milton Keynes, Reading and Hull, differing frequencies have been documented for like. In Hull, the middle class used like more often than the working class did. In the other cities, gender or social class patterns were not found. An increase in the frequency of the focus maker like has been generally associated with youth culture, and Golato (2000) shows quite convincingly that quoting in narratives is a marker of international youth culture. She found that the English quotative be like is equivalent to German so, “which is mostly realized in the forms of und ich so and und er so” (Golato 2000: 52). This quotative pattern is also analyzed in Auer (2006) as one type of syntactic use of the item, but Golato goes beyond the static linguistic description to suggest that “so is used as a demonstrative deictic marker referring to the performative aspect of the quote” (2000: 45). The contemporary use of so in ich so versus er so or of be like seems to be as an “embodied quote” which “allow the speaker to give a visual representation of what was going on during the event described” (2000: 45) Golato’s point is that the “new” construction is used for interactive reasons: the quote is “transformed into a performance to be watched and listened to by the coparticipant” (2000: 45). The verbal expression is thus framed by discourse and interactional needs and the speakers performing the event are quoting by changing their “footing” (Goffman 1981: 128).

47. Areal variation and discourse In addition to the pragmatic role of quotatives, Macaulay (1991) has analyzed a whole “set of discourse markers,” in particular oh, well, (you) see, I mean, you know, (you) ken, (of) course, anyway, in fact, now, here and mind you. He shows all twelve to have important functions in managing discourse continuity, and further he finds classbased differences in their use in particular functions and even positions (e.g., of course in the sense of ‘surely you understand that’ overwhelmingly occurs in initial position in the lower-class interviews [1991: 167], while for the middle class it is more integrated into the grammatical middle field; stylistically the lower class thus have a more “pragmatic”, the middle class a more “syntactic, elaborative” mode in using this discourse marker). Macaulay’s work on the communicative and social functions of these discourse markers demonstrates not only that “there are social class differences in the use of discourse markers, but also that there are other differences related to genre … [and] to individual style” (1991: 176). The point to be remembered is that capturing variation in discourse, whether linked to geography or not, requires a consideration of the contextually embedded functions discourse elements perform.

3.2. Particles Particles are metacommunicative devices which encode relevant pragmatic meaning in verbal interactions. In line with modern lifestyles and needs in interpersonal communication, their use can change and their changing regional distribution may also induce sociolinguistic change. Such a sociolinguistic change has occurred in northern Germany with respect to the use of the evidentials eben and halt. Eichhoff’s (1977/2000) areal distribution of the modal particles eben (northern Germany) and halt (southern Germany) into two “dichotomic worlds” is outdated, and has been corrected by the recent investigations of Elspass and Möller (2005). Their internet atlas shows that both modal particles (eben and halt) currently coexist all over Germany, at least in the major cities. From a sociolinguistic point of view, simple coexistence of these two particles is only a very rough description of the present distribution. Social forces govern different frequencies of distribution. It would be interesting to know more about the social conditioning of this variation in discourse and about what is driving this process, but one important driving factor has been the reunification of Germany, leading Dittmar (2000) to study the spread of halt to East Berlin. In an exclusively eben area (northern East Germany, still quite isolated from West Germany at the time), halt found its way into the eastern part of the city once the Berlin Wall fell in 1989. West Berlin had been a walled island up until 9 November 1989, and many of the Berliners who left this isolation for West Germany were replaced by southern Germans, in particular Swabians, in the seventies and eighties. They brought with them the modal particle halt, which they used as a subjective correlate for evidence marking in coexistence with eben as a marker for objective evidence (es hat eben drei Grad minus draussen ‘it’s minus three degrees outside’). In a corpus of 25 recordings of West Berliners and 31 of East Berliners, West Berliners used halt much more often than the East Berliners did. Dittmar (2000) distinguishes four types of particle users: (a) those who use only eben (23); (b) those who use only halt (2); (c) those who use both, in more or less equal numbers (27); (d) those who use both of them, even in combination (6). Types (a) and (c) predominate.

873

874

VIII. Structural domains: Methodological problems The particles eben and halt, whether used alone or in combination with each other, appear in most of the cases in evaluative or argumentative stretches of discourse (i. e., talking about and evaluating East⫺West problems). Speakers who use both particles are motivated to do so because it gives them the stylistic choice between being softer or harder, more personal or more distant, more friendly and charming or less. What can we infer from these results? The West Berlin speakers not only use eben and halt predominately in alternating ways (co-occurrence), they also use halt more often than eben. I consider the very frequent use of halt (alongside eben) to be an indicator of the communicative surplus value of this epistemic “softener” in interpersonal discourse. This argument is supported by the following sociolinguistic reflections. “Eben-only” users are able to express all of their pragmatic intentions with this particle. Their only disadvantage is that those who additionally use halt (and expect its use in communication) often consider the utterances of “eben-only” users to be more formal and distant and therefore less personal and friendly. The reverse may be true for those (southern) speakers who use only halt, with one major exception: all German speakers have to use eben as an adjective or temporal adverb. Furthermore, they also use it also as a sentence equivalent, in that one can react to a statement in an elliptical way with ja eben, but not with *(ja) halt. This is not the only argument indicating that halt and eben are not synonymous: the terms can, moreover, be combined in patterns of eben halt, which would not be possible if they were synonyms. It follows from this sociolinguistic argument that those who use both modal particles do so for reasons of epistemic differentiation (mitigating or reinforcing the evidence of their speech acts) and stylistic flexibility. In this sense, the use of both particles offers advantages in register use. Dittmar (2000) predicts that speakers in East Berlin and Brandenburg, for whom (up until 1989) the norm of eben(t) was the only valid option, will increasingly use the particle halt in order to differentiate their registers and to make their personal motives easier to understand, or to react to the pressure of the linguistic market. The adoption of halt meets personal and stylistic needs related to becoming a more “unique individual”. There may also be a new job-related need to appear friendlier and to become more influential by using a particle which carries the prestige of the Western linguistic market. Those East Berlin speakers who used the modal particle halt frequently (i) were young (and paid attention to Western prestige patterns) and (ii) had already worked or lived in the Western part of the city, and had adapted to Western patterns in the course of intensive daily contacts. The take-up of the modal particle halt was accompanied by patterns of non-normative use. The particle halt was occasionally used in ungrammatical syntactic positions (initial and final part of the utterance). The process of learning to use it correctly has parallels in second-language learning. The deviant use of the particle indicated a certain period of cognitive insecurity in the new communication community.

4. Variation in discourse: Prospects and perspectives Recent investigations on variation in discourse open a window onto the pragmatic meaning of variation in the process of conversational interaction (studies reported on in section 2), they open another window onto the comparison of particular discourse segments

47. Areal variation and discourse across discourse types and genres (lists, formulas, routines) and of regional expressions for particular communicative functions. With respect to variation in the course of conversational interaction, the pragmatic functions of types of code switching or code shifting should be compared ⫺ in more than some isolated conversations ⫺ with respect to types of (interactional) encounters (native⫺non-native, expert⫺layman, professional⫺non-professional) and to types of communicative events/genres (Günthner and Knoblauch 1995). There is a lack of more explicit criteria for comparing functions in both dimensions. Prosodic features seem to be prominent devices for these purposes. Looking at discourse segments on the other hand, regional research centers should work together in comparing expressions, routines, lists, particles and discourse markers in their corpora with respect to their positions in turns and their communicative functions in, e.g., direct and indirect speech, evaluations or comments. Areal variation in discourse is a very young research field still lacking results comparable to those obtained in other fields of sociolinguistic description. It could, however, be shown in this short overview that it is becoming increasingly important in sociolinguistics ⫺ as documented by the increasing number of challenging new descriptions and attempts to explain them.

5. Reerences Ammon, Ulrich, Hans Bickel, Jakob Ebner, Ruth Esterhammer, Markus Gasser, Lorenz Hofer, Birte Kellermeier-Rehbein, Heinrich Löffler, Doris Mangott, Hans Moser, Robert Schläpfer, Michael Schloßmacher, Regula Schmidlin, Günter Vallaster (eds.) 2004 Variantenwörterbuch des Deutschen: Die Standardsprache in Österreich, der Schweiz und Deutschland sowie in Liechtenstein, Luxemburg, Ostbelgien und Südtirol. Berlin/New York: de Gruyter. Auer, Peter 1986 Konversationelle Standard/Dialekt-Kontinua (Code-Shifting). Deutsche Sprache 14: 97⫺124. Auer, Peter 1992 Introduction: John Gumperz’ approach to contextualization. In: Peter Auer and Aldo Di Luzio (eds.), The Contextualization of Language, 1⫺38. Amsterdam: Benjamins. Auer, Peter 1995 The pragmatics of code-switching: A sequential approach. In: Lesley Milroy and Pieter Muysken (eds.), One Speaker, Two Languages: Cross-Disciplinary Perspectives on CodeSwitching, 115⫺135. Cambridge: Cambridge University Press. Auer, Peter (ed.) 1998 Code-switching in conversation: Language, Interaction and Identity. London/New York: Routledge. Auer Peter 2006 Construction Grammar meets conversation: Einige Überlegungen am Beispiel von ‘so’Konstruktionen. In: Susanne Günthner and Wolfgang Imo (eds.), Konstruktionen in der Interaktion, 291⫺314. Berlin: de Gruyter. Brinker, Klaus, Gerd Antos, Wolfgang Heinemann and Sven F. Sager (eds.) 2001 Text- und Gesprächslinguistik / Linguistics of Text and Conversation, vol. 2. (Handbücher zur Sprach- und Kommunikationswissenschaft, 16.2) Berlin/New York: de Gruyter.

875

876

VIII. Structural domains: Methodological problems Chambers, James K. and Peter Trudgill 1980 Dialectology. Cambridge: Cambridge University Press. Cheshire, Jenny, Paul Kerswill and Ann Williams 2006 Phonology, grammar, and discourse in dialect convergence. In: Peter Auer, Frans Hinskens and Paul Kerswill (eds.), Dialect Change: Convergence and Divergence in European Languages, 135⫺167. Cambridge: Cambridge University Press. Denkler, Markus 2007 Code-switching in Gesprächen münsterländischer Dialektsprecher. Zur Sprachvariation beim konversationellen Erzählen. Zeitschrift für Dialektologie und Linguistik 74(2⫹3): 164⫺195. Dittmar, Norbert and Doris Hädrich 1988 Gibt es die Berliner Schnauze? Schlagfertigkeit und Berliner Stilregister im linguistischen Kreuzverhör. In: Norbert Dittmar and Peter Schlobinski (eds.), Wandlungen einer Stadtsprache: Berlinisch in Vergangenheit und Gegenwart, 103⫺144. Berlin: Colloquium. Dittmar, Norbert and Peter Schlobinski (eds.) 1988 The Sociolinguistics of Urban Vernaculars: Case Studies and their Evaluation. Berlin/New York: de Gruyter. Dittmar, Norbert 2000 Sozialer Umbruch und Sprachwandel am Beispiel der Modalpartikeln halt und eben in der Berliner Kommunikationsgemeinschaft nach der “Wende”. In: Peter Auer and Heiko Hausendorff (eds.), Kommunikation in gesellschaftlichen Umbruchsituationen, 199⫺234. Tübingen: Niemeyer. Dubois, Sylvie and David Sankoff 2001 The variationist approach toward discourse structural effects and socio-interactional dynamics. In: Deborah Schiffrin, Deborah Tannen and Heidi E. Hamilton (eds.), The Handbook of Discourse Analysis, 282⫺303. Oxford/Malden: Blackwell. Elspaß, Stephan and Robert Möller 2005⫺ Atlas zur deutschen Alltagssprache (AdA). Available at ; last accessed 25 March 2009. Eichhoff, Jürgen 1977/2000 Wortatlas der deutschen Umgangssprachen. 4 vols. Berne/Munich: Francke. Erman, Britt 1993 Female and male usage of pragmatic expressions in same-sex and mixed-sex interaction. Language Variation and Change 4: 217⫺234. Goffman, Erving 1981 Forms of Talk. Philadelphia, PA: Philadelphia University Press. Golato, Andrea 2000 Und ich so / und er so [and I’m like / and he’s like]: An innovative German quotative for reporting on embodied actions. Journal of Pragmatics, 32(1): 29⫺54. Gumperz, John J. 1982 Discourse strategies. (Studies in Interactional Sociolinguistics 1.) Cambridge: Cambridge University Press. Gumperz, John J. and Dell Hymes (eds.) 1964 The Ethnography of Communication. American Anthropologist 66(6), Part 2 [Special issue]. Günthner, Susanne and Hubert Knoblauch 1995 Culturally patterned speaking practices ⫺ The analysis of communicative genres. Pragmatics 5(1): 1⫺32. Holmes, Janet 1995 Women, Men and Politeness. Harlow: Longman. Hove, Ingrid 2008 Gesprochene Standardsprache im Deutschschweizer Alltag. Erste Ergebnisse eines Forschungsprojekts. In: Helen Christen and Evelyn Ziegler (eds.), Sprechen, Schreiben,

47. Areal variation and discourse

877

Hören ⫺ Zur Produktion und Perzeption von Dialekt und Standardsprache zu Beginn des 21. Jahrhunderts, 83⫺100. Vienna: Praesens. Hymes, Dell 2003 [1972] Models of the interaction of language and social life. In: Paulston, Christina B. and G. Richard Tucker (eds.), Sociolinguistics: The Essential Readings, 30⫺47. Malden/ Oxford: Blackwell. [Originally published in: Dell Hymes and John Gumperz (eds.), Directions in Sociolinguistics. The Ethnography of Communication, 35⫺71. New York: Holt, Rinehart and Winston.] Hymes, Dell (ed.) 1964 Language in Culture and Society. A Reader in Linguistics and Anthropology. New York: HarperCollins. Kallmeyer, Werner (ed.) 1994 Kommunikation in der Stadt, vol. 1: Exemplarische Analysen des Sprachverhaltens in Mannheim. Berlin/New York: de Gruyter. Kallmeyer, Werner and Inken Keim 1994a Phonologische Variation in der Filsbachwelt. In: Kallmeyer (ed.): 141⫺249. Kallmeyer, Werner and Inken Keim 1994b Formelhaftes Sprechen in der Filsbachwelt. In: Kallmeyer (ed.), 250⫺317 Labov, William, Sharon Ash and Charles Boberg 2005 The Atlas of North American English. Phonetics, Phonology and Sound Change. Berlin/ New York: Mouton de Gruyter. Levinson, Stephen 1983 Pragmatics. (Cambridge Textbooks in Linguistics.) Cambridge: Cambridge University Press Levinson, Stephen 1988 Conceptual problems in the study of regional and cultural style. In: Dittmar and Schlobinski (eds.), 161⫺190. Macaulay, Ronald S. 1991 Locating Dialect in Discourse: The Language of Honest Men and Bonnie Lassies in Ayr. (Oxford Series in Sociolinguistics 1.) Oxford/New York: Oxford University Press. Schiffrin, Deborah 1994 Approaches to Discourse. Oxford: Blackwell. Schiffrin, Deborah 2001 Discourse markers: Language, meaning, and context. In: Deborah Schiffrin, Deborah Tannen and Heidi E. Hamilton (eds.), The Handbook of Discourse Analysis, 54⫺75. Oxford/Malden: Blackwell. Schlieben-Lange, Brigitte and Harald Weydt 1978 Für eine Pragmatisierung der Dialektologie. Zeitschrift für Germanistische Linguistik 6(3): 257⫺282. Schlobinski, Peter 1988 Code-Switching im Berlinischen. In: Norbert Dittmar and Peter Schlobinski (eds.), Wandlungen einer Stadtsprache: Berlinisch in Vergangenheit und Gegenwart, 83⫺102. Berlin: Colloquium. Schwitalla, Johannes 1994 Sprachliche Ausdrucksformen für soziale Identität beim Erzählen. Beobachtungen zu vier Gruppen in Vogelstang. In Kallmeyer (ed.), 510⫺577 Weinreich, Uriel 1954 Is a structural dialectology possible? Word 14: 388⫺400.

Norbert Dittmar, Berlin (Germany)

Index A absolute vs. relative space 45 Acadian French 380, 381, 386⫺388 accent 150, 232, 540, 544⫺545, 745, 754, 788⫺790, 792, 842, 844, 846 ⫺ Cockney J Cockney accent ⫺ English J English accents ⫺ foreign J foreign accent ⫺ regional J regional accent ⫺ social J social accent ⫺ working-class J working-class accent acceptability task 518⫺520, 522⫺523 accommodation 20, 167, 213, 216, 251ⴚ252, 254, 318, 341, 362, 376, 457⫺458, 462, 464, 481, 531⫺532, 614, 651 acrolect 232, 342, 445 adjacency 380, 811ⴚ812, 814, 850, 868 adolescents 150, 155, 170, 192, 271, 365, 538, 540, 687, 697, 865, 867 alternation (phonological) 365, 368, 560, 776, 805 American English 39, 156, 174, 214, 245, 250, 461, 463, 504, 541, 556, 560, 586, 766, 767, 823, 824 apparent-time analysis 134⫺144, 202, 205⫺ 206, 218, 247, 251, 253, 662, 680 architecture of language 227, 229, 236ⴚ238, 376, 706 areal typology 436, 855ⴚ856, 858 assimilation 76, 130, 333, 368, 457, 775, 777, 796, 815 atlas ⫺ dialect J dialect atlas ⫺ language J language atlas ⫺ linguistic J linguistic atlas ⫺ Sprachatlas J Deutscher Sprachatlas attitudes J language attitudes Ausgleichssprache 113, 277 Australian English 243, 445, 455, 460, 461, 793 Austria 89, 228, 230, 281⫺283, 288⫺289, 303, 593, 608, 750, 775, 809, 866, 870 Austrian German 228, 279, 289, 593, 608 authenticity 10, 35, 39, 514⫺515, 613, 621, 623, 641, 725, 727, 731, 733⫺736, 741, 749, 752

B background variable 532, 534, 535, 546 basilect 232, 260, 267, 358, 445, 455, 458 belonging 34, 47, 51ⴚ53, 55, 60⫺61, 242, 535, 538, 540, 741, 752

Berlin 250, 280⫺281, 347, 576, 579, 791⫺ 793, 867⫺868, 871, 873ⴚ874 bidimensional 671⫺672 ⫺ approach 584 ⫺ dialectology 669⫺670 ⫺ method 674, 676 bilingualism 37, 42, 118, 127, 175, 341, 344, 376, 379, 385⫺387, 393⫺394, 396, 398⫺ 402, 411⫺412, 424⫺427, 444, 454, 470, 475, 479, 481, 485⫺486, 490, 687, 692, 699, 707, 735, 747, 812⫺813 Bohnenberger, Karl 96, 100ⴚ101 borrowing 39, 63, 83, 242ⴚ244, 250, 253, 286, 318, 358⫺359, 391⫺392, 397, 399⫺ 400, 424⫺425, 427, 475, 479⫺481, 483⫺ 486, 489⫺490, 555, 564, 606, 636 boundary ⫺ dialect J dialect boundary ⫺ social J social boundary ⫺ structural J structural boundary Bourdieu, Pierre 19ⴚ22, 38, 62, 165, 168, 621, 728 British English 39, 149, 156, 174, 175, 214, 229, 243, 319, 461, 788, 823, 824, 844, 845, 855, 856

C Canadian French 243, 377, 393, 402, 404, 411, 486, 724ⴚ736 Catalan 303, 334, 381, 383ⴚ385, 480, 484, 490 cartography 122, 155, 567⫺569, 571⫺578, 582, 585⫺586, 709 central place theory 6, 146 city dialect 315, 317, 320, 325⫺327 cluster ⫺ style J style cluster cluster analysis 183⫺184, 190, 303, 770⫺771 Cockney accent 822 code shifting 77, 211, 340, 744, 865, 868, 869 code switching 46, 262, 341, 426ⴚ427, 481, 484, 687, 746, 747, 753, 822, 869, 870 cognitive map J mental map colonialism 441, 452 colonization 3, 111, 114, 121, 333, 335, 337⫺ 338, 385, 452, 459, 464

880

Index commodification 10, 155, 448, 730, 742, 745 communicative genre 11, 246, 502, 701, 753, 865, 866, 875 communicative space 388, 472ⴚ475, 635 community of practice 636 comparability 428, 512ⴚ513, 576, 701, 828, 858 completion task 523 concentration zones J speech levels concept of variety 211, 214ⴚ215 conservation 30, 481, 503, 616, 721 contact ⫺ dialect J dialect contact ⫺ language J language contact contact-induced change 391ⴚ394, 396, 398ⴚ 399, 407, 410, 412, 436 contact variety 395, 398, 400, 401, 406, 407, 469, 691, 859 contact zone 282, 284⫺288 contagion/contagious diffusion 148, 317, 320⫺323, 651, 812 continuity 48, 454, 854ⴚ855 convergence 156, 213, 242ⴚ247, 250⫺253, 290, 303⫺304, 358, 429, 478, 480ⴚ482, 484, 486, 489 ⫺ horizontal J horizontal convergence ⫺ linguistic J linguistic convergence ⫺ vertical J vertical convergence ⫺ zone J zone convergence conversational interaction 868, 874⫺875 conversational style 717 corpus linguistics 132, 398, 402, 495, 498, 502⫺503, 507⫺508, 513⫺514, 596, 603, 830, 845, 850, 871 correlational sociolinguistics 623 creole 232, 432ⴚ433, 440ⴚ448, 454⫺455, 463⫺464, 469, 747, 852 creole space 442, 445ⴚ446 cultural space (Kulturraum) 109, 110, 121, 452, 809

D Dach(sprache) 231, 277, 285⫺287, 291 data collection J methods of data collection demotization 267⫺269 Denmark 263, 265⫺266, 271, 538, 632 dependent variable 530, 532, 533, 534, 536, 573, 828, 829 destandardization 300⫺301, 304, 356 deterritorialization 52, 55, 447

Detroit 155, 172, 325 Deutsche Dialektgeographie 92ⴚ93, 96, 579 Deutscher Sprachatlas 75, 89, 91, 93, 122, 205, 515⫺516, 524, 579 ⫺ institute 211 diachronic, diachrony 79, 205, 211, 227, 247, 410, 614, 617, 762, 817 diaglossia 230, 282, 290⫺291, 305, 367 dialect ⫺ city J city dialect ⫺ new J new dialect ⫺ primary J primary dialect ⫺ secondary J secondary dialect ⫺ USA J USA dialects dialectality 676⫺678 dialect atlas 89⫺91, 97, 202, 204⫺206, 774, 779 dialect boundary 23, 28, 30, 36, 39, 88, 91, 94⫺95, 97ⴚ101, 108, 120, 126, 151, 154, 156, 174, 179⫺182, 186, 188, 191, 196, 209⫺210, 217, 236, 246, 275⫺276, 284, 287, 291, 375, 382, 595, 656, 663ⴚ665, 678, 789, 810 dialect contact 316, 328, 455 dialect description 71, 88 dialect dictionary 71, 593, 595⫺596, 601⫺ 602, 825 dialect feature 155, 266, 326, 749 ⫺ primary J primary dialect feature ⫺ secondary J secondary dialect feature ⫺ tertiary J tertiary dialect feature dialect geography 24, 87ⴚ101, 107, 109, 114, 126, 203, 205, 315, 328, 335, 513⫺517, 608, 614, 616, 664, 762, 764, 771, 772 dialect grammar 70, 80, 616ⴚ619 dialect leveling 12, 19, 27, 75, 242ⴚ244, 260ⴚ263, 266, 268⫺269, 271⫺272, 276, 301, 324, 328, 336, 342, 357ⴚ362, 469, 677 dialect lexicography 592⫺593 ⫺ electronic J electronic dialect lexicography dialect map 188, 574, 771, 837 dialect mixture 80, 83, 243, 269, 458 dialectology 78, 87, 108⫺109, 122, 128, 375, 470, 550, 562, 601, 698, 724, 805, 841 ⫺ European J European dialectology ⫺ generative J generative dialectology ⫺ German J German dialectology ⫺ history of J history of dialectology ⫺ modern J modern dialectology ⫺ perceptualJ perceptual dialectology ⫺ pluridimensional J pluridimensional dialectology

Index ⫺ social J social dialectology ⫺ structural J structural dialectology ⫺ traditional J traditional dialectology ⫺ urban J urban dialectology dialectometry 236, 551, 561, 563, 577, 581, 582, 665, 769, 771 dialect syntax J syntax dialect user vs. standard user 24, 26 Dialektabbau 276, 288, 304 Dialektumbau 276 diaphasic 227, 228, 229, 231, 234ⴚ235, 236, 237, 299⫺300, 585, 613, 709 diaphasic variety 228, 234ⴚ235 diastratic 227, 228, 229, 232ⴚ234, 236, 237, 275, 283, 289ⴚ290, 296, 299⫺300, 471, 585, 613, 669, 672, 673, 676, 709 diastratic variety 228, 232ⴚ234 diatopic 227, 228, 229, 230ⴚ232, 233, 236, 275, 277, 290, 296, 299⫺300, 301, 304, 305, 308, 339, 469, 471, 585, 593, 594, 596, 606, 613, 616, 669, 673, 676, 709, 710, 772, 822, 835 diatopic variety 228, 230ⴚ232, 296 dictionary J dialect dictionary diffusion 6, 143, 145, 146, 148ⴚ151, 253, 320, 326, 328, 362, 367, 368, 379, 396, 399, 420, 423, 425, 433, 464, 649ⴚ650, 654⫺657, 658⫺660, 660⫺662, 663⫺664, 664⫺665, 678, 679, 681, 805, 814, 817 ⫺ hierarchical J hierarchical diffusion ⫺ lexical J lexical diffusion ⫺ wave J contagion/contagious diffusion diffusion model 150, 616, 654ⴚ657, 664 digital space 58⫺62 Digital Wenker Atlas (DiWA) 89, 202, 204, 205, 207, 603, 672, 680, 854 digitalization 52 diglossia 282, 283, 291, 355 direct method 79, 80, 87, 92, 94⫺95, 98, 100, 502ⴚ504, 515ⴚ516, 543, 580 discontinuous language space 335 discourse 10ⴚ11, 20, 155, 246, 253, 396, 397, 398, 399, 404, 410, 422⫺423, 435, 489, 668, 823, 838, 865, 866ⴚ867, 868, 871, 873, 874, 875 ⫺ media J media discourse discourse markers 64, 246, 482, 691, 870, 872, 873 divergence 4, 75, 156, 213⫺214, 241ⴚ242, 244, 246, 251, 252, 275, 276, 277⫺291, 325, 355, 384, 478, 481⫺482, 485, 488, 489, 490, 834, 869, 872

881 Dutch 96, 136, 137, 138, 180, 181, 214, 236, 243, 276, 283ⴚ285, 291, 300, 320, 356, 359, 361, 362, 365, 366, 368⫺370, 383, 393, 441, 442, 453, 457, 480, 514, 523, 556, 596, 746, 766, 771, 777, 825, 829, 830⫺834, 842, 848, 853, 858 dynamic language atlases (sprachdynamische Atlanten) 202, 204ⴚ206, 215, 219

E East German 279, 288, 873 edit distance 552ⴚ554, 556, 557 electronic dialect lexicography 508, 514, 600ⴚ608 elicitation 79, 396, 512, 523⫺524, 536, 545, 671, 825, 843, 845, 846, 858 ⫺ oral J oral elicitation ⫺ written J written elicitation empirical quality standards 495, 508 ⫺ objectivity J objectivity ⫺ reliability J reliability ⫺ representativity J representativity ⫺ validity J validity enclaves 118, 334, 335, 337, 338, 340, 489 English 39, 42, 61, 63, 129, 137, 144, 145, 156, 174, 175, 190, 230, 252, 269, 315, 341, 386, 393, 395, 401, 404⫺406, 409, 421, 427, 441, 452, 454, 479, 480, 484⫺485, 486⫺487, 540, 620, 664⫺665, 735, 748, 750, 805, 806, 809, 839, 844⫺846, 849, 852, 853⫺859 ⫺ American J American English ⫺ Australian J Australian English ⫺ British J British English ⫺ New Zealand J New Zealand English ⫺ Quebec J Quebec English ⫺ Tristan da Cunha J Tristan da Cunha English English accents 189, 269, 542 environmental determinism 3 essentialism 700⫺701 ethnography 47, 59, 62, 122, 128, 163, 164ⴚ 166, 168, 169, 170, 174, 625, 634, 636, 639, 728, 730, 869 ⫺ multi-sited J multi-sited ethnography European tradition 45, 202, 276 evaluation 111, 149, 174, 209, 219, 253, 290, 298, 300, 327, 336, 508⫺509, 529, 538ⴚ539, 541ⴚ546, 657, 777, 778, 871 experiment 495⫺496, 502, 525, 529ⴚ531, 532⫺546

882

Index

F feminist 7, 9, 10, 11 Finland 276, 385⫺386 Fischer, Hermann 87, 94, 96ⴚ98, 99, 101, 574 Flanders 95, 96, 284, 286, 288, 320⫺321, 323⫺328, 356⫺371, 382, 383, 833, 834, 843, 853, 857 Flemish 283, 286, 303, 320⫺321, 324⫺326, 328, 356⫺362, 364⫺366, 368, 369, 380⫺ 383, 833, 842, 853 focusing 206, 301, 323, 455, 457, 460, 777 folk 4, 34, 113, 180, 190, 195, 699 folk notion 179, 181, 184, 190, 196, 541 foreign accent 175, 363, 473, 534, 542, 690, 750 formal, formality 25, 26, 53, 149, 167, 174, 231, 234, 237, 262, 268, 347, 357, 360, 385, 456, 536, 537, 823, 832, 865, 869, 874 France 24, 110, 118, 204, 285, 286, 319, 338, 358, 380, 383, 384, 387, 443, 452, 455, 468, 479, 568, 577, 578, 729, 730, 733, 872 francophone Canada J Canadian French French 24, 39, 42, 111, 190, 227, 245, 277, 285⫺287, 338, 356, 377, 384, 441, 445, 452, 479, 485, 592, 805, 812, 822, 833, 858 ⫺ Acadian J Acadian French ⫺ Canadian J Canadian French frequency 77, 94, 98, 134, 151, 213, 214, 232, 236, 245, 252, 253, 366, 397, 403, 404, 410, 430, 455, 456, 486, 489, 450, 508, 625, 695, 750, 753, 832, 833, 842, 844, 845, 853, 872 Fronterizo 706⫺722 full variety 203, 215, 216, 217 functional stability 355

G gender 1, 7, 8, 9, 79, 143, 144, 148, 155, 164⫺165, 167, 190, 206, 233, 316, 474, 488, 494, 497, 507, 583, 584, 632, 636, 638⫺639 generative dialectology 129ⴚ138, 229, 623 genre ⫺ communicative J communicative genre ⫺ media J media genre geographic adjacency 380, 811 geographic(al) space 8, 35, 45, 117, 245, 375, 376, 724, 773, 809, 810, 811 geographic(al) variation 126, 227, 231, 315, 505, 616, 822

geography 1ⴚ12, 35, 51, 143, 144, 145, 147, 317, 375ⴚ376, 574, 633, 750, 772 ⫺ dialect J dialect geography ⫺ humanistic J humanistic geography ⫺ linguistic J linguistic geography ⫺ Marxist-influenced J Marxist-influenced geography geolinguistics 145, 147, 155, 205, 375, 574, 581, 583, 586, 669, 678 German 75, 81, 90, 205, 207, 216, 228, 277, 279, 289, 300, 305, 333, 337, 479, 480, 487⫺ 488, 497, 555, 579, 592, 767, 775, 795, 807, 809, 812, 857 ⫺ Austrian J Austrian German ⫺ East J East German ⫺ Swiss J Swiss German ⫺ West J West German German dialectology 94, 120, 237, 854 Germany 57, 94, 100, 108, 109, 112, 116, 118, 121, 204, 206, 207, 216, 230, 250, 251, 252, 276, 277, 279, 284, 285, 286, 380, 452, 499, 569, 580, 582, 584, 747, 750, 865, 873 Gillie´ron, Jules 92, 109, 204, 315, 502, 515, 574, 576ⴚ578, 580, 584, 614, 616, 664, 761 globalization 12, 41⫺42, 51, 54⫺56, 63, 445, 447⫺448, 544, 632ⴚ635 glocalization 28, 55 grammar ⫺ dialect J dialect grammar ⫺ regional J regional grammar grammaticalization 344, 423, 457, 680, 691, 777, 839, 858

H Haag, Karl 83, 94, 98ⴚ100 Herder, Johann Gottfried 4, 34⫺35 heterogeneity 39, 147, 226, 339, 398, 458, 569, 575, 577, 580, 584, 618, 619, 621, 623, 698, 741, 754 heteroglossia 37, 743, 746 hierarchical diffusion 6, 148, 317⫺323, 324, 326, 814 high variety (H-variety) 281, 296, 297, 358, 363, 365 history of ⫺ dialectology 1, 205, 613, 620, 623 ⫺ linguistics 204, 577, 582, 587, 724 ⫺ sociolinguistics 1, 108, 613⫺616, 617, 619⫺623, 624 homogeneity 33, 434, 622, 687, 698

Index horizontal convergence 242ⴚ247, 250⫺253, 278 horizontal stability 355 humanistic geography 7, 10, 11 hybrid form 149, 265, 269, 270 hypercorrection 215, 217, 269, 299, 307, 457, 660, 764 hyperdialectalism 299

I identity 7, 9, 12, 40, 55, 56, 63, 65, 119, 148, 149, 175, 188, 195, 236, 237, 251, 275, 289, 325, 333, 337, 348, 370, 444, 446, 448, 453, 472, 474, 490, 657, 687, 689, 725⫺733, 741, 745, 751, 752, 753 ⫺ local J local identity ⫺ regional J regional identity ⫺ rural J rural identity ⫺ social J social identity ⫺ urban J urban identity identity marker 155, 306, 325, 326, 346, 360, 748, 814, 822 imperialism 3, 116, 452 implication 217, 245, 303, 445, 446, 653, 656, 658, 660, 664, 817 independent variable 406, 409, 500, 530, 532ⴚ533, 625, 822, 828, 829, 835 indirect method 79, 87⫺88, 92, 96, 496, 499, 505⫺507, 515ⴚ516, 522, 544, 579 inflection 80, 88, 208, 210, 432⫺433, 482, 674, 804, 805, 806, 808, 811, 815 informal, informality 25, 53, 149, 167, 175, 237, 268, 357, 369, 471, 536, 537, 689, 692, 696, 698, 742, 749, 832, 865, 869, 872 innovation 6, 7, 63, 76, 77, 97, 99, 126, 143, 145, 146, 148ⴚ151, 203, 205, 208, 209, 219, 317, 318, 320, 321, 322, 324, 325, 336, 396ⴚ 397, 431, 447, 457, 576, 649, 650, 651, 654ⴚ 660, 662, 663, 679, 711, 714, 715, 716, 717, 719, 721, 854 intensity 764, 787⫺788, 789, 797 interference 23, 231, 283, 285, 303, 368, 399, 410, 427, 475, 516, 517 internet 51, 54, 59, 60, 61, 447, 471, 506⫺ 507, 635, 825 interview 206, 495, 502, 503, 515, 516ⴚ520, 522, 529, 535ⴚ537, 539, 641, 858, 872 intonation 268, 333, 553, 644, 787, 790ⴚ795, 797, 798, 799

883 isogloss 89, 91, 94, 126, 128, 151, 154, 203, 207, 210, 215, 219, 236, 242, 244, 246⫺247, 253, 275, 513, 525, 575, 578, 615, 656, 658, 662, 663ⴚ664, 665, 673, 674, 761, 762, 769⫺770, 794, 796⫺797, 837, 853, 854

J Jaberg, Karl 204, 580, 583, 618, 761 Jud, Jakob 204, 580, 583, 761

K koine´, koineization 113, 152, 153, 232, 243, 253, 277, 287, 301, 324, 328, 336, 339, 340, 346, 359, 368, 445, 453, 455, 458, 460, 463, 469 kulturmorphologische Dialektologie 107ⴚ110, 113, 114ⴚ117, 120⫺122

L language ⫺ architecture of J architecture of language ⫺ and place 7, 65, 155, 632⫺633, 645 ⫺ marginal J marginal language ⫺ regional J regional language ⫺ transnational J transnational languages language atlas 90, 202, 204⫺206, 215, 219, 572, 575, 580, 583, 586 language attitudes 188, 337, 348, 442, 529, 537⫺540, 545⫺546, 676, 696⫺698, 843 language change 29, 35, 73⫺74, 77, 87⫺88, 93, 100⫺101, 108, 147, 155, 203, 206ⴚ207, 209⫺210, 219⫺220, 242, 247, 251⫺252, 265, 316, 321, 325, 344, 355, 358⫺359, 366⫺367, 424ⴚ426, 436, 456, 469, 649⫺ 651, 657, 662, 809 language contact 118, 206, 244, 251, 284, 337, 341, 344⫺345, 379, 392, 399, 409, 411⫺412, 419⫺420, 427ⴚ432, 447, 453, 455, 460, 463, 467, 469, 471, 475, 478ⴚ484, 490, 698, 706, 775⫺767, 805, 808, 811ⴚ813, 855, 859 language enclave J Sprachinsel language evaluation 111, 149, 174, 182, 209, 216, 218⫺219, 280, 290, 297⫺301, 305⫺ 307, 319, 327, 336, 442, 508⫺509, 538⫺541, 541ⴚ546, 869, 871

884

Index language galaxy 42 language ideology 11, 34, 38⫺40, 43, 55⫺56, 62⫺63, 108, 110, 120, 266, 269, 271, 614, 618, 641⫺642, 646, 687, 726, 735, 741 language industry 734⫺736 language island J Sprachinsel language loyalty 118, 150, 328, 344, 482, 489, 625 language maintenance 101, 214, 322⫺323, 341, 345ⴚ347, 424⫺426, 458, 474, 483 language map 33, 36, 88, 379, 384, 502, 505, 574, 578, 855 language pocket J Sprachinsel language space 1, 6, 8, 228, 231, 260, 271, 281, 290, 376 language space ideology 107 language typology 334, 419ⴚ420, 422⫺423, 434, 453, 571, 840ⴚ843, 856⫺858 language use 21, 27, 129, 164, 167, 169, 228, 234⫺235, 267, 269, 271, 285, 341, 376, 379, 396, 505, 514, 529, 532ⴚ533, 536, 538, 578, 583⫺584, 640, 642, 687⫺689, 699, 750, 822 language variation 126⫺127, 129, 138, 155⫺ 156, 205, 226⫺227, 229⫺230, 234, 266, 296, 316, 375, 461, 497, 513⫺514, 529, 613⫺614, 633, 686, 698⫺699, 702, 745, 750, 868⫺869 lectal variation 821⫺826, 831 leveling 12, 75, 242ⴚ243, 301, 324, 357, 370, 455⫺457, 460 ⫺ of dialects J dialect leveling ⫺ rural J rural leveling Levenshtein distance 552, 554⫺557, 559, 770 lexical diffusion 75, 82ⴚ83, 250, 269, 364, 657, 663 lexical variation 127, 136, 506⫺507, 821⫺ 829, 832, 835, 846, 866 lingua franca 232, 243, 281, 289, 368, 431, 441, 447, 454, 471 linguistic atlas 91⫺92, 204⫺206, 513⫺517, 567ⴚ568, 571ⴚ572, 574, 576, 581⫺582, 616, 669, 677, 761 linguistic convergence 342 linguistic dynamics (Sprachdynamik) 201⫺ 203, 211⫺212, 218⫺219, 585, 669⫺670, 680⫺681 linguistic geography 87, 95, 98, 100, 155, 375, 507, 581, 709, 761 linguistic market 19, 21⫺22, 26, 30, 265, 730, 874 linguistics ⫺ history of J history of linguistics

linguistic space 121, 188, 227, 237, 251, 275, 277, 291, 336, 376, 446, 470ⴚ471, 576, 584, 696 linguistic stability 355, 358, 360, 370, 481 linguistic variable 128, 129, 164, 226, 229ⴚ 230, 238, 243ⴚ247, 398, 514, 516, 517, 551, 577, 625, 627 local identity 55, 118, 150, 151, 169ⴚ170, 172, 325, 360, 458, 538, 742, 745, 749 location-based services 60 Locke, John 34⫺35 low variety (L-variety) 277, 281, 297, 298, 357 Luxemburg 242, 361, 595, 597, 606⫺607

M macrosynchronization 213ⴚ214, 216 map, mapping 33, 35, 40⫺41, 53, 89, 94, 98, 126⫺128, 164⫺165, 179⫺182, 184, 190, 204⫺205, 247, 502⫺503, 505, 567⫺587, 672, 709⫺722, 769, 865 ⫺ dialect J dialect map ⫺ language J language map ⫺ mental J mental map Marburg School 88, 94, 96, 505, 771 marginal language 205, 377⫺379, 387, 399, 440ⴚ441, 443, 448 Marxism 6⫺11 Marxist-influenced geography 7, 8ⴚ9 mass media 9, 214, 741, 750 meaning in variation 163ⴚ169, 869 measuring, measurement 245, 253, 323, 394, 494ⴚ497, 499, 524, 530, 532⫺533, 544ⴚ 546, 550ⴚ564, 569, 639, 671, 677, 762, 764ⴚ770, 789, 791 media ⫺ new J new media ⫺ mass J mass media media discourse 740⫺741, 743ⴚ744, 750 media genre 744ⴚ746, 749⫺753 mental map 184ⴚ185, 188⫺189, 191, 195⫺ 196, 539⫺541 mesolect 232, 444, 463 mesosynchronization 202, 213, 216, 219 methods of data collection ⫺ data elicitation J elicitation ⫺ direct J direct method ⫺ indirect J indirect method microparametric variation 840 microsynchronization 212⫺214

Index microvariation 129⫺130, 132⫺133, 135⫺ 137, 139, 817 migrants, migration 50⫺51, 55, 251, 322⫺ 323, 338⫺340, 380⫺381, 443, 452, 468⫺ 470, 688, 726 minority 334ⴚ335, 338, 346, 376, 378⫺381, 383, 387, 468⫺471, 474⫺476 Mischung J dialect mixture Mitzka, Walther 89, 111, 113, 120⫺122, 650 mobile positioning 60 mobility 11, 50⫺51, 152, 251, 315⫺316, 321, 327, 632⫺633 modern dialectology 838 modernity 29ⴚ30, 34⫺35, 39 morphological exponent 807⫺808, 815 morphology 79⫺80, 90, 126, 130, 135, 137, 204, 210, 243, 252, 285, 290, 301, 345, 365, 366, 392, 394, 403, 433, 480, 482, 486, 503, 506⫺507, 578, 582, 600, 616, 671, 674, 804⫺810, 814ⴚ816, 826, 844, 851 morphosyntactic variable 868, 872 movement (in space) 52, 251, 338⫺340, 442, 451⫺452, 811 ⫺ population J population movements multidimensional scaling 190, 556, 770 multilingualism 42, 133, 490, 687⫺702 multiple grammars model 133 multi-sited ethnography 726, 730

N narrowcasting 742, 744, 746, 752 national standard variety 22, 23, 24 , 25, 26, 108 , 111, 112, 113, 114, 214, 216, 217, 262, 263, 267, 269, 300, 301, 302, 308, 323, 356 nationalism 2⫺3, 34, 50, 109, 118, 725⫺729, 736 nation-state 3⫺4, 33, 35⫺40, 375, 379 navigation systems 2, 10, 58ⴚ60 neogrammarian 70ⴚ84, 108, 245 Netherlands 95⫺96, 180, 276, 283⫺284, 288, 303, 327, 356⫺357, 359, 361, 366, 517, 832ⴚ834, 843 network J social network new dialect 300⫺301, 324, 455, 457, 460 new media 740 new standard variety 305 New Zealand English 456, 458, 460, 461 non-convergence 480ⴚ482, 489 Norway 26, 94⫺95, 263⫺264, 268⫺272

885

O objectivity 495 obtrusive(ness) 530⫺531 Occitan 24, 383ⴚ385 onomasiology 824, 832, 835⫺836 Optimality Theory 129, 132, 135, 138, 252, 491, 775, 777, 840 oral elicitation 514⫺515, 516ⴚ520 oral performance 155, 169, 196, 202, 212, 220, 251, 494, 615, 620, 622⫺623, 641, 645, 751⫺752, 869, 872 origo 45 Ortsgrammatik 79ⴚ82, 616 Ortsloyalität 251, 328, 360⫺361, 363⫺365, 367 overseas variety 453⫺455

P paradigm 622, 805, 807 particle 246, 691, 844, 851, 873ⴚ874 Paul, Hermann 70, 71, 72ⴚ78, 111, 202, 212 perceptual dialectology 179⫺182, 192⫺193, 539 phenomenology 10, 45⫺46, 53⫺55 phonetics 760, 764, 779, 787 phonology 77, 80, 129⫺130, 132, 135, 290, 368, 760, 769, 771, 774⫺777, 779, 787 pidgin 440ⴚ448, 453⫺454 pidginization 441, 483 place J language and place pluridimensional 206, 568, 578, 583⫺585, 669, 704, 709 pluridimensional dialectology 583, 706 political space 33 political territories 33, 99, 110, 116⫺117, 120⫺121, 276, 289, 382, 725 population movements 120⫺121, 338ⴚ340, 442, 451⫺452 Portuguese 243, 338, 393, 441, 452, 471, 485, 487, 585, 706ⴚ711, 713⫺714, 716⫺722 ⫺ Riograndense J Riograndense Portuguese ⫺ Uruguayan J Uruguayan Portuguese postmodernism 10⫺11 practice J social practice pragmatics 290⫺291, 494, 497, 668, 866 primary dialect 231, 252, 290, 336 primary dialect feature 252, 280, 290, 336, 362

886

Index pronunciation distance 553, 559⫺563 prosody 691, 786ⴚ789 psychometrics 562

Q quality criteria J empirical quality standards Quebec 401⫺403, 407, 469, 726, 728⫺733 Quebec English 401⫺402, 407 questionnaire 79, 89, 94⫺96, 502⫺503, 506⫺ 507, 514ⴚ517, 520⫺524, 538, 671, 843⫺846

R reallocation (of variants) 308, 456 real-time analysis 202, 204ⴚ205, 207, 218, 253, 404, 680, 832 reevaluation of varieties 297⫺299 regiolect 152⫺153, 203, 216⫺218, 232, 296⫺ 298, 302ⴚ304, 306⫺308, 369 region 3, 5, 10, 147, 153 regional accent (Regionalakzent) 217⫺218, 232, 307, 363, 368⫺369, 504, 506, 542, 544, 680 regional grammar 88, 93, 202, 502 regional identity 267, 270, 271, 328, 334, 345, 360, 749 regional language 94, 203, 211, 215ⴚ220, 305, 334, 385, 504⫺506, 676, 678, 679 register 41⫺42, 234ⴚ235, 502⫺503 regularization 39, 365⫺366, 456⫺457 Reihenschritt(theorie) 252, 653 reliability 495ⴚ496, 512, 516, 524, 530, 618 relict zone 678, 679 repertoire 231, 260, 290⫺291, 474, 775⫺776 representativity 495, 498, 501, 508⫺509 residual zone 431⫺432 restructuring of varieties 276, 297ⴚ308 Rhenish fan (Rheinischer Fächer) 243, 248⫺ 249 Riograndense Portuguese 708, 710, 717 Romance tradition 515, 580, 761 roof J Dach(sprache) rule-based 134 ⫺ phonology 129⫺131 ⫺ syntax 129, 135⫺136 rural identity 814 rural leveling 79, 148⫺149, 316⫺319, 323, 326ⴚ328 rurbanization 28, 728

S salient, saliency, salience 209, 252, 301, 700, 869 sampling 399, 534, 701 secondary dialect 231 secondary dialect feature 336, 362⫺363 sectoral variety 211, 215 semantics 79, 127, 596, 826, 828 semasiology 581, 582, 596, 604, 824 shifting J code shifting similarity 179ⴚ183, 275, 551, 554, 770, 836 simplification 301, 344, 366, 402, 427, 456ⴚ 457, 463 Sinitic 430 slang 237, 447, 639, 697, 700, 702, 743 social accent 542, 544 social boundary 11, 28, 56 social dialectology 615, 619ⴚ621, 625, 752 social factors 79, 420, 423ⴚ424, 615 social meaning 117, 167, 176, 543, 869 social identity 18, 19, 52, 53, 226, 633, 639, 752, 868 social network 6, 8⫺9, 50⫺51, 63, 65, 151, 165, 170ⴚ171, 234, 321ⴚ323, 328, 425, 474⫺475, 625ⴚ626 social practice 7⫺8, 11, 19ⴚ20, 29⫺30, 43⫺ 46, 143, 151, 154⫺155, 163ⴚ167, 169⫺171, 176, 447⫺448, 633, 635ⴚ642, 644⫺646, 728, 736, 865 social science 5, 144, 508, 615, 622, 725 social space 21, 27, 163, 532 social style 646 social variation 169, 232, 316, 583, 620, 822 sociolects 233, 584, 822 sociolexicology 823ⴚ824, 825, 826, 835 sociolinguistic variable 229, 638, 823, 824, 826, 835 sociolinguistics ⫺ correlational J correlational sociolinguistics ⫺ history of J history of sociolinguistics ⫺ variationist J variationist sociolinguistics sound change (Lautwandel) 74, 76, 77, 78, 82ⴚ84, 99⫺100, 170, 245, 655, 767, 773, 779 Southeast Asia 41, 42 space ⫺ absolute vs. relative J absolute vs. relative space ⫺ communicative J communicative space ⫺ creole J creole space ⫺ cultural (Kulturraum) J cultural space

Index ⫺ digital J digital space ⫺ discontinuous language J discontinuous language space ⫺ geographic(al) J geographic(al) space ⫺ language J language space ⫺ linguistic J linguistic space ⫺ political J political space ⫺ social J social space ⫺ topographic J topographic space ⫺ transnational J transnational spaces ⫺ variation J variation space ⫺ virtual J virtual space spatial analysis 575 spatialization 38, 39, 40 spatial variation 143, 764, 821, 822 speech levels (Sprechlagen) 41ⴚ42, 215, 217, 218, 303, 306 Sprachatlas J Deutscher Sprachatlas Sprachbund 427ⴚ428, 812 Sprachinsel 118, 332ⴚ348, 381, 489, 807⫺ 808, 811 Sprachdynamik 211 Sprechlagen J speech levels spontaneity, spontaneous 76, 399, 441, 512, 514, 520, 650, 654, 717, 825, 858 spread zone 431⫺432 stability ⫺ functional J functional stability ⫺ horizontal J horizontal stability ⫺ linguistic J linguistic stability ⫺ structural J structural stability ⫺ vertical J vertical stability Stamm, Stämme 108, 118ⴚ121 standard variety 22ⴚ23, 24, 26, 29, 217, 277, 297, 290, 296, 299ⴚ300, 301, 302, 305, 307, 308, 345, 346, 356, 410, 856 ⫺ national J national standard variety ⫺ new J new standard variety standardization 24, 38ⴚ39, 262, 301 St. Helena 461, 463 stimulus variable 531ⴚ534, 535, 540, 541 string distance 552 structural boundary 209, 214, 217, 296, 674⫺676, 678 structural complexity 432, 444, 859 structural dialectology 127⫺128 structuralism 70, 71, 126, 127 structural stability 355, 356, 365ⴚ366 study of variation 164, 169, 176, 866 style 155, 166ⴚ169, 173⫺176, 234ⴚ235, 271, 325, 503, 541⫺542, 544, 638⫺642, 646, 696, 701, 717, 742⫺743, 751⫺754, 865⫺867

887 ⫺ conversational J conversational style ⫺ social J social style style cluster 638ⴚ640, 646 substandard 57, 216, 230, 237ⴚ238, 252, 259, 284, 291, 302, 357, 368, 369, 708, 717, 718, 719, 867, 870 superimposition (of standard variety) 289, 304, 305, 306, 307 supra-local structure 152 Swedish 95, 380, 688, 689, 690, 693, 694, 695, 697, 699, 700, 806 ⫺ in Finland 385, 386 Swiss German 74, 216, 243⫺245, 247, 252, 279, 282, 289, 359, 360, 471, 518, 521⫺522, 524, 649, 651⫺652, 658, 659, 660, 763, 772⫺773, 788, 792, 817, 847, 848, 870 switching J code switching synchronization 211ⴚ213, 215, 216, 220, 676, 678, 681 ⫺ macro- J macrosynchronization ⫺ meso- J mesosynchronization ⫺ micro- J microsynchronization syncretism 210, 807 synonymy 579, 827 syntax 79, 113, 126, 135⫺138, 243, 290, 345, 365, 367, 392, 423, 425, 432, 482, 483, 490, 496, 506, 507, 513, 578, 582, 616, 689, 815, 826, 831, 837ⴚ859

T temporal 788⫺789 territorial principle 381⫺383, 385 territory 33, 37, 51, 56, 60, 65, 99, 114, 116, 117, 121, 144, 168, 171, 186, 193, 268, 276, 301, 375, 377, 381ⴚ383, 385, 386, 387, 388, 471, 712, 725, 746, 747, 794, 811 ⫺ political J political territories tertiary dialect feature 368 theory of language change J language change tone accent 674, 787, 796, 798, 808 tone accent boundary 674 topographic space 810 traditional dialectology 126, 128, 201, 275⫺ 276, 315⫺316, 531, 616, 618, 641, 663 transidiomatic practices 52, 62ⴚ64, 65 transition(al) zone 99, 246, 269, 361, 579, 673, 774, 796, 805, 815, 843, 846, 853ⴚ854 translation task 518ⴚ519, 522, 524, 843 transmigrants, transmigration 51, 54

888

Index transnational languages 50⫺52, 55⫺56, 58⫺ 65, 443, 687, 729 transnational spaces 50⫺51 transnationalism 50, 725, 726, 729, 736 Tristan da Cunha English 457, 458, 461ⴚ463 typology 334, 344, 378ⴚ381, 419ⴚ440, 581, 838, 839, 840, 841, 842, 843, 855ⴚ857, 858 ⫺ areal J areal typology

U urban 6, 11, 25, 28, 40, 79, 111, 115, 121, 143, 145ⴚ146, 147, 148, 149, 151, 154, 155, 156, 157, 171, 172, 232, 246, 259, 270, 280, 281, 290, 302, 315ⴚ332, 333, 337, 340, 345, 346, 347, 348, 382, 445, 447, 448, 470, 475, 614, 618, 619, 625, 688, 689, 698, 701, 702, 718, 727, 743, 748, 750, 814, 867, 868 urban identity 325⫺326 urban space 156, 688, 698, 702 urban dialectology 145, 316, 475 urbanization 28, 316, 323, 327, 328, 424, 618, 728 Uruguayan Portuguese 485, 707⫺722 USA dialects 230

V validity 26, 91, 181, 299, 302, 366, 436, 471, 494, 495, 496, 497ⴚ498, 499, 501, 503, 504, 506, 507, 509, 530, 544, 562, 575, 618, 680 variability 206, 226, 259, 303, 391, 394⫺395, 397, 400, 410⫺411, 445⫺446, 456, 458, 460, 494, 502, 571, 537, 583⫺584, 657, 662ⴚ663, 676, variable ⫺ background J background variable ⫺ dependent J dependent variable ⫺ independent J independent variable ⫺ linguistic J linguistic variable ⫺ morphosyntactic J morphosyntactic variable ⫺ sociolinguistic J sociolinguistic variable ⫺ stimulus J stimulus variable variable rule 132, 143, 400ⴚ409, 622 variation ⫺ geographic(al) J geographic(al) variation ⫺ lectal J lectal variation ⫺ lexical J lexical variation ⫺ meaning in J meaning in variation

⫺ microparametric J microparametric variation ⫺ micro- J microvariation ⫺ spatial J spatial variation ⫺ social J social variation ⫺ study of J study of variation variationist sociolinguistics 145, 148, 398, 531, 535, 536, 614, 635, 805 variation space 621 variation theory 163 variety 22, 112, 128, 168, 174, 202, 203, 205, 211, 214ⴚ215, 218, 226, 227, 228, 229, 235, 236, 237, 242ⴚ244, 251, 252, 253, 259, 261, 266, 271, 275, 277, 289, 290ⴚ291, 295ⴚ296, 297, 299, 302, 307, 334, 339, 341, 343, 347, 357, 367, 369, 370, 393, 395, 434, 445, 453ⴚ 454, 455ⴚ456, 458, 463, 464, 472, 474, 499, 615, 688, 779, 795, 811, 833, 845, 852, 855, 856, 859 ⫺ concept of J concept of variety ⫺ contact J contact variety ⫺ diaphasic J diaphasic variety ⫺ diastratic J diastratic variety ⫺ diatopic J diatopic variety ⫺ full J full variety ⫺ high (H-variety) J high variety (H-variety) ⫺ low (L-variety) J low variety (L-variety) ⫺ national standard J national standard variety ⫺ new standard J new standard variety ⫺ overseas J overseas variety ⫺ reevaluation of J reevaluation of varieties ⫺ restructuring of J restructuring of varieties ⫺ sectoral J sectoral variety ⫺ standard J standard variety ⫺ vernacular J vernacular varieties variety complexes 708 variety formation 213, 215, 216, 217, 265, 301, 453 Verdichtungsbereiche J speech levels vernacular, vernacularization 23, 26, 27, 166, 167, 232, 243, 245, 280, 281, 283, 286, 291, 296, 297ⴚ298, 317, 318, 323, 325, 337, 339, 347, 394, 395, 399, 446, 448, 504, 593, 621, 626, 632, 641ⴚ642, 645, 696, 741, 742, 744, 745, 746, 747, 748, 749, 750, 751, 754, 844, 846, 867, 869 vernacular varieties (lantsprachen) 174, 297, 299, 302, 621, 625, 747, 749 vertical convergence 259ⴚ260, 265, 269, 271⫺272 vertical stability 355

Index virtual community 61 virtual space 743, 746 vocalism 130, 284 voice 191ⴚ195, 248, 250, 252, 256, 319, 366, 456, 457, 517, 541, 543, 544, 545, 546, 558, 559, 617, 638, 677, 678, 679, 742, 744, 745, 746, 747, 748, 752ⴚ754, 764, 788 voice placement 193 voice quality 195, 644, 787, 790 Volk 34, 107ⴚ122 Volksforschung 108, 110, 111

W wave diffusion J contagion/contagious diffusion Wenker, Georg 87, 88ⴚ91, 92, 97, 204, 574, 575ⴚ576, 577, 578, 579, 583, 616, 760, 761 West German 278, 279, 873 working-class accent 318

889 Wrede, Ferdinand 87, 88, 89, 91ⴚ94, 97, 100, 101, 108, 120, 121, 505, 579 written elicitation 514⫺515, 520ⴚ525 Württemberg School 88, 96ⴚ101

Y youth 24, 47, 61, 213, 246, 396, 697, 872

Z zone 57, 151, 152, 185, 209, 219, 228, 278, 286, 296, 303, 357, 379, 380, 382, 383, 662, 707, 843, 847, 854 ⫺ contact J contact zone ⫺ relict J relict zone ⫺ residual J residual zone ⫺ spread J spread zone ⫺ transition(al) J transition(al) zone zone convergence 427⫺429