Competing Models of Linguistic Change: Evolution and Beyond (Current Issues in Linguistic Theory)

Competing Models of Linguistic Change AMSTERDAM STUDIES IN THE THEORY AND HISTORY OF LINGUISTIC SCIENCE General Edito...

Author: Ole Nedergaard Thomsen

282 downloads 1912 Views 2MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Competing Models of Linguistic Change

AMSTERDAM STUDIES IN THE THEORY AND HISTORY OF LINGUISTIC SCIENCE General Editor E.F.K. KOERNER (Zentrum für Allgemeine Sprachwissenschaft, Typologie und Universalienforschung, Berlin) Series IV – CURRENT ISSUES IN LINGUISTIC THEORY Advisory Editorial Board Lyle Campbell (Salt Lake City); Sheila Embleton (Toronto) Brian D. Joseph (Columbus, Ohio); John E. Joseph (Edinburgh) Manfred Krifka (Berlin); E. Wyn Roberts (Vancouver, B.C.) Joseph C. Salmons (Madison, Wis.); Hans-Jürgen Sasse (Köln)

Volume 279

Ole Nedergaard Thomsen (ed.) Competing Models of Linguistic Change. Evolution and beyond.

Competing Models of Linguistic Change Evolution and beyond

Edited by

Ole Nedergaard Thomsen Roskilde University

JOHN BENJAMINS PUBLISHING COMPANY AMSTERDAM/PHILADELPHIA

4-

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences — Permanence of Paper for Printed Library Materials, ANSI Z39.48-1984.

Library of Congress Cataloging-in-Publication Data Competing models of linguistic change : evolution and beyond / edited by Ole Nedergaard Thomsen. p. cm. -- (Amsterdam studies in the theory and history of linguistic science. Series IV, Current issues in linguistic theory, ISSN 0304-0763 ; v. 279) Includes bibliographical references and index. 1. Linguistic change. P142.C66 2006 417/.7--dc22 2006049278 ISBN 90 272 4794 3 (Hb; alk. paper) © 2006 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. • P.O.Box 36224 • 1020 ME Amsterdam • The Netherlands John Benjamins North America • P.O.Box 27519 • Philadelphia PA 19118-0519 • USA

CONTENTS

Introduction Ole Nedergaard Thomsen I:

……………………………………………..1

General considerations of language change

The non-linear nature of diachronic change Michael Fortescue ...………………………………………………….17 Explanations, or...?—Some metatheoretical reflections on a prevalent tradition within historical linguistics Brit Mæhlum ………………………………………………………….33 Quantifying the functional load of phonemic oppositions, distinctive features, and suprasegmentals Dinoj Surendran & Partha Niyogi …………………………………...43 II: The concept of ‘evolution’ as an explanatory model of language change: Pro et contra Synchrony, diachrony, and evolution Henning Andersen ……………………………………………………59 The relevance of an evolutionary model to historical linguistics William Croft …………………………………………………………91 III: Functional factors in ‘evolution’: Functional motivation of selection Grammaticalization of indirect object cross-reference in Spanish as a case of drift Silvia Becerra Bascuñán ……………………………………………133 The role of functional factors in language change: An evolutionary approach Guido Seiler ...………………………………………………………163

vi

CONTENTS

Computational modelling of prototypicality in language change: Neutralization to schwa, default logic, and the history of the German noun Miguel Vázquez-Larruscaín .………………………………………..183 IV: Cognitive perspectives on semantico-syntactic change: Mental Grammar, Cognitive Grammar, and Dynamic Syntax From propositional syntax in Old Russian to situational syntax in Modern Russian Per Durst-Andersen …………………………………………………211 Construal operations in semantic change: The case of abstract nouns Lena Ekberg …………………………………………………………235 Clitic placement in Old and Modern Spanish: A dynamic account Miriam Bouzouita & Ruth Kempson .……………………………….253 V: Grammaticalization studies: Content and evolution Grammaticalisation as content reanalysis: The modal character of the Danish s-passive Lars Heltoft ………………………………………………………….269 Aspect and animacy in the history of Russian: Developing the idea of parallel grammaticalization Jens Nørgård-Sørensen ……………………………………………..289 VI: Integrating theories of language and language change Towards an integrated functional-pragmatic theory of language and language change. In commemoration of Eugenio Coseriu (1921-2002) Ole Nedergaard Thomsen …………………………………………...307 Index

………….…………………………………………………………..339

Contributors

……...………………………………………………………343

INTRODUCTION The present volume, Competing Models of Linguistic Change. Evolution and beyond, is essentially a selection from the contributions to Section 4, “Different models of linguistic change”, of The 16th International Conference on Historical Linguistics, held at the University of Copenhagen, 11–15 August 1 2003 (ICHL 2003). The section was centered around the competitive plenary papers by Henning Andersen and William Croft, respectively, on linguistic change as ‘evolution’, and the title of this volume has accordingly been chosen to reflect this theme and the discussion of the two competing stances on the viability of applying to phenomena of linguistic change — phenomena of a humanistic nature (involving freedom and creativity) — a theory of evolution originating from the natural sciences (dealing with chance and necessity). Of course, right from the beginning of modern (historical) linguistics, in the 19th century (cf. August Schleicher 1863; Diderichsen 1966a [1958], 1966b [1959]; Keller 1988; Janda & Joseph 2003), there has been an interest in (metaphorical or real) biological parallels to linguistic change — cf., trivially, the Stammbaum model of the Neogrammarians, with the parallel between linguistic reconstruction of a protolanguage and the Darwinian biological reconstruction of the ‘origin of species’. But only with the explicit and exact proposals by William Croft (2000) has it become possible to truly advance with an evolutionary model of language change, not as a natural phenomenon, but within a domain-neutral model of a ‘generalized analysis of selection’ (Hull 1988, 2001). However, an oppositional contention, at least going as far back as Coseriu (1952, 1957), and eloquently formulated by Henning Andersen (this volume), takes it that cultural phenomena could not possibly be handled (observed, described, and ‘explained’, i.e., understood) in the same way as natural phenomena, that is, that a (cultural) ‘memetics’2 could not be con1

A selection of papers from the general program has been published as volume 257 of the present series, Current Issues in Linguistic Theory (cf. Fortescue, Skafte Jensen, Mogensen & Schøsler, eds. 2005). 2 We are talking about the theory of ‘memes’, roughly “memory patterns”; from Greek mimeme “something imitated”, cf. mimesis “imitation”; and/or mneme “memory”.

2

COMPETING MODELS OF LINGUISTIC CHANGE

ceived of as dealing with the same type of phenomenon as (biological) genetics, however generally understood.3 While I was organizing the section on models of linguistic change, I received the sad news of the death in October 2002 of the aforementioned great functional-structural and historical linguist Eugenio Coseriu (1921–2002). Given his still utmost relevant theory of linguistic change, which dates back to (at least) the mid fifties of the previous century (Coseriu 1952, 1957), I felt an obligation to commemorate his extraordinary genius. Therefore, I asked Henning Andersen, as an expert on Coseriu, and also already an invited keynote lecturer for the conference, to offer a commemorational sketch of Coseriu’s theory, as a preliminary to the section on ‘different models’. The outline was felicitously delivered, but regrettably it could not be included in the volume. However, fortunately, the commemoration of Coseriu is present in a couple of the papers. I hope that, retrospectively, the importance of Coseriu’s seminal contributions to an integrated theory of language and language change will become more widely recognized, and even be used as a foundation. The section on ‘different models’ — apart from the plenary debate between Andersen and Croft on the (ir)relevance of the concept of evolution as an explanatory model — exemplified various different current trends in the theory of linguistic change, some of them represented in the present collec4 tion. The papers The papers that are represented in the volume have been arranged into groups, corresponding to underlying themes and motives. In the following I shall briefly sketch their contents and implications. More concretely, the volume has been composed as revolving around the central theme of the (ir-) relevance of the concept of evolution as an explanatory model of linguistic change: The central, plenary papers (see Section 1.2) have been placed most 1.

Cf. Coseriu (1974 [1957]: 154n.7): “[...] die kulturellen Objekte haben eine historische Entwicklung und keine ‘Evolution’ wie die Naturobjekte.” Cf. Coseriu (p. 246): “[...] Das System entwickelt sich nicht im Sinne einer ‘Evolution’, sondern wird durch die Sprecher in Übereinstimmung mit ihren Ausdrucksnotwendigkeiten geschaffen.” It should, however, not go unnoticed that even Coseriu and Andersen use “evolutionary” terms like ‘selection’ and ‘mutation’ in their theories, see Nedergaard Thomsen (this volume). 4 Some of them, unfortunately, could not, for various reasons, be included in the volume: (Radical) Construction Grammar, Minimalist Program (projecting functional categories), Lexical Diffusion, Punctuated Equilibrium (R.M.W. Dixon’s model), Usage-Based Models (the role of frequency), mathematical models, and the Comparative Method. 3

INTRODUCTION

3

prominently in section II of the volume, after a general onset section (I) (see Section 1.1). After the central section follow three sections (III–V, see Sections 1.3–1.5) more or less stemming from studies of ‘evolutionary’ changes, viz. III on functional factors, IV on cognitive factors, and V on grammaticalization. The volume is rounded off by a coda (VI, see Section 1.6) by the editor, going beyond the concept of evolution, and focusing on a functional-pragmatic understanding of language change — based on Coseriu’s Aristotelian and von Humboldtian conception of language as ‘process’ (creation; cf. Process philosophy, see Fortescue, this volume). General considerations of language change The first two contributions concern the understanding of language change on a general theoretical level, not being especially concerned with discussing language change as an evolutionary process, in the ‘naturalist’ sense. The third paper is methodological in nature, dealing with quantitative methods of observation and description (as necessary prerequisites for sound theoretical explanation). The second and third papers deal with the two structuralist notions, ‘push–pull’ chains and ‘functional load’, respectively. In the first paper, “The non-linear nature of diachronic change”, Michael Fortescue, on the most general (theoretical/philosophical) level proposes a Whiteheadian ‘Process’ theory of language and linguistic change (thus placing himself in the von Humboldtian–Coserian tradition). He distinguishes, as is only natural for a functionalist theory, between the language of the individual and that of the collective, the community of speakers. The language of the community of speakers, for instance English, “is only an abstraction from the varieties of linguistic behavior of communities of English speakers” (p. 18n.2). The relation between the (functional) language(s) of the individuals and the (historical) language of the speech community is one of ‘emergence’. Change goes on in the community, but is driven by the single (innovative) speech acts of the individual speakers. That is, change is an ‘invisible hand’ process, in Keller’s (e.g., 1988) use of Adam Smith’s notion — and the historical language a phenomenon of the ‘third kind’. It is argued that gradual change spreading between individuals is in fact compatible with saccade-like changes in the shared system. Fortescue proposes a concept of ‘virtual occasions’ on the level of the speech community to handle the holistic, non-linear nature of linguistic change. Brit Mæhlum, in her paper “Explanations, or ...? Some metatheoretical reflections on a prevalent tradition within historical linguistics”, discusses — and rejects — proposals for explanations of linguistic change by way of system-internal factors (exemplified in her paper by ‘pull/push-chains’ in the 1.1

4


evolution of the Norwegian vowel system). She concludes that a language is not and could not be a self-regulating, “cybernetic” system, but is rather a potential of individual speakers: it is not a structure sui generis, with immanent /autonomous factors “explaining” language change, nor can change be explained by psychological factors of optimality with respect to the psychological “implementation” of the language in the minds/brains of the single speakers; and lastly, change (towards symmetry) cannot be “explained” in an instrumentalist way, seeing symmetry and simplicity as desiderata of science itself (cf. the hocus pocus branch of American structural linguistics) — and not as properties of languages (psychologically “implemented” or not). Regarding these proposals, why would a non-optimal system arise in the first place?, she asks. She opts for a (re-)introduction onto the scene of historical linguistics of the language users themselves and the socio-cultural phenomena they participate in as explanantia (cf. Joseph 1992), a trend she sees growing still stronger in current historical linguistics in recent years (exemplified by Andersen, Anttila, Shapiro, and Keller, but also, albeit to a lesser degree, Croft, whose theory is based on concepts from the ‘exact’ sciences). By seeing language change as a socio-cultural phenomenon, originating with the individual speaker-listeners, Mæhlum places herself in the Coserian tradition (and sociolinguistics). Dinoj Surendran & Partha Niyogi, in the third contribution, “Quantifying the functional load of phonemic oppositions, distinctive features, and suprasegmentals”, outline a general method for providing quantitative data on the degree to which a language ‘relies on’ different sorts of linguistic features — this reliance is the functional load (or, yield) of the given features. The proposed (computational–mathematical) method can be used as a tool in historical linguistics in general when basing itself on corpus investigations (the ‘particular’ level in investigations of language change). Specifically, it can be used in an attempt to answer the Prague school question, (to which degree) are important, entrenched features less prone to change (e.g., merge) than less important ones? Eventually, an answer to this question will be crucial to an understanding (and explanation) of (the functioning of) language. 1.2 The concept of ‘evolution’ as an explanatory model of language change: Pro et contra As stated above, it has (once again) become fashionable in linguistics to conceive of language change in terms of the biological conception of evolution, even if in a generalized format, as in Croft’s (2000) theory, Theory of Utterance Selection, which is an application and development of the evolutionary biologist David Hull’s (1988, 2001) general theory for evolutionary

INTRODUCTION

5

phenomena of any kind, termed Generalized Analysis of Selection. Trivially, when construing language change — and other kinds of cultural change as well — as comparable to biological evolution, it is evident that we have at our disposal some specific ‘functional’ concepts (in the artificial intelligence sense of ‘functional’) that are illuminating for an understanding of the phenomenon. Centrally, we have as interacting organism, or ‘interactor’, a single, historical individual natural language user possessing a set of linguistic universals (i.e., his language ‘faculty’, allowing for possible typological variation) and a specific manifestation of it, his specific language (i.e., his communicative competence, grammar) — a specific set of ‘rules’ (norms, conventions) which makes it possible for him to speak with other, at least to some extent similar natural language users. That is, the interactor is not “alone”: a normal interaction requires at least two interactors. The interactors are inter-communicatively related and constitute speech communities, a speech community being a specific population of interactors held together by their ability and habit of, on one level, communicating with and understanding one another via a linguistic convention, and, on another, meta-level, of transmitting this conventional praxis of communicating. The transmission can be under-stood as ‘replication’. The entity being replicated — the so-called ‘replicator’ — is then the linguistic convention (a rule, form, etc.). Crucially, replication is not always identical with ‘copying’ (in the sense of identity between original and copy) in that oftentimes ‘altered replication’ occurs whereby linguistic variation is generated — the natural language users make ‘innovations’, i.e., create variant replicators. The different variants “compete” in the ‘environment’ for ‘selection’ by the interactors, i.e., for ‘differential replication’ (linguistic ‘propagation’). In this way, ‘lineages’ of conventions are generated. Each natural language user on the one hand produces and receives a finite extensional set, or corpus, of actual utterances, of usage that includes ungrammatical utterances, performance errors, etc. — that is, the ‘corpus’ is a historical entity. On the other hand, his competence potentially ‘generates’ an infinite intensional set of (potential) ‘grammatical’ expressions. The ‘language’ is the corpus of the actual utterances of the speech community (in a given historical time segment). This corpus manifests the linguistic conventions of the speech community — the historical community grammar (the union set of individual speakers’ grammars). A part of this corpus forms the input, in language acquisition, to the ‘abduction’ by the individual of his individual grammar (Andersen 1973), a process whose premises are the language universals as well as utterance contexts. The corpus of present and past, realized, community grammars is an extensional set manifested in accordance with the language universals of the single individual speakers. The language universals, on

6


the other hand, ‘generate’ an intensional set of (potential) languages (including never actualized languages, and languages which may be actualized in the future, e.g. new creoles or new daughter languages of existing languages). The two contributions within this section, the one by Henning Andersen and the one by William Croft, both deal with the (biological or generalized) conception of ‘evolution’ as an explanatorily relevant framework for understanding language change, the former rejecting its relevance, the latter defining and defending it. This is the central competing issue of the volume. Andersen — in his contribution “Synchrony, diachrony, and evolution” — alluding to the titles of the central master piece of Coseriu (1957) and his seminal paper Coseriu (1968) — rejects like Coseriu himself the blind application to linguistic matters of models imported from other disciplines, especially the natural sciences, because each subject matter requires its own theory: language, being a cultural phenomenon (characterized by creativity and freedom), cannot be explained by a theory developed to explain the evolution of living species (this kind of ‘evolution’ being characterized by chance and necessity). Andersen draws upon the important distinction, introduced into historical linguistics by Coseriu (1957), of the historical/particular level (actuality), the empirical/general level (potentiality), and the theoreticl/rational level (conceptual necessity) — a distinction similar to that of the three kinds of adequacy of Chomskyan linguistics, viz. observational, descriptive, and explanatory adequacy. On the historical level, a language (speaker’s competence) is, in the synchronic dimension, a ‘practice of speaking’ (manifested in speech acts), and, in the diachronic dimension, a ‘tradition of speaking’ (manifested in transmissional links backwards and forwards in time); on this level the language historian describes actual linguistic changes. On the general level the historical linguist studies generalized change schemas (e.g. grammaticalization) and laws of change. Both the historical and the general level concern the how of language change, whereas the rational level is devoted to the why of language change, which is investigated by the theoretician of historical linguistics. Andersen stresses that linguistic change is ‘Lamarckian’, whereas biological evolution is, of course, ‘Darwinian’. Language change is governed by reanalysis (abduction) — a (fallible, but rational) logical inference, evolution by replication — a blind, irrational chance mechanism. Grammar formation is a creative activity (cf. Coseriu’s emphasis on creativity in his theory of language and language change), an example of the ‘over-performance’ of the human mind, as Andersen succinctly puts it. Only linguistic ‘evolution’ (change) is subject to the process of adaptation (found in neologisms and extensions): the interactor may create variant replicators that will fit the environment better (the motivation) and thus will get differentially replicated

INTRODUCTION

7

(i.e., adopted). Thus, crucially, innovation and adoption may be purposeful activities. Croft, for his part, in “The relevance of an evolutionary model to historical linguistics”, emphasizes that his evolutionary model of linguistic change, Theory of Utterance Selection (Croft 2000), is an instantiation of the general — and generalized — model of ‘evolution’, Generalized Analysis of Selection (GAS), mentioned above. This implies that ‘functions’ (in the AI sense of domain-neutral concepts) are utilized, that is, concepts not bound to biology — or any other discipline for that matter. It is a crucial feature of GAS that each element in it must have an instantiation in a given domain of the phenomenal world for ‘evolution’ (to be said) to exist there — and for the theory to explain it. ‘Evolution’ concerns the (temporal) relation (of copying, or ‘replication’) between a succession of originals and replicates, such that in a sequential chain the former member is the original, and the latter member is the replicate (copy). The replicate then comes to function as original, and so on and so forth. The generalized function of subject of replication — of being replicated (original/copy) — is ‘replicator’, in language: tokens of linguistic structure in utterances (‘linguemes’). It is essential that replication is (necessarily) not a truthful process — identical replication is only the limiting case. Rather, one has to recognize, on a first level, ‘altered replication’ and, on the second level, ‘differential replication’, where copies different from the originals may be produced and selected, and passed on. Evolution, then, is a twostep process, with two kinds and levels of replication, viz., firstly, as fundamental basis of evolution, the level of ‘altered replication’ where variation in a population is produced, and, secondly, as the second source of evolution, the level of ‘differential replication’, or (especially, natural) ‘selection’, where some variants (replicates) become less frequent whereas others gain in frequency, owing to ‘environmental interaction’. The subjects of interactions, or ‘interactors’, are normally organisms. Organisms form a species, i.e. a ‘population’, contracted by interbreeding (‘intercourse’) and productive isolation from other species. Interbreeding produces new organisms (interactors) and new replicators. A population of interactors is a historical entity (a token, not a type), bounded in time and space by the split of an ancestral population and a split of an offspring generation. Croft applies this general theory of selection, GAS, in an explanation of language change. He only recognizes real, historical entities (tokens), viz. community languages, e.g., Hungarian, and the psychologically real (“functional”) languages (grammars) of the language users. The linguistic instantiation of GAS is his Theory of Utterance Selection, TUS, mentioned above. TUS has the welcome result, says Croft, of avoiding false analogies and teleological mechanisms — Croft in this

8


way crucially diverging from the Coseriu–Andersen teleological theories. However, like Coseriu, Croft also wants his model to be an integrated framework for language and language change (the former he terms Radical Construction Grammar). And, likewise like Coseriu (and Jakobson), he stresses the theoretical necessity of explaining when change is absent (identical replication), not only when it is present (altered and differential replication): absence of change is conceived of as the obligation felt by the speaker to behave in conformity with convention (which solves a crucial coordination problem). The basis for the existence of linguistic ‘evolution’ is, firstly, the essential characteristic of any language, viz. innovation: the production of 5 variation, and, secondly, its propagation through the population of interactors. According to Croft, variation is functional (in the sense of making replicators fit the communicative environment better), whereas propagation is social — the linguistic instantiation of GAS thus being different from the biological one, and from Coseriu’s and Andersen’s contention, namely that the same factors are operative in innovation as well as in propagation (selection). This controversy is relevant for the next two papers. 1.3 Functional factors in ‘evolution’: Functional motivation of selection The next two papers, contributions 6 and 7, by Silvia Becerra Bascuñán and Guido Seiler, respectively, investigate the role of functional factors in linguistic ‘evolution’, the former with respect to the long-term kind termed ‘drift’ (since Sapir 1921), a theme prominent in the writings of Coseriu (Coseriu 1957, 1968) and Andersen (see the references in Andersen, this volume), the latter with respect to Croft’s theory as summarized above. Silvia Becerra Bascuñán, in her paper “Grammaticalization of indirect object cross-reference in Spanish as a case of drift”, on the basis of an extensive empirical corpus investigation of the development of cross-reference and head marking of indirect object in the grammar of Spanish, confirms the existence of the phenomenon of Drift, and of the specific theory proposed by Andersen (see the references to Andersen, this vol.), whereby drift has structure and direction. More specifically, Spanish has generalized (reanalyzed) its inherited type — Latin’s cross-reference and head marking of subject arguments — to also covering the primary grammatical relations of direct and, especially, indirect object. The evolution of cross-reference of the indirect object, i.e., the grammaticalization of the atonic dative pronouns as head 5

A feature also heavily focused upon by Coseriu, e.g., in his abstractive process from variance to invariance, from Speech to Norm, to System, to Type, to Universals (Coseriu 1952, 1957, 1968, 1974).

INTRODUCTION

9

markers of indirect object, is an almost one thousand year long process of actualization of the reanalysis (generalization), which is now, at the beginning of the 21st century, almost at its end — ‘full’ grammaticalization. The investigation of 20th century Spanish shows that the evolution is more advanced in the unmarked categories of the ‘architecture’ of the ‘historical language’: it is more advanced in the lower strata and younger generations of the population, more advanced in the more conversational, informal speech styles, and more advanced with prototypical main clauses with referential direct objects and personal, human, or animate IOs in foregrounded, empathetic functions. Her specific finding is that the instigating forces in the global evolution were local markedness shifts whereby marked categories (e.g., emphasis and empathy) came to be ‘inflated’, or ‘de-valued’. That is, there were external, pragmatic (functional) motivations behind the internal, long-term structural development. Guido Seiler, in his contribution on Prepositional Dative Marking variation in Upper German, “The role of functional factors in language change: An evolutionary approach”, pleas for the relevance of functional factors in the conception of variant selection and propagation (i.e., adoption and diffusion) —at variance with Croft’s contention, mentioned above, that propagation, or spread through communicative networks, is conditioned solely by sociolinguistic factors, not by factors of code efficiency. Seiler specifically demonstrates that innovative variants are adopted in a speaker’s functional grammar owing to their functional values — they compete for adoption, and get adopted if they are better ‘fit’ — by being better adapted. Seiler introduces the important notion of implementation for the integration of a variant in the individual’s functional language and its changing status therein. Thus, in addition to the spreading of a variant inter-individually (social propagation, in the environment of the speech community), there is its intra-individual implementation, that is, its adoption in the functional language of the individual speaker, it’s evolving systematic arrangements with other elements of the functional language. Some variants, which have been ‘implemented’ in response to functional factors, may become obligatory and thus lose their functional motivation (as was also seen in Becerra Bascuñán’s study on grammaticalization, mentioned above, where the head markers of indirect object at some time in their evolution were selected owing to their indexation of empathy, which value was lost later). In chapter 8, by Miguel Vázquez-Larruscaín, entitled “Computational modeling of prototypicality in language change: Neutralization to schwa, default logic, and the history of the German noun”, the evolution of the German noun paradigms is investigated. The development is understood as a ‘drift’

10


towards morphological prototypicality, the target. In the case of German, syntactic gender and prosodic type are the crucial determinants in the evolution of the declensional system, the initial stage being represented by a rather arbitrary system. The view is that morphology and phonology are integrated components of a grammar (as would be expected in a functional–structural theory). 1.4 Cognitive perspectives on semantico–syntactic change: Mental Grammar, Cognitive Grammar, and Dynamic Syntax The next three contributions are ‘cognitive’ in nature, however stemming from three radically different conceptions of ‘cognitive’ linguistics, viz. Mental Grammar, Cognitive Grammar, and Dynamic Syntax. Contribution 9 by Per Durst-Andersen, “From propositional syntax in Old Russian to situational syntax in Modern Russian”, is written within his framework of Mental Grammar (see the references in his contribution), which is, in fact, in the vein of the functional structuralism of Coseriu, Jakobson, and Andersen. Durst-Andersen defends a theory whereby the semantic contents of 6 lexical and grammatical categories and constructions are dual (unitary) cognitive structures, involving on the one hand a — perceptual — ‘image’ side, referring to (‘picturing’) the/a ‘world’, and a — conceptual — ‘idea’ side, representing the conceptualization or description of it: each content side of a linguistic entity is thus an ‘image-idea pair’, as he terms it. Accordingly, the content of a common noun, for instance, is a cognitive entity, which pairs a ‘prototypicalized’ figure (cf. ‘image’) and a prototypicalized description of it (‘idea’). Lexical-typologically, a language is either image-centered, as French, or idea-centered, as Danish — they specify either the image or the idea, respectively, in naming the nominal concept. Also verbs are image-idea pairs, i.e., they name verbal contents, which in the image dimension represent a figure–ground relationship — termed ‘ground situation’, and in the idea dimension its description — called a ‘ground proposition’; ‘ground’ in both contexts suggesting lexical, i.e., potential (basic) meaning. Sentential-syntactic structure — on its content side — is then either situation-centered, as in Modern Russian, or proposition-centered, as in Old Russian. The paper investi6

Notice that Durst-Andersen subscribes to the Aristotelian–von Humboldtian–Coserian– Jakobsonian conception of meaning as unitary, cf. Coseriu (1971:219–220): “Ein einheitlicher einzelsprachlicher Inhalt ist nicht etwa eine sekundär gewonnene Synthese von Bezeichnungsmöglichkeiten, sondern, umgekehrt, etwas intuitiv Primäres und sprachlich Unmittelbares [...] so ist die Konstruktion mit X nicht etwa die algebraische Summe von “Instrument”, “Stoff”, “Begleitung” usw., sondern sie entspricht einer primären und viel allgemieneren Funktion, die dann verschiedene Anwendungsmöglichkeiten zuläßt.”

INTRODUCTION

11

gates the development from Old Russian’s proposition-based syntax to Modern Russian’s situation-based syntax, a change in syntactization strategy. The emergence of the grammatical category of verbal aspect (situational) has been crucial for this development of situational syntax. (Modern) Russian is thus ‘super-typologically’ coherent: it is ‘reality-based’, both in its construal of the verbal categories (word classes and their functional categories/operators) and in its content syntax. Representing quite another conception of cognitive linguistics, Lena Ekberg in the 10th contribution to the present volume, “Construal operations in semantic change: The case of abstract nouns” gives a (Langackerian) Cognitive Grammar account of the development of abstract nouns in Swedish. She demonstrates how diverse construal operations on the underlying semantic representation of a lexical expression may generate semantic variation, either polysemy or contextual meaning variation, and this variation may lead to change. Thus change is engendered by the individual cognitive subject/speaker, but is present virtually in the underlying schematic structures and the construal operations themselves. In contribution 11 by Miriam Bouzouita & Ruth Kempson, “Clitic placement in Old and Modern Spanish: A dynamic account”, the development of clitic placement from Old Spanish to Modern Spanish is studied, with the aim of understanding or explaining, within the model of Dynamic Syntax — a dynamic model of syntax which is integrated with language use (processing), not only the synchronies of Old and Modern Spanish, but also the diachronic changes that took place between the two states. The changes are shown to be formal consequences of progressive lexical simplification, where the indirect object is taking the lead as against the direct object (by having fewer restrictions). Grammaticalization studies: Content, typology, and evolutive change The next two contributions deal with aspects of grammaticalization and grammaticalization theory. They both stress the semantic foundation of grammaticalization and typology (cf. also Heltoft, Nørgård-Sørensen & Schøsler 2005). By proposing a conception of ‘parallel grammaticalization’ based on a semantic conception of linguistic type, they could, in fact, be classified as belonging within the functional structuralist camp, prominently represented by Coseriu (cf. his principle of semanticity and his concept of the linguistic type as a functional level of relations of coherence of the grammatical systems). Lars Heltoft, in “Grammaticalisation as content reanalysis: The modal character of the Danish s-passive”, makes a case for semantics as the crucial basis of grammaticalization, showing that the grammaticalization of a 1.5

12


grammatical category, in this case passive in Danish, may involve, rather than semantic bleaching (the traditional conception of grammaticalization as a reduction phenomenon), the addition of polysemy, which may be reanalyzed as an instance of a different grammatical category, in this case mood. His specific contribution within this field of research is his explanation of the reanalysis in the history of Danish of the synthetic s-passive voice as additionally coding a ‘subjective’ mood — indicating that the evidential source of the proposition is not the speaker himself, but another intention or consciousness (the analytic, periphrastic blive-passive then involving ‘objective’ mood, indicating the speaker as source). This is evidently an example of a type-conforming grammaticalization, ‘parallel’ with the grammaticalization of the Danish subject as a modal and categorical subject (cf. Heltoft 2001) — both grammaticalizations actualize the ‘supertype’ of Danish as explicitly coding the interpersonal dimension (factors of the speaker and the addressee). In the 13th contribution, “Aspect and animacy in the history of Russian: Developing the idea of parallel grammaticalization”, Jens Nørgård-Sørensen, basing his investigation on Durst-Andersen’s theory of contentive ‘supertypology’ (for references, see Durst-Andersen, this volume), stresses the need to focus on semantically-based linguistic types in historical linguistics. More specifically, he proposes a theory of what he terms parallel grammaticalization, whereby the development of a nominal category in a language, in the present case animacy in Russian, may go in tandem with the development of a verbal category, in this case aspect in Rusian (the determinant category), owing to a common general, or abstract, motivating principle, to the effect that there has to be a correlation (homology, congruity) between the two major grammatical subsystems, the nominal and the verbal. Nørgård-Sørensen shows that Modern Russian’s aspect-based linguistic type developed out of an original type based on the aktionsart distinction between ‘action’ (as an indivisible whole) and ‘non-action’: A distinction was introduced (in the late 14th century) within (telic) actions between an ‘activity’ part and a ‘state’ part, giving rise to activity-profiling imperfective aspect and resultant state-profiling perfective aspect. The introduction of this verbal distinction was paralleled by the (presumably) simultaneous reanalysis of nouns as belonging to either the animate (cf. actor/activity in the verbal subsystem) or the inanimate (cf. nonactor/state) class, in the same way as simplex verbs (denoting non-actions) belong to either the activity or the state subclasses. 1.5 Integrating theories of language and language change The last contribution to the volume, “Towards an integrated functional– pragmatic theory of language and language change: In commemoration of Eu-

INTRODUCTION

13

genio Coseriu (1921–2002)”, by Ole Nedergaard Thomsen, concerns a further development of Eugenio Coseriu’s ‘Process’ (’ ε ν έ ργ ε ι α) theory of language and language change (1952, 1957, 1988 [1982]), and inspired by suggestions in Andersen (1989). It represents an attempt to understand (explain) language change (and acquisition) as socio-psychological ‘creation’, as constituted by meta-communicative speech acts by single individual speaker-listeners, establishing in dialogue — for each other — as perlocutionary out-come traditions of speaking, that is, dia-lects — ever-emergent linguistic conventions. Language and linguistic change are thus part-and-parcel of communication. The theory proposed is strictly deducible from speech act theory (primarily that of Searle’s) and legal philosophy (primarily the theory of directives and norms of the Danish philosopher of jurisprudence and language, Alf Ross, cf., e.g., Ross 1968). The unitary model is based on a sociopsychological conception of legislation, rather than the ‘naturalist’ model of evolution discussed in this volume. Concluding remarks It is my hope that the present volume will inspire further — however, collaborative, rather than competitive — work in the direction of an interdisciplinary and integrative (historical) linguistics that sees language in its proper perspective, as an intersection, or unity, between nature and culture, and individual mind and society, and that takes variation and change as the basic conditions of language as practice and tradition. In the words of Jakobson (1973: 52), 2.

The genetic code, the primary manifestation of life, and, on the other hand, language (the universal endowment of humanity) and its momentous leap from genetics to civilization, are the two fundamental stores of information transmitted from ancestry to progeny, the molecular heredity and the verbal legacy as a necessary prerequisite of cultural tradition.

Coseriu (1988 [1982]:149) gave a — to my mind — succinct formulation of the relation between nature and culture with respect to language: A language [...] does not exist as an object or an organism, and thus it does not have an organic continuity independent of the consciousness of its speakers. A language is an historically given “technique” of speaking: it exists only as a tradition [i.e. culture] of the ability to speak [i.e. nature], that is, as a traditional technical knowledge, or as a “competence” which has been handed down by and to the individual members of language communities. Thus, what is interpreted as “linguistic change” is not a process of change in language products [...] but rather the creation of language traditions, the historical objectivization of what has been

14


produced in speech; that is to say, nothing other than language as it is being created.

Roskilde, June 2006

Ole Nedergaard Thomsen REFERENCES

Andersen, Henning. 1973. “Abductive and deductive change”. Language 49.765–793. Andersen, Henning. 1989. “Understanding linguistic innovations”. Language Change: Contributions to the study of its causes ed. by Leiv Egil Breivik & Ernst Håkon Jahr, 5–25. Berlin & New York: Mouton de Gruyter. Coseriu, Eugenio. 1952. Sistema, norma y habla (con un resumen en alemán). Montevideo: Universidad de la Republica, Faculdad de Humanidades y Sciencias. (Reprinted in Teoría del lenguaje y lingüística general: Cinco estudios, 11–113. Madrid: Gredos, 3 rd rev. and corr. ed. 1973.) Coseriu, Eugenio. 1957. Sincronía, diacronía e historia: El problema del cambio lingüístico. Montevideo: Universidad de la Republica, Faculdad de Humanidades y Sciencias. (1974 [1957]. Synchronie, Diachronie und Geschichte: Das Problem des Sprachwandels, translated by Helga Sohre. Munich: Fink.) Coseriu, Eugenio. 1968. “Sincronía, diacronía y tipología”. Actas del XI Congreso Internacional de Lingüística y Filología Romanicas, Madrid 1965, vol. 1, 269–281. (German translation: “Synchronie, Diachronie und Typologie“. Energeia und Ergon. Sprachliche Variation — Sprachgeschichte — Sprachtypologie ed. by Jörn Albrecht, Jens Lüdtke & Harald Thun, vol. I:173–184. Tübingen: Gunter Narr, 1988.) Coseriu, Eugenio. 1971. Sprache, Strukturen und Funktionen: XII Aufsätze. Tübingen: Gunter Narr. Coseriu, Eugenio. 1974. “Les universaux linguistiques (et les autres)”. Proceedings of the 11th International Congress of Linguists ed. by Luigi Heilmann, vol. I:43–73. Bologna: Societa editrice Il Mulino. (Spanish translation: “Los universales del lenguaje (y los otros)”. Gramática, semantica, universales: Estudios de lingüística funcional, 148–205. Madrid: Gredos, 1978.) Coseriu, Eugenio. 1982. “Linguistic change does not exist”. UCLA Conference on Causality and Linguistic Change, Los Angeles, May 1982. (Repr. in: Energeia und Ergon. Sprachliche Variation — Sprachgeschichte — Sprachtypologie ed. by Jörn Albrect, Jens Lüdtke & Harald Thun, vol. I:147–157. Tübingen: Gunter Narr, 1988.) Croft, William. 2000. Explaining Language Change: An evolutionary approach. London & New York: Longman.

INTRODUCTION

15

Diderichsen, Paul. [1958] 1966a. “Udvikling og Struktur i Sprogvidenskaben”. Helhed og Struktur ed. by Paul Diderichsen, 276–307. Copenhagen: G.E.C. Gad. Diderichsen, Paul. [1959] 1966b. “Darwin og Sprogvidenskaben”. Helhed og Struktur ed. by Paul Diderichsen, 326–339. Copenhagen: G.E.C. Gad. Fortescue, Michael, Eva Skafte Jensen, Jens Erik Mogensen & Lene Schøsler, eds. 2005. Historical Linguistics 2003: Selected papers from the 16th International Conference on Historical Linguistics, Copenhagen, 11–15 August 2003. (= Current Issues of Linguistic Theory, 257.) Amsterdam & Philadelphia: John Benjamins. Heltoft, Lars. 2001. “Recasting Danish subjects. Case system, word order and subject development”. Grammatical Relations in Change ed. by Jan Terje Faarlund, 170–204. Amsterdam & Philadelphia: John Benjamins. Heltoft, Lars, Jens Nørgård-Sørensen & Lene Schøsler, eds. 2005. Grammatikalisering og struktur. Copenhagen: Museum Tusculanum Press. Hull, David L. 1988. Science as a Process: An evolutionary account of the social and conceptual development of science. Chicago: University of Chicago Press. Hull, David L. 2001. Science and Selection: Essays on biological evolution and the philosophy of science. Cambridge: Cambridge University Press. Jakobson, Roman. 1973. Main Trends in the Science of Language. London: Allen & Unwin. Janda, Richard D. & Brian D. Joseph. 2003. “On language, change, and language change — Or, Of history, linguistics, and historical linguistics”. The Handbook of Historical Linguistics ed. by Brian D. Joseph & Richard D. Janda, 3–180. Oxford & Malden, Mass.: Blackwell. Joseph, Brian D. 1992. “Diachronic explanation: Putting speakers back into the picture”. Explanation in Historical Linguistics ed. by Garry W. Davis & Gregory K. Iverson, 123–144. Amsterdam & Philadelphia: John Benjamins. Keller, Rudi. 1988. “Zu einem evolutionären Sprachbegriff”. Energeia und Ergon. Vol. II: Das sprachteoretische Denken Eugenio Coserius in der Diskussion ed. by Harald Thun, 143–158. Tübingen: Gunter Narr. Ross, Alf. 1968. Directives and Norms. London: Routledge. Schleicher, August. 1863. Die Darwinsche Theorie und die Sprachwissenschaft: Offenes Sendschreiben an Herrn Dr. Ernst Häckel, a.o. Professor der Zoologie und Director des zoologischen Museums an der Universität Jena. Weimar: Hermann Böhlau, 31 pp., 1 table (2nd and 3rd ed., 1873; repr. in Nova Acta Leopoldina N.F. 42, No.218, 377/378-393 [1975].) Schleicher, August. 1869 [1863]. Darwinism Tested by the Science of Language. Transl. from the German, with preface and additional notes, by Alexander V.W. Bikkers. London: J. C. Hotten, 69 pp., 1 table. (Repr. in Linguistics and Evolutionary Theory: Three essays by August Schleicher, Ernst Haeckel, and Wilhelm Bleek ed. by E.F.K. Koerner, 1–71. Amsterdam & Philadelphia: John Benjamins, 1983.)

THE NON-LINEAR NATURE OF DIACHRONIC CHANGE MICHAEL FORTESCUE University of Copenhagen The Whiteheadian framework The ‘Process’ view of language that I have presented in Fortescue (2001) not only concerns the level of the individual speaker/hearer, it also concerns that of the community of speakers. It is here that the various forms of diachronic change characterizing languages through time, such as phonetic weakening and strengthening, semantic expansions or contractions, grammaticalization, markedness shift and general drift, are manifest. I shall argue in the following that such processes are inherently non-linear in the same way that they are on the individual cognitive level. There are other commonalities between the two levels as well as differences: diachronic change can also be analysed in terms of ‘prehensions’, ‘eternal objects’, and even (understood rightly) ‘subjective aims’, without opting for an outright teleological view of the evolution of languages.1. The most important difference is precisely that the 1.

1

For those who find the term ‘eternal object’ difficult to swallow (and there are many!), I suggest replacing it with figura in the Hjelmslevian sense. They are only ‘eternal’ in the sense of being repeatable and therefore communicable. ‘New’ eternal objects are emerging all the time from actual occasions of communication, only some of which are repeated and maintained long enough — and by a stable and sizeable enough portion of a speech community — to become absorbed as conventionalized elements in the complex eternal object that constitutes the individual human language concerned. As regards the term ‘prehension’, this can be replaced if you wish by ‘reaction to’. Here, in a nutshell is the essence of Whitehead’s philosophy of ‘organism’: the basic elements in it are self-organizing ‘actual occasions’, which each ‘enjoy’ one quantum of time. These ‘drops of experience’ come together from the data of their immediate past, then ‘perish’ at once to become objective data for subsequent occasions. They consist of processes of ‘concrescence’, whereby the inherited data are grasped by ‘prehensions’ (or ‘feelings’) and successively integrated according to a ‘subjective aim’. This aim is partly determined by the occasion’s perspective on relevant ‘ eternal objects’ (‘forms of definiteness’) and partly by the nature of the data itself. It strives towards the ‘satisfaction’ of the occasion (the achievement of maximal unity and intensity of integration of its prehensions) and grows more differentiated and explicit as the concrescence proceeds. I list in the Appendix

18

MICHAEL FORTESCUE

prehensions relevant to the higher level have little to do with the teleological purposes of the individual speaker/hearer. The ‘subjects’ of such prehensions must be whole communities of speakers; they are collective and thus only indirectly reflect the communicative intentions of individual speakers. There is from a Whiteheadian perspective no problem in speaking of the collective ‘conceptual’ prehension by a community of speakers of the complex eternal object defining the language that its speakers have in common, nor of the collective ‘physical’ prehension by a community of socio-historical variation, i.e. those sociolinguistic, dialectal, contact-induced, and language-external cultural factors manifest in the linguistic productions of both influential and peripheral members from its own ranks that have a bearing on diachronic 2 change. ‘Prehension of’ does not presuppose individual ‘consciousness of’. Communities — as nexūs — display their own patterns of change, related to the processes within individual speaker/hearers in a way similar to the way the teleological linguistic behaviour of the individual is in turn related to the collective behaviour of neurons at a still lower level — that is, in a 3 relationship of emergence. At the interface between the collective and the individual nexus levels — where language acquisition is played out — there is a continual interplay back and forth, as every new speaker learns to apply the eternal objects maintained at the collective level (extractable from input-incontext) to his or her own communicative needs, meeting and adapting to ever new social circumstances. As is well known, variation in the details of this ‘match’ helps fuel diachronic change at the collective level. More specifically, one might suggest that the ubiquitous asymmetry of ‘markedness’ displayed by languages and by their diachronic pathways of change indirectly reflects the the types of eternal object and prehensions relevant (as I see it) to discussions of language. 2 It is misleading, however, to talk in this context of the ‘evolution’ of individual languages (thus harking back to the romantic notion of the language as organism). Although, as Croft has shown, one can indeed argue for a certain analogy with Darwinian evolution (Croft 2000), there is no selection process involved between eternal objects (languages) as such, only between societies (communities) in which they find ingression. ‘English’ is only an abstraction from the varieties of linguistic behaviour of communities of English speakers. One may talk metaphorically of selection amongst token utterances in the manner of Croft, but this mixes society levels — that of token exchanges between individuals and that of the abstracted types that make up the shared object ‘English’ at the community level. 3 From the point of view of the individual neuron one might say (to stretch the analogy a bit) that overall changes in its environment are also of an inscrutable ‘hidden hand’ nature vis-à-vis its own internal ‘subjective aim’ (namely to fire when the necessary input conditions are met). Processes occur within nexū s that in turn constrain their realization. Nexūs are not themselves fixed and repeatable (only the eternal objects defining their structure are that), but their levels (types) must be distinct in terms of the relationship of their constituent occasions to each other and to the relevant eternal objects — including types of subjective aim.

THE NON-LINEAR NATURE OF DIACHRONIC CHANGE

19

prehensions of individual speaker/hearers, where every positive prehension — for instance of a word — is associated with a negative one heightening its contrast with ‘what it is not’. This may in turn reflect the organization of cortical networks of neural columns with their alternating layers of excitation and inhibition (cf. Calvin 1996). The basic unit of process, the ‘concrescence’ (made up of successive prehensions), is the same at all levels, although the levels themselves cannot be reduced from higher to lower or vice versa.4 I shall return below to the question I have sidestepped here as to whether the natural unit of such higher-level concrescences is simply a higher level ‘actual occasion’ or something else. What might it mean to say that diachronic linguistic processes are essentially non-linear, just as mental processes in the minds of individual speakers are? It would surely mean that the ‘input’ to such processes is by necessity multiple and to some degree indeterminate, though the output (the change) is determinate and of a single, known type. This is of course the way historical linguists often envisage linguistic change: the idea of multiple factors mutually enhancing or inhibiting each other and of the general non-predictability of individual linguistic change (despite widespread tendencies and drifts) is not at all unorthodox today. However, just as ‘Pattern’ and ‘Process’ are all too often confused in synchronic analyses, there is often a confusion between the Pattern and the Process side of diachronic change such that regular, stepwise changes in a language system as such are assumed to be linked directly to similar changes in individual speakers in the community concerned. This caveat does not mask a reversion to a Saussurian exaggeration of the mutual independence of synchrony and diachrony — on the contrary, Pattern is obviously the result of Process and Process is constrained by Pattern. Specifically as regards the question of the non-linearity of diachronic change, I would simply go further than most historical linguists (though no further than 5 emergentists like Hopper ) and claim that all linguistic changes have multiple 4

Note that in the higher (‘intellectual’) kinds of prehension that typify the individual speaker/hearer further superimposed levels of societal embedding may be drawn upon or absorbed, heightening the contrast still further (compare Andersen 2001:47 on the hierarchical relationship manifest in diachrony between different kinds of opposition on the scale exclusive > inclusive > contrary > contradictory > converse). ‘Actualization’ of a new linguistic pattern in diachrony can in general be equated with the emergence (by ‘conceptual reversion’) of new combinations of eternal objects within the overall complex eternal object (e.g. ‘English’) that finds ingression in the community of its users. 5 Hopper has recently applied his conception of emergentism — originally applied to the pathways of grammaticalization — to the origin of lexical items, whereby continual erosion of form is balanced by processes of renewal ‘in response to the collective needs of discourse’. A specific example he provides (Hopper 1990:153) is that of the ‘demorphologization’ of class

20

MICHAEL FORTESCUE

causes, although only one — or even none — may appear immediately relevant (e.g., a simple unconditioned change regarded as the result of a purely chance ‘mutation’ of norm). By this I do not mean just that it is whole (sub)systems that change (as in the manner of system-internal pull- or pushchains), but that any change in system is going to be linked — however indirectly — to multiple causes within the community of speakers in which it comes to be accepted. The ‘actualization’ of such a unitary system change is typically ‘messy’ as regards its extension in time and space as the innovation spreads through the community. In good Whiteheadian fashion, I am suggesting here a ‘both/and’ perspective on diachronic change, one that links the systematic, step-wise shift from one system stage to the next to the ‘chaotic’ diffusion of the innovation through a community of individual speakers. This latter process is itself largely ‘indeterminate’, both as regards the distribution of an innovation in the community of speakers and as regards the overlapping registers and preferences and habits that cause the change to actualize to a variable degree in the usage of individual speakers. In fact the very notion of a community of speakers is problematic and indeterminate ‘around the edges’ — one need think only of the phenomenon of bilingualism and the ‘belonging’ of one and the same individual to distinct or overlapping communities. Yet diachronic change does in hindsight appear to display a high degree of regularity and ‘motivation’. How is this to be reconciled with the truism (as it is seen today) that all diachronic change emerges from a ‘pool’ of synchronic variation? The integration into the common language code of the community of some of the potential contrasts within that variation involves piecemeal and largely unconscious processes of diffusion and gradual ‘codification’. Such consensual (if tacit) codification — however unstable through time — must continually result for the language to be passed on from generation to generation. That ‘what is passed on’ is subject to disturbances to purely ‘internal’ genetic inheritance prefixes in the Australian language Olgolo, where these items — originally articles — have been absorbed into the lexical substance of the nouns to which they attach to compensate (to some degree) for the loss of initial consonants in that language, thus restoring more normal CV syllable phonotactics and, in the process, gradually losing their distinct semantic function. The causes of such ‘hidden hand’ developments can be said to be by necessity multiple, since neither the phonological erosion of initial consonants nor the attachment of classificatory prefixes on their own is a sufficient cause of the change — only their conjunction (within the overall ‘type’ of the language concerned) can be said to be that. One could also mention Bates & MacWhinney’s (1987) competition model: for them grammaticalization is an adaptive process, the on-going result of continual competition between rival factors and the emergence of ‘new’ solutions whereby expressability may be increased through the compacting and automaticization of old ones.


21

caused by contact with other communities of speakers is also a commonplace — convergence between neighbouring languages may arise through the primary mechanism of bilingualism, considerably complicating historical reconstruction. But is it always necessary to resort to convergence to explain ‘obscured’ reconstruction situations? Might the inherent grain of ‘chaos’ in all diachronic processes based on the social utilization of a fluctuating pool of variation — including partial changes checked and/or abandoned — not be sufficient to render reconstruction indeterminate beyond a certain time depth without invoking community-external input at all? Yet certain specific changes (rather than others) occur again and again, independently across the world’s languages, given analogous starting points, and related languages may display (to some degree) predictable internal ‘drift’ reflecting their common origins as they split into dialects and then separate languages. How can these two aspects — on the one hand the unpredictable ‘chaotic’ side of diachrony and on the other its predictable, generalizable side — both be accommodated within one framework and related in turn to cognitive processes within individual speakers? Examples of non-linear diachronic change Before sketching a Whiteheadian answer to these far-reaching questions, let me dwell for a while on some specific examples of diachronic change that appear to entail non-linear processes, for example vowel harmony in Uralic and Paleosiberian languages. Perhaps the most problematical area for reconstruction in the Uralic family is the nature of its vowel correspondences (cf. Abondolo 1998:16ff., especially as regards the controversial issue of vowel quantity in the proto-language), and a recent book (Marcantonio 2002) has in fact claimed that the persisting problems in this area demand that the whole 6 notion of the unity of this assumed family should be reassessed. Palatal vowel harmony characterizes the whole family (though a few modern languages have lost it again), and it is assigned by Janhunen (1982:26) to the proto-language. This has complicated the task of reconstruction (it often not being possible to determine the original quality of the second syllable of stems), but it is widely accepted that the direction of assimilation-at-a-distance involved here has generally been progressive, that is that the harmony of the whole word, front or back, has always been determined primarily by the vowel of the first syllable of the stem. On top of this, ‘vertical’ high/low harmony was apparently 2.

6

She prefers a model of waves of innovation spreading across the whole Uralic — and neighbouring — areas to a traditional Stammbaum one. The source of individual innovations could have been either internal or external to the family. Most Uralicists, it is probably true to say, regard this viewpoint as overstated, the great time depth involved being the major source of the difficulty of reconstruction rather.

22

MICHAEL FORTESCUE

involved at the proto-stage as regards the choice of (a reduced array of) second or third syllable vowels, and at least one full-scale circular shift involving the whole vowel system is known to have affected the Permic languages in particular, seriously disturbing the original vowel harmony situation there. Regressive rather than progressive assimilation has also been adduced to explain specific correlations en route between proto-Uralic and the various lower branches of the family, for instance the change of proto-Uralic */u/ to protoSamoyedic // before a second syllable //, of proto-Ugric */i/ to proto-ObUgric /o/ before second syllable /å/, and of proto-Finno-Permic */i/ to protoPermic /e/ in open syllables before second syllable /ä/ (Sammallahti 1988). In Estonian central // probably arose this way too, namely from mid vowel /e/ (short or long) before a back vowel (except when in absolute initial position), as in mk < *meekka “sword”, vras < *veeras “foreign”, vrkk < *verkko “net”, and prn < *perna “spleen”. It may well be that this factor — regressive assimilation — has been underestimated in reconstructing the vowel system of proto-Uralic (PU) itself. Thus the illabial back and/or low vowel */ï/ reconstructed by Janhunen (for which Rédei (1988) has */a/ rather) seems to be required by certain Ob-Ugric and Samoyedic correlates with Finno-Permic /a/, but may actually represent the result of regressive assimilation (as in hypothetical *appï “father (-in-law)” corresponding to proto-Samoyedic pand hence reconstructed by Janhunen as *ïppï). This direction of development is in fact the most likely on typological grounds, going as it does from an unmarked segment (/a/) to a marked one (/ï/) in a specific marked environment (rather like umlaut in Germanic).7 In the vowel harmony system of Chukotko-Kamchatkan languages, observe, the assimilation goes both ways quite consistently (though the system is not of the same ‘palatal’ sort), i.e. is bidirectional, and bidirectional adjustment of sequences of stem syllables may also have been involved in the early prehistory of Uralic. Bergsland’s reconstruction of proto-Eskimo-Aleut, which recognized patterns of assimilation — both progressive and regressive — of certain successive vowel pairs in Aleut compared with Eskimo, solved a particularly recalcitrant problem of reconstruction in a language family that does not even display regular vowel harmony (Bergsland 1986:28ff.). Thus, for instance, the sequence */u/ plus /a/ often goes to /a/ plus /a/, as in tana-X “land” < *nuna. In 7

Marcantonio concurs — she also points out (2002:97f.) that only one of Janhunen’s sets actually covers all Ob-Ugric and Sammoyedic and that there are reflections of both vowels in dialect forms in all these languages (Hungarian goes mainly with Finno-Permic rather). Note also that Sammallahti (following Janhunen) does not have any PU stems reconstructed with initial /a/ (or /å/) at all.


23

effect, any vowel in Aleut can correspond to any vowel in Eskimo (although one-to-one full vowel correspondences probably dominate statistically). The precise conditioning factors are by no means clear, yet the reconstruction of the proto-system (which was much like that of conservative Eskimo languages today) is not in doubt and is generally accepted by Eskimologists. I have not mentioned contributing external factors yet, but these may well have been important in producing the overall picture of Uralic vowel correspondences. Turkic incursions from the south along the Ob and Yeniseian valleys, for example, may have left traces in the phonology as well as in the 8 lexicon of the Ob-Ugric and Samoyedic branches of the family. Purely internal convergence between these two branches of Uralic appears to have further complicated the picture, with innovations diffusing across the border between them. It can be argued that such contact phenomena resulting ultimately in convergence always proceed in a non-linear/chaotic manner. This would at all events appear to be the case in Australia, where the mesh-like relationship amongst all its native languages has prompted a radical departure from the traditional Stammbaum family tree picture of language relations on the part of Dixon (1997). His ‘punctuated equilibrium’ model of alternating convergence and ‘punctuation’ episodes completely undermines attempts to reconstruct unitary proto-languages. The results are not only unpredictable but also unravelable beyond a certain time depth. More recently this model has been applied to Indo-European by Renfrew (1999), who sees a convergence episode in the Balkans intermediate to the spread of agriculture into Europe from Asia Minor and the formation of the individual Indo-European families (which involved more local convergence processes via language shift in peripheral areas of the continent). Thomason & Kaufman (1988) have shown in their case studies of the historical interaction between English and French and between Russian and Uralic just how complex the results of such interaction can be.9 I have argued myself for a mesh-like situation in Mesolithic Siberia (Fortescue 1998). Often 8

In fact the hypothetical */a/ > /ï/ change affecting especially Samoyedic and Ob-Ugric mentioned above could have had such a source outside the family, perhaps in Turkic, where the same alternation is found today in certain dialects adjacent to Uralic. 9 According to Thomason & Kaufman (1988) language shift must take place relatively rapidly if it is to produce marked substratum effects (via imperfect learning), so at least some kinds of convergence episodes must — pace Dixon — be short-lived, and, conversely, I suspect, some punctuation episodes can continue undisturbed for a long time. This is indeed the conclusion of Watkins (2001). It is worth bearing this in mind when attempting to distinguish loans from common inherited lexicon — contrast for instance Germanic (including Scandinavian) loans in Balto-Finnish, which are perfectly integrated into Finnic phonology, with Koryak ones in Itelmen, where ‘new’ phonemes (or old ones in new distributions) occur, reflecting the relative recentness (and speed) of intense contact there.

24

MICHAEL FORTESCUE

the diachronic changes relatable to such episodes of contact and convergence suggest multiple causes (a point emphasized by Thomason & Kaufman 1988: 57), internal as well as external, the two reinforcing each other, as in the case of Slavic palatalization, which seems to combine the natural tendency to phonetic palatalization of alveolars before front vowels with areal/substratum influence from neighbouring Uralic languages in which this trait is widespread.10 Even more complex causal scenarios have been suggested for some widespread areal phenomena, such as that of the diffusion of a whole set of parallel polysemic functions of one or more morphemes originally meaning “acquire” in South-East Asia to potentiality or manner complement constructions. Enfield (2001:280ff.) suggests that this might involve both genetic and areal factors, but also ‘conceptual naturalness’, for example by reflecting common grammaticalization pathways. All these factors may have conspired in extending the distribution of the morphemes concerned (and their calques). The point of all this is that the factors behind individual changes — both phonological and semantic, sociological and typological — are more often than not multiple. They may moreover affect larger units than individual phonemes or morphemes — for instance whole syllabic sequences in phonological changes, where the changing unit may well be disyllabic, a potential change in a single segment being coupled to the whole embedding syllable and its immediate context. In semantico-syntactic change whole constructions may be involved, that is depend both on the semantics of individual ingredient morphemes and on the broader morphosyntactic possibilities of the language (and/or a surrounding linguistic area). In other words, such changes reveal the essential non-linearity of diachronic processes. Types of prehension involved in diachronic change As I began by stating, it can be argued that some of the same basic processes as at the individual cognitive level are involved also in diachrony, namely physical and conceptual prehensions; only the ‘societal’ level is different here, namely the level of the community of speakers. Physical prehensions at this level would be external factors affecting the individual language, i.e. 3.

10

A purely language-internal phonological change can also have multiple causes — this I suspect is the case with, for example, the fact that the Mandarin negative morpheme ba- has a falling tone (if its tone is not lost entirely) before syllables with any tone except the falling (‘first’) tone, before which it becomes rising. This seems to combine the natural phonetic tendency for a neutral or unstressed syllable to rise somewhat preceding a relatively high pitch (or one starting high) with a factor of iconicity: the tone of the negative morpheme will always move in the opposite direction (or be followed by a change in direction) to that of the following verbal morpheme.


25

historical, social and technological changes and conflicts, whereas conceptual prehensions would be internal ones of natural ‘drift’, the system aligning to universals (certain species of ‘eternal object’) that reflect — if only indirectly — the structure and teleology of human organisms. Communities do not, on the other hand, collectively recognize or imagine propositions and judgments of a higher order of complexity since these require the internal organization of individual speaker/hearers both to produce and to comprehend — they specifically require the ‘indicative’ prehension of logical subjects of propositions, and communities do not produce collective utterances about concrete nexūs (except perhaps via the projections of individual propagandists or demagogues). One may nevertheless question whether the unit of such processes is the concrescence of a single ‘actual occasion’, the concept lying at the core of the Whiteheadian framework — as regards the level of the individual self-organizing organism at least. Would an actual occasion at the level of the community display a ‘subjective aim’ for instance? If so, it would have to be — from from the perspective of the individual speaker — of the ‘hidden hand’ sort, much like the causal nexus behind the organized activity of ant colonies (see Keller 1999). Such an unconscious ‘subjective aim’ might in essence be no different from that of concrescences at any nexus level, i.e. be manifest as a general striving towards a state of maximal intensity of contrasts (‘balanced harmony’ as Whitehead calls it). This could be interpreted as the basic driving force behind ‘drift’, drawing language systems towards a state of maximal economy and expressiveness in the integration of the multiple strands of their historical input ‘data’ (and thus enabling optimal inter-communication amongst their constituent speakers); it may or may not be mediated by a constant pull towards ‘type’ (in the manner of Coseriu) or towards reduction of markedness in 11 the manner of Andersen (2001). Given the complexity and multiple compet11

Actually I would prefer to describe this in terms of the general ‘preparedness’ of a given language for a change of a certain type as opposed to another possible one, depending both on its prior history and on its relevant synchronic patterning. Stages in known grammaticalization routes — such as that leading from lexeme to adposition to case affix — may thus be “jumped over” or another turning altogether taken at a certain point in such a chain. Take for instance the fact that there is no sign in Mandarin Chinese of adpositions (from verbs) going further to case affixes since affixation as such is not a feature of Chinese — instead it appears to be developing applicative constructions (a head-marking pattern) with these morphemes, parallel with its common compound verb construction. Thus compare: (i)

tā gěi wo măi yibĕ n shū he give me buy one.CLASS book “He bought a book for me.”

26

MICHAEL FORTESCUE

ing functionality of language this goal can never be fully or permanently reached but only endlessly strived after. One might even talk of the collective ‘subjective form’ of a macro-concrescence at this level; this further Whiteheadian term, referring to the ‘how’ of an actual occasion’s self-organization, would provide an overall cultural dimension to the evolution of the individual language in the manner of Humboldt. If one were to accept the reality of such macro-units, however, they would have to be acknowledged to be both more loosely and more abstractly organized than the actual occasion at the level of the individual organism. Note that Whitehead himself distinguished ‘higher’ and ‘lower’ grades of actual occasions according to the societies/nexūs in which they are embedded, but ‘higher’ refers here to a specific kind of (potentially conscious) complexity of organization, not to scale. The four kinds of actual occasion he mentions are: those in empty space, non-living occasions, living occasions, and conscious ones (Whitehead 1978: 177ff.). Communities 12 of speakers correspond to none of these. For Whitehead they are nexūs, not (ii)

wǒmài-gei tā sān běn I sell-give him three CLASS “I sold him three books.”

shū book

In Chukchi, on the other hand, a language making widespread use of incorporating structures and of case affixes both on nouns and on verbs (forming gerundives), certain verbal stems have gone directly to form new case affixes without traversing a stage of adpositions. Thus kopragt() “to the net”, from stem kupre- “net” plus jt- “go for” (plus an old dominant vowel harmony allative case marker - that only surfaces today in one dialect) — identical historically with “going for the net”. This is despite the fact that the language does have a few postpositions (following locative case-marked nominals). 12 The strictly ‘monadic’ or ‘quantumized’ interpretation of Whitehead’s philosophy as being based on only one kind of actual occasion operative on all levels of nexus (as assumed by Lucas 1989:160f.), is surely untenable. It is certainly inconsistent with a ‘wheels-withinwheels’ view of the complexity of biological and cognitive life — organisms consist after all of neurons and these of molecules composed of atoms, etc., each level typified by its own kinds, scales and speeds of concrescent processes, ranging from the minimal flash of the unstable atomic particle’s lifeline to — arguably — the single concrescent unfolding of the entire cosmos, all functioning ‘in parallel’ at any one moment in time. Nevertheless, it is easy to see why the later Whitehead (after the introduction of his ‘epochal’ theory of time in Process and Reality) might be interpreted this way, since he focussed largely on the individual organism (for similar reasons he is often accused — unjustly — of ‘panpsychism’). He stressed on the one hand that it is the organization of the nexus itself that determines the mode of unfolding of actual occasions at that level and, on the other, that there is a certain basic design to actual occasions at all levels (they all involve prehensions, subjective aims, satisfaction, etc.). But he also insisted that actual occasions have their ‘grades’ and differing internal structure (op. cit.:18f.). They differ in particular as regards the relationship between their ‘mental’ and ‘physical’ poles (bear in mind that the former concerns the prehension of — or reaction to — pure form/pattern, and has nothing to do with consciousness). One might


27

the actual occasions constituting them — he refers, for example, to the idealized nexus of ‘all speakers of Greek’, held together and defined by the unifying eternal object ‘Greek’. A hypothetical actual occasion on the level of the community of speakers would — as on any other level — have to be defined by multiple input and single output conditions (with no further external input occurring between the two), i.e. be defined in terms of a unitary change in the system. It would also have its own conceptual pole, where it prehends the complex eternal object of the language concerned, and a physical pole, namely the nexus of its constituent speakers. It is the shared eternal object that defines the nexus. Again as on the level of the individual speaker/hearer, such an actual occasion would fill an indeterminate duration in time. Any change in the system takes time to spread throughout the nexus, but the internal stages of a unitary concrescence towards its ‘satisfaction’ (the change in the system) could nevertheless be stated logically. A single phonological or semantic change may also interact with others, and polysemy is ubiquitous, as is dialectal and individual variation. Changes tend to spread through the nexus of speakers at varying rates, with a creative ‘cutting edge’ of innovation introduced by creative individuals or groups and not yet fully accepted (or encoded ‘emically’) by all members of the community. In fact, the precise delimitation of the speech community itself is, as I have claimed above, also to some degree indeterminate. If the resultant changes themselves can only be defined in terms of the system, i.e. the relevant eternal object, how can this Whiteheadian perspective on diachronic change be brought into line with diffusion and ‘wave theory’? The only way available is apparently to accept that the eternal object (e.g. some standard form of English) remains constant between one quantum leap and the next but that the nexus in which it finds ingression varies continuously over time as a wider and wider range of speakers adopts successive spreading changes. So while an actual occasion at the level of the individual speaker is determined by its physical pole conditions of input and the various subsidiary eternal objects drawn into its concrescence via prehensions at its conceptual pole, at the level of the community of speakers the unitary occasion would be determined by a single conceptual pole change associated with a varying extent of anchoring at its physical pole (the nexus of its speakers) as it works out its own ‘actualization’. But this contrasts both with smaller scale (e.g. neural or atomic) levels and with still broader scale physical world levels, since an actual occasion here would be at a higher level of abstraction than occasions on either the smaller or larger scales. It would only be ‘virtual’, the indirect result of the envisage the recurring distribution of this ‘ground plan’ of actual occasions across different grains or levels of reality as analogous with fractals (self-similarity ‘all the way down’).

28

MICHAEL FORTESCUE

communicative activity of individual speakers. My conclusion is in fact that the notion ‘virtual (rather than actual) occasion’ is precisely what is needed to fill out the Whiteheadian scheme of things at this point and provide a conceptual handle on the holistic, non-linear 13 nature of language change. Crucially, it does not refer to self-organizing organisms, which is what the term ‘actual occasion’ on the level of living beings refers to, but to something more abstract (on the other hand, it does share something with Whitehead’s two lower grades of actual occasion listed above — ‘abstract’ should not be confused with ‘higher’ in this connection). Compare Coseriu’s term ‘virtual language’ (virtuelle Sprache) applying to the psychological aspects of language process (Coseriu 1975:82). This term is not quite the same as what I mean by ‘virtual occasion’ at the level of the community of speakers, since it also covers individual mental activity by a speaker, but it nevertheless points in the same direction.14 One may thus say that an eternal object like English changes through time via successive ‘virtual occasions’, consisting in turn of ‘virtual prehensions’ and ‘virtual subjective aims’. For the historical linguist (as pragmatic realist) such virtual processes are perfectly real. Since the prehensions concerned are both of the eternal object (Coseriu’s ‘System’) and of the nexus of speakers, this is quite compatible with Coseriu’s claim that diachronic change takes place via the ‘Norm’ of the society speaking the language rather than directly via its more abstract and idealized functional ‘System’ — in other words by individual speakers choosing from a pool of available synchronic variation just those variants that are given by society as most normative in that time and situation (or by their choosing to ignore them). One can regard ‘Norm’ as the interface between System and individual usage (habla/parole), and the notion of ‘virtual occasion’ precisely integrates the two levels. Norm is an ‘unstable state of equilibrium of the System’ (Coseriu 1975:97) and a change of Norm is usually gradual, spreading 13

One might consider the possibility also of still broader scale ‘virtual occasions’ beyond those defining individual diachronic changes. What, for instance, is the status of long-term changes like the alternating convergence and punctuation episodes described by Dixon? What of the changing relation of language systems to general ‘type(s)’? There is in principle no problem here within the Whiteheadian scheme of things since ‘eternal objects’ may refer to/find ingression in any scale of actuality. There are in theory eternal objects defining types of change at any level, from the microscopic to the macrocosmic. Eternal objects can be said to be hierarchically organized like Chinese boxes — the difficulty here is merely terminological, since this intrinsic ordering does not really belong to the eternal objects themselves (as Whitehead put it), but to the real world of nexūs of concrescent occasions they reflect. 14 For Coseriu langue in both its ‘social’ and its ‘individual’ aspects is ‘virtual’, in so far as it is purely ‘mental’ — as opposed to ‘impulses to speak’ and their realization in concrete speech acts and their enduring products.


29

and overlapping. But from the point of view of System, change can still be step-wise; indeed it must be so since it is the eternal object of the System at any point in time that defines the extent of the community forming its physical basis. Diachronic change starts with change spreading among individual speakers but ends with a change to the eternal object itself. Multiple changes (often competing and criss-crossing) may be sweeping through the community of speakers and affecting in turn the complex layering of Norm, but the result to System (the ‘satisfaction’ of each successive virtual occasion) is a unitary adjustment of its constituent contrasts, either pertaining to single segments, to larger structural units, or to whole subsidiary systems. This may not be quite how Whitehead himself would have seen it, but recall that for him eternal objects are pure potentiality, emerging from actual occasions and their mutual relationships, so they represent already the locus of all ‘virtual’ reality. It is at all events on this abstract plane of eternal objects, of pure Pattern, that the link between the processes of individual cognition and those of diachronic change is to be found, namely in the concept of the ’virtual occasion’, abstracted from (but structurally reflecting) ‘actual occasions’ of individual linguistic activity. REFERENCES Abondolo, Daniel. 1998. The Uralic Languages. London: Routledge. Andersen, Henning, ed. 2001. Actualization: Linguistic change in progress. Amsterdam & Philadelphia: John Benjamins. Bates, Elizabeth & Brian MacWhinney. 1987. “A functionalist approach to the acquisition of grammar”. Functionalism in Linguistics ed. by René Dirven & Vilém Fried, 209–264. Amsterdam & Philadelphia: John Benjamins. Bergsland, Knut. 1986. “Comparative Eskimo-Aleut phonology and lexicon”. Journal de la Société Finno-Ougrienne 80.63–137. Calvin, William H. 1996. The Cerebral Code. Cambridge, Mass. & London: A Bradford Book (The MIT Press). Coseriu, Eugenio. 1975. Sprachtheorie und allgemeine Sprachwissenschaft. Munich: Wilhelm Fink Verlag. (Orig. ed.: Teoría del lenguaje y lingüística general. Madrid: Gredos, 1962.) Croft, William. 2000. Explaining Language Change: An evolutionary approach. Harlow, Essex: Longman. Dixon, R.M.W. 1997. The Rise and Fall of Languages. Cambridge & New York: Cambridge University Press. Enfield, Nick J. 2001. “On genetic and areal linguistics in mainland South-East Asia: parallel polyfunctionality of ‘acquire’”. Areal Diffusion and Genetic Inheritance ed. by A.Y. Aikhenvald & R.M.W. Dixon, 255–290. Oxford: Oxford University Press.

30

MICHAEL FORTESCUE

Fortescue, Michael. 1998. Language Relations across Bering Strait. London & New York: Cassell Academic. Fortescue, Michael. 2001. Pattern and Process: A Whiteheadian perspective on linguistics. Amsterdam & Philadelphia: John Benjamins. Hopper, Paul J.1990. “Where do words come from?”. Studies in Typology and Diachrony for Joseph H. Greenberg ed. by William Croft, Keith Denning & Suzanne Kemmer, 151–160. Amsterdam & Philadelphia: John Benjamins. Janhunen, Juha. 1982. “On the structure of Proto-Uralic”. Finnisch-Ugrische Forschungen 44.23–42. Keller, Rudi. 1990. Sprachwandel: Von der unsichtbaren Hand in der Sprache. Tübingen: Francke. Lucas, George R. 1989. The Rehabilitation of Whitehead. Albany: State University of New York Press. Marcantonio, Angela. 2002. The Uralic Language Family. (= Publications of the Philological Society, 35). Oxford: Blackwell. Rédei, Karolj. 1988. Uralisches Etymologisches Wörterbuch. Wiesbaden: Harrassowitz. Renfrew, Colin. 1999. “Time depth, convergence theory, and innovation in Proto-Indo-European: ‘Old Europe’ as a PIE linguistic area”. The Journal of Indo-European Studies 27:3/4.257–293 Sammallahti, Pekka. 1988. “Historical phonology of the Uralic languages”. The Uralic Languages ed. by Denis Sinor, 478–535. Leiden & New York: Brill. Thomason, Sarah Grey & Terrence Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Watkins, Calvert. 2001. “An Indo-European linguistic area: ancient America”. Areal Diffusion and Genetic Inheritance ed. by Alexandra Y. Aikhenvald & R.M.W. Dixon, 44–63. Oxford: Oxford University Press. Whitehead, Alfred North. 1978. Process and Reality: An essay in cosmology (corrected ed.). New York: The Free Press. (Original ed. New York: Macmillan, 1929.)


APPENDIX Species of eternal object (‘figurae’) relevant to language as Pattern A. Concepts i. predicates (of activity, state, quality, etc.) ii. logical subject types B. Relational/structural patterns i. morphosyntactic ii. phonological iii. (co)referential iv. semantic C. Types of subjective form i. discourse modulation functions ii. epistemic stance iii. validatory attitudes D. Intentional types i. discourse acts ii. strategies (and patterns of inference) iii. behavioural norms iv. background communicative principles E. Nexus types i. contexts of situation ii. social/historical embedding iii. signs (word types, constructions, etc.) Types of prehension involved in language as Process 1. Physical 2. Conceptual 3. Hybrid 4. Indicative 5. Propositional 6. Intuitive judgments 7. Anticipatory

31

EXPLANATIONS, OR ...? SOME METATHEORETICAL REFLECTIONS ON A PREVALENT TRADITION WITHIN HISTORICAL LINGUISTICS

BRIT MÆHLUM Universitetet i Trondheim Introduction This article’s point of departure is the so-called ‘Scandinavian vowel shift’, i.e. the qualitative changes that went on within the row of long, back vowels in the Norwegian and Swedish region after the Old Norse period. My real agenda however, is located within the area of philosophy of science. While referring to the more or less codified explanatory principles behind the Scandinavian sound change, I will analyse certain aspects of the dominant research strategies within historical linguistics: which are the prevalent explanations within this scientific tradition, and why have these principles become the authoritative ones? One of the essential questions is therefore, where did the scientific ideals derive from that have been so successful and influential within the discipline? And have these ideals been sufficient in giving us the answers we have been asking for, basically, why did the changes take place? Let us have a brief look at the changes that restructured the long, back vowels in the Norwegian and Swedish but not the Danish area in the post-Old Norse period; see Figure 1. 1.

FRONT

BACK

FRONT

CENTRE

BACK

/i:/

/y:/

/u:/

/i:/

/y:/

/:/

/u:/

/e:/

/ø:/

/o:/

/e:/

/ø:/

/:/

/æ:/

/æ:/

a: Old Norse long vowels

/o:/

b: Post-Old Norse long vowels

Figure 1: Post Old-Norse ‘vowel shift’.

34

BRIT MÆHLUM

Figure 1 shows the vowel system before (a) and after (b) the ‘vowel shift’ had been working. We can see there was an internal movement affecting the back vowel qualities, with the result that some phonemes acquired a ”higher” quality, while one of them — the Old Norse — became more fronted and got into the new central position. The dominating and, to a certain extent, authoritative explanation of these changes over the last half century has been André Martinet’s (1952) principle of phonological ‘chain movements’. A chain movement refers to a situation where a series of sound changes takes place, each one influencing the next. Perhaps the most famous incident of such a process is ‘The Great Vowel Shift’ in English, which is often cited as a classic example of a chain movement. Regarding the Scandinavian vowel shift, there has been controversy as to whether the successive movement in the vowel qualities was a push-chain or a drag-chain process, respectively (see, e.g. Eliasson 1983; Torp 1982:55ff.). This concerns where in the chain the changes were initiated and the direction of the subsequent movements. In a push-chain the whole process is assumed to begin at the bottom or back end of the chain, and each sound then “pushes” its closest neighbour out of place. Conversely, in a drag-chain, the first movement is assumed to start at the top or front end of the chain, and empty slots are left in the chain, which other sounds move up to fill. The arguments in favour of one or the other of these principles related to the Scandinavian vowel shift are of little relevance for my purpose here. The important aspect in this context is the theoretical presuppositions underlying these modes of explanation. And within such a general frame, push-chain and drag-chain are virtually similar processes. I will return more specifically to the inner nature of chain movements later (see Section 3). Firstly, however, there are certain aspects of linguistic historiography and its relationship to the general history of science that I want to focus on. Within the frame of philosophy of science Within linguistics in general and in historical linguistics in particular, there has without doubt, been a strong tradition to analyse language as an autonomous structure. Language has been considered as a natural object, above all as a living organism, more or less analogous to different kinds of physical or biological structures. In short, it is the so-called “autonomous linguistics” (see e.g., Newmeyer 1986), which has possessed an indisputable hegemony within the linguistic field — as opposed to paradigms founded more explicitly within the humanities. The opposition between these two lines has also been characterised as a contrast between a ‘speaker-free’ versus a ‘speaker-centred’ linguistics (see Lass 1980:122). 2.

EXPLANATIONS?

— SOME METATHEORETICAL REFLECTIONS

35

Not surprisingly, it is within the general history of science we find the presuppositions for the hegemonic position of the autonomous, “speaker-free” tradition. Basically, this has to do with the internal relationship between natural sciences and humanities. A highly abbreviated version of this rather wellknown history shows how the exact natural sciences had the position as the guiding star in the scientific world at the beginning of the 19th century, the initial phase of modern humanities in the Western world. Thus, from the very beginning natural sciences held a position as the scientific ideal for the newly established disciplines within humanities — such as linguistics. Within this climate of opinion, it was not only legitimate, but rather a qualification of the “scientificness” of a new paradigm, to consider its own objects of inquiry as analogous with natural objects, inspired by practices within chemistry, geology, biology, comparative anatomy, mathematics (cf. Koerner 1995; Öhman 1994). Related to language, these analogies with natural sciences appeared to be particularly successful, because — to quote Roger Lass (1980:112): “[…] some aspects of language can be formally represented in such a way that it LOOKS LIKE an autonomous object (whether it ‘really is’ or not)”. This stock of ideas was promoted throughout the 19th century and had a prosperous period during most of the 20th century, too, sustained by an omnipresent positivism, or at least positivistic inspired ideals, which underpinned linguistics in general. According to such ideals, language can — or should be — treated as an objective and “unbiased” structure, something that in certain respects lives its own life, like a natural or organic being. If we then take a look at the types of explanations that have been prevalent within the field of historical linguistics throughout most of this period, they complete this picture in a neat way. As a consequence of the focus on language as a formal, natural, and objective system, there has been a strong inclination to look for various internal, autonomous forces that have influenced these natural systems, and thus brought about changes. The explanans (i.e. the explaining factor) has in other words been located in different internal linguistic domains. This means that the explanation of a linguistic change has been sought within language itself — as a capacity of an inherent property, a certain disposition, or an inner drift. This strong impact from natural sciences has in general had numerous consequences for the preferred explanations within historical linguistics — as, for example, the inclination to search for causal, predictable, and linear mechanisms behind the linguistic changes. There is no time here to go further into these topics, but a highly relevant perspective, however, is to highlight the strong continuity in these theoretical foundations, from the beginning of the

36

BRIT MÆHLUM

1800’s and until today. Lightfoot (1999:254) has given the following characteristics of this ongoing practice: The classical view of science has dominated work on language change now for 200 years [...]. Many nineteenth-century scholars [...] believed that they had discovered laws of change [...]. Twentieth-century researchers have sought historical universals, diachronic continua, theories of change, theories of grammaticalization, and whatnot. In all cases, they were looking for straight-line explanations for changes, which would allow long-term predictions [...]. This desire for a longterm determinist theory, under which there are endogenous reasons for a language to change and to become better in some way, occurs even in recent generative work. (my emphases, B.M.)

Back to the chain movements — but still within a theoretical frame How do the chain movements fit into this scientific landscape? I would say, perfectly well. The explanatory power of the chain movement principle is first and foremost based on some kind of auto-mechanical principle within language itself; a successive row of changes carried through as a consequence of the inherent oppositions within the phonological system. The following is an application of the drag-chain model to the Scandinavian vowel shift, with focus on the explanatory mechanisms the principle of drag-chain seems to presuppose. Figure 2 gives an idea of how these movements may have been going on in the row of long back vowels, according to a conventional dragchain model. 3.

FRONT

[CENTRE]

/i:/

/y:/

/e:/

/ø:/

←

BACK

(3)

/u:/ ↑ (2)

/o:/ ↑ (1)

/æ:/

/:/

Figure 2: A representation of the drag-chain-principle.

The initial change, number (1) in the model, is assumed to have occurred when Old Norse , which at this time was a back vowel, takes a horizontal step forward in the system, where (in the Old Norse period) there was an empty

EXPLANATIONS?


37

slot. Thus the gets a new and central quality, viz. /:/. With this step, however, a new empty slot is created within the system; i.e. where the was before. This structural ”gap” is in the next round filled, when the moves one step upwards and adopts he former -qualities. Finally, the long , which in Old Norse was a very low, rounded and back vowel, makes a corresponding movement into the gap of the former . And the whole transformation is eventually fulfilled, as a process of three successive stages. The essential aspect here is why these three changes are assumed to have happened at all — according to the logic of the drag-chain mechanisms. Here we have arrived at one of the crucial points connected to these explanations. The whole idea of drag-chain movements is based upon a principle where sound systems at certain points in history appear to have been asymmetrical or disharmonic in some way — reflected as holes in the systems, or so-called ‘structural gaps’. Such gaps are, however, rather quickly eliminated, due to mechanisms that have the function of re-establishing a maximum harmonic and symmetrical structure. And for the sake of clarity, a drag-chain process is supposed to have such a harmonising function. This way of explaining language change thus seems to be a paradigmatic example of a teleological explanation. The change is supposed to be purposeful or directional in a specific way, and explanans is found in the goal, i.e. telos, of the whole process. A number of controversial questions are actualised by teleological explanations in linguistics (see e.g., Vincent 1978; Lass 1980: 64ff.; Itkonen 1983:31–34; Eliasson 1983; and for a recent overview, Croft 2000:66 ff.). Here I will approach one of these questions: one of the basic dimensions of explanations like these is that they seem to ascribe to language a will of its own. The linguistic structure seems to possess control of its own future, an authority to be in command of its own fate. Accordingly, it may be a legitimate inference to characterise the mechanisms effectuating linguistic change as some kind of language-internal auto-mechanism. However, there are other ontological alternatives, too, and I will discuss a couple of them. Rather than locating the explaining mechanisms within the linguistic structure itself, it could be a plausible solution describing them as properties in the mental apparatus of the language users, a psychological or cognitive drift, helping to make language as good and beneficial as possible. Both these alternatives, however, the language-internal as well as the psychological, leave us with a number of annoying questions. The idea of inherent mechanisms like these is based on the assumption that these processes are general, in the sense of natural and universal, and thus relevant for any language. Regarding the supposed harmonising function of drag-chain processes, it is highly relevant then to ask why some languages come to that unhappy situ-

38

BRIT MÆHLUM

ation in the first place, so that they diverge from such ideal harmonious structures. And how anomalous and asymmetrical must a situation become, before processes like drag-chain or push-chain mechanisms come into effect? And not least, how do we handle the irritating fact that not all languages in the world fulfil these prescribed harmonious orders? In addition to these annoying aspects, there is a troublesome therapeutic dimension underlying these types of explanations, as have been emphasised by a number of linguists (see, specifically, Lass 1980; but also Lightfoot 1999). With a drag-chain mechanism there is a hole in the system that is being repaired, i.e. treated with a special kind of curative therapy, whereas with a push-chain process there is a prophylactic arrangement going on to avoid unhappy mergers in the vowel system. Whether we locate these highly intelligent mechanisms within the linguistic system or within the mental capacity of the language users, the logical problems seem to be insuperable. There is, however, a third alternative in this ontological discussion. According to this viewpoint, the control and steering mechanisms in the chain movements have little or nothing to do with the object of inquiry itself, but rather with the person who is doing the analysis, i.e. the linguist. This means that the idea of harmonious structures may be seen as an epistemological and heuristic need in the scientist’s mind and psychology, more than characterising inherent properties of the object of study. In other words, this has to do with the meta-language and metaphors that we are using in our scientific practice. Within historical linguistics it is, paradoxically enough, the strongly positivistic oriented Roger Lass who has conducted some admirable inquiries about these “non-positivistic” topics. One of Lass’ claims is that what makes an explanation attractive and successful, is some criteria of simplicity, intelligibility, and not least aesthetics: I think that we often judge [...] the ‘explanatory’ power of a statement or model on the basis of the PLEASURE, of a very special kind, that it affords us [...]. This pleasure is essentially ‘architectonic’, the structure we impose on the chaos that confronts us is beautiful in some way, it makes things cohere that otherwise would not, and it gives us a sense of having transcended the primal disorder (Lass 1980:157).

What Lass focuses on here, is the risk we run as scientists in ascribing to our data certain patterns and qualities that may be regarded as natural properties. We are then in danger of ignoring the fact that these qualities may be, first and foremost, intellectual, heuristic constructions, helping us to make order out of a more or less chaotic reality. If we apply these ideas to the idea of drag chainprocesses, a number of questions become urgent: What are the independent

EXPLANATIONS?


39

data affirming that certain phonological structures are “symmetrical” or “harmonious”? Are such characteristics really naturally given qualities? And if they are, which are then the criteria for claiming that something has become “asymmetrical” or “disharmonious”? If we take a look at Figure 1 in the preceding (Section 1), is the post-Old Norse vowel system actually more harmonious than it was before the change? In my opinion it could just as well be the other way around; in several respects the structure of the Old Norse system looks more neat and symmetrical than it was after the changes. The Norwegian philosopher Dagfinn Føllesdal (cf. Føllesdal 1973) has discussed the vigorous inclination within science more generally to try to “put everything” into well-formed and harmonious patterns. It seems to be an “[…] incurable realistic tendency” in a simple theory, he says. “Whether we like it or not, we feel that a simple harmonious theory reflects an ontological order […]” (Føllesdal 1973:261). At the same time it is important to underline that such theories have considerable significance in scientific endeavour. The intrinsic danger turns up at the moment when the limits between theory and reality become diffuse or non-existent. Then the theory has attained a status analogous with reality, or it may even have displaced reality. I will not assert that it is such an ontological displacement that has taken place in connection with the chain movements, but it could be worthwhile to consider the possibility. Maybe push-chain and drag-chain processes first of all represent some mental “maps”, telling us theoretical, idealised “macro-stories” of linguistic change, but at the same time making us inclined to ignore all the “micro-stories” that constitute these narratives. The Swedish linguist Gun Widmark (1998) has characterised this dilemma in the following way, related to the phenomenon of chain movements: the crucial theoretical problem with these changes is what is happening during the process, on the way from one station to another. Linguists are used to neglecting more or less the trip itself, and focusing only on the beginning and the end, she says (Widmark 1998:95). Concluding remarks Concludingly, I will briefly comment on the explanatory power of models like the push-chain and the drag-chain models. Do mechanisms like these really give us an explanation of, for example, the Scandinavian vowel shift? In the light of the previous discussion, my answer here may be rather obvious. But let us for a moment take for granted the existence of the mechanisms that are supposed to be involved in the chain movements. Whether we localise these mechanisms within language itself or in the mental capacity of the language users, we are all the same confronted with a theoretical dilemma: interpreting a linguistic change with the label ‘chain movement’ is first and 4.

40

BRIT MÆHLUM

foremost a description of the processes that may have been going on. As such, these depictions give us an account of a number of what and how questions connected to the process of change, but nothing about the why. Why was the whole process initiated in the first place; what kind of motives were at work? And why was the innovation in the second run transmitted and spread so that it actually became a real change? This is a criticism that has been put forward more generally against the autonomous tradition within historical linguistics (see, e.g., McMahon 1994:47): one has more or less neglected the actuation problem (why a change begins) as well as the implementation or transmission problem (how the change is spread). Instead, one has been satisfied with producing a set of elegant, descriptive, language-internal principles. And as a consequence of some kind of analytical contamination, these descriptions have been given the status of explanations. Within such an autonomous tradition, explanandum and explanans both seem to derive from the same internal linguistic domain. This is often considered a dubious practice in itself (see, e.g., Faarlund 1990:38–39; Keller 1994: 75–76). But it appears to be even more troublesome if one takes into consideration a fundamental comprehension shared among more and more linguists, namely the conviction that the motives behind language change are genuinely cultural phenomena. The consequence of this view is that linguistic change has to be considered as a historical process, located within a socio-cultural frame, and not as some kind of natural process subordinated to laws of nature. If this is accepted as a basic presupposition, it should also prevent us from regarding language as an autonomous entity. Language cannot be detached from its users or the situations in which they use it. On the contrary, linguists have to bring people back into the language — in order to understand why language is changing. This way of analysing linguistic change represents a genuinely critical alternative to what has been “mainstream” linguistics for the last two hundred years. In the last decades, however, this alternative way of practicing historical linguistics has been accentuated by an increasing number of scientists, represented by, for example, Michael Shapiro (1991), Henning Andersen (1973, 1989), Raimo Anttila (1993), and Rudi Keller (1994).1 A unifying char1

This list may be expanded by names like William Croft and Salikoko S. Mufwene and their contributions Explaining Language Change. An evolutionary approach (2000) and The Ecology of Language Evolution (2001), respectively. There is, however, a decisive difference between these two and Shapiro et al. in the preceding, for although both Croft and Mufwene have what I would call a sociolinguistics-like and humanistic approach in their way of handling language change, they at the same time use actively and very explicitly certain biological processes as both a device and a theoretical framework in their analyses. This may be looked upon

EXPLANATIONS?


41

acteristic of these somewhat differing approaches is their attention on language change as something that needs to be analysed within a historical, semiotic, 2 and thus hermeneutic context. The motives behind linguistic change have to be receded from language-internal, structural levels and instead situated within domains where the users of the language are focused. According to such an approach, it is not language itself that changes, but rather the language users who change their way of using it. The crucial factors enabling us to explain the phenomenon “language change” have, accordingly, to be localised to the social and intentional nature of human beings, and not to the language system as a genetic and biological entity. REFERENCES Andersen, Henning. 1973. “Abductive and deductive change”. Language 49. 765–793. Andersen, Henning. 1989. “Understanding linguistic innovations”. Language Change: Contributions to the study of its causes ed. by Leiv Egil Breivik & Ernst Håkon Jahr, 5–27. Berlin & New York: Mouton de Gruyter. Anttila, Raimo. 1993. “Change and Metatheory at the beginning of the 1990s: the primacy of history”. Historical Linguistics: Problems and perspectives ed. by Charles Jones, 42–73. London & New York: Longman. Croft, William. 2000. Explaining Language Change. An evolutionary approach. London & New York: Longman. Eliasson, Stig 1983. “Is sound change teleological? The case of the Central Scandinavian vowel shift ”. Manuscript. Føllesdal, Dagfinn. 1973. “Vitenskap og virkelighetsforståelse”. Kirke og kultur 78.258–272. Faarlund, Jan Terje. 1990. Syntactic Change: Toward a theory of historical syntax. Berlin & New York: Mouton de Gruyter. Itkonen, Esa. 1983. Causality in Linguistic Theory: A critical investigation into the philosophical and methodological foundations of ‘non-autonomous’ linguistics. Bloomington: Indiana University Press. Keller, Rudi. 1994. On Language Change: The invisible hand in language. London & New York: Routledge. Koerner, E.F. Konrad.1995. “The natural science impact on theory formation in 19th and 20th century linguistics”. Professing Linguistic Historiography ed. by E.F. Konrad Koerner, 47–76. Amsterdam & Philadelphia: John Benjamins. as merely an analogical or metaphorical way of talking about some already developed theoretical constructs (cf. Croft 2001:11). But at the same time, it is hardly possible not to consider such an approach as an opportunistic way of legitimising one’s own research strategy — by appealing to certain established theories within the natural sciences, in this case evolutionary theory. 2 For a further discussion and analysis of these approaches, see Mæhlum (1999:154 –214).

42

BRIT MÆHLUM

Lass, Roger. 1980. On Explaining Language Change. Cambridge: Cambridge University Press. Lass, Roger. 1997. Historical Linguistics and Language Change. Cambridge: Cambridge University Press. Lightfoot, David. 1999. The Development of Language: Acquisition, change and evolution. Oxford: Blackwell. Martinet, André. 1952. “Function, structure, and sound change”. Word 8.1–32. McMahon, April M.S. 1994. Understanding Language Change. Cambridge: Cambridge University Press. Mufwene, Salikoko S. 2001. The Ecology of Language Evolution. Cambridge: Cambridge University Press. Mæhlum, Brit. 1999. Mellom Skylla og Kharybdis: Forklaringsbegrepet i historisk språkvitenskap. Oslo: Novus. Newmeyer, Frederick J. 1986. The Politics of Linguistics. Chicago: The University of Chicago Press. Shapiro, Michael. 1991. The sense of Change: Language as history. Bloomington: Indiana University Press. Torp, Arne. 1982. Norsk og nordisk før og nå. Oslo, Bergen & Tromsø: Universitetsforlaget. Vincent, Nigel. 1978. “Is sound change teleological?”. Recent Developments in Historical Phonology ed. by Jacek Ficiak, 409–430. The Hague: Mouton. Widmark, Gun. 1998. Stora vokaldansen. Om kvantitativa och kvalitativa förändringar i fornsvenskans vokalsystem. Uppsala: Kungl. Gustav Adolfs Akademien för svensk folkkultur. Öhman, Sven. 1994. “Two hundred years of “scientificness” in linguistics”. Proceedings of the 14th Scandinavian Conference of Linguistics and The 8th Conference of Nordic and General Linguistics, 16–21 August 1993 ed. by Jens Allwood & al., vol. 1, 17–30. (= Gothenburg Papers in Theoretical Linguistics, 69.) Göteborg: University of Göteborg.

QUANTIFYING THE FUNCTIONAL LOAD OF PHONEMIC OPPOSITIONS, DISTINCTIVE FEATURES, AND SUPRASEGMENTALS DINOJ SURENDRAN & PARTHA NIYOGI University of Chicago Introduction Languages convey information using several methods, and rely to different extents on different methods. The amount of reliance of a language on a method is termed the ‘functional load’ of the method in the language. The term goes back to early Prague School days (Mathesius 1929; Jakobson 1931; Trubetzkoy 1939), though then it was usually taken to refer only to the importance of phonemic contrasts, particularly binary oppositions. We recently described a general framework to find the functional load (FL) of phonemic oppositions, distinctive features, suprasegmentals, and other phonological contrasts (Surendran & Niyogi 2003). It is a generalization of previous work on quantifying functional load in linguistics (Hockett 1955; Greenberg 1959; Wang 1967) and automatic speech recognition (Carter 1987). While still an approximation, it has already produced results not obtainable with previous definitions of functional load. For instance, Surendran & Levow (2004) found that the functional load of tone in Mandarin is as high as that of vowels. This means it is at least as important to identify the tone of a Mandarin syllable, as it is to identify its vowels. King (1967) notes that Mathesius (1931:148) “regarded functional load as one part of a complete phonological description of a language along with the roster of phonemes, phonemic variants, distinctive features, and the rest.” We agree with this view. While we have an interest in any role functional load might have in sound change, our primary concern here is that a historical linguist who wants to investigate such a role has the computational tools to do so. The outline of this contribution is as follows. First, in Section 2, we give an example of how functional load values can be used to investigate a hypothesis regarding sound change. Then, in Sections 3 and 4, we describe a framework for functional load in increasing levels of generality, starting with the 1.

44

DINOJ SURENDRAN & PARTHA NIYOGI

limited form proposed by Hockett (1955). Several examples, both abstract and empirical, are provided. Example: Testing the Martinet Hypothesis in a Cantonese merger One factor determining whether phonemes x and y merge in a language is the perceptual distance between them. Another factor, suggested by Martinet (1933, also see Peeters 1992), is the functional load FL(x,y) of the x–y opposition, i.e. how much the language relies on telling apart x and y. Martinet hypothesized that a high FL(x,y) leads to a lower likelihood of a merger. The only computational investigation of this hypothesis thus far is that of King (1967), who found no evidence that it was true. Doubts have been raised to his methodology (Hockett 1967), and to his overly harsh conclusion that the hypothesis was false. However, while King’s work had limitations, it was done in a time of limited computing resources and was a major advance in talking about functional load qualitatively. Sadly, it was not followed up. A full test of the Martinet Hypothesis requires examples of mergers in different languages, with appropriate (pre-merger) corpora for each case. We only have one example, but this suffices for illustrative purposes. In the second half of the 20th century, n merged with l in Cantonese in word-initial position (Zee 1999). For such a recent merger, corpus data is available. We used a word-frequency list derived from CANCORP (Lee & al. 1996), a corpus of Cantonese adult–child speech, which has coded n and l as they would have occurred before the merger. It is not a large corpus, and its nature means that there is a higher percentage of shorter words than is normal. However, it is appropriate since mergers are most likely to occur as children learn a language. Leaving definitions for later, we obtained the value 0.00090 for FL(n,l), where the n–l opposition was only lost in word-initial position. Such a small number might tempt one to conclude that this is indeed an example of the loss of a contrast with low functional load. However, that would be premature, as the absolute value for the load of a contrast is meaningless by itself. It can only become meaningful when compared to loads of other contrasts. Table 1 (in Appendix A) shows the FL values for all binary consonantal oppositions in Cantonese, when the opposition was lost only in word-initial position. This gives a much better sense of how small or large FL(n,l) is. However, ‘much better’ does not mean ‘definite’, and linguistic knowledge is required to interpret the data. The key question is which of the 171 oppositions of Table 1 should be compared to the n–l opposition. Consider the following possibilities: 2.

QUANTIFYING FUNCTIONAL LOAD

45

1. All 171 oppositions are comparable. Of these, 121 (74%) have a lower FL than the n–l opposition. Thus, the n–l opposition had a moderately high importance compared with consonantal oppositions. 2. On the other hand, several of those pairs seem irrelevant for the purpose of mergers. Perhaps only those pairs that are likely to merge should be considered. While it is not clear what ‘likely to merge’ means, let us suppose for argument’s sake that only consonants that have a place of articulation in common (consonants with secondary articulations have two places) can merge. In this case, only coronal consonants should be considered, namely h h n, l, t, t , s, ts, ts . Of the 21 binary coronal oppositions, 10 have a larger functional load than the n–l opposition, and 10 have a smaller functional load. Thus, the n–l opposition was of average importance compared to other coronal oppositions. 3. Yet a third point to bear in mind for interpretative purposes is that the phoneme that vanished in the n–l merger was n. Resorting to blatant anthropomorphism for a moment, if n had to disappear (in word-initial position), why did it have to merge with l rather than with some other consonant? In this case, only consider the 18 oppositions of the form n–x, where x is any consonant other than n. Of these, only FL(n,m)=0.00091 is higher than FL(n,l). Even when allowing for random variation in the FL values obtained, it is clear that the n–l opposition was very important compared to binary oppositions involving n and other consonants. There are, of course, other possible interpretations. The key point to note is that functional load values should be interpreted with respect to other functional load values, and the choice of ‘other’ makes a difference. The most conservative conclusion based on the above observations is that this is an example of the loss of a binary opposition with non-low functional load. More examples in other languages must be analyzed before we can make further generalizations. We hope we have at least whetted the reader’s appetite for functional load data. Defining the functional load of binary oppositions Binary oppositions of phonemes are the most intuitive kind of phonological contrast. As Meyerstein (1970) noted in his survey of functional load, this was the only type of contrast most linguists attempted to quantify. 3.

46


Perhaps the most common definition of FL(x,y) — the functional load of the x–y opposition — is the number of minimal word pairs that are distinguished solely by the opposition. The major flaw with this definition is that it ignores word frequency. Besides, it is not generalizable to a form that takes into account syllable and word structure, or suprasegmentals. We shall say no more about it. Hockett’s definition The first definition of FL(x,y) that took word frequency into account was that of Hockett (1955). He did not actually perform any computations with this definition, although Wang (1967) did. The definition was based on the information theoretic methods introduced by Shannon (1951), and assumes that language is a sequence of phonemes whose entropy can be computed. This sequence is infinite, representing all possible utterances in the language. We can associate with a language/sequence L a positive real number H(L) representing how much information L transmits. Suppose x and y are phonemes in L. If they cannot be distinguished, then each occurrence of x or y in L can be replaced by the occurrence of a new (archi-)phoneme to get a new language Lxy. Then the functional load of the x–y opposition is as in (1): 3.1

(1)

FL(x,y) = [ H(L) – H(L xy) ] / H(L)

This can be interpreted as the fraction of information lost by L when the x–y opposition is lost. Computational details It is not possible to use (1) in practice. We now give the details of how it can be made usable, taking care to note the additional parameters that are required. To find the entropy H(L) of language/sequence L, we have to assume that L is generated by a stationary and ergodic stochastic process (Cover & Thomas, 1991). This assumption is not true, but is true enough for our purposes. We need it because the entropy of a sequence is a meaningless concept — one can only compute the entropy of a stationary and ergodic stochastic process. Therefore, we define H(L) to be the entropy of this process or, more precisely, the entropy of the process’s stationary distribution. Intuitively, this can be thought of as follows: suppose there are two native speakers of L in a room. When one speaks, i.e. produces a sequence of 3.2

QUANTIFYING FUNCTIONAL LOAD

47

phonemes, the other one listens. Suppose the listener fails to understand a phoneme and has to guess its identity based on her knowledge of L. H(L) refers to the uncertainty in guessing; the higher it is, the harder it is to guess the phoneme, and the less redundant L is. Unfortunately, we will never have access to all possible utterances in L, only a finite subset of them. This means we must make more assumptions: that L is generated by a k-order Markov process, for some finite non-negative integer k. This means that the probability distribution on any phoneme of L depends on the k phonemes that occurred before it. In our speaker–listener analog above, this means that the only knowledge of L that the listener can use to guess the identity of a phoneme is the identity of the k phonemes preceding it and the distribution of (k + 1)-grams in L. An ngram simply refers to a sequence of n units, in this case phonemes. The uncertainty in guessing, with this limitation, is denoted by Hk (L), and decreases as k increases. A classic theorem of Shannon (1951) shows that H k(L) approaches H(L) as k becomes infinite. The finite subset of L that we have access to is called a corpus, S. This is a large, finite sequence of phonemes. As S could be any subset of L, we have to speak of H kS(L) instead of Hk(L). If Xk+1 is the set of all possible (k + 1)-grams and Dk+1 is the probability distribution on X k+1, so that each (k + 1)-gram x in X has probability p(x), then (2)

HkS (L) = [– x X p(x) log2 p(x) ] / (k + 1)

There are several ways of estimating Dk+1 from S. The simplest is based on unsmoothed counts of (k + 1)-grams in S. Suppose c(x) is the number of times that (k + 1)-gram x appears in S, and c(Xk+1) = xXk+1 c(x), then (3)

p(x) = c(x) / c(Xk+1)

To illustrate, suppose we have a toy language K with phonemes a, u and t. All we know about K is in a corpus S = “atuattatuatatautuaattuua”. If we assume K is generated by a 1-order Markov process, then X2 = {aa, at, au, ta, tt, tu, ua, ut, uu} and c(aa) = 1, c(at) = 6, c(au) = 1, c(ta) = 3, c(tt) = 2, c(tu) = 4, c(ua) = 4, c(ut) = 1, c(uu) = 1. The sum of these counts is c(X2) = 23. D2 is estimated from these counts: p(aa) =1/23, p(at) = 6/23, etc. Finally H1,S = (1/2) [ (1/23) log2 (23/1) + (6/23) log 2 (23/6) + … + (1/23) log2 (23/1) ] = 2.86 / 2 = 1.43. In other words, a computationally feasible version of (1) is (4): (4)

FLkS (x,y) = [ H kS (L) – HkS.xy (L xy) ] / HkS (L)

48


S.xy is the corpus S with each occurrence of x or y replaced by that of a new phoneme. It represents Lxy in the same way that S represents L. FL kS(x,y) can no longer be interpreted as the fraction of information lost when the x–y opposition is lost, as such an interpretation would only be true if L was generated by a k-order Markov process. However, by comparing several values obtained with the same parameters, as we did with the Cantonese merger example of the previous section, we can interpret this value relatively. Returning to our toy example, suppose we want to know the functional load of the a–u opposition with the same k and S. We create a new corpus S.au with each a or u replaced by a new phoneme V. Then S.au = “VtVVttVtVVtVtVVtVVVttVVV”, c(Vt) = 7, c(VV) = 7, c(tt) = 2, c(tV) = 7, and eventually H1,S.au = (1/2) [ (7/23) log2 (23/7) + (7/23) log2 (23/7) + (2/23) log2 (23/2) + (7/23) log2 (23/7)] = 1.87/2 = 0.94. Then the functional load FL1,S(a,u) = (1.43 – 0.94) / 1.43 = 0.34. Robustness to k (Markov order) It would be nice to have some assurance that the values used for k and S in (4) make little difference to our interpretation of the values we get. Surprisingly, there has been no mention, let alone study, of this problem in the functional load literature. This may be because it is mathematically clear that different choices of k and S (e.g. different k for the same S) result in different FL values. However, there is a loophole. We have already said that FL values should be interpreted relative to other FL values. Once we accept this relativity, then preliminary experiments suggest that interpretations are often robust to different choices of k and S. For example, we computed the functional load of all consonantal oppositions in English with k=0 and k=3 using the ICSI subset of the Switchboard corpus (Godfrey & al. 1992; Greenberg 1996) of hand-transcribed spontaneous telephone conversations of US English speakers. Figure 1 (in Appendix B) shows how FL 0,Swbd(x,y) and FL3,Swbd(x,y) compare for all pairs of consonants x and y. The correlation is above 0.9 (p gostit’ “stay/live with smb. as a guest”. The meaning “be a commercial traveller, trade” could never survive, because the Actor’s role activity (the guest trading) is no longer defined in relation to the guest’s necessary counterpart, i.e. the host, but instead in relation to a buyer or customer whereby the activity becomes quite different. The meaning “receive as smb. as a guest” which reflects a transitive use in OR would not make any sense in MR: MR does not have ambivalent verbs — that is to say, it is impossible in MR to have a verb where the description “be guest” occurs in the entailment-structure as well as in the autonomous state description. The MR Case System At first sight it seems as if there are no appreciable differences between the OR and the MR case systems. The structure of the original case system in OR seems to be intact in MR, i.e. we still have a distinction between direct cases (the nominative and the accusative) and oblique cases (the genitive, the dative, the instrumental). The fact that the locative has fallen out of the system of non-governed, ‘pure’ cases in favour of the system of prepositionally governed cases does not seem drastic in itself, but when added to this that the 3.4

230

PER DURST-ANDERSEN

accusative and the dative have lost their ability to name directions without a preposition the situation changes. If we, moreover, consider all those verbs that originally took a ‘pure’ oblique case (for instance, smotriti + GEN “look at”), but in MR take a preposition (e.g., smotret’ na + ACC), we get the feeling that the prepositional case system has been extended in its use and redefined (the accusative and the locative are contact cases — the accusative is dynamic, the locative static, all other cases are non-contact cases — the instrumental is static, the two others are both dynamic and static). These changes cannot but be related to changes in the system of ‘pure’ cases. In other words, we might be talking about a qualitative leap, rather than about quantitative changes. This is confirmed by verbs that originally governed an oblique case, but in MR take an accusative object. This suggests that the notion of lexical transitivity has entered the picture. Many of the verbs that originally governed the genitive, including all perception verbs (e.g. sluš ati “listen”; videti “see”; dozirati “keep an eye on smb.”; primeč ati “notice”, etc.) have become transitive (sluš at’, videt’, primetit’, etc). In other words, the very turning of your eyes or ears towards something has been reinterpreted: originally it denoted an extrovertive viewpoint where the endpoint (the distant thing) was profiled; in MR it denotes something transitive: Y is not q-theme Y is the distant thing (10) X looks at Y Y exists on L Y is q-figure and q-theme As illustrated by (10), the speaker is facing a dilemma because Y can be understood in two ways: 1. Y can be viewed from the point of view of the perceiver X whereby the direction goes from X to Y, i.e. from something near to something distant (the extrovertive viewpoint). In this understanding the verb profiles the endpoint itself, i.e. Y is the goal of X’s visual activity. This interpretation yields the genitive case and fits nicely into the structure of the oblique case system of OR; 2. Y can be viewed as a static entity that exists on a certain location, i.e. Y has become an independent entity that has its own life completely isolated from the visual activity. In this understanding Y is figure and the location itself is ground. This reinterpretation has taken place within the framework of the direct case system of MR, and since Y is q-figure and therefore secondary figure, Y receives the accusative case, the case for secondary subjects, q-themes. This suggests that the original direct case system has

FROM PROPOSITIONAL TO SITUATIONAL SYNTAX IN RUSSIAN

231

been reinterpreted from being a propositionally oriented system operating with underlying subjects within ground-propositions to becoming a referentially oriented system operating with figure-ground constellations in pictures that mirror situations in reality. The following example from MR illustrates this: Characterization X keeps away from Y/ GEN (11) X is afraid of Y X keeps away from Y/ ACC Situation description In MR bojat’sja “be afraid of” may take an object in the genitive or in the accusative case (originally it governed the genitive). When it appears with a transitive object, it gives a situation description which is felt by Russians to be a concrete description. When it occurs with the genitive case, the idea of having a figure appearing on a ground is automatically blocked. The result is that we automatically leave the level of situation description and instead gets a characterization of the person in question which reflects the speaker’s own inner world and not the real external world. This description is felt by Russians to be abstract. What we witness is that the direct case system of MR has obtained the same external frame of reference as the oblique case system in OR, i.e. it has become a question of local existence (there is a figure-ground constellation) vs. no local existence (there is no figure-ground constellation). The original structure is intact, but whereas the oblique case system of OR was alone in describing external relations between participants from various viewpoints, in MR it has received a partner in the shape of the direct case system. Both systems speak now in the same tongue — the entire system has achieved harmony. Earlier the direct case system involved a choice between p-theme and q-theme, now it involves a choice between p-figure and q-figure: p-figure yields the nominative case, q-figure yields the accusative case. If there is no figure-ground constellation, the genitive case, being the oblique case per se, takes over (see also below). 3.5 Syntactic structures It is crucial to understand that not only the system and structure of Russian cases (cf. Durst-Andersen 1996), but also its syntactic patterns are grounded in the observable structures of our real world (where we deal with notions like picture and figure-ground) and not in the unobservable structures of our thoughts (where we deal with notions like proposition, logical subject and predicate). That is why a situation without a kernel corresponding to Eng. It

232

PER DURST-ANDERSEN

darkens will be named by a Russian verb, Temneet, which does not take a subject or another argument for that matter (the ground-situation of the verb will have no p-figure and therefore there will be no p-theme). One can simply say that when there is no figure-ground constellation, there can be no subject. This rule knows of no exceptions in MR. In contradistinction to that situation, English is — just like OR — based on the principle of propositional syntax, but in English it has been completely generalized in demanding a subject in any sentence. And that is also why a Russian verb which prototypically takes a subject in the nominative case, e.g., byt' “be” obligatorily marks “loss of figure”, if, for instance, X is not home in the de facto situation referred to by the verb. This is done by putting the underlying subject in the genitive case, e.g. Ego ne bylo doma “He was not at home”. The nominative and the accusative can only be used if there is local reference. In that way MR can be said to be determined by the prototype structures of reality and their various manifestations in de facto situations, and not the other way around, where sentence structures can be said to put a certain form on structures in our referential reality — which is the case in, for instance, English. Here it and there in socalled thetic sentences have no reference at all, i.e. they do not refer to anything in our referential reality — they only serve the obligatory function of topic (i.e. in English there must be a surface sentential subject). Moreover, it applies to all types of construction where what appears to be subject in many other languages, for instance, in English and in Danish, turns out to be direct objects in MR, cf. He was killed by a grenade and Ego (ACC) ubilo granatoj (INSTR). The point is that when there is no p-figure (corresponding to Actor), there can be no surface subject in the nominative case. Since the person was qfigure (corresponding to Undergoer) in the situation referred to, he receives the accusative case marking. That the Gestaltist notions of figure-ground indeed play a crucial role in MR appears from the fact that imati + ACC, which was the normal construction for “to have smth.”, was preserved in MR as imet’, but it was outranged by a new existential construction, viz. “smth. exists with smb.”, e.g. U nego bila ž ena “(lit.) With him was wife”, i.e. “He had a wife”. The existential construction is the normal way of expressing HAVE-relations — imet’ in MR means “to have at one’s disposal”. The existential construction was introduced and gained more and more ground because it fits the figureground-like way of thinking (for a contrastive-typological analysis of English and Russian, see Durst-Andersen 2002). Conclusion: The role of aspect in the changes When wondering why the above mentioned qualitative leap took place, the first thing that comes to one’s mind is the introduction of the category of 4.

FROM PROPOSITIONAL TO SITUATIONAL SYNTAX IN RUSSIAN

233

aspect. But one may also wonder why aspect was introduced in the first place, the process of which started 600 years ago and which has not ended yet, although we are close to the terminating point. I see a natural connection between the naming strategy and the lexicalization patterns of OR, on the one hand, and the rise of the grammatical category of aspect, on the other. OR had chosen to name the ground-propositional structures and had a systematic lexicalization pattern telling that when naming an activity or an action the state description should be specified — exactly the state description motivates the common expression of the various meanings of a single lexeme. This is tantamount to saying that from a lexical point of view all dynamic verbs foreground the state description, but background the very activity description. As I see it, the described lexicalization pattern in OR paved the way for the occurrence of the grammatical category of aspect which, as a matter of fact, just repeats what has been standard at the lexical level and transfers it to the grammatical level. The perfective aspect inherently belongs to action verbs, is the marked and prominent member of the opposition; it asserts the state description and presupposes the activity description. In doing so, the perfective aspect creates a state situation which is foregrounded and an activity situation which is backgrounded. The connection could not have been more clear. Legitimizing lexical procedures on a higher level is thus just another way of defining grammaticalization. With the introduction of aspect the original lexicalization pattern had to be changed. This was done by eliminating all ambivalent verbs by creating a whole range of prefixed variants of the original simplex verb. For instance, MR created obokrast’/obkradyvat’ “rob smb.”, nakrast’/nakradyvat’ “steal a lot”, vykrast’/vykradyvat’ “pinch, nick”, raskrast’/raskradyvat’ “loot, clean out”, and others, in order to be able to describe different “stealing-situations” from the point of view of the resulting state. Krast’ was preserved for the “normal” stealing-situation, but a perfective partner was created, viz. ukrast’, in order to complete the dual notion of an action. All these verbs are possession-based action verbs and are all transitive. They follow the new naming strategy and lexicalization pattern since they name the action from the point of view of its image structures by focusing on the state, but on the condition that the state is an effect of the same prototypical activity. Likewise, I see the reinterpretation of the original direct case system as being a result of the introduction of the category of aspect. The final result is that, on the one hand, we find harmony between the direct and the oblique case subsystems, and between lexicalization patterns and aspect, and on the other, between aspect and the syntactization mechanism which is governed by the situations themselves.

234

PER DURST-ANDERSEN

Acknowledgments I wish to thank Henning Andersen who has been of great help to me in my attempt to grasp the system of Old Russian, and Jan Erik Nielsen who helped me in finding material and data. REFERENCES Baron, Irène. 2002. “Leksikalske prototyper i dansk”. Sprog og typer i kontrast: Forskelligartede danske bidrag til en forståelse af sproglig typologi ed. by Per Durst-Andersen, 1–21. Copenhagen: Copenhagen Business School. Bermel, Neil. 1997. Context and the Lexicon in the Development of Russian Aspect. Berkeley: Berkeley University Press. Durst-Andersen, Per. 1992. Mental Grammar: Russian aspect and related issues. Columbus, Ohio: Slavica. Durst-Andersen, Per. 1996. “Russian case as mood”. Journal of Slavic Linguistics 4.177–273. Durst-Andersen, Per. 2000. “The English progressive as picture description”. Acta Linguistica Hafniensia 32.45–103. Durst-Andersen, Per. 2002. “Russian and English as two distinct subtypes of accusative languages”. Scando-Slavica 48.102–126. Durst-Andersen, Per & Michael Herslund. 1996. “The syntax of Danish verbs: Lexical and syntacic transitivity”. Content, Expression, and Structure. Studies in Danish Functional Grammar ed. by Elisabeth Engberg-Pedersen & al., 65–102. Amsterdam & Philadelpha: John Benjamins. Guillaume, Gustave. 1965 [1929]. Temps et verbe. Paris: Champion. Hewson, John & Vit Bubenik. 1997. Tense and Aspect in Indo-European Languages: Theory, typology, diachrony. Amsterdam & Philadelphia: John Benjamins. Ivanov, Vjacheslav V. 1983. Istorič eskaja grammatika ruskogo jazyka. (Izd. vtoroe.) Moscow: “Prosveš č enie”. Krys’ko, Vadim B. 1994. Razvitie kategorii oduš evlennosti v istorii russkogo jazyka. Moscow: Lyceum. Krys’ko, Vadim B. 1997. Istorič eskij sintaksis russkogo jazyka. Ob’ekt i perexodnos’. Moscow: Lyceum. Lunt, Horace G. 1970. Concise Dictionary of Old Russian (11th–17th Centuries). Munich: Wilhelm Fink Verlag. Nørgård-Sørensen, Jens. 1997. “Tense, aspect and verbal derivation in the language of the Novgorod birch bark letters”. Russian Linguistics 21.1–21. Sreznevskij, Izmail I. 1989 [1893]. Slovar’ drevnerusskogo jazyka. Moscow. Stecenko, Anatolij N. 1972. Istoriceskij sintaksis russkogo jazyka. Moscow: “Vysš aja š kola”.

CONSTRUAL OPERATIONS IN SEMANTIC CHANGE THE CASE OF ABSTRACT NOUNS

LENA EKBERG University of Lund Introduction This paper deals with lexical variation and semantic change of abstract nouns.1 The theoretical framework is Cognitive Semantics as presented in works by Langacker (1987, 1991), Talmy (2000), and others. We assume that at least a subset of the regular sense developments is due to cognitively based processes — so called ‘construal operations’ (Croft & Wood 2000) — operating on a semantic input, i.e., the schematic structure constituting the foundation of the more detailed lexical meaning. As for relational predicates, e.g. prepositions and verbs, transformations on the basic schematic structure, commonly referred to as the image schema, have proved to be an important device of generating new meaning variants (Ekberg 2001, 2004). For nouns, (also) other types of operations can be assumed to be relevant in relating meaning variants to each other. In this paper I argue for the impact of construal operations by means of analyzing a selection of Swedish abstract nouns, namely tillfälle “occasion”, mening “opinion”, uppgift “information”, and fråga “question”. The nouns chosen show various degrees of relatedness to verbs that are synchronically used with the same meaning. As a consequence, they show various degrees of abstractness in terms of different content ontologies (Paradis 2005). In short, there are abstract entities that are located in time and space, i.e. events (actions, processes or states), and those that are not located in time and space, e.g. abstract phenomena such as idea, opportunity. A noun referring to an event re1.

The reported study is part of the project The Dictionary of the World, of which the overarching aim is to describe and explain regularities in lexical polysemy. The members of the project are, in addition to the author, Jerker Järborg and Kerstin Norén (project leader). The project is financed by The Bank of Sweden Tercentenary Foundation. I am grateful to my project colleagues and to Carita Paradis and Christer Platzack for valuable comments and stimulating discussions. I am solely responsible for remaining inconsistencies. 1

236

LENA EKBERG

lates to a situation profiled as a THING (Langacker 1987) but has the same situation type characteristics as the related verb (Paradis 2005). Crucial for the semantic account of these (and other) lexemes is the notion of profile–base asymmetry. The profile is what the lexeme designates within a given cognitive domain, i.e., the base. In other words, the profile is equal to those portions of the base that the semantic predication invokes (Langacker 1987). As will be argued, the profile of verbal nouns includes an event structure (E), embedded in the holistic conceptualization characteristic of a nominal. The concept of an event comprises activities, processes and states.2 The different event types represent different elaborations of an event structure; thus E will be short for a semantic entity with a relational internal structure, e.g. one of the various elaborations of the basic action chain underlying the semantics of all four nouns in question. The thesis elaborated here is that the process of ‘attention’ plays a crucial role in relating meaning variants of verbal nouns to each other. Attention is taken to be one of four basic cognitive abilities manifested in language by way of continuously occurring processes — the others being ‘comparison’ (the foundation of metaphorical relations), ‘choice of perspective’ and ‘constitution’ (Gestalt), respectively. Croft & Wood (2000) distinguish five subtypes of attention, which are partly overlapping. The most important ones for a lexical description of verbal nouns are ‘selection’, ‘scalar adjustment’ and ‘schematization’. Selection amounts to focusing on only part(s) of an entity or an experience and disregarding the rest, whereas scalar adjustment (‘abstraction’, in terms of Langacker 1987) is construing an event with various degrees of abstraction/specificity. On this view, schematization — the choice of viewing something by means of a more encompassing category — is one end of the abstraction cline. Cognitive linguists see construal operations of the type presented above as structural devices of the lexicon. The assumption made here is that the same types of operations that underlie the structure of the lexicon also underlie the structure of lexical networks, i.e. the category formed by the various meanings of a single lexeme. Furthermore semantic change and synchronic lexical variation is claimed to be two sides of the same coin. The tracks of semantic change may simultaneously be looked upon as meaning relations within a lexical network.

2

Activities and processes have the feature [+dynamic] in common but differ with respect to agentivity — activities are [+agent] while processes are [–agent]. State predicates are both non-dynamic and non-agentive (see further Schmid 2000).

237

CONSTRUAL OPERATIONS IN SEMANTIC CHANGE

Lexical variation and semantic change of some abstract nouns Tillfälle “occasion” One of the synchronic uses of tillfälle is “occasion”, designating a single point of time, as in (1). 2. 2.1

(1)

vid det tillfället ... at that occasion .... “on that occasion ...”

The conceptualization of a single point of time presupposes an event (E) taking place at the specific moment (see Figure 1). 3 The event constitutes the LM (the reference point) in relation to the focused point of time (PT), which is the TR. I assume that the TR–LM-relation is hosted within the profile of a region, since the “surface” meaning of tillfälle is a holistic situation connecting a point of 4 time (PT) with a particular event (E), cf. Figure 1.

E/TR Profile

Time Base PT/LM Figure 1: The schematic structure of “occasion”.

The example in (2) illustrates the profile–base-relation inherent in the semantics of tillfälle. (2)

3

Vid ett tillfälle UNDER MÄTNINGARNA bröts strömmen. at one occasion during measurements.the switched.off.PASS current.the “One one occasion DURING THE MEASUREMENTS the current was switched off.”

In Schmid (2000: 79), tillfälle functions as a shell noun with an inherent semantic gap that is filled by an unspecific ‘event’. 4 Here, and henceforth, the profile is given in boldface and the focused entity within the profile is designated by a filled circle.

238

LENA EKBERG

Here the base is explicitly referred to as a number of temporally sequential events, bound to a sequence of points in time, out of which a particular point of 5 time is profiled by the noun phrase ett tillfälle “one occasion”. In (3) however, the meaning of the particular event is focused, at the cost of temporal meaning. Note that the point of time (September the 28th) is explicitly stated and consequently not focused in the use of the noun. As shown in Figure 2, the event corresponds to the TR (in boldface) and the point of time to the LM (underlined). (3)

Detta historiska tillfälle inträffar den 28. september. this historic event occurs the 28th September “This historic event will occur the 28th of September.”

The meaning in (3) thus represents an interchanged TR–LM-relation within the profiled region compared to the one in Figure 1, namely: E/TR

Profile

Time Base PT/LM

Figure 2: The schematic structure of “event”.

A further step from the pure temporal meaning is when tillfälle corresponds to “opportunity”: (4)

bereda någon tillfälle att framföra sina synpunkter give someone opportunity to express his/her views “to give someone opportunity to express his/her views”

This meaning assumes — just as the meaning “occasion”— a course of time, which includes a specific point of time when the possible event may be realized. In other words, the relation between the elements hosted within the pro-

5

The — sometimes slightly manipulated — examples are mainly taken from a computerized corpus named SUC/LBAB (see Järborg 1999).

239


file is the same as in the temporal meaning but the setting is virtual instead of actual.6 In Figure 3, the virtual setting is represented by broken lines. E/TR

Profile

Time Base PT/LM Figure 3: The schematic structure of “opportunity”.

Summarizing, the synchronic meaning category of tillfälle includes the meanings “occasion”, “event”, and “opportunity”. “Event” is related to “occasion” through change of focus within the profile. The meaning “opportunity” instead retains the relation between point of time and event as profiled by the temporal meaning but changes the setting of the relation. Note that whereas “opportunity” intuitively is a separate sense, 7 the temporal and the eventive meanings are rather variants of the same sense. As I have argued “occasion” and “event” in fact presuppose each other (and are often difficult to separate in actual use). In other words, they seem to be different ‘facets’ of the same meaning, conventionally associated with the lexical item (see Paradis 2003; Cruse 2003). The boundary between “occasion” and “event” on the one hand and “opportunity” on the other is motivated by the different bases (or settings) decisive of the interpretations of the profile. The temporal and the eventive meanings are interpreted within the same setting, whereas “opportunity” is interpreted within a different setting. The history of tillfälle however adds further meaning variants to the lexical network. The following variants are attested in Old Swedish (Söderwall 1884–1918): “event” (5), “opportunity” (6), “occasion” (7), “cause” (8), and “accident” (9). Of these, “accident” seems to be the earliest meaning, cf. the parallel formations MLG tôval, MHG zuoval “accident”, and Lat. accidens (from ad- “to” and cadere “fall”, Hellquist 1957).

6

Setting is defined as a “global, inclusive region within which an event unfolds or a situation obtains” (Langacker 1991:553). 7 ‘Meaning’ is used as a superordinate term, not distinguishing between contextually triggered interpretations and well-entrenched lexical meaning variants (‘senses’).

240

LENA EKBERG

(5)

ciprianus […] taladhe til biscopin all sin tilfälle Ciprianus [...] spoke to bishop.the all his events “Ciprianus ... told the bishop everything that had happened to him”

(6)

han sökir tilfälle mik fördriffwa oc dräpa he seeks opportunity me to.drive.away and kill “he seeks an opportunity to drive me away and kill me”

(7)

jak gaf thik tillfälle I gave you occasion “I gave you occasion”

(8)

at ond gerning var hans dödz tilfälle that evil deed was his death. GEN cause “that evil deed caused his death”

(9)

aff eno tilfälle of one accident “because of an accident”

As I will argue, the meaning “accident” — paraphrased as “a random cause of something that has happened” — in fact includes all the other meaning variants of tillfälle. From “accident” we may proceed to “cause”, by way of ‘generalization’ of the cause-notion, suppressing the feature “random” and leaving “cause” as the only profiled content. The semantic structure of “cause” (C), as well as “accident”,8 in turn presupposes the concept of “event” (E) — bound to a point of time (PT), as diagrammed in Figure 4, below. By varying the structure in Figure 4, also the other meaning variants (cf. (5), (6), and (7) above) are derived. First, “event” is related to “cause” through ‘promotion’ of the event-notion (E) in the profile, resulting in a schematic structure equal to the one in Figure 2. Further, “occasion” is related to “event” through promotion of the notion point of time (PT), resulting in the schematic structure of Figure 1. In both cases the focus of attention is changed. Finally, “opportunity’ is related to both “occasion” and “event” by changing the setting of the entire event-structure, cf. Figure 3.

8

It is assumed that “accident” and “cause” share the same schematic structure. The feature “random”, distinguishing the two meaning variants, is taken to be part of the lexical specification.


241

Profile C

E PT

Time Base

Figure 4: The schematic structure of “cause”/“accident”.

As shown above the set of meaning variants in Old Swedish also includes the ones retained in Modern Swedish, i.e. “occasion”, “event”, and “opportunity”. Thus both diachronically and synchronically the meaning variants of tillfälle stay within the realms of the complex meaning attested in the earliest use of the word. Assuming that the concept “cause” includes an event, bound to a point of time, provides an explanation of the coexistence of “cause” and “opportunity” (i.e. a possible point of time bound to a possible event) within the same lexical network. There are in fact further examples of this coexistence. Swedish, Norwegian anledning — formed in accordance with German Anleitung “cause” 9 (orig. “guidance”) — shows, cross-linguistically, the same ‘lexical intraference’ (with a term borrowed from Croft 2000) as tillfälle. In Modern Swedish the central meaning is “reason” whereas in Norwegian the central meaning is “opportunity”. Thus a possible explanation of this cross-linguistic polysemy is that “opportunity” is derived from “reason” — which meaning is present also in Norwegian — by way of interpreting the event-structure (presupposed by “cause”) in a virtual setting. This derivation will thus be a parallel to the semantic development of tillfälle. 2.2 Mening “opinion” The noun mening is derived from the verb mena “(to) mean; think”, and one of the salient synchronic meanings is “view, opinion”, as in (10): (10) Enligt min mening … according.to my view …. “According to my view …” 9

Germ. Anleitung in turn is derived from the verb anleiten “(to) guide” (composed of an “to” and leiden “(to) lead”). The cause-sense — actually, “what has led to something” — thus seems to be a natural extension of the concrete sense of the verb.

242

LENA EKBERG

The schematic structure underlying the more specific lexical meaning of the noun presupposes an elaborated variant of the schematic model of an action chain (Langacker 1991:283). As the profiled entity — LM1, in Figure 5 — is interpreted within the event frame of a communication situation, the Trajector (TR) matches the sender whereas LM2 corresponds to the addressee. LM1 is equal to the “message” — here specified as the “opinion” of the TR — transferred from TR to LM2.

TR

LM1

LM2

“opinion” COMMUNICATION Figure 5: The schematic structure of “opinion”.

In addition to the meaning “opinion”, the word shows a number of semantically related meanings, namely “intention” (11a), “purpose” (11b), “sense” (11c), and “sentence” (11d). (11) a. Meningen är att förstärka partimedlemmarnas inflytande […] intention.the is to.reinforce party.members.the.GEN influence “The intention is to reinforce the influence of the party members […]” b. Lydnad är meningen med människans liv på jorden. obedience is purpose.the with man.the.GEN life on earth.the “Obedience is the purpose of man’s life on earth.” c. i ordets verkliga mening in word.the.GEN real sense “in the word’s real sense“ d. en lång mening a long sentence / “a long sentence”


243

Opinions are typically expressed by means of verbal communication, i.e. in words and sentences, which (in our culture at least) are conceptualized as some sort of containers (Reddy 1979). In terms of this metaphorical thinking the linguistic form corresponds to the embracing container whereas the linguistic meaning is seen as the content of the container. There is thus a clear connection between “opinion” (10) and “sense” (11c) in that part of the background knowledge of the meaning “opinion” is highlighted (or profiled) in the meaning variant “sense”. A verbally expressed opinion is not only an instance of a linguistic sense (wrapped up in words and sentences), but it also manifests intention and purpose. Thus “intention” and “purpose” are linked to both “opinion” and “sense”, respectively. Further, the use of mening in (11b) is related to the metaphorical use in (11c) — life is just like a word conceptualized as a container but whereas the content of life is equal to the purpose of life, the content of a word is equal to its sense. Also the meaning “sentence” (11d) — although clearly separated from the variants in (11a–c) — fits well into the base knowledge of the sense “opinion”. Sentences are in fact not only the typical means to verbally express an opinion but the way to do it. That is, sentences represent the form side of a linguistic expression used to express an opinion. Interestingly, also Eng. sentence shows an earlier meaning “meaning, sense”, “opinion”, motivated by its origin in Lat. sententia “mental feeling, opinion” (ODEE: 809). Together, all these meaning variants — “sense”, “intention”, “purpose”, and “sentence” — are parts of the base, referred to as COMMUNICATION in Figure 5, along with other semantic aspects which are not lexically realized. The interconnections between the various elements in the base are illustrated in Figure 6 below. Lexically realized meanings are given within double quotation marks. In each of the uses in (11) above only one of the elements in the base is focused — that is has entered the profile — whereas the others are backgrounded.

244

LENA EKBERG

Profile:

Mening “opinion”

Base:

[linguistic expression]

[symbolic unit]

[form]

“sentence”

[functional unit]

[meaning]

“intention”

“purpose”

“sense”

Figure 6: Overview of the interconnections in the base of Sw. mening “opinion”.

Uppgift “information” The noun uppgift has two well-entrenched meanings, namely “information” vs. “task”. The meaning “information” can be paraphrased as “a message containing a statement”, showing that uppgift is viewed as a container (cf. 2.2.). In addition, the verbal message per se is conceptualized as a bounded object, which can be given (12a), received (12b), and possessed (12c). 2.3

(12) a. uppgift skall lämnas om hur många dagar […] information shall leave.PASS about how many days […] “information should be left about how many days […]” b. att han fått uppgiften om yxhugget that he got information.the about cut.of.axe.the “that he got the information about the cut of the axe” c. Har du några uppgifter på det? have you any information.PLUR on that “Have you any information about that?” The range of uses of uppgift supports the idea that verbal messages are conceived of in terms of the conduit metaphor, shown in the examples below where “the content” of uppgift is underlined.


245

(13) a. Framställningen bör innehålla uppgift om övriga report.the ought to.contain information about other omständigheter. circumstances “The report ought to contain information about other circumstances.” b. att uppgiften skall avse näringsidkarens registrerade firma that information.the shall concern manufacturer.the. GEN registered firm “that the information should concern the manufacturer’s registered firm” In some uses, however, only the content is focused — disregarding the embracing verbal form. For instance, this is the case when uppgift is combined with the verb stämma “to be correct”: (14) Stämmer uppgiften? is.correct information.the “Is the information correct?” Compare also the example in (15) where uppgift takes an attribute that modifies the truth of the verbal content: (15) tvivelaktiga / äkta uppgifter doubtful / genuine information.PLUR “doubtful / genuine information” In other uses the focus is shifted the other way around, so to speak, that is, to the form of the verbal information, disregarding the conceptual content. In (16) this has led to an almost concrete sense of the word uppgift. (16) De uppgifter som förr sparades på papper finns i dag […] those information. PLUR that previously saved.PASS on paper is. found today […] “The information that previously was saved on paper is found today…”

246

LENA EKBERG

Given the semantic description of uppgift — “a message containing a statement” — form and content are different facets of the meaning (see 2.1.). Thus the different uses focusing on either the form or the content — or referring to both — are contextual meaning variants, not different senses. Out of context, both aspects (form and content) are simultaneously referred to. Another meaning of uppgift, clearly distinct from “information”, is “task”. This meaning can be paraphrased as “a certain activity, with a certain 10 purpose, that is expected to be executed by a given person, organization etc.”. (17) a. Åklagarens uppgift är att försöka visa att […] prosecutor.the.GEN task is to try to.show that […] “The prosecutor’s task is to try to show that […]” b. Musikernas uppgift är den svåraste. musician.PLUR.the.GEN task is the most.difficult “The musician’s task is the most difficult one.” At first glance the two meanings of uppgift seem to be unrelated, but I will argue that there is a cognitively plausible way of connecting the two. Uppgift appeared with the meaning “information” in the 17th century, derived from the complex Swedish verb upgiva “to give away”, “inform”. In the 18th century the noun appears also with the meaning “task”, which is a translation loan from German Aufgabe with the same meaning (NEO). Common for both meanings, however, is the conception of an action chain (see 2.2.) underlying the semantics of the derivational verb base “to give away”. When used to mean “information” (Figure 7, below) the profiled entity — LM1 — is equal to the message transferred from TR to LM2. On the other hand, when used to mean “task” (Figure 8, below), LM1 is identical to “a piece of activity” that, also, is transferred (given) to LM2. The underlying schematic structure is identical, but the boundary of the profile is different. LM1 is still the entity focused on, but also LM2 is hosted within the profile — reflecting the fact that “task” refers to a goal-oriented action. The lexical content of LM1 is due to the different cognitive domains (bases) in which the noun uppgift is interpreted.

10

The definition is taken from the lexical database GLDB/SDB (see Järborg 1999).

247


TR

LM1

LM2

“statement” COMMUNICATION Figure 7: The schematic structure of “information”.

TR

LM1

LM2

“task” ACTIVITY Figure 8: The schematic structure of “task”.

The representations in Figure 7 and 8, respectively, are assumed to be different specifications of the schematic structure underlying the verb upgiva “to give away”, of which the noun uppgift is derived. Summarizing, the meaning category of uppgift includes both the meaning “information” and the related but distinct (i.e. lexical) meaning “task”. The former meaning has contextual variants focusing on either the form of the verbal message or the content expressed. The meaning “task” is linked to “information” (and its variants) by means of a different construal of the profile, in turn related to different bases of the underlying schematic structure. Fråga “question” In Modern Swedish the central meaning of fråga is “question”, which according to SAOB is elaborated as “request for information about a state of affairs of which one is ignorant” (my translation). The complex meaning is de2.4

248

LENA EKBERG

rived from the verb fråga “to ask”, elaborated as “x asks for information from z”. The underlying schematic structure is thus the same as for uppgift. That is, the schematic structures of fråga and uppgift, respectively, are variants of the general action chain (see 2.2, 2.3.), where the TR transfers an entity (LM1) to a recipient (LM2). In Figure 9, illustrating the schematic structure of “question”, the profiled entity LM1 is the “question”, which is understood in relation to the cognitive domain STATE OF AFFAIRS. The entire action chain in turn is related to the domain COMMUNICATION .

TR

LM1

LM2

“question” COMMUNICATION Figure 9: The schematic structure of “question”.

What the meaning variants of fråga have in common is that they all select elements in either of the bases — STATE OF AFFAIRS vs. COMMUNICATION — and promote them to the profile. The most salient meaning variants of fråga are given below, numbered F1–F5. The central meaning is referred to as F. F1: “state of affairs of which one is ignorant”, which further develops into “problem, matters”, cf. the examples in (18): (18) a. frågan är hur vi ska göra med X question.the is how we shall do with X “the question is what to do with X” b. lösa en fråga to dissolve a question “to dissolve a question (i.e. a problem)” Meaning derivation: F > F1, as a result of promotion of elements in the base STATE OF AFFAIRS into the profile, combined with a conventionalized prag-


249

matic inference. The inference is based — hypothetically — on the reasoning that a state of affairs of which one is ignorant easily may become problematic. F2: “state of affairs”; exemplified by: (19) ekonomiska, sociala, politiska frågor economical, social, political matters “economical, social, political matters” Meaning derivation: F1 > F2, as a result of generalization of meaning F1 (“state of affairs of which one is ignorant”). F3: “discussion; consideration (about a state of affairs)”, illustrated by an example from the 18th century: (20) Tå fråga är om Ledamots uteslutande böra altid Tolf Ledamöter vara närvarande. (SAOB) when consideration is about member.GEN exclusion ought always twelve members be present “When exclusion of a member is considered twelve members should always be present.” Meaning derivation: F > F3, as a result of selection of elements in the base COMMUNICATION. The meaning F3 relates to the reciprocal communication, of which the questioning is part. F4: “argument, debate, dispute” (18th–19th c.), exemplified by: (21) Det är bekant, hvad fråga varit emellan Philosophiska Sedelärare, om en enda princip för Sedeläran (SAOB) it is known what dispute been between philosophical moral. philosophers, about one single principle for moral.philosophy.the “It is known what dispute has been between philosophical moral philosophers, about one single principle for the moral philosophy” Meaning derivation: F3 > F4, as a result of pragmatic inference. The motivation for this assumption is that “discussion” often implies “disagreement”. F5: “doubt”, exemplified by: (22) Frågan är om fästingar verkligen orsakar hjärnhinne-inflammation. question.the is whether ticks really cause meningitis “The question is whether ticks really cause meningitis.”

250

LENA EKBERG

Meaning derivation: F > F5 as a result of selection of elements in the base STATE OF AFFAIRS, combined with pragmatic inference. The reasoning goes as follows: it is likely that one may be doubtful about a state of affairs of which one is ignorant. The developmental network of fråga is summarized in Figure 10. F

F1

F3

F2

F4

F5

Figure 10: The developmental network of fråga “question”.

As for the meaning variants F1, F3, F5 they can all be traced back to meaning F — through the process of selection of elements in either of the two bases of the semantic predication. The variant F2 in fact instantiates the meaning of an entire domain (STATE OF AFFAIRS), present in the variant F1 (and F). Summary and conclusion The present account of the semantics of four abstract nouns leads to the following — tentative — conclusions. Firstly, the semantic pole of a noun profiling a region may host a schematized relational structure. More precisely this schematic structure is what Talmy (2000) refers to as an ‘event frame’ with particular characteristics depending on the conceptual content of the noun. The event structure embedded in the profile is inherited from a related verb, either this verb is synchronically present (uppgift) or not (tillfälle). However, the embedded event is invariably construed as a whole unit, giving rise to a holistic conceptualization. Secondly, it has been demonstrated that processes generating semantic variation and change operate on the schematized structure underlying the lexical representation of a linguistic expression. Among these processes are ‘selection’ of elements of the profile vs. the base, and ‘change of setting’ of the entire event structure. ‘Selection’ amounts to the foregrounding of x, where x is either part of the profile or the base. In the latter case, the elements in the profile is backgrounded — that is, put in the base — and substituted by the new profiled element(s). ‘Generalization’ and ‘promotion’ are taken to be subcases of selection. The former amounts to a case where the foregrounding of x leads to the 3.


251

elimination of other elements in the profile, leaving x as the only profiled content. The latter amounts to promotion of elements from the base to the profile. ‘Change of setting’ is assumed to be a general construal mechanism — a transformational device — giving rise to new meaning variants. The prerequisites of meaning variation of a lexeme are intrinsic in the underlying schematic structure as well as in the construal operations that may apply to that structure. Thus every instance of semantic change and variation — either resulting in polysemy or contextual meaning variation — is motivated by the possibilities of varying a given schematized structure by means of general and cognitively motivated construal operations. REFERENCES Croft, William. 1998. “The structure of events and the structure of language”. The New Psychology of Language: Cognitive and functional approaches to language structure ed. by Michael Tomasello, 67–92. London: Lawrence Erlbaum Associates. Croft, William, 2000. Explaining Language Change: An evolutionary approach. Harlow, Essex: Longman. Croft, William & Esther J. Wood. 2000: “Construal operations in linguistics and artificial intelligence.” Meaning and Cognition: A multidisciplinary approach ed. by Liliana Albertazzi, 51–78. Amsterdam & Philadelphia: John Benjamins. Cruse, D. Alan. 2003. “The construal of sense boundaries.” Revue de Sémantique et Pragmatique 12.101–119. Ekberg, Lena. 2001. “Transformations on the Path-schema and a minimal lexicon”. Studia Linguistica 55:3.301–323. Ekberg, Lena. 2004. “Transformations on image schemas and cross-linguistic polysemy”. Nordlund: Småskrifter från Institutionen för nordiska språk. 24.25–46. Hellquist, Elof. 1957 [1922]. Svensk etymologisk ordbok. (3rd ed.) Lund: Gleerups förlag. Järborg, Jerker. 1999. “Lexikon i konfrontation”. Research Reports from the Department of Swedish, Göteborg University. Göteborg. GU-ISS-99-6. (http://svenska.gu.se/~svedk/resrapp/konfront.pdf) Langacker, Ronald W. 1987. Foundations of Cognitive Grammar, vol. 1: Theoretical Prerequisites. Stanford, Calif.: Stanford University Press. Langacker, Ronald W. 1991. Foundations of Cognitive Grammar, vol. 2: Descriptive Application. Stanford, Calif.: Stanford University Press. NEO = Nationalencyklopedins ordbok. 1996. Språkdata, Göteborg & Bra Böcker, Höganäs. ODEE = C. T. Onions, ed. 1966. The Oxford Dictionary of English Etymology. Oxford: Clarendon Press.

252

LENA EKBERG

Paradis, Carita. 2003. “Where does metonymy stop? Senses, facets and active zones”. The Department of English in Lund: Working Papers in Linguistics 3.77–91. (Also in Metaphor and Symbol 19:4.245–264.) Paradis, Carita. 2005. “Ontologies and and construals in lexical semantics.” Axiomathes 15.541–573. Reddy, Michael. 1979. “The conduit metaphor”. Metaphor and Thought ed. by Andrew Ortony, 284–324. Cambridge: Cambridge University Press. SAOB = Ordbok över svenska språket utgiven av Svenska Akademien. 1889–. Lund. Schmid, Hans-Jörg. 2000. English Abstract Nouns as Conceptual Shells: From corpus to cognition. (= Topics in English Linguistics, 34.) Berlin & New York: Mouton de Gruyter. Söderwall, K. F. 1884–1918. Ordbok öfwer Svenska Medeltidsspråket. Lund. Talmy, Leonard. 2000. Toward a Cognitive Semantics, 2 vols. Cambridge, Mass.: The MIT Press.

CLITIC PLACEMENT IN OLD AND MODERN SPANISH A DYNAMIC ACCOUNT

MIRIAM BOUZOUITA & RUTH KEMPSON Kings College, London Introduction This paper addresses clitic placement in Old Spanish (12th century–mid 15th century) in finite main clause environments with the goal of using Dynamic Syntax (DS; Kempson & al. 2001; Cann & al. 2005) to explore a diachronic account of the differences found between the clitic distribution in Old Spanish (OSp) and Modern Spanish (MSp). The aim of this paper is twofold. On the one hand, we provide a synchronic account for both Old and Modern Spanish clitics. In addition, we model the diachronic changes that took place, which can be represented as progressive lexical simplifications. 1.

Clitic placement in Old Spanish Unlike MSp, OSp clitics occur both preverbally (proclisis) and postverbally (enclisis). Preverbal clitics need not be adjacent to the verb, a phenomenon called ‘interpolation’. Further, unlike MSp, no clitic occurs first in a sentence. This restriction on sentence-initial clitics is known as the ToblerMussafia law. 2.

Strict proclisis constructions: Negation, wh-questions and focus There are only three kinds of root clauses in which OSp clitics continuously occur in a proclitic position throughout the whole period between the 12th and the 16th century, namely main clauses with negation, wh-questions, and construals containing a left-dislocated focused object NP which is not coreferential with the clitic (Nieuwenhuijsen 1999; Granberg 1988:131–136 inter alia):1 2.1

1

For each OSp example we will provide the century from which it comes between brackets. For visual clarity, the clitic will be in bold.

254

MIRIAM BOUZOUITA & RUTH KEMPSON

(1)

Ove famne e nom diestes a comer had-I hunger and not-CL gave-you to eat “I was hungry and you did not feed me.” (XIII; Granberg 1988: 131)

(2)

Qui los podrie contar? who CL could-he count “Who could count them?” (XIII; Fontana 1993:270)

(3)

A to linnaje la daré to your lineage CL give-will-I “I will give it to your lineage.” (XIII; Granberg 1988:135)

2.1.1 Strict enclisis constructions: Verb-initial and vocative clauses. The root clause environments in which clitics appear in enclisis in the 13th century are verb-initial clauses and clauses commencing with a vocative (Nieuwenhuijsen 2 1999; Granberg 1988:152–155, inter alia). (4)

Respondiol don Pelayo en guisa […] answered-he-CL don Pelayo in this-way “Don Pelayo responded him […].” (XIII; Fontana 1996:41)

(5)

Amigos, ruegouos por Dios […] friends, beg-I-CL for God “Friends, I beg you for God’s sake […].” (XIII; Granberg 1988: 152)

As concerns the shift from enclisis to proclisis, which was taking place in the Middle Ages, the earliest examples of sentence-initial clitics are found in the 15th century (1438) and they occur in direct style dialogue, which indicates that, in colloquial language, sentence-initial clitics may have been widely used.3 As with verb-initial main clauses, paratactic root clauses commencing 2

Proclisis is found in clauses commencing with a vocative if the vocative element is preceded by an adverb that normally is followed by a preverbal clitic such as, for example agora “now”. Since we decided to limit the scope of this account by excluding adverbs, these vocative proclitic examples have not been taken into account. In this preliminary account, coordination is also excluded. 3 As Granberg (1988:246) noted, it is difficult to determine when exactly this shift started taking place on basis of written corpora, due to the existence of prescriptive rules:

CLITIC PLACEMENT IN OLD AND MODERN SPANISH

255

with a verb seem to admit proclitic object clitics in the 16th century. The same applies for clauses commencing with a vocative expression. Pro-/enclitic constructions: Clitic left-dislocation and preverbal subjects In some root environments, OSp clitics oscillate between pro- and enclisis, namely in Clitic Left-Dislocation structures (CLLD) and in clauses containing an expressed preverbal subject. As concerns the CLLD constructions, in the 13th and 14th c., object clitics occur predominantly in post-verbal positions unless the dislocated NP contains a quantifier such as todos “all” or am(b)os “both” (Rivero 1986; Ramsden 1963:85–86; Granberg 1988:146–152): 2.2

(6)

E estas pazesi traxolasi marutas […] and these peaces brought-he-CL Marutas “And Marutas brought this peace […].” (XIII; Fontana 1993:266)

(7)

Con el su manto a amasi lasi cubrió with the his mantle to both CL covered-he “He covered both with his mantle.” (XIII; Ramsden 1963:86)

Aside from the CLLD constructions involving todos/ambos, the earliest cases in which CLLD occurs with proclitic object clitics are found in the 15th c. With respect to clitic behavior after preverbal subjects, Granberg (1988: 195–227) showed that the position of the OSp clitic is determined by the presence or absence of emphatic stress on the subject, as in Modern Galician: clitics occur postverbally unless the subject is highlighted by emphatic stress.4 (8)

E yo and I

donna doña

Maria Maria

que uendi who sold-I

la the

sobre above

dicha eredad la otorgo […] spoken inheritance CL confer_on-I “And I, doña Maria who sold the above mentioned inheritance confer it on […].” (XIII; Granberg 1988:205) “You cannot say […] that what some careless people or foreigners say, commencing their discourse with those enclitics like in: te vas? […] or se va, which is intolerable, but the proper use is vaste?, voime, vase, vanse; that all are warned with this […]. The rule says that if the verb introduced the discourse, the pronouns are postponed to it […].” (Our translation from Correas 1626 apud Granberg 1988:246) 4 Example (8), taken from a legal document, contains an oath while (9) proceeds from a document that recounts a series of events.

256


(9)

el arçipreste fiçolo assi the archpriest did-he-CL like-this “The archpriest did it like this.” (XIII; Granberg 1988:209)

In these construals as well, there is a shift from enclisis to proclisis starting in the 14th century. Dynamic Syntax analysis In this section we will give a DS analysis for OSp object clitics. We will also outline a possible analysis for Renaissance Spanish (RSp) and sketch an account of the OSp–RSp shift. 3.

A brief introduction to Dynamic Syntax The novelty of DS is that it is a grammar formalism that reflects the dynamics of parsing. According to DS, the parsing of a natural language string is a monotonic tree growth process defined over the left-to-right sequence of the words, where the tree to be established as the overall goal represents one possible interpretation of that string and has some logical form of type t (Ty(t)) decorating its root node. This overall goal — represented as ?Ty(t) — of establishing a propositional formula as interpretation at the root node of a tree, e.g. Fo(Sneeze(Ruth))) as interpretation for Ruth sneezes, will be achieved through the annotations of the nodes of the tree, each node being decorated with a subterm of the formula. The tree is arrived at by imposing additional subgoals such as ?Ty(e  t) for a predicate, ?Ty(e) for an individual-denoting expression, and then using the words in the string in order to create the appropriate decorations. Transitions from one partial tree to another are licensed by the interaction of lexical, computational and pragmatic rules. At any interim stage, the tree will be in some way not fully specified. The primary tools required for this account of clitics are the concepts of ‘underspecified formula’ value, ‘unfixed node’, and ‘linked structure’. Underspecified formula values are projected from pronouns, which are lexically defined as providing a metavariable place-holding device instead of a full content expression. This placeholder is then assigned a value either from context or during the construction process. 3.1

3.1.1 Building unfixed nodes from the left periphery. Unfixed nodes are used to represent positional underspecification at an early stage of the parsing process, as the structural analogue of pronoun construal: the structural relation of the unfixed node to the other nodes in the tree structure is not known at the point at which the left-peripheral expression is parsed. This analysis is used to


257

replace accounts of long-distance dependency and other movement phenomena: the left-peripheral position marks the introduction of the appropriate term into the structure albeit at an unfixed node, the position of the ‘trace’ in movement accounts is the point in the left-right interpretation process at which the contribution of the unfixed node is determined. The introduction of an unfixed node is ensured by the ‘*Adjunction’ rule which introduces a node that is characterized as being dominated by a node a (Tn(a)) and requiring a fixed tree node position (? xTn(x)). This new node is required to be decorated by an argument node (marked ?Ty(e)).5 Sentences containing left-dislocated items such as Mary, John likes will be analyzed in these terms as in Figure 1, which represents the point after the verb has been processed:

Figure 1: Parsing Mary, John likes.

With the pointer at the open object node which the lexical actions of the verb have introduced, there remain some outstanding requirements: a requirement for a fixed tree node position (?xTn(x)) on the unfixed node and a requirement for a construal of type e (?Ty(e)) on the object node. These two requirements can be solved simultaneously by unifying the not yet annotated object node and the unfixed node through a process called ‘Merge’. Once all terminal nodes are decorated, rules for evaluating the tree provide annotations for nonterminal nodes, leading to a tree that is decorated at the top node with the formula value Fo(Like(Mary)(John)) of type t (Ty(t)). 3.1.2 Building linked structures from the left periphery. Another primary tool required for DS is the concept of ‘linked structures’. These structures involve the development of two separate trees connected by a ‘link relation’, with a requirement for a shared term in each of the two trees. The rule of ‘Link Ad5

Tn(a) means, by definition, “somewhere dominating the current node is a tree node with label a”. Ty(e) picks out individual-denoting expressions. In any partial node, there is a pointer ◊indicating the node currently being developed.

258


junction’ introduces a new tree across a ‘link relation’ from a head node in some partial tree structure to a new tree with requirement ?Ty(t).6 It will also impose a requirement on the local tree for a copy of the head node formula α somewhere within the newly introduced tree, forcing an anaphoric link between the two structures. This concept of ‘linked structure’ is used for relative clauses and topic constructions. In the case of relative clauses, it is, in English, the relative pronoun which provides the required copy. Figure 2 shows the partial tree for the construal of a relative clause at the point at which the relative pronoun who has been uttered and parsed.

Figure 2: Parsing Mary, who in Mary, who smokes, is sick.

Note that in this example the link relation is built from the node annotated with formula Fo(Mary) to a new tree, the top node of which is connected to an unfixed node, introduced by the process of *Adjunction. The parsing of the relative pronoun ensures a flow of information between the two trees and will therefore fulfill the requirement for a copy of the head formula Fo(Mary) (?Fo(Mary)) since it constitutes an anaphoric device. In the case of topic structures, the pronoun has to provide the copy, as in Ruth Kempson, she talks too fast. 3.1.3 Building unfixed nodes and linked structures from the right periphery. Both these strategies of building paired linked structures and introducing unfixed nodes may also be applied in the later stages of processing an utterance. Building linked structures at the right periphery is used to model the ‘background topic’ effect in sentences such as She talks too fast, Ruth Kempson. In these construals, the pronoun, having lexically introduced a place-holding metavariable as a decoration on the subject node, is identified contextually. Then the subsequent full NP is interpreted as providing the decoration for a separate linked tree. This formula and the metavariable already identified are 6

The modal operator and its inverse describe this relation.


259

interpreted as coreferential, thus ensuring that the paired linked structures share a term. The use of unfixed nodes at a late stage of the interpretation process can be illustrated by English expletive pronouns, as in It is likely that I am wrong. In these construals, the pronoun introduces a place-holding device (the lexically provided metavariable). This is replaced during the construction process by the rule of Merge, which applies after the sequence I am wrong has been used to decorate an unfixed node, unifying that propositional structure with the subject node, thereby updating the metavariable. The updating of the unfixed node within the structure does not itself require the presence of an anaphoric device, but the interaction of Merge, and updating a metavariable is not precluded, in this case allowing the replacement of that metavariable with structure. The only difference between a regular anaphoric device and the expletive pronoun is that the metavariable which the latter projects can be replaced by such tree-structure rather than merely some formula value. This difference is secured by presuming that the substitution of the metavariable projected by regular anaphoric expressions must be a terminal node in the resultant tree (Cann & al. 2005). Since this is a minor difference in lexical specification, we can model the development of anaphoric expressions into expletive devices by simple loss of this restriction. 3.2 A Dynamic Syntax account of Old Spanish clitics 3.2.1 Proclisis triggers: Unfixed nodes. The main claim about proclisis in OSp, we propose, is that if a clitic follows an expression interpreted as decorating a left-peripheral unfixed node, the clitic will appear in proclisis. 7 Given the independent motivation for unfixed nodes in DS analyses, this claim leads us to expect two environments in particular as inducing strict proclisis: (i) whquestions and (ii) left-dislocated focused NP constructions.8 3.2.1.1 Wh-questions. English wh-questions in which the wh-expression appears sentence-initially as in Who did Bill see?, are modeled in DS as projecting a metavariable WH that annotates an unfixed node (Kempson & al. 2001:150–189). As in Mary, John likes, the unfixed node will merge with the object node. OSp wh-questions are very similar to English ones:

7

This account is developed in detail in Bouzouita (2002, 2005). Negation also induces proclisis. We will not provide a detailed analysis for negation since this issue has not been addressed yet in the DS framework. For now, we will use the feature [NEG +] to indicate the presence of a negation operator in a clause. 8

260


(10) ¿Qué me darás? what CL give-will-you “What will you give me?” (XIII; Granberg 1988:132) (11) Qui los podrie contar? who CL could-he count “Who could count them?” (XIII; Fontana 1993:270) Example (10) contains a wh-element which functions as an object while the wh-expression in (11) is a subject. Since this object wh-expression projects a left-dislocated element, it will annotate an unfixed node, as in English. Initial wh-elements that function as subjects will also decorate unfixed nodes in OSp, due to the fact that, unlike English, OSp is a subject pro-drop language which can be represented formally in DS by letting expressed subjects annotate either unfixed nodes or linked structures while the parsing of the verb will project the full predicate-argument structure. Given the analysis of wh-initial questions as decorating an unfixed node, proclisis seems to be triggered by the presence in the tree of a left-peripheral unfixed node. Accordingly, we will define the preverbal clitic as itself introducing lexically a locally unfixed node, in addition to an already present unfixed node.

Figure 3: Clitics decorating a locally unfixed node.

As shown in Figure 3, the object metavariable (Fo(U)) from the clitic in (11) decorates the locally unfixed node and will unify with the object node that has been introduced by the parsing of the verb.9 The subject metavariable (Fo(WH)), on the other hand, will merge with the subject node.

9

This tree display ignores the auxiliary podrie for simplicity.


261

3.2.1.2 Focused NP constructions. OSp sentences containing leftdislocated focused NPs will be analyzed in the same way as English focus constructions, as in Mary, John likes (see Figure 1). There we saw that *Adjunction and Merge were used to model so-called ‘focus movement’. Likewise, OSp focused NPs will decorate unfixed nodes which will unify with the direct or indirect object node projected by the verb. These phenomena then also corroborate the hypothesis that OSp clitics occur preverbally when preceded 10 by a left-peripheral expression taken to decorate an unfixed node. 3.2.2 Enclisis triggers. Unlike the proclisis case, enclitic pronouns need no special statement: like postverbal object NPs, they decorate a fixed object node. We will look at (i) verb-initial root clauses, and (ii) main clauses introduced by vocative phrases. 3.2.2.1 Verb-initial root clauses. In the 13th c., the Vfinite –CL configuration is the predominant one for root clauses. Since in this configuration the clitics appear in the position in which full NPs usually occur, we propose that these OSp clitics decorate in the tree the fixed object node on which the pointer has been left and which has been introduced lexically by the verb. As regards paratactic main clauses commencing with a verb, we consider them as distinct trees of type t. In these cases we expect that the clitic will appear in enclisis (as long as there is no negation marker) since prior to parsing the verb, the tree is by definition empty and so will not have an unfixed node, which is the trigger for proclisis. 3.2.2.2 Root clauses commencing with vocatives. In our view, main clauses commencing with vocative phrases can be analyzed, like CLLD structures, as instances of linked structures:11 (12) Amigosi,ruegouosi por Dios […] friends, beg-I-CL for God “Friends, I beg you for God’s sake […].” (XIII; Granberg 1988: 152)

10

Further evidence for this account is strict proclisis in OSp relative clauses. See Section 2.1.2 for an account of English relative pronouns as decorating an unfixed node. 11 Although the vocative expression and the clitic are coreferential in example (12), this is not always the case. Such instances are then analyzed as linked structures which do not impose a requirement for a copy of the head node formula.

262


Recall the concept of linked structures. Topic constructions are analyzed as a pair of linked trees: the left-peripheral NP projects a tree of type e (with a root node annotated by a formula α) which is linked to the top node of another tree of type t, the development of which is constrained by the requirement for the occurrence of the formula αsomewhere in this tree. We propose, that is, the same analysis for OSp CLLD structures and the vocative cases involving coreferential clitics. With respect to OSp clitic placement, the discussed vocative constructions and verb-initial root clauses both corroborate indirectly the hypothesis that proclisis takes place after any expression that induces the construction of a left-peripheral unfixed node since these construals do not involve unfixed nodes as tools for their account and, as expected on this analysis, are expressed using postverbal clitics. 3.2.3 The mixed situation: Pro- and enclisis triggers. The mixed situation is displayed by CLLD structures and preverbal subjects. Again the distribution is as the DS analysis would lead us to expect. 3.2.3.1 Clitic left-dislocations. The account so far proposed treats OSp CLLD constructions as a pair of linked structures (see (6)). This analysis, however, does not apply for the CLLD construals involving a quantifier such as todos “all” or am(b)os “both” (see (7)). In our view, these left-dislocated quantifiers are emphasized and thus, need to be analyzed as decorating unfixed nodes which will unify with the coreferential metavariable associated with the clitic through the application of Merge. Consequently, the OSp clitics arise in proclisis. 3.2.3.2 Expressed preverbal subjects. This assumption leads us to expect with Granberg (1988:195–227) that proclisis arises when the expressed preverbal subject is emphasized while enclisis appears when the stress lies on the verb. We have already seen that subject NPs in subject pro-drop languages can either decorate a linked structure or an unfixed node. The difference from English is that no pronoun will be required to establish the link-required copy, as it is the verb and its lexical specification that will ensure the presence in the tree of a metavariable in subject position, exactly as though a morphologically expressed pronoun were present. This metavariable will duly be replaced by a term which is identical to whatever decorates the linked structure, fulfilling its requirement for a shared term. So we expect subject NPs to be able to be construed as either focused or as a topic-presenting structure without any explicit anaphoric devices being present. The only reflex of this distinction in


263

construal will be the positioning of the object clitics, these being proclitic in the focused case, enclitic otherwise. The mixed effect is thus expected. 3.2.4 Old Spanish clitics 3.2.4.1 Lexical entry. In the light of the general proclisis pattern of occurring immediately after any left-peripheral expression that induces the construction of an unfixed node, we propose the following lexical entry for OSp clitic lo as illustrative of the general pattern. See figure 4.

Figure 4: Lexical entry for OSp accusative clitic.

This lexical entry makes pro- and enclisis complementary since it not only induces the construction of an unfixed node to be decorated with a metavariable in the presence of either a negative marker or an already constructed unfixed node but it also explicitly states that enclitic positions in those environments are ungrammatical. The trigger condition T is a specification ensuring that this use of the clitic only decorates a fixed node in a structure (one with an immediately dominating node). Notice how nothing additional needs to be said to ensure the Tobler-Mussafia law, which becomes in this account epiphenomenal. In a partial structure with only a predicate (as provided by the verb), 12 enclisis will be obligatory. 12

Unlike other accounts (e.g. Rivero 1986), this analysis does not reduce the Tobler-Mussafia effect to a purely phonological phenomenon.

264


3.2.4.2 Interpolation and multiple *Adjunction. In order to account for the interpolation cases, we apparently need multiple applications of the process of *Adjunction. Consider the following example of double interpolation: (13) Qui lo fer non quisiesse […] who CL do not wanted-he “Who were to refuse to do it […].” (XIII; Sánchez Lancis 1992: 327) Qui, the wh-element, decorates an unfixed node introduced by *Adjunction and, thus, triggers proclisis. Given the presence of this unfixed node, the clitic also annotates an unfixed node but this second unfixed node is induced by the lexical actions of the clitic. However, this unfixed node is in turn apparently followed by the construction of another unfixed node which the left-dislocated infinitive fer “do” will decorate, and which is also introduced computationally by *Adjunction. 13 Sketch for Renaissance clitics As mentioned earlier, during the Middle Ages a shift from enclisis towards proclisis took place and seems to have been almost completed at the beginning of the 16th century. Basically, all the OSp cases in which enclisis appeared during the Middle Ages acquired the possibility of occurring in proclisis as well. In other words, the restrictions that were imposed on the occurrence of proclitic pronouns in the 13th century no longer applied in the 16th century. Notice how, in the following lexical entry (cf., Figure 5), these subconditions for proclisis — which were present in the lexical entry for OSp clitics — are dropped, hence generalizing proclisis. Nonetheless, the restrictions for enclisis did not vanish. The diachronic shift from enclisis towards proclisis is thus modeled in this account as the simplification of the lexical characterization of the clitic. 14 However, it remains to be seen what triggered these lexical simplifications. 3.3

13

These data are problematic for Cann & al. (2005) who claim that only one node within a tree can be unfixed at any one stage in the construction process. However, they propose a range of types of structural underspecification, and it remains possible that (13) can be analyzed as involving more than one type of Adjunction operation. See Bouzouita (2005, forthcoming) for an account along these lines. 14 See Bouzouita (2005) where different hypotheses are explored.


265

Figure 5: Lexical entry for RSp accusative clitic.

3.4 A Dynamic Syntax account of Modern Spanish clitics 3.4.1 Modern Spanish: Standard Spanish data. In Modern (Standard) Spanish (MSp), only proclisis has survived the passage of time. (14) Lo

vi, a Juan saw-I to Juan “I saw Juan.” CL

vi saw-I “I saw him.”

(15) *(Lo) CL

(16) Le

a él to him

di el libro gave-I the book “I gave Juan the book.” CL

a Juan to Juan

As regards accusative clitic doubling, MSp can double full NPs, if an intonational break precedes the full NP.15 There is a specificity restriction: only refer16 entially construed NPs are allowed. However, doubling with overt, so-called strong pronouns él/ella, is obligatory. With the dative counterpart, there is e15

In this MSp account, we have omitted CLLD structures, since the main focus of this paper is the diachronic change in clitic placement in finite main clauses. 16 Indefinites are also allowed, but, for simplicity, we shall take these here to be name-like under this interpretation (Fodor & Sag 1982), though see Kempson & Meyer-Viol (2004).

266


qually obligatory doubling with strong pronouns, but with full NP clitic doubling, there is no specificity restriction and no requirement for an intonational break. 3.4.2 Lexical entries for Modern Spanish. The analysis we propose for MSp clitics involves a further lexical simplification: compared to RSp, the enclisis trigger is lost. Therefore, in MSp, clitics decorate an unfixed node irrespective of whether any other (unfixed) node has already been introduced, as in RSp.

Figure 6: Lexical entries for MSp accusative and dative clitic.

Modeling the dative clitic pronoun as lacking one restriction retained by the accusative clitic can bring out the distinction between dative and accusative. The accusative clitic remains a full pronoun, only decorating a node in the tree that is terminal.17 All instances of clitic doubling must therefore involve a pair of linked trees, linked only through the sharing of a term, a sharing which is secured through the presence of the pronoun. The specificity restriction follows, as quantifiers cannot bind across from one tree structure to another, so not across linked trees. The dative, on the other hand, lacks this terminal node restriction. This allows it to decorate an unfixed node which can then unify with any other node that may get introduced into the structure, whatever the complexity of the structure with which it unifies.18 So, for example, in a derivation such as (16) in which the clitic pronoun precedes the verb and the full doubled NP follows it, the clitic will be taken to decorate an unfixed node which then merges with the indirect-argument node projected by the verb, and this node may then unify with some node introduced by late application of 17

It remains to be seen whether this difference is due to a diachronic change or whether it stems from some inherent difference between the accusative and dative case, the dative case signaling a semantic relation whose status as adjunct or argument is unclear. 18 In this paper, names have been analyzed as semantically simple terms. In a fuller account, these would be analyzed as uniquely referring terms with internal structure.


267

*Adjunction. Hence the lack of any necessity of an intonational break.19 It is notable that on this analysis, there is no need to invoke ambiguity between two uses of the dative clitic as a regular anaphoric expression on the one hand and some quasi-agreement device on the other, since the analysis of the dative clitic as lacking a terminal node restriction, immediately predicts its broader distribution than the accusative clitic which retains such a restriction. The idiosyncrasy of the obligatory doubling with a strong pronoun across all dialects is captured by analyzing these strong pronouns as lexically defined only to decorate a linked structure or an unfixed node, in effect encoding their use for specialized pragmatic effects. Despite appearances, they never decorate a node as introduced into the structure through lexical specification by the verb. Hence the obligatory presence of the clitic, itself decorating an unfixed node, but merging with the appropriate argument node of the verb’s specification, and expressing the basis for the shared-term restriction imposed by the link relation. Conclusion Though this short sketch is no more than the beginning of an account of clitics, the general advantage of the dynamic perspective of DS is apparent. With just three concepts of tree growth, all independently motivated, we have provided a principled account of the heterogeneous positioning of OSp clitics; and thus, unlike other accounts, we do not have to invoke additional machinery or special projections. Furthermore, we outlined an account of the diachronic change in clitic placement in finite main clause environments as a process of lexical simplification, namely the reduction of complexity of possible triggers, which leads to a contracted distribution, since enclisis is considered ungrammatical in MSp. The difference between the MSp accusative and dative clitic has been captured by the absence of a terminal node restriction for the dative clitic, which is the basis of its more generalized distribution in clitic doubling. This analysis signally does not necessitate the invocation of ambiguity in the clitic itself. 4.

19

In Río de la Plata Spanish (Uruguay/Argentina), the accusative clitic is also losing this terminal node restriction, no longer requiring any intonational break between the string containing the clitic pronoun and that of the doubled NP, but nevertheless retaining a specificity restriction (Suñer 1988).

268


REFERENCES Bouzouita, Miriam. 2002. Clitic Placement in Old and Modern Spanish: A dynamic account. MSc Dissertation, King’s College London. Bouzouita, Miriam. 2005. “Variation and change in Spanish clitic placement: From Medieval to Renaissance Spanish”. Paper presented at the Dynamic Syntax Workshop, King’s College London, May 2005. Bouzouita, Miriam. Forthcoming. “Variation and change in Spanish clitic placement”. Papers in Linguistics of the University of Manchester Proceedings of the 14th Postgraduate Conference in Linguistics, 4th March 2005, The University of Manchester ed. by Nuria Yañez-Bouza. Manchester: Manchester University Press. Cann, Ronnie, Ruth Kempson & Lutz Marten. 2005. The Dynamics of Language. Oxford: Elsevier. Fodor, Janet Dean & Ivan A. Sag. 1982. “Referential and quantificational indefinites”. Linguistics and Philosophy 5.355–398. Fontana, Josep M. 1993. Phrase Structure and the Syntax of Clitics in the History of Spanish. PhD Dissertation, University of Pennsylvania. (http:// repository.upenn.edu/dissertations/AAI9331867/.) Fontana, Josep M. 1996. “Phonology and syntax in the interpretation of the Tobler-Mussafia Law”. Approaching Second: Second position clitics and related phenomena ed. by Aaron L. Halpern & Arnold M. Zwicky, 41–83. Stanford, Calif.: CSLI. Granberg, Robert Arthur. 1988. Object Pronoun Position in Medieval and Early Modern Spanish. PhD Dissertation, University of California Los Angeles. Ann Arbor, Mich.: United Microfilms International. Kempson, Ruth, Wilfried Meyer-Viol & Dov Gabbay. 2001. Dynamic Syntax: The flow of Language Understanding. Oxford: Blackwell. Kempson, Ruth & Wilfried Meyer-Viol. 2004. “Indefinites and scope choice”. Descriptions and Beyond ed. by Marga Reimer & Anne Bezuidenhout, 558–84. Oxford: Oxford University Press. Nieuwenhuijsen, Dorien. 1999. “Cambios en la colocación de los pronombres átonos en la historia del español”. (= Estudios de Lingüística Española, 5.) ISSN: 1139-8736, Depósito Legal: B-3919s8-99. (http://elies.rediris.es/ elies5/.) Ramsden, Herbert. 1963. Weak-pronoun Position in the Early Romance Languages. Manchester: Manchester University Press. Rivero, María Luisa. 1986. “La tipología de los pronombres átonos en el español medieval y el español actual”. Anuario de Lingüística Hispánica 2.197–220. Sánchez Lancis, Carlos. 1992. “La interpolación de complementos entre el pronombre personal átono y el verbo en español medieval”. XXe Congrès International de Linguistique et Philologie Romanes, Zurich, 6–11 April 1992 ed. by Gerold Hilty, vol. II, 324–334. Tübingen: Franke Verlag. Suñer, Margarita. 1988. “The role of agreement in clitic doubled constructions”. Natural Language and Linguistic Theory 6.391–434.

GRAMMATICALISATION AS CONTENT REANALYSIS THE MODAL CHARACTER OF THE DANISH S-PASSIVE

LARS HELTOFT Roskilde University Introduction The history of the inflectional passives of the Scandinavian languages represents a particularly interesting chain of grammaticalisation processes, from prehistoric times to around 1600. The more recent changes, undergone particularly by Danish, do not lend themselves to analyses along the lines of the traditional models of grammaticalisation (as summarised in e.g. Hopper & Traugott 1993, Heine, Claudi, & Hünnemeyer 1991), nor can they be accounted for within antistructuralist models (see for instance Bybee, Pagliuca & Perkins 1993). Rather, they will call for quite different, especially structural models that include ‘content reanalysis’ as a central factor. The changes studied in the present contribution do not represent a semantic bleaching process, but a process that adds more polysemy to a morphological form and thereupon reinterprets it as a mood, in addition to its being a member of the category of voice. Nor does this process involve any other changes at the level of expression than analogical ones and simple generalisations. The first part of this development is a classical reinterpretation of a lexical category, the reflexive pronoun, as an inflectional category. Let us restrict ourselves to the 3rd person of the reflexive pronoun, Germanic *sik, Gothic sik, OIcel. sik, as in Old Icelandic geyma sik “guard oneself, hide oneself” > geymask (i.e. geyma-sk), an example of the so-called middle or mediopassive inflectional form of Old Scandinavian verbs. In East Norse, the mediopassive endings reduced to -s and were even more tightly grammaticalised in passives, e.g. Swedish hittas (hitta-s) “be found”, dödas (döda-s) “be killed”, and skjutas (skjuta-s) “be shot”. The second part of the development, the one that I shall focus on in this contribution, is a process that results in a split between Danish and Swedish with respect to the passive. Danish undergoes a thorough restructuring, a semantic reorganisation of already existing inflectional paradigms. 1.

270

LARS HELTOFT

The classical model of grammaticalisation will not provide us with an adequate description of this change since the change makes use of already existing inflectional expression material and reorganises it semantically. Rather, the change calls for a model the core of which is ‘content reanalysis’. Many content reanalyses do follow the lines sketched in the classical model, but others certainly do not. What is common to all processes that lead to grammatical status, or lead to changes in grammatical status, is content change and thereby paradigmatic change (cf. Heltoft 1996, Heltoft & al. 2005). An understanding of grammatical change that is adequate also to the facts scrutinised in the present case study is the one proposed in Andersen (2001a). However, as far as I can judge, Andersen’s specific typology of morphological change (cf. Andersen 1980) does not offer any parallels to the present type of content reorganisation. The results of the changes: The Danish and the Swedish s-forms To understand the development of the inflectional passive in -s in Danish we shall need an overview of the differences between Danish and Swedish in this respect. Danish shows a particularly interesting reanalysis of the s-passive that added functions of mood to this form. Example (1) illustrates this modal character of the ‘s-mood’: 2.

(1)

For Paludan-Müller er der ingen tvivl. Anne Meinstrup var en gammel, stiv, hovmodig kvinde uden situationsfornemmelse. Hun kæftede op og var derved selv skyld i ”To Paludan-Müller’s mind there is no doubt. A.M. was an old, stubborn, haughty woman with no sense of situation. She shot her mouth off and was therefore herself the reason at hun blev myrdet. why she become.PAST murder.PERF.PTC why she was murdered.” Grev Christoffer var helt uden skyld. I hans fremstilling “Duke Christoffer had no guilt at all. In his exposition myrdes Fru Anne Meinstrup murder.PRES.S -MOOD Lady A.M. Lady A. M. is murdered a second time.”

for anden gang. a second time

GRAMMATICALISATION AS CONTENT REANALYSIS

271

Hendes ry som agtværdig kvinde hugges Her reputation as an honourable woman cut.PRES.S -MOOD lige så meget sønder, som hendes legeme to pieces just as much as her body

blev det become.PAST it

den dag på tinget i Ringsted. that day at the thing in R. “Her reputation as an honourable woman is just as much cut to pieces as her body was that day at the thing in Ringsted.” (Saxo 4, 1998:37) We shall need to explicate a number of subtle distinctions here. In the above text the present tense of the s-passive is used to mark a sentence as a part of the exposition by a narrator different from the speaker (viz. the historian PaludanMüller), as part of the intention objectified in his text. The two occurrences of the periphrastic passive, on the other hand, are used in a simple constative sense to indicate that the descriptions are the speaker’s own. This purely descriptive sense is not possible with the Modern Danish s-passive. At this point we find a striking difference between Modern Danish and Modern Swedish. Swedish has retained the simple descriptive sense of the spassive, as in the following example from Heltoft (1996): (2)

Hampus Broberg och Helena Hansson Hampus Broberg and Helena Hansson

häktades […] arrest.PAST.S-PASS

fem minutter över tio nästa morgon. five minutes past ten the next morning “Hampus Broberg and Helena Hansson were arrested five minutes past ten the next morning.” A Danish translator would have to use the constative periphrastic passive, (3): (3)

Hampus Broberg og Helena Hansson blev Hampus Broberg og Helena Hansson become.PAST anholdt arrest.PERF.PTC

[...] fem minutter over ti næste morgen. five minutes past ten the next morning.

This picture is confirmed in the Swedish (3’), but more complexity is added.

272

LARS HELTOFT

(3’) “Mrs. Heiberg’s two languages. One is the official one, appears in her memoirs. One life. The other one appeared in her life, men anas i det som skars bort i memoarerna. but is faintly seen in what cut.PAST.S-PASS out in the memoirs but is faintly seen in what is cut out in the memoirs.” (Regn Liv, 325) (3’’) men anes i det der blev skåret væk i memoirerne. but is faintly seen in what become.PAST cut.PERF.PTC out in the memoirs (Regn Liv, Danish translation, 89) In Danish, not only is it necessary to use the (descriptive) periphrastic passive form owing to the semantics here, but also ablaut verbs (strong verbs) have no (non-descriptive) inflectional counterpart to the past of the periphrastic passive in the first place. Swedish has skars “was cut”, but there is no Danish *skares. (Notice that, for reasons to become clearer later, the Danish form would be expected to come out as disyllabic, not monosyllabic as in Swedish). This is the general picture: Swedish bars “was carried”, förbjöds “was forbidden”, greps “was taken, arrested”, sköts “was shot”, sjöngs “was sung”, but Danish *bares, *forbødes, *grebes, *skødes, *sanges. The weak verbs, however, do have an inflectional counterpart to the past of the periphrastic passive: afsluttedes “was finished”, forfremmedes “was promoted”, fragtedes “was transported”, hentedes “was brought”, hørtes “was heard”, protesteredes “protests were put forth”, sendtes “was sent”, serveredes “was served”, svaredes “was answered”, uddeltes “was delegated”, ødelagdes “was destroyed”, åbnedes “was opened”, etc. Any exposition, let alone explanation, of this development must account for this apparent defect or asymmetry of the Danish system. Later I shall offer such an account, but before that we shall need a few extra examples of the modal uses of the s-passive, see (4)–(7) from Heltoft (1996) and Heltoft & Falster Jakobsen (1996). A frequent use or sense of the s-mood of the passive is the generic or normative one, compare (4) and (5): (4)

tales der dansk i Skåne? speak.PRES.S-MOOD there Danish in Scania? “Is Danish spoken in Scania?”


(5)

bliver der talt become.PRES there speak.PERF.PTC “Is Danish spoken in Scania?”

273

dansk i Skåne? Danish in Scania?’

Example (4) is the normative reading: the question raised is whether Danish is the linguistic norm in Scania; whereas (5) is the simple descriptive reading, asking whether it occurs that Danish is spoken there. The examples in (6) and (7) constitute another next-to-minimal pair: (6)

FIOLTEATRET LUKKES The Fiol Theatre close.PRES.S-MOOD “The Fiol Theatre is to be closed.”

(7)

FIOLTEATRET BLIVER LUKKET The Fiol Theatre become.PRES close.PERF.PTC “The Fiol Theatre will be closed.”

The two pairs of examples in (4)–(5) and (6)–(7) illustrate the difference between the s-passive and the perifrastic passive in Danish. Both are passive voices, but they do not differ along parameters normally associated with voice, neither with respect to subject demotion nor with object promotion. Heltoft & Falster Jakobsen (1996) have suggested and developed their own version of an oral suggestion by Durst-Andersen that this opposition should be treated as a mood distinction. Example (6) is a newspaper headline. The sense of lukkes “will be closed” or “is to be closed” (the ‘s-mood’ of the passive paradigm) is modal in so far as the utterance involves a point of view or an intention different from the speaker’s point of view or intention. The closing of the theatre is somebody else’s plan or intention or possibly reported from somebody else, just like in example (1) where the s-passive indicated that the story of the murder of Lady Meinstrup was the intention of an inserted narrator. The example in (7), on the other hand, places the intention solely with the speaker: it is a simple declarative prediction, the future sense being the normal interpretation of dynamic verbs of change in the present tense. The view of Heltoft & Falster Jakobsen (1996) is that “this opposition has to do with the question of whether the consciousness of other persons is involved or not”, and they “speak of the involvement of particular intentions, speaker and non-speaker” (quoted from summary in Heltoft 1996: 488). The Danish s-mood has grammaticalised what, in the terminology of Bech (1952), we could call an ‘external modal factor’, that is, in the present context, a non-speaker bound modality. It follows that the s-mood cannot be

274

LARS HELTOFT

used in describing simple factual events, since such simple constatives are only related to the speaker.1 Middle voice and passive voice The Old Scandinavian middle or mediopassive is one category with four contextually determined and fully productive senses (cf. Dyvik 1980; Heltoft 1996). Its four semantic subcategories are the following: 3.

a. reflexive: geymask “guard oneself, hide”, setjask “sit down”, etc. b. reciprocal: berjask “fight one another”, mœtask “meet”, etc. c. intransitive: andask “die”, takask “begin”, týnask “perish”, etc. d. passive: búask “be equipped”, synjask “be denied”, etc. The passive reading of the mediopassive is a relatively late development, possibly formed on the basis of learned Latin models. It coexists with the common Germanic periphrastic passive, of proto-Germanic origin. In Swedish, the s-passive became the unmarked option, whereas the periphrastic passive seems restricted to descriptions of situational events and processes. This means that Modern Swedish is closer to the medieval situation than Modern Danish (Brøndum-Nielsen 1928–1974). In both Danish and Swedish, the mediopassive is no longer a single category. The reflexive sense has been restricted to the full reflexive pronominal construction V + reflexive pronoun. Restricting ourselves to Modern Danish, this language has gemme sig “hide (oneself)” and sætte sig “sit down”, etc., but forms like *gemmes and *sættes have no reflexive sense. The reciprocal and the intransitive senses are now unproductive, leaving behind large classes of verba deponentia, with no active counterparts, semantically distinct from real passives, but inflectionally identical. Danish has reciprocals such as enes “agree”, mødes “meet (each other)”, skændes “quarrel”, slås “fight”, and intransitives such as dages “break (of the day)”, grønnes “turn green”, slukkes “die out (of fire, light)”, tændes “to light, catch fire”, ældes “grow old”, and ændres “change”. 2

1

In Swedish, the s-passive is in a classical sense the unmarked member of the category. The periphrastic passive is normally used to describe single events and is the marked, or ‘intensive’, member of the category; the s-passive has the descriptive sense and use as well, but in addition, it has a set of generic, normative, and hortative senses. The Danish system is less straightforward with respect to markedness, but could probably be viewed as an exclusive opposition, i.e. an opposition between a so-called ‘+term’ and a ‘–term’ (Hjelmslev 1935). 2 Hopper & Traugott (1993:153) get this situation wrong.


275

These unproductive classes must be kept apart from the still productive passive sense. Traditional Danish grammar (e.g. Oxenvad 1934) has confused these issues, e.g. by claiming that s-passives should be agentless. The following example (8) has been analysed as an agentless passive, when in fact it is an intransitive. (8)

Stjernerne tændtes stars.DEF .PL light.PAST-S “The stars lit.”

In brief, there is no reason to believe that inflectional and periphrastic passives should differ along the parameters of voice. The probable outset of the split between the voice systems of Danish and Swedish is a reading of the passive with a general and therefore very often deleted agent. Such general readings involve generally accepted knowledge, and therefore speaker-external consciousness, in the sense that other people than the speaker would vouch for their truth, too. They are readily interpreted as normative. Normative readings, too, involve speaker-external viewpoints, since a norm derives its modality from say, society, not from the speaker. The existence of a speaker-external modal factor has been generalised to ‘viewpoints’ in general (consciousness and intentions), including the viewpoint of particular non-speakers. The semantic chain of generalisation and reinterpretation in Danish can be assumed to obey the following pattern (cf. Heltoft 1996): (9)

1. Generalisation (to include the modal senses in the meaning potential of the s-form): Passive > General passive > Normative sense > Non-speaker bound sense: speaker-external views and knowledge > Views and intentions of particular non-speakers, including inserted narrators. 2. Reinterpretation (only modal senses preserved): General passive > Normative sense > Non-speaker bound sense: speaker-external views and knowledge > Views and intentions of particular non-speakers, including inserted narrators.

The Swedish pattern does not involve any specific senses of non-speaker boundness: (10) Passive > General passive > Normative sense

276

LARS HELTOFT

The lack of the s-mood in the past tense of strong verbs in Danish As stated above, Danish ablaut verbs (strong verbs) have no inflectional counterpart to the past of the periphrastic passive. This has no immediate connection with tense semantics, since Danish readily forms inflectional past tenses of weak verbs (see Section 2). Due to their highly specific senses, past tense forms will be expected to occur at a low frequency rate and to occur in specific or marked contexts, but the point is that examples are relatively easy to find or that such contexts come readily to mind. 4.

(11) […] den anklagede skulle tage et stykke glødende jern og bære det ni skridt [...]. Bagefter blev han iført en vante, som become.PAST dressed in a glove, which forsegledes seal.PAST.S -MOOD

og først and not

aftoges take.off. PAST.S-MOOD

efter nogle dages forløb. “[…] the accused was to take up a piece of fervent iron and carry it nine steps […]. Afterwards he had a glove put on, which was sealed and only taken off after some days.” (Skalk 1994.4:28) The text containing the example in (11) is a description of the medieval procedure of the ordeal by fire. There are three past passives. The first one is the periphrastic mood of iføre “clothe, dress in”, the second the s-mood of forsegle “seal”. The third one — which at first glance is a counterexample to the claim that Danish does not form any past s-moods from strong verbs — is the past smood from a strong verb (of the 6th ablaut class, cf. below) aftage “take off”. Let us take an overview of the morphological situation, see Table 1. Conjugation classes 1st class

2nd class

3rd class

4th class

Infinitive bide ”bite” gribe ”catch” ride ”ride” forbyde ”forbid” lyve ”lie” skyde ”shoot” drikke ”drink” stjæle ”steal” synge ”sing” bære ”carry” skære ”cut”

Present s-Passive bides gribes rides forbydes lyves skydes drikkes stjæles synges bæres skæres

Past s-Passive *bedes *grebes *redes *forbødes *løjes *skødes *drakkes *stjales *sanges *bares *skares

Swedish Past s-Pass. bets greps reds förbjöds ljögs sköts dracks stals sjöngs bars skars


5th class 6th class

277

bede ”ask for” bedes *bades bads give ”give” gives *gaves gavs jage ”hunt, expel” jages [weak ] joges lade ”let” lades lodes (läts) tage “take” tages Toges togs Table 1: Overview of the morphology of the 6 conjugation classes.

The 6th class is an obvious exception adding extra complexity to the pattern to be explained below. It is deviant in allowing past s-mood forms, as already seen in (11). Similar descriptions of norms of the past coded by the s-mood of strong class 6 verbs are easy to construct, cf. (12)–(13): (12) Gravide tjenestepiger joges bort uden pardon. pregnant handmaids drive.PAST.S-MOOD away without mercy “Pregnant handmaids were driven away without mercy.” (13) Vanskabte børn toges Handicapped children take.PAST.S-MOOD

fra moderen, from their mother,

udsattes og efterlodes til en grum skæbne. expose.PAST.S-MOOD and leave.PAST.S-MOOD for a cruel fate ”Handicapped children were taken from their mothers, were exposed and left for a cruel fate.” At first glance, (14) is a possible exception to the pattern that class 5 does not form the s-mood. (14) Der gaves ingen anden udvej. there give.PAST.S-MOOD no other way out “There was no other way out.” Other seemingly possible exceptional forms (i.e. s-moods of strong verbs) are also in need of comment, see examples (15) and (16). (15) Der fandtes ikke mere mad i huset. there find.PAST.S-MOOD no more food in the house “There was no more food to be found in the house.” (16) Det sås hvordan sygdommen havde taget på ham. it see.PAST.S-MOOD how the disease had weakened him “It was apparent how the disease had weakened him.”

278

LARS HELTOFT

In (16), the s-mood conveys the sense of what was generally obvious about the health of the person described — it is not a description of a (single) situation, and thus the periphrastic form is not expected. The s-passives in the examples in (15)–(16) have a normative or generic sense, which is also found in Swedish (Holm 1952). This is the result of a development that Danish and Swedish have in common. The generic or normative sense does not refer to particular events. And even though the forms are passive, they never have a proper passive sense or function: gaves “existed, was at hand”, but not “was given”; fandtes “existed”, but not “was found, was discovered”; sås “could be seen”, but not “was spotted, observed, seen”. Note also that strong verbs form the original middle past tenses with the reciprocal and intransitive senses without any problem, see the list in (A). (A)

slås “fight each other” rives “fight scratching each other” strides “struggle” omgås “see each other”

slås rives strides omgås

sloges reves stredes omgikkes

In Modern Danish, there is no interpretation of these forms as passives: sloges “*was beaten”, reves “*was torn, demolished” (but Swedish: revs “was torn, demolished”), omgikkes “*was surrounded, circumvented”. It is beyond doubt that the problem is one of content, not of morphophonemic restrictions. A traditional account The outline by Brøndum-Nielsen (1928–1974) of the development of the passive system stresses another morphological feature, namely the generalisation in Danish of polysyllabic structure in the s-passive. In Old Scandinavian, there were two sets of s-passive-formation: one that we shall name the ‘thematic’ class, and another, the ‘athematic’ class. The inflectional system is given in Tables 2 (thematic) and 3 (athematic). These classes are fully distinct in the perfect participle (added for convenience). The thematic formation adds an extra syllable to the present tense indicative (‘extrasyllabic’) (Table 2). 5.

Thematic ōn-verbs ī an-verbs ē n-verbs

Infinitive kalla “call” tapa “lose” hitta “find” høra ”hear” sigia ”say” hafua “have”

Present Indic. kall-a-s tap-a-s hitt -i-r hør-e-s sig-i-s hafu-i-s

Present Subj. kall-i-s tap-i-s hitt-i-r hør-e-s sig-i-s hafu-i-s

Perf. Ptc. kallathær tapathær hittær hørthær sagat hafuat

Table 2: Thematic Old Danish weak verb classes, 3.SG . passive.


279

The result is that, synchronically, we find syncretisms between the indicative and the subjunctive forms of ī an- and ē n-verbs. With the ōn-verbs, the -a- element is the indicative or default marker, contrasting with -i-/-e-, the subjunctive marker. We have skipped the past tense in Table 2, since all weak verbs have thematic structure. Notice in passing that in later stages of Danish, also in the dominant Zealandic norm, all vowels in non-stressed syllables reduce to schwa. Later paradigms will not have vowel differences in the subjunctive vs. indicative opposition. The jan-verbs and the strong verbs are athematic, see Table 3. Athematic jan-verbs

strong verbs

Infinitive dylia “conceal” læggia “lay” kræfia “demand” synia ”deny” luka “shut” dræpa “kill” taka “take” gifua “give” swæris “swear”

Pres. Indic. dyl -s læg-s kræf-s syn-s luk-s dræp-s tak-s gif-s swær-s

Pres. Subj. dyl-i-s lægg-i-s kræf-i-s syn-i-s luk-i-s dræp-i-s tak-i-s gifu-i-s swær-i-s

Past Indic. – – – – lauk-s drap-s tok-s gaf-s sor-s

Table 3: Athematic Old Danish verb classes, 3.SG , passive.

In these classes, the indicative mood has the zero-form, in contrast to the -iformant of the subjunctive. In present-day Danish, we find extrasyllabic structure everywhere. Swedish however, has preserved and in fact reinforced the original syllabic differences; see the list of originally thematic verbs in (B) below. Notice that the productivity of the athematic formation in Swedish has resulted in reinterpretation of ī an-verbs and -ē n-verbs as jan-verbs: hörs for older hö-res, sägs for older säges. In (C) a list of originally athematic verbs is given. (B)

Danish kaldes tabes hittes høres siges haves

Swedish kallas tapas hittas hörs sägs havs

(C)

Danish forbydes dræbes tages/toges gives sværges

Swedish förbjuds/förbjöds dräps/draps tags/togs givs/gavs svärs/svors

In Brøndum-Nielsen, the strategy of the exposition is based on the assumption of analogical extension of the ‘extrasyllabic’ expression in Danish. Extension is found first in the (unmarked) present tense, later also in the (marked) past

280

LARS HELTOFT

tense. This is as it should be, as predicted by the markedness theory of Andersen (2001a, b), and also Timberlake (1977). We cannot in this context relate in detail the geographical spread of the ‘extrasyllabic’ form to the historical development. The spread in the present tense is earlier in Jutlandic texts, later in Zealandic texts, and only sporadically found around 1500 in Scanic, the most conservative Danish dialect. Brøndum-Nielsen (1928–1974) provides an extensive overview of older and younger forms in the manuscripts. Strikingly, however, the synchronic relevance of the category of the mediopassive — of which the passive was but one variant — is never called into question. The mediopassive is assumed to be a synchronic fact of both the input language (13th c. Danish) and the target language (16th c. written norm). I have checked all references to the change of the present and past forms of the strong verbs. Extensive samples are quoted of the present, including many passives such as the following (orthography normalised): drikkes, drives, forlades, holdes, lades, skrives, sjunges, lægges, sættes, overvindes. But with the past tense of the strong verbs, the situation is different, and examples are relatively scarce. Brøndum-Nielsen (1928–1974) claims that the analogical past tense form is not very common, but of those quoted almost all are reflexive, reciprocal, or intransitive uses, see (17)–(20). (17) then frwe this lady

gawes ther ille veder express.PAST.REFL her grief over this

(18) vor herre omgickiss Our Lord was:together.PAST.RECIPR

met mennisken with men

(19) tha ieg drogis met ianemund vndher olye træeth when I fought.PAST.RECIPR with Ianemund under the olive tree (20) ther thet om sidher befandes when at last it feel.PAST.S-PASS “when at last it could be felt” (of a figurative birth: the sense is the normative or generic sense mentioned in (13–15)) Of the 13–14 instances referred to by Brøndum-Nielsen (1928–1974), only one is a candidate for the status as a descriptive, non-normative ‘extrasyllabic’ passive. The conclusion to be drawn is that the strong verbs do not at all form ‘extra-syllabic’ past passives. The analogical change is found only with intransitives and reciprocals.

281


The internal structure of the paradigm To find the reason for this apparent breakdown of the mediopassive category, we shall scrutinise the internal structure of the paradigm of the strong verbs. For ease of exposition, I have again normalised the orthography to modern standards. In comparison with the old system, this late medieval Zealandic system has undergone further merging due to the reduction of all secondary vowels to schwa. The result of this is that present plural and past plural forms can no longer be distinguished from the subjunctive. Only the present singular and the past singular maintain a clear distinction between the traditional moods. Note that in the past tense, stem formation (ablaut) is part of the indicative marking system. I have put these crucial forms in boldface. For classes 4–5, the etymological Scandinavian form of the ablaut stem for the past plural and for the subjunctive has a long a, rendered ā. I have preferred to retain this etymological notation, since late Middle Danish orthography does not always allow us to determine whether this āhas been retained or has followed the general rounding process to å, older Danish orthography aa, Swedish å (cf. Swedish forms like gåvo past.pl.ind. ‘gave’ and vore past.subj. ‘were’). It is also uncertain whether the form *sangs is attested, but similar verbs of this class have forms like draks ‘was drunk’, fands ‘was found’. Everywhere, it is the indicative singular against the rest; see Table 4: a–f. 6.

Table 4: a–b Voice PRES. IND. SG. PRES. IND. PL. PRES. SUBJ. PAST .IND.SG. PAST .IND.PL. PAST .SUBJ.

Table 4a: Class 1 Active griber gribe gribe greb gribe/grebe gribe

Passive gribs gribes gribes grebs gribes gribes

Table 4: c–d Voice PRES. IND. SG. PRES. IND. PL. PRES. SUBJ. PAST IND. SG. PAST IND. PL. PAST SUBJ .

Table 4c: Class 3 Active sjunger sjunge sjunge sang sunge sunge

Passive sjungs sjunges sjunges sangs sunges sunges

Table 4b: Class 2 Active forbiuder forbiude forbiude forbød forbude/forbøde forbude

Table 4d: Class 4 Active skær(er) skære skære skar skāre skāre

Passive forbiuds forbiudes forbiudes forbøds forbudes forbudes

Passive skærs skæres skæres skars skāres skāres

282

LARS HELTOFT

Table 4: e–f Voice PRES. IND. SG. PRES. IND. PL. PRES. SUBJ. PAST IND. SG. PAST IND. PL. PAST SUBJ .

Table 4e: Class 5 Active giver give give gav gāve gāve

Table 4f: Class 6 Active tager tage tage tog toge toge

Passive gives gives gives gavs gā ves gā ves

Passive tags tages tages togs toges toges

Table 4: a–f: Conjugation classes 1–6.

Now, let us assume a reinterpretation of the s-forms as a kind of mood, as sketched in (9) above, with a modal factor different from the speaker, and not necessarily including the speaker. And let us again look at the old present tense forms of jan-verbs and strong verbs, cf. Table 5: a–b: Table 5a: jan-verbs dylia “conceal” læggia “lay” kræfia “demand” Table 5b: strong verbs luka “shut” dræpa “kill” taka “take” gifua “give” swæris “swear”

PRES .IND.PASS .

PRES .SUBJ. PASS.

Modern Danish

dyl-s læg-s kræf-s

dyl-i-s lægg-i-s kræf-i-s

dølges lægges kræves

PRES .IND.PASS .

PRES .SUBJ. PASS.

Modern Danish

luk-s dræp-s tak-s gif-s swær-s

luk-i-s dræp-i-s tak-i-s gifu-i-s swær-i-s

lukkes dræbes tages gives sværges

Table 5: a: Passive forms of jan-verbs. b: Passive forms of strong verbs.

For this part of the paradigms, it will be a straightforward assumption that what we find here is not analogy, but the replacement of the unmarked indicative passives by the subjunctive form to comply with the new content. Indicative forms disappear, and in contrast to the unquestioned traditional assumption by Brøndum-Nielsen, the resulting form is not an indicative. Diachronically, the passive subjunctive forms, all originally of the structure in (21): (21) subjunctive marker in -i- + passive marker in –s


283

have been resegmented (schwa, here represented as -e-, replaces the etymological formant -i) as in (22): (22) -e- + -s > -es As stressed above, this resegmentation went hand in hand with the reanalysis of the -s as part of a modal form; in fact, this is the expression side of the coin. This leaves the full paradigms above in a kind of stalemate: the vocalic form of the past -s-mood is a non-indicative that need not be speaker-bound, and thus does not combine with marked indicative stems; and for various semantic reasons, the subjunctive stems are not available either. Let us go through each type in detail: For the 1st class (type gribe), the stem-type greb- is a marked indicative. The form *grebes cannot be formed as a passive since it is inconsistent. The past subjunctive gribes is a full merger with the present tense form and therefore unlikely. For the classes 2–3 (type forbiude, siunge) the stem-type forbød- and sang- are similarly marked indicatives. The forms *forbødes and *sanges cannot be formed since they are inconsistent. The stems forbud- and sung- are neutral in the sense that they form both the past indicative plural and the subjunctive mood in general. These stems are not marked indicatives, and we would therefore not expect them to be inconsistent with the subjunctive form. For the classes 4–5 (type skære, give), the stem-type skar- and gav- were originally marked as indicative singular by having short stem vowels. As such, they are marked indicatives, and the forms *skares, *gaves cannot be formed consistently. The past subjunctive stems skār-, gāv- are also found in the past indicative plural form; they are therefore neutral with respect to mood, and no inconsistency should arise. The rounded output vowel forms like *skaares (skåres), *gaaves (gåves) seem to offer a clearcut opposition. For the 6th class (type tage), the stem-type tog- differs from the rest of the classes, since it is neutral with respect to mood. Thus, the form toges can be formed without any inconsistency. To some degree, we find preservation of the simple -s, but since the -s is elsewhere part of the new modal formant, such indicative -s-formants are isolated, and reduced to marked, archaic genres and contexts, all in accordance with Andersen’s markedness theory (Andersen 2001a). Seventeenth century examples from Brøndum-Nielsen (1928–1974) are biblical. At one end, class 1 leads to either inconsistency or to extensive presentpast mergers. On the given inherited base, this class cannot form a distinct spassive.

284

LARS HELTOFT

At the other end, we find class 6 with no internal obstacles to the form toges. Classes 2–5 offer seemingly neutral forms based on the ablaut stems in u- and -ā-/-aa- (å), respectively: forbudes and sunges; skāres/skaares and gāves/gaaves. These forms would seem to be available as new past s-mood forms. Until now, I have stayed faithful to the traditional categorisation by Brøndum-Nielsen, but a clue to the situation described may be the fact that the so-called past subjunctive is not a past tense form at all and is therefore not available as a tense counterpart to the present tense s-mood. Bjerrum (1967) offers a convincing description of the past subjunctive as a potentialis form, neutral with respect to the present vs. past distinction and therefore not a past tense counterpart to the present subjunctive. Bjerrum supplies compelling examples of this function of the traditional subjunctive, which I shall from now on simply name potentialis (POT ) for the misleading ‘past subjunctive’, (23): (23) æn thot kunæn waræ fræls oc mannæn waræ but even if the wife be.POT free and the man be.POT “but even if the wife might be free and the man might be annøgth oc han hafthæ børn with hennæ, tha wæræ enthralled, and he had children with her then be.PRES.SUBJ a thrall, and he should have children with her, then shall børnæn frælsæ. the children free the children be free.” Where the Scanic dialect preserves both a present and a past subjunctive, with the same set of hypothetical, jussive and optative content functions, the Zealandic dialect has developed a non-temporal division of labour between the two subjunctive forms. The past subjunctive is no longer a past, but a general hypothetical mood, neutral with respect to tense. Very convincing is the following example (24). (24) Æn warthær thæt But if so happens sculdæ must.POT .

a fyrstæ thing at hin ær swaræ at the first thing that he who answer

at waræ that be.POT

a staddær oc han gripæ present and he seize.POT


førræ oc duldæ thingit before and object:to.POT the thing

285

æn han tokæ than he take.POT

til swaræ oc buthæ iarn i gen […]. to answer and offer.POT iron in return (or, instead) “But if so happens at the first thing that he who is to stand trial that he is present and he speaks up in objection to (the legality of) the thing before he is tried, and should he offer the ordeal by fire instead […].” (EL 147,1) This past subjunctive form is not a tense form at the stage of development in question. It is a general hypothetical mood, neutral to tense. The original opposition within the voice system between skars and skāres, bøds and budes, sangs and sunges, is one of mood, not of tense. Already for this reason, the socalled past subjunctive is not a candidate for the function as a past tense of the new s-mood. And secondly, its hypothetical, or potentialis, function is not semantically identical to the content of the s-mood. The potentialis function is a traditional subjective mood, in the sense that the modal factor is with the speaker, but as stressed above, the s-mood has an objective modal factor (the modal factor is not with the speaker). A summary of the changes and clashes involved will be useful: 1. No stems marked as indicatives can form a base for the change. Ablaut classes 1–5 have marked indicative singular stems (grep-, bøth-, sang-, skar-, gav-) and these are not available. 2. All past subjunctive forms have turned into potentialis forms and thereby lost their temporal opposition to the present tense. Forms like gribes, budes, sunges, skāres, gāves are by way of this change turned into passive (tense-neutral) potentialis. For two reasons, such forms are not candidates for the role as past tenses of the s-mood, a. they are not past tenses, and b. they are part of a subjective mood system, but the past needed is an objective mood. 3. Stems of the 6th ablaut class (in -ō-, for instance tog-, lod-, jog-) have no built-in opposition of mood. These stems are therefore open as the basis for reinterpreted forms in -es, as are the stem systems of the weak verbs. All weak verbs of late medieval Danish show a merger in the past tense between indicative and subjunctive: kast-ede and hør-te are neutral with 3 respect to mood. Verbs of the 6th ablaut class form their s-mood in exactly the same way as the weak verbs, cf. (25): 3

This indicative–subjunctive merger is generalised in modern Danish to the present situation where the former categories of indicative and subjunctive have been replaced by one overall category of proximity with respect to the world described. The modern past forms are [–

286

LARS HELTOFT

(25) Neutral past tense: form or stem + -es kast-ede- + -es > kastede-s hør-te+ -es > hørte-s tog+ -es > tog-es Where a neutral past tense stem is available, the s-mood can be readily formed. Conclusion The development of the Danish s-passive, then, can only be understood on the basis of a semantic analysis of the input and output systems of the change. We cannot presuppose the relevance of the middle voice for both Old Scandinavian and later Danish. Only after the analysis of the modern s-passive as a mood can we find the clues to the differences in development between Danish and Swedish: in semantics and use and in morphological reshaping. And only after separating the unproductive remnants of the middle from the smood can we see the important differences involved in this change. The theoretical vocabulary of Brøndum-Nielsen (1928–1974) allows only analogy and sound change as the basic dynamic factors. But only the fate of the reciprocal and intransitive uses can be described as expression regulations alone, that is, as analogical levelling. The s-passive departs from the mediopassive to form in Danish an objective mood. In the light of content-oriented grammaticalisation theory, the formation of the s-mood stands out as a classical example of semantic reanalysis that leads to a marked result. The present tense subjunctive of the s-passive is reinterpreted as an objective mood at the expense of the indicative. The second part of the process is the generalisation of this s-mood into the past tense, the marked term of the tense system. If I have got the substance right of the systems involved, this serves as an example of an actualisation process that leads in part to miscarriage, due to semantic systems in conflict. Actualisation processes, then, must be seen on the background of the system they are about to change, and in two senses. First, we must have detailed knowledge of the system prior to the process of change, and secondly, we must follow the process of spread from unmarked to marked contexts. Some such spreads may encounter problems due to limitations and semantic obstacles created by the systematic oppositions found in the marked context. 7.

proximal] (or, should you so wish, [+distal]). Proximity in the past tense must then be interpreted in use as either temporal distance or distance in the sense of a hypothetical or counterfactual scenario.


287

REFERENCES Andersen, Henning. 1980. “Morphological change: Towards a typology”. Historical Morphology ed. by Jacek Fisiak, 1–50. The Hague: Mouton. Andersen, Henning. 2001a. “Markedness and the theory of linguistic change”. Actualization: Linguistic change in progress ed. by Henning Andersen, 22–57. Amsterdam & Philadelphia: John Benjamins. Andersen, Henning, ed. 2001b. Actualization: Linguistic change in progress. Amsterdam & Philadelphia: John Benjamins. Bech, Gunnar. 1952. Semantische Entwicklungsgeschichte der hochdeutschen Modalverba. (= Det Kongelige Danske Videnskabernes Selskab. HistoriskFilologiske Meddelelser, 32:6.) København: Ejnar Munksgaard. Bjerrum, Anders. 1966 [1954]. Grammatik over skånske lov. Copenhagen: Københavns Universitets fond til tilvejebringelse af læremidler. Bjerrum, Anders. 1967. Grammatik over de sjællandske Love. Copenhagen: Københavns Universitets fond til tilvejebringelse af læremidler. Brøndum-Nielsen, Johannes, in collab. with Karl Martin Nielsen. 1928–1974. Gammeldansk Grammatik I–VIII. Copenhagen: J. H. Schultz/Akademisk Forlag. Bybee, Joan, William Pagliuca & Revere Perkins. 1993. The Evolution of Grammar: Tense, aspect, and modality in the languages of the world. Chicago & London: The University of Chicago Press. Dyvik, Helge. 1980. ”Har gammelnorsk passiv?”. The Nordic Languages and Modern Linguistics 4. (= Proceedings of the Fourth International Conference of Nordic and General Linguistics, Oslo 1980.) ed. by Even Hovdhaugen, 81–107. Oslo: Universitetsforlaget. Heine, Bernd, Ulrike Claudi & Friederike Hünnemeyer. 1991. Grammaticalization: A conceptual network. Chicago: University of Chicago Press. Heltoft, Lars. 1996. “Paradigms, word order and grammaticalization”. Content, Expression and Structure: Studies in Danish Functional Grammar ed. by Elisabeth Engberg-Pedersen, Michael Fortescue, Peter Harder, Lars Heltoft & Lisbeth Falster Jakobsen, 469–494. Amsterdam & Philadelphia: John Benjamins. Heltoft, Lars & Lisbeth Falster Jakobsen. 1996. “Danish passives and subject positions as a mood system”. Content, Expression and Structure: Studies in Danish Functional Grammar ed. by Elisabeth Engberg-Pedersen, Michael Fortescue, Peter Harder, Lars Heltoft & Lisbeth Falster Jakobsen, 199–234. Amsterdam & Philadelphia: John Benjamins. Heltoft, Lars, Jens Nørgård-Sørensen & Lene Schøsler. 2005. “Grammatikalisering som strukturforandring”. Grammatikalisering og struktur ed. by Lars Heltoft, Jens Nørgård-Sørensen & Lene Schøsler, 9–30. Copenhagen: Museum Tusculanum Press. Hjelmslev, Louis. 1935. La Catégorie des cas: Étude de grammaire générale. Première partie. (= Acta Jutlandica VII.) Århus: Universitetsforlaget i Aarhus. Holm, Gösta. 1952. Om s-passivum i svenskan. Lund: Gleerup.

288

LARS HELTOFT

Hopper, Paul J. & Elizabeth Closs Traugott. 1993. Grammaticalization. Cambridge: Cambridge University Press. (2nd ed. 2003) Lehmann, Christian. 1985. “Grammaticalization: Synchronic variation and diachronic change”. Lingua e Stile 20.303–18. Lehmann, C. 1992. “Word order change by grammaticalization”. Internal and External Factors in Syntactic Change ed. by Marinel Gerritsen & Dieter Stein, 95–416. (= Trends in Linguistics: Studies and Monographs, 61.) Berlin & New York: Mouton de Gruyter. Oxenvad, Erik. 1934. “Om brug af passiv på dansk”. Studier tilegnede Verner Dahlerup paa Femoghalvfjerdsaarsdagen ed. by Poul Andersen, Aage Hansen and Peter Skautrup, 80–88. Århus: Universitetsforlaget i Aarhus. Timberlake, Alan. 1977. “Reanalysis and actualization in syntactic change”. Mechanisms in Syntactic Change ed. by Charles N. Li, 141–177. Austin: The University of Texas Press.

ASPECT AND ANIMACY IN THE HISTORY OF RUSSIAN DEVELOPING THE IDEA OF PARALLEL GRAMMATICALIZATION

JENS NØRGÅRD-SØRENSEN University of Copenhagen Introduction This study proceeds from two basic facts. First, languages undergo constant changes which influence all layers of their structure. Secondly, languages tend to develop into types.1 The classical typological distinction between isolating, agglutinative, inflectional, and polysynthetic languages is based on the observation that languages may exhibit internal harmony in their morphological structure. Studies in language typology are mostly based on formal morphological and syntactic features. A remarkable exception is Durst-Andersen (1995) who proposes a typology grounded on semantic universals. In the present study we shall proceed from this point of view and assume that semantic factors play a crucial role in language typology and, consequently, in language change. The two facts pointed out above, namely (i) that languages undergo constant changes, and (ii) tend to develop into types, seem to be generally accepted by linguists of various persuasions, but obvious consequences are often neglected. In historical linguistics, particular instances of change in one and the same language are mostly treated as fairly isolated phenomena. However, given the above-mentioned facts, one should expect various simultaneous changes to be mutually interconnected, even if they appear to be independent of each other and as affecting different parts of speech, or different morphological and syntactic patterns. In the case of simultaneous changes one should look for a common motivating principle. This motivation could be a basic, highly generalized semantic distinction, as I will show in this contribution. 1.

1

In a tentative, fairly informal formulation: that a language belongs to a specific type means that there is a common principle determining or at least somehow penetrating its grammatical structure and lexical patterns.

290

JENS NØRGÅRD-SØRENSEN

The verbal category of aspect and the nominal category of animacy are usually not treated as interconnected, either in diachronic or in synchronic investigations. However, that nominal grammar in Modern Russian (MR) can be shown to mirror verbal grammar — with aspect as the determining category — has been suggested by Durst-Andersen (1996:199–206) who crucially points to animacy as the nominal counterpart of aspect. This view has been supported by detailed studies of the grammatical features of MR nouns (cf. Nørgård-Sørensen 1997c:344–345 where I resume the arguments in favour of treating animacy as the dominant grammatical category of MR nouns). If it is true that animacy is linked with aspect in MR, these two categories should also be linked in the diachrony of the language. In the following I shall outline the historical development of aspect and animacy in Russian. I shall also comment on the syntactic category of transitivity which has often — and justly — been linked with aspect. I shall show that the histories of aspect and animacy exhibit certain striking similarities. I shall further demonstrate that these similarities can be linked with a common semantic principle. For a process of this kind — where an underlying, highly generalized semantic feature motivates parallel paths of development of apparently unconnected categories — I have proposed the term ‘parallel grammaticalization’ (cf. Nørgård-Sørensen 1997b:96–97 where I first pointed to the connection between aspect and animacy from a historical perspective, however without the details and arguments which will be presented in the following). Animacy Animacy is a gender category. Any MR noun is inherently either animate or inanimate, and this grammatical feature is reflected in the inflection both of the noun itself and of subordinate words. Thus, animacy satisfies Hockett’s (1958:231) classical definition of gender, “Genders are classes of nouns reflected in the behavior of associated words”. Additionally, the traditional Indo-European gender distinction between masculine, feminine, and neuter is fully preserved in MR. Accordingly, MR has two overlapping, in principle mutually independent gender categories. However, in order to avoid confusion I shall refer to the two different categories of gender in the traditional way, as ‘animacy’ and ‘gender’, respectively. Animacy is the younger category which at a certain point was established as a supplement to gender. Both being gender categories, in the general sense, they can hardly remain fully unconnected, and, as already mentioned, there is no doubt that in MR animacy has taken precedence: in assignment patterns it sometimes dominates gender while the reverse is never the case (NørgårdSørensen 1997c:339–340). It follows that animacy is now to be considered the 2.

291

PARALLEL GRAMMATICALIZATION

primary, classificatory (inherent) category of Russian nouns. Unlike gender, which depends on both formal and semantic features of nouns, animacy is fully semantically based: animate nouns are those denoting living beings, other nouns are inanimate. In the inflection of the noun, animacy is reflected in the choice of accusative desinence: the accusative of animate nouns inherits the genitive desinence, the accusative of inanimate nouns the nominative desinence. This pattern is characteristic of the so-called 1st declension masculine (historically the masculine o-stems) and of all plurals, cf. Table 1 where the arrows show the direction of desinence inheritance.

Table 1: MR noun declension (only NOM, ACC and GEN; generally simplified).

Note that animacy is marked in all plural nouns in Russian (unlike several other Slavic languages). This is the background for claiming that animacy is a feature of all Russian nouns, that is, of the noun as a part of speech. I shall now sketch the development leading to the establishment of animacy. As it appears, we should be looking for instances of genitive forms in positions where the accusative would be expected. The genitive in accusative position first appeared with masculine singular o-stems denoting living beings, compare (1) and (2) (cited from Krys’ko 1994:31), both from the Codex Laurentianus (1377) which is the oldest preserved copy of the Old Russian (OR) Nestor Chronicle from the beginning of the 12th century. (1)

Izjaslavŭ ž e [...] posla muž ĭ svoi. Izjaslav but [...] sent man.ACC(= NOM) his.ACC(=NOM) “But Izjaslav sent his man.”

(2)

[...] posla k nemu zlato [...] i muž a sent to him gold.ACC [...] and man.GEN “[...] sent gold and a wise man to him”

mudra wise.GEN

292


In (2), the direct object phrase consists of (i) a neuter noun zlato with the desinence -o which in this context unequivocally reflects the accusative, and (ii) a masculine noun muž a with an attributive adjective mudra, both with the genitive desinence -a. In (1), on the other hand, we find the same noun muž ĭ “man” in a form that can be identified as the “historical” accusative, that is, the accusative that should be predicted from regular sound changes. As a result of sound changes in prehistoric time the Slavic syllable was restructured fairly drastically, and a number of desinences, inherited from PIE, were reduced. The nominative and accusative singular of the o-stems were left identical (Table 2). Case NOM ACC

PIE desinences -os -om

Old Bulgarian (reflecting Common Slavic) vlǔk-ǔ2 “wolf” vlǔk-ǔ

Table 2: o-stems: masculine, singular.

It is generally assumed that this nominative-accusative syncretism led to the expansion of the genitive in accusative function (see, for instance, Iordanidi 1995:284). The o-stems contained a large number of frequent person-denoting nouns, and in the common sentence type, that is Subject–Verbal predicate– Direct object, there would, in principle, be no way to distinguish the subject from the object if both were o-stems. A situation where the subject is not formally distinguished from the object is likely to cause structural disharmony in a language that marks subject, object, and other constituents by case. The problem did not really arise if one noun was animate and the other inanimate. Animacy would provide sufficient information for an unequivocal functional differentiation: the animate noun would be perceived as the subject, the inanimate as the object. However, if both nouns were animate, and especially if both denoted persons, there would be no formal indication of their syntactic status. It should be emphasized that in this and most other varieties of Slavic the word order does not play any role in signaling the syntactic function of noun phrases. Word order had merely communicative functions (topic–focus articulation). The “solution” to the problem was to use the genitive as direct object case. Such a use of the genitive was not alien to the native speakers of that time. In older varieties of Slavic a fairly large number of verbs appeared regularly with a genitive complement (Krys’ko 1997:194–198), a fact which 2

In nominal declension the correspondent of the vowel ǔafter palatal consonant was ĭ , for instance, muž ĭ“man”. ǔand ĭwere in older varieties of Slavic so-called reduced vowels, back and front respectively.


293

must have facilitated this development. The genitive as the case of the firstcomplement was able to spread to the instances considered. This explanation of the development is generally recognized as highly plausible, but certain consequences have been ignored. It has not been sufficiently acknowledged that in the beginning the development merely represented a change of usage rules (spread of the genitive by analogy) and not a case of grammaticalization (in the sense of a reanalysis affecting the grammatical system). Only when the genitive form of animate nouns was reinterpreted as also representing the accusative, we may speak of a new gender distinction, viz. that of animate–inanimate, having been established. This step in the development should be identified as a situation where the new animate accusative (identical to the genitive) clearly expands, eventually causing the old accusative (identical to the nominative) to disappear. In the following I shall attempt to pinpoint the approximate time of this crucial reanalysis. There is further evidence supporting the explanation of the development as motivated by the phonological merging of the nominative and the accusative. The genitive in accusative function first appeared in direct object position and only later in other accusative functions, for instance, in connection with prepositions (Krys’ko 1994:172). This indicates that the “critical” subject-predicate-object structure was actually where the development started, and that it started, not as a reanalysis of the grammatical system, but as an adjustment of the distribution of the case forms within existing valency patterns. So far I have mostly referred to relative chronology. Below follows an overview (in Table 3) of the absolute chronology of the development of the genitive in accusative function (mainly based on Krys’ko 1997). Period Common Slavic (before 800) 11th c. 14th c. 15th c. 18th c.

Gender-Number-Animacy MASC.SING. ANIM (mainly o-stems, some ŭ-stems and i-stems) MASC.SING. ANIM (o-stems) and MASC. PLUR. ANIM (o-stems) MASC.SING. ANIM (o-stems) and all PLUR. ANIM Informal language: MASC.SING. ANIM and all PLUR. ANIM Formal language: MASC.SING. ANIM and all PLUR. ANIM

Case form Genitive or historical accusative Genitive or historical accusative Genitive or historical accusative Genitive Genitive

Table 3: The history of the genitive form in accusative function.

This overview calls for a few comments. First, by a “historical accusative” I mean an accusative form matching

294


the predictions of regular sound developments, cf. muž ĭ“man” in (1) — as opposed to the genitive form muž a in (2). Secondly, the genitive in accusative function appeared in all classes of singular masculine nouns with phonological merging of the nominative and the accusative (Krys’ko 1994:35, 51–52, 68). These were primarily o-stems, but they also included some ǔ-stems and i-stems. In early historical times the ǔstems were adopted to the o-stem pattern and ceased to exist as an independent declension. The i-stems lost their masculine nouns and were basically left as a feminine declension. This allows us, for reasons of simplification and clarity, to treat the development as an o-stem phenomenon. Thirdly, as it appears, the genitive in accusative function spread from the singular animate o-stems first to the plural animate o-stems (11th c.) and later to all plural animate nouns (14th c.). The reason for the latter step was that the former distinct plural declensions merged into one. We can leave out the details of this development; it suffices to point out that by the 14th century there were practically no plural desinences left identifying the nouns as belonging to one particular declension of the ones still distinguished in the singular. As a consequence of the merging of the plural declensions, the “competition” of the two accusatives, hitherto restricted to the plural o-stems, spread to all plural nouns. I shall now present a tentative interpretation of the data (inspired by theoretical considerations in Andersen 2001a, 2001b). In the historical process we can identify two motivating factors. First, the merging of the nominative and the accusative in the singular in prehistoric Slavic motivated the genitive as first-complement marker to spread to the cases considered, cf. (2) as opposed to (1). This should be seen as an adjustment of usage rules rather than as a reinterpretation of the system. The change neither affected the case system nor the gender system as such. Second, the merging of all plural declensions in 14th century Russian had the drastic consequence of extending the genitive in accusative function to the entire noun lexicon. In itself this was also an adoption of usage rules, but, as I see it, it made possible two mutually interconnected reinterpretations (collective reanalyses): (a) The genitive form could be reinterpreted as the (only) accusative form of animate nouns in the plural and in the singular of masculine o-stems (= 1st declension in MR). (b) As a consequence of (a) a new gender distinction, that of animate– inanimate, appeared.


295

Note that none of these reinterpretations affected the case system as such. The reinterpretations did not mark a change in the function of individual cases. In relation to case, these changes were surface phenomena: certain forms were replaced by others. On the other hand the two reinterpretations marked a change in the gender system: animacy was established as an additional gender distinction. Aspect Having sketched the development of animacy, I shall now turn to aspect. It is a well-known fact that verbal categories underwent fundamental changes in the history of Russian. Protoslavic and Common Slavic had a tense system, relatively closely resembling that of PIE, cf. Table 4. 3.

Simple forms Present Aorist Imperfect

Composed forms ( AUX + participle) Perfect Pluperfect I Pluperfect II

Table 4: Tense in Common Slavic (reflected in Old Bulgarian and formal registers of OR).

In addition there were various less stable composed forms conveying future and futurum exactum. They are not essential to our argument and need not concern us in the following. Let us turn to MR and compare its verbal grammar to that of Common Slavic. Of the tense forms enumerated in Table 4 there are only two left: the present and the perfect, the latter usually called the “preterite” since it is the only past tense form. On the other hand there is a fully grammaticalized category of aspect. Aspect has often been taken to be a category in OR as well, presumably because the verbal derivation system is formally identical in OR and MR. This system can be illustrated schematically in the following slightly simplified way, cf. Figure 1.

Figure 1: Verbal derivation in OR and MR (simplified).

296


In MR, this derivation system serves the double purpose of deriving new verb lexemes and new aspect pairs. An aspect pair can be derived in various ways. The dominant productive model is the one illustrated in Figure 2.

Figure 2: MR: Derivation of aspectual pairs.

By prefixation, cf. govorit’ > u-govorit’, a new perfective verb lexeme is formed. An imperfective partner verb is derived by suffixation, cf. u-govorit’ > u-govar-iva-t’. According to Durst-Andersen (1992), on whose theory I will base my presentation of MR aspect, one can distinguish between three lexical classes of verbs: ‘state’ verbs (e.g., “stand”, “sit”), ‘activity’ verbs (e.g., “work”, “walk”), and ‘action’ verbs (e.g., “give”, “persuade”). The argument runs — in a somewhat oversimplified form — as follows: the distinction between activity, state and action has universal status, since it is based on a description of how human beings conceptualise real situations in different types of images, or “pictures”. States correspond to ‘stable pictures’ (with the figure being stable in relation to the ground), and activities correspond to ‘unstable pictures’ (with the figure moving in relation to the ground). Unlike a state or an activity, an action is complex, in that it corresponds to a combination of an activity and a state (the state being understood as the natural effect of the activity in question, referred to as ‘telecity’ in Durst-Andersen 1992.). In MR the so-called aspect pairs are verbs denoting actions (rather than activities and states). When referring to a single action, the perfective partner is used to assert the state description, presupposing the activity leading to this state, cf. Figure 3.


Activity

→

297

Assertion ↓ State

Figure 3: The perfective aspect partner.

The imperfective partner is used to assert that the activity obtains, thus treating the state as a standard implicature, cf. Figure 4.3 Assertion ↓ Activity

→

State

Figure 4: The imperfective aspect partner.

We can now give the following summary of the particularly characteristic features of MR as a language based on aspect: (i) A verb form must be either perfective or imperfective (as a grammatical category aspect is obligatory). (ii) While state verbs and activity verbs are unpaired and imperfective tantum, action verbs are paired. An action verb is formally a combination of an imperfective and a perfective verb, cf. the box in Figure 2. This means that when a new perfective action verb occurs (most frequently in the form of a prefixed verb), an imperfective partner must also be formed. (iii) Action verbs are often transitive. Having outlined the grammar of the verb in OR and MR, I shall now trace the development leading from the old to the new system. The above-mentioned simplification of the tense system, leaving the perfect as the only past tense form, took place everywhere in North Slavic (a common name for East and West Slavic). The aorist, the imperfect and (for a short period of time) the pluperfect were gradually restricted to more and more marked functions (semantically and stylistically) before eventually disappearing. In the West Slavic language Polish, they all seem to have disappeared fairly early. Even in the earliest texts from the 14th century (presumably copies of older texts) the perfect is already close to the status of a general preterite. 3

As regards the interpretation of procedurals (Aktionsarten) of the types kurit’ “smoke” – pokurit’ “smoke a little” and kurit’ – kurivat’ “smoke repeatedly”, see Durst-Andersen (1992: 172–178).

298


The pluperfect is rare and functionally not distinguishable from the perfect; of the aorist and the imperfect there are altogether only 26 occurrences, some of them questionable (Klemensiewicz & al. 1955:367–373). In Russian, on the other hand, only the pluperfect seems to have disappeared early. The imperfect and especially the aorist appeared regularly till about the middle of the 18th century, however only in formal registers. However, when considering the Russian data it should be remembered that Russian — unlike Polish — was under heavy influence of Old Church Slavonic up till as late as the 18th century. Old Church Slavonic, which is in fact a variety of Old Bulgarian, came to Russia in connection with the adoption of Christianity in the 10th century as the Slavic language to which the Bible and other both religious and secular texts had been first translated. Naturally, Old Church Slavonic was established as the norm that in principle should be observed in all varieties of written language. This meant that scribes had to use the original tense forms, cf. Table 4, which were fully preserved in the Bulgarian-based Old Church Slavonic. The Old Church Slavonic influence makes it extremely complicated to investigate the development of the Russian tense/aspect system. However, from about 1950 and onwards archaeologists have excavated almost a thousand letters written on birch bark in Novgorod and other ancient Russian cities. These are almost all informal messages on everyday matters. For this reason, the authors refrain from observing the Old Church Slavonic norm and obviously stick to their oral language. In the birch bark letters, even in the earliest ones from the 11th century, the perfect is totally dominant, also in contexts where one would expect another tense form according to the Old Church Slavonic norm.4 This indicates that the perfect was established as the general preterite earlier than hitherto believed (cf. Remneva 1995:138). It is reasonable to assume that this took place before 1000. This conclusion would be in harmony with the Polish data, and with the fact that the change took place before North Slavic was finally split into East and West Slavic. The question is now, when was aspect established in Russian? In Nørgård-Sørensen (1997a, 1997b) I have argued that this question is best investigated on the basis of contexts where the MR aspect forms have a marked reading. This is the case with certain negated perfective forms. When a MR perfective verb is negated, the negation will only have the asserted state as its scope. The activity, presupposed by the perfective aspect, will not be affected by the negation, cf. Figure 5, which is the “negated correspondent” of Figure 3. 4

The few aorists and imperfectives in the birch bark letters appear almost exclusively in the few letters with a more official content, e.g. in wills.


Activity

→

299

Negation ↓ ⌐State

Figure 5: Perfective verbs: negation scope.

If, on the other hand, an imperfective verb is negated, the negation will logically be directed towards the activity, but since the state is implicated by the activity, the negation will extend its scope to the state as well, cf. Figure 6, which is the “negated correspondent” of Figure 4. Negation ↓ ↓ ⌐Activity → ⌐State Figure 6: Imperfective verbs: negation scope.

Thus, in order to negate the action as a whole one must use the imperfective verb in MR. In connection with negation the imperfective verb is the neutral, unmarked choice. The limited negation scope in connection with perfective verbs has consequences for the pragmatic interpretation of the given utterance. A perfective imperative must be interpreted as a request not to let the expected state come true (however, presupposing that the activity takes place). Pragmatically this can only be interpreted as a warning. A warning is issued in a situation when an activity potentially leading to a non-desired state is already somehow taking place. A negated perfective verb in the present tense also has a marked reading. It carries the presupposition that the activity, being outside the scope of the negation, takes place, but it also asserts that the expected state (the natural result of the activity) does not obtain. Pragmatically this corresponds to readings like, “does not succeed in …, after all does not …”, etc. In Nørgård-Sørensen (1997a, 1997b) I have provided evidence that in OR (represented by the birch bark letters, in order to avoid the influence of Church Slavonic as far as possible) the prefixed verbs did not convey the marked readings characteristic of the formally corresponding MR aspectual forms, cf. (3) and (4). (3)

Až ĭvodja po:g: rublja pro-da, ali nevodja nĭ pro-dai if lead? for 3 roubles sell-IMP , if not lead? not sell-IMP “If (?), then sell for three roubles; if not (?), then don’t sell.” (Birch bark letter 65, 1281–1299)

300


(4)

Čemu why

ne not

vo-soleš i send-PRS .2.SG

č eto ti what you.DAT

esemo be.PRS.1.SG

vodala kovati [...] given forge.INF “Why don’t you send me what I have given you to forge?” A ne And not

sestra sister

ja I.NOM

vamo, you.DAT,

ož e tako dě laete, if so do.PRS .2.PL

ne is-pravitĭ mi nič eto ž e. not make-PRS.2.PL me.DAT nothing “And I won’t be your sister, if you behave like that and don’t do anything for me.” (Birch bark letter 644, 1110–1130) In (3) we find two imperatives of prefixed verbs of which the latter is negated. The context does not allow for the “warning” reading which would be the only possibility in connection with the formally corresponding MR form. In (4) we find two negated present tense forms of prefixed verbs but again they cannot be interpreted in line with the corresponding MR forms. Rather, they represent unmarked negations of the actions referred to. We can conclude that the MR aspect system was alien to OR. The question is now, how and when was the modern aspect system established? Looking for an answer to this question, we should recall a characteristic feature of the modern aspect system: that any action verb (telecity verb) must be paired, cf. from Figure 2: u-govorit’ (PF) / u-govar-iva-t’ (IPF ) ‘persuade’. This means that when the modern aspect system was established, the derivation of imperfective partners from perfective action verbs became compulsory. As illustrated in Figure 2, imperfective aspect partners are mainly derived by means of suffixation. For this reason, an investigation of the historical development of suffixation patterns, in particular the use of the suffix -iva-/ivaj-, could be a key to the history of aspect. In Table 5 below, I have sketched the development of verbal suffixation with a special emphasis on -iva-/-ivaj-. The statistics supporting the conclusion about the remarkable increase of the number of -iva-/-ivaj- derivatives can be extracted from Silina (1987:198) (based on the card catalogue of the Old Russian Dictionary 114 c., a catalogue including practically all texts from the given period), cf. Table 6 below.

301


Period Comm. Slavic (before 800) 11th c.

Suffixes Several suffixes, including -va-/-vaj-, -ova-/-ovaj-, New suffix, -iva-/-ivaj-

14th c.

Remarkable increase of the number of -iva-/-ivaj- derivatives Automatic derivation of imperf. verbs by means of -iva-/-ivaj-. Final elimination of the aorist.

18th c.

Meaning/function of suffixes Activity, iterativity, etc. Initially purely iterative, later also indicating activity, iterativity, etc. Stylistically neutral Imperfective

Table 5: Development of verbal suffixes, in particular -iva-/-ivaj- (based on Silina 1987). Period 12th c. 13th c. 13th−14th c. 14th c. 1400–1425

Verb lexemes 11 23 16 112 63

First-time registered 11 22 13 83 41

Number of tokens 14 32 18 189 106

Table 6: Verbs with the suffix -iva-/-ivaj-.

Unfortunately Silina does not provide information on the quantity of texts from the individual periods. So the numbers are not directly comparable. On the other hand, she comments on the use of -iva-/-ivaj- in various registers. Till about 1300, -iva-/-ivaj- was restricted to certain genres of secular literature; later it became common in all text types. Thus, the suffix obtained an unmarked status at the time when it started to be used much more frequently. It is natural to assume that the increase of the number of -iva-/-ivaj- derivatives in the 14th century is caused by a structural change — in the form of a collective reanalysis — establishing the aspect system. However, this is not all that has to be said about this change. The change might have been of a more fundamental kind. I shall now attempt to provide an overall interpretation of all the data discussed in the light of the theory of parallel grammaticalization. Parallel Grammaticalization In the introduction I claimed that it is possible to identify a common semantic distinction underlying both aspect and animacy. It is now time to be more specific on this point. As appears from the presentation above, the Modern Russian aspect system is based on a distinction between action and nonaction, the action being perceived as consisting of two parts, the activity and the state. The activity/state distinction is reflected grammatically at least at two levels: (a) in lexicalization (activity vs. state verbs); (b) in the basic aspect meaning, the imperfective verb asserting the activity and the perfective verb 4.

302


asserting the state. The nominal correspondent of the activity/state distinction is obviously that of actor/non-actor. And this is precisely the distinction that has been grammaticalized as the category of animacy. There is thus a semantic parallel between aspect and animacy. This parallel is unlikely to be accidental. It should be seen as a generalized semantic distinction penetrating different layers of grammar and lexicon and thus as the result of parallel grammaticalization. If we look at the history of the two categories we find further evidence in favour of this idea: the development of aspect and the development of animacy exhibit certain similarities. I shall conclude my presentation by outlining these processes and point out the similarities. Common Slavic (before 800): As a result of regular sound changes the nominative and accusative singular (o-stems, ǔ-stems and i-stems) merge, thus bringing about an extended use of the genitive as first complement. This should be seen as an externally motivated change of usage rules. In the same period there are several productive verbal prefixes and suffixes. New action verbs may be derived by prefixation; non-action verbs, that is verbs denoting activity, iterativity etc., are derived by suffixation (Nørgård-Sørensen 1997a). Note, however, that the presumably universal conceptual split of the action into its constituent parts, activity and state, still did not play any role in verbal grammar. North Slavic (10th c.): The perfect achieves the status of a general preterite, leaving the remaining past tenses as peripheral, restricted to more and more marked contexts in the following process of actualization. This change must be seen as the result of an internally motivated reanalysis of the tense system. It is not quite clear what the motivation for this reanalysis might have been. A highly likely motivating factor is the development of the above-mentioned productive prefixation and suffixation system with its distinction of action and nonaction verbs. Russian (11th c.): A new suffix: -iva-/-ivaj- appears as the result of a reinterpretation of the border between root and suffix in suffixed verbs derived from verbs with an -i- as thematic vowel, for instance, poč i-ti “rest” > poč i-vati, being reinterpreted into poč -iva-ti. In the beginning the -iva-/-ivaj- verbs seem to have been used exclusively as iterative verbs, but later they obtained the broader function as a marker of non-action. Being very productive, -iva-/ivaj- began ousting other suffixes in this function. A possible interpretation of the development up to this point is that at a certain time in the prehistoric period a grammatical distinction between action and non-action description was established. It is a question of definition whether such a system should be described as an aspect system, but it was certainly


303

not identical to the MR aspect system, since it treated the action as an indivisible whole (the discussion of negation above provides decisive evidence for this fact). In the beginning the action/non-action distinction was expressed syntactically: action was largely marked by the presence of a direct object in the accusative or — more generally — a first complement in some oblique case, but most verbs could also be used intransitively (Krys’ko 1995: 465–466, 1997; Durst-Andersen 2005:62–69). Later, in an actualization process up to about 1400, the action/non-action distinction became more and more lexicalized as a result of very productive verbal derivation: lexical action verbs were formed by prefixation, non-action verbs by suffixation. As already mentioned, it can be assumed that the establishment of the action/non-action distinction triggered the decline of the multi-tense system, but possible evidence in favour of this conclusion must be considered lost in prehistoric time. In the same period and later — until about 1400 — the genitive form spread as first complement case of animate nouns. This was an externally motivated adjustment of usage rules and not linked to the simultaneous development of the verbs. Russian (14th c.): The formerly distinct plural declensions merged into one, and as a result the possibility of using the genitive as first-complement case with animate nouns spread to all plural nouns. As already mentioned, this was also an externally motivated adjustment of usage rules and did not involve any reanalysis, neither of case, nor of gender. I believe that we are here at a decisive turning point. The development described so far made possible two reanalyses which — from the point of view of parallel grammaticalization — actually constituted one integral reanalysis. First, verbal prefixation and suffixation, which had been used to express action vs non-action, were reanalysed as markers of perfective and imperfective aspect respectively. This involved a conceptual reinterpretation of the action from an indivisible whole to a complex concept consisting of two constituents, the activity and the state, linked by ‘telicity’. In other words, it markedly strengthened the grammatical status of the activity/state distinction. Second, the genitive form of animate nouns in the 1st declension singular (formerly, o-stems) and in all plurals, which had been used along with the historical accusative in direct object and other accusative functions, was reanalysed as the accusative proper of these nouns. As to content, this involved the establishment of animacy as a new gender distinction. Both of these reanalyses must have taken place in the 14th century, presumably towards the end of the century. This is supported by the evidence presented above. In the 15th century, the historical accusative of animate nouns disappeared from informal registers where the new accusative (identical to the

304


genitive) became the only option, and around the end of the 14th century we see a rapid increase in the number and occurrences of -iva-/-ivaj- derivatives in all registers. The fact that the changes were first actualized in unmarked contexts (informal registers) confirms that we are dealing with internally motivated reanalyses, that is, a reshaping of the grammatical system and not just of the usage rules (Andersen 2001a:33). In view of the fact that the two reanalyses are motivated by a common semantic distinction (activity/state, corresponding to actor/non-actor), it should not be considered accidental that they coincide in time. They represent one integral reanalysis of major parts of the grammar, that is, a case of ‘parallel grammaticalization’. The reason why it took place at this particular time in history is that this was the point when the previous development of both verbal and nominal grammar had provided the premises. In the given situation the reanalyses did not at first call for any drastic adjustment of usage rules. The consequences only became clear in the following actualization process. Conclusion In this case study I have tried to demonstrate that it is worthwhile looking for semantic links between apparently disconnected grammaticalization processes. It is reasonable to assume that grammaticalization is parallel to a much larger extent than has hitherto been acknowledged and that the proposed theory of ‘Parallel Grammaticalization’ can contribute to our understanding of how and why languages develop into types. 5.

REFERENCES Andersen, Henning. 2001a. “Markedness and the theory of linguistic change”. Actualization: Linguistic change in progress ed. by Henning Andersen, 21–57. Amsterdam & Philadelphia: John Benjamins. Andersen, Henning. 2001b. “Actualization and the (uni)directionality of change”. Actualization: Linguistic change in progress ed. by Henning Andersen, 225–249. Amsterdam & Philadelphia: John Benjamins. Durst-Andersen, Per. 1992. Mental Grammar: Russian aspect and related issues. Columbus, Ohio: Slavica Publishers. Durst-Andersen, Per. 1995. “Mental’naja grammatika i lingvistič eskie supertipy”. Voprosy jazykoznanija 6.30–43. Durst-Andersen, Per. 1996. “Russian case as mood.” Journal of Slavic Linguistics 4:2.177–273.


305

Durst-Andersen, Per. 2005. “Fra oldrussisk til moderne russisk. Leksikaliseringsmønstre og syntaktiseringsmekanismer”. Grammatikalisering og struktur ed. by Lars Heltoft, Jens Nørgård-Sørensen & Lene Schøsler, 57– 84. Copenhagen: Museum Tusculanum. Hockett, Charles F. 1958. A Course in Modern Linguistics. New York: Macmillan. Iordanidi, Sof’ja I. 1995. “Suš č estvitel’noe”. Drevnerusskaja grammatika XIIXIII vv. ed. by Vjač eslav V. Ivanov, 170–294. Moscow: Nauka. Klemensiewicz, Zenon, Tadeusz Lehr-Spł awiński & Stanisł aw Urbańczyk. 1955. Gramatyka historyczna ję zyka polskiego. Warszawa: Pań stwowe wydawnictwa naukowe. Krys’ko, Vadim B. 1994. Razvitie kategorii oduš evlennosti v istorii russkogo jazyka. Moscow: Lyceum. Krys’ko, Vadim B. 1995. “Zalogovye otnoš enija”. Drevnerusskaja grammatika XII–XIII vv. ed. by Vjačeslav V. Ivanov, 465–506. Moscow: Nauka. Krys’ko, Vadim B. 1997. Istorič eskij sintaksis russkogo jazyka. Ob”jekt i perechodnost’. Moscow: Indrik. Nørgård-Sørensen, Jens. 1997a. “Tense, aspect and verbal derivation in the language of the Novgorod birch bark letters”. Russian Linguistics 21:1.1– 21. Nørgård-Sørensen, Jens. 1997b. “Vido-vremenye formy v drevnenovgorodskom dialekte v sopostavlenii s sovremennym russkim jazykom”. Trudy aspektologič eskogo seminara Filologičeskogo Fakul’teta MGU im. M. V. Lomonosova ed. by Marina Ju. Čertkova, vol. 2, 83–98. Moscow: Izdatel’stvo Moskovskogo Universiteta. Nørgård-Sørensen, Jens. 1997c. “Russian nominal morphology: A hierarchy of semantic features”. Formale Slavistik (= Leipziger Schriften zur Kultur-, Literatur-, Sprach- und Übersetzungswissenschaft, vol. 7.) ed. by Uwe Junghanns & Gerhild Zybatow, 335–346. Frankfurt am Main: Vervuert Verlag. Remneva, Marina L. 1995. Istorija russkogo literaturnogo jazyka. Moscow: Filologija. Silina, Vera B. 1987. “Specifika vyraž enija vidovych različ ij v drevnerusskom literaturnom jazyke”. Drevnerusskij literaturnyj jazyk v ego otnoš enii k staroslavjanskomu ed. by Lidija P. Žukovskaja, 196–208. Moscow: Nauka.

TOWARDS AN INTEGRATED FUNCTIONAL–PRAGMATIC THEORY OF LANGUAGE AND LANGUAGE CHANGE IN COMMEMORATION OF EUGENIO COSERIU (1921–2002)

OLE NEDERGAARD THOMSEN Roskilde University Introduction: An integrated functional–pragmatic theory of language and language change I shall commence this paper with a quotation (1) from the Mexican Nobel Prize winner for Literature in 1990, Octavio Paz (1914–1998), which is highly relevant to my aim because it expresses a particularly clear functional– pragmatic view on language and language change, one that is consonant with Coseriu’s theory (dealt with in Sections 2 and 3 below, and in Nedergaard Thomsen, in prep.) on the one hand, and on the other, the further development I shall propose in the latter part of this paper (Section 4), a development which is based on speech act theory and legal philosophy. (I shall give the original wording in Spanish in the main text and supply a translation into English in a footnote. This citational praxis is followed throughout the paper.) 1.

(1) “[...] Todas las sociedades humanas comienzan y terminan con el intercambio verbal, con el decir y el escuchar. La vida de cada hombre es un largo y doble aprendizaje: decir y escuchar. [...] Empezamos escuchando a la gente que nos rodea y así comenzamos a hablar con ellos y con nosotros mismos. [...] Este aprendizaje insensiblemente nos inserta en una historia: somos los descendientes no solo de una familia sino de un grupo, una tribu o una nación. A su vez el pasado nos proyecta en el futuro: somos los padres y los abuelos de otras generaciones que, a través de nosotros, aprenderán el arte de convivencia humana: saber decir y saber escuchar. El lenguaje nos da el sentimiento y la conciencia de pertenecer a una comunidad. El espacio se ensancha y el tiempo se alarga: estamos unidos por la lengua a una tierra y a un tiempo. Somos una historia. [...] La lengua es de todos y de nadie. [...] [N]uestra lengua [...] posee un conjunto de reglas pero esas reglas son flexibles y están sujetas a los usos y a las costumbres [...]. El idioma vive en perpetuo cambio y movimiento; esos cambios aseguran su continuidad y ese movimiento su permanencia. [...] [El] proceso [del lenguaje humano] es imprevisible y no esta fijado de antemano; es una diaria invención, el resultado de una contínua adaptación a las circonstancias y a los cambios de aquellos que, al usarlo, lo

308

OLE NEDERGAARD THOMSEN inventan: los hombres. [...]” 1 (Octavio Paz: “Neustra lengua” La Jornada 1997-0408)

The theory of ‘linguistic change’ (understood in Coseriu’s 1979, 1988 [1982], 1980b sense) is part and parcel of the theory of language — ‘language’ being an abstraction over individual, specific languages. A specific language is a concrete, real historical ‘entity’.2 It is a communicational instrument applied by an individual on an object level in normal communication, i.e. in discourse production and comprehension, and as such it is a static product (ergon), a cultural instrument taken for granted — a ‘premise’ in language use. On a parallel, meta-level this instrument is being renewed: made and remade, created and recreated, formed and re-formed, by its users, adapting it to the ever changing circumstances of language use (as mentioned in the quotation 3 above). To reiterate, it is “ever-evolving”, in that the language users (re-) create it continually — not only when acquiring it but also in the constant negotiation of its “current” norms, the norms “in force”, and in this sense it is a dynamic, processual entity (energeia), an “issue”, an agendum to be dealt with — a provisional ‘conclusion’ in ‘meta-usage’. But it is important to understand that it is primarily ‘process’ and can only be a ‘product’ insofar as it is primarily ‘process’. This was stressed by Coseriu (1957), see (2):

1

All human societies start and end with verbal interchange, with speaking and with listening. The life of each human being is a large and double apprenticeship: to know how to speak and to know how to listen. We begin listening to the people who are around us and in this way we begin speaking with those and with ourselves. [...] This apprenticeship insensibly inserts us in a history: we are the descendants not only of a family but also of a group, a tribe, or a nation. In its turn, the past projects us into the future: we are the parents and grandparents of other generations who, through us, will learn the art of human living together: to know how to speak and to know how to listen. Language lets us feel and recognize that we pertain to a community. In a greater spatio-temporal perspective the language unites us with a location and a time. We are a history. [...] The language belongs to everyone and to no one. [...] Our language possesses a set of rules but those rules are flexible and are subject to uses and customs [...]. The language lives in perpetual change and movement; the changes secure its continuity and the movement its permanence. [...] The process of the human language is unpredictable and is not fixed on beforehand; it is a daily invention, a result of a continuous adaptation to the circumstances and the changes by those who, by using it, invent it: the human beings. [...]. (Translation ONT) 2 Cf. Henri Delacroix (1924:128–129) cited in Jakobson (1973), “A language is a historical variation on the great human theme of Language.” Cf. also Chomsky’s view that man speaks the Human Language. 3 The paper is thus based on Coseriu’s (1988 [1982]) “dynamic conception of language as creativity (’ ε νέ ργ ε ι α)”— as “a two-fold productivity: productivity as regards the produced ‘objects’ and productivity as regards the production of the corresponding procedures of production (which themselves can be ‘produced’)”.

AN INTEGRATED THEORY OF LANGUAGE AND LANGUAGE CHANGE

309

(2) “[…] la lengua funciona y se da concretamente en el hablar. Tomar como base de la teoría de la lengua este hecho significa partir de la conocida afirmacion de Humboldt de que el lenguaje no es ’έ ργ ονsino ’ εν έ ργ ε ι α[...]. Realmente, y no en algún sentido metafórico, el lenguaje es actividad, y no producto. Más aún: sólo porque es actividad y se conoce como tal, puede abstraerse y estudiarse también como «producto» [...]. En efecto, para recordar una distinción aristotélica, una actividad puede considerarse: a) como tal, κ ατ ’’ ε ν έ ργε ι α; b) como actividad en potencia, κατ αδύν αμι ν ; y como actividad realizada en sus productos, κατ ’’ έ ρ γον . No se trata, evidentemente, de tres realidades distintas, sino de tres aspectos, mejor 4 dicho, de tres modos de considerar la misma realidad.” (Coseriu 1957:44–45)

Coseriu’s basic conception of language as Process is Aristotelian, roughly similar to the Peircean conception of reality as centrally Process — ‘secondness’, peripherally Product — ‘firstness’ — and Pattern — ‘thirdness’. It is also consonant with the Whiteheadian Process conception dealt with in Fortescue (2000, this vol.). Notice Coseriu’s basis in von Humboldt’s conception of language (cf. Coseriu 1988), which is also crucial to Chomsky (cf. Chomsky 1974 [1972]). Each single natural language user has his own language instrument as his private, personal, internal(lized) property, and as such it is an idio-lect, the languge of a single speaker–listener. However, at the same time he conceives of his instrument as being the “same as” — functionally identical with, that is, shared by each and every other member of his (section of his) speech community — or he offers it to every other member he communicates with for its adoption (or acquisition), and as such it is or becomes public property, an impersonal, inter-subjective convention, a social, collective agreement. Accordingly, it is basically not an individual-psychological ‘competence’ (in the Chomskyan sense), but a social-psychological ‘dia-lect’ (however not in the sense of de Saussure’s langue of the masse parlante; cf. Paz’s words above, “la lengua es de todos y de nadie”; see also Harder 2003). Coseriu

4

A language functions and exists concretely in speaking. Taking this fact as the fundamental premise in linguistics means taking as point of departure the famous claim of Humboldt that language is not product but process (creative activity). Really, and not metaphorically speaking, language is activity, not product. But more importantly, only by being an activity in itself is it possible to abstract it and study it also as a product. Recalling an Aristotelian distinction, an activity can be considered as such, as process (creative activity); as potential activity (pattern); and as activity realized in its products. Evidently, this is not a matter of three different realities, but of three different aspects, or modes of appearance of the same reality, namely language as creative activity. (Translation ONT)

310

OLE NEDERGAARD THOMSEN

(1957) emphasizes this conception of language as a social-psychological, cultural phenomenon of the individual, as in (3):5 (3) “En cuanto saber transmisible (y no simple «habilidad» estrictamente personal), el saber lingüístico es cultura. Esto significa que la lengua, además de fundar y «reflejar» la cultura no-lingüística — además de ser, como decía Hegel, «la actualidad [effectividad: Wirklichkeit] de la cultura [...]» —, es ella misma cultura [...]. En efecto, el hombre no sólo tiene conocimiento de las cosas por medio del lenguaje, sino que tiene también conocimiento del lenguaje [...]. En este sentido, el «aspecto cultural» de la lengua es la lengua misma como saber lingüístico. [...] Finalmente, en cuanto saber común de varios o de muchos hablantes, el saber lingüístico es interindividual o social; y en cuanto saber tradicional (y no universal), es un saber histórico. Por ello, precisamente, el punto de vista histórico puede adoptarse sin contradicción también con respecto a la lengua sincrónica: desde el punto de vista histórico (no diacrónico), la lengua sincrónica es un sistema actual de tradiciones lingüísticas antiguas y recientes [...].” (Coseriu 1957:60–61)6

According to the above contentions, the de Saussurean dichotomy between synchrony and diachrony is solely the case of two points of view on the same reality, namely the language instrument, which is assuredly a historical, processual entity. Therefore, it is possible to speak of an integrated theory of language and language change. More precisely, the language instrument is “vectorial”: a dynamic entity targeting an (ever) emergent synchronic state, and being constrained by panchronic universals. 5

Coseriu distinguishes (cf. quotation (11), below), in a way comparable to Chomsky’s i- and e-language, between the speakers’ functional languages and the historical language of the speech community. Thus, I do not really understand Andersen’s (2001a:236, fn. 9) contention that “[...] Coseriu does not operate with ‘speakers’ grammars’; his ‘norms’ are explicitly characterized as a speech community’s “historical norms” [...] Coseriu’s ‘norms’ thus cannot [to, del., ONT] be equated with the usage rules we may wish to posit for individual speakers’ grammars. [...]”. In fact, Coseriu distinguishes between the (functional) norms of the single speakers and the historical norms of the speech community (and likewise for the other levels of the ‘idiomatic’ language). 6 By being transmittable knowledge (and not simply a strictly personal capability), linguistic knowledge is culture. That means that besides being the basis of and “reflecting” nonlinguistic culture — besides being the “actuality [reality] of culture”, as Hegel puts it — it is in itself culture. Actually, man does not only have knowledge of the world by means of language, he also has knowledge of language. In this way, the “cultural aspect” of language is the language itself as linguistic knowledge. Finally, by being knowledge common to various or many speakers, linguistic knowledge is inter-individual, or social; by being traditional (and not universal) knowledge, it is historical knowledge. Therefore, precisely, the historical point of view can be adopted without contradiction also with respect to the synchronic language: from the historical (not diachronic) point of view, the synchronic language is an actual system of old and recent linguistic traditions. (Translation ONT)


311

A language, as claimed above, is a changeable and changing (negotiated) cultural instrument, and thus a concrete not an abstract entity. By being negotiable and negotiated, in language meta-usage, it is an ‘intentional’ cultural phenomenon: it is being ‘directed at’ in the phenomenological and Searlean sense of being a phenomenological ‘object’ (cf. the references below). Thus a basic feature of language is its reflexivity (‘ipsity’, ‘self-directedness’; cf. also Harris 1998). Another basic feature is its functionality, as stressed in (4), again from Coseriu (1957; cf. also Harder 2003): (4) “[…] la lengua no pertenece al orden causal sino al orden final [...], a los hechos que se determinan por su función. Si se entiende la lengua funcionalmente, primero como función y luego como sistema — y es así como hay que entenderla, pues la lengua no funciona porque es sistema, sino, al contrario, es sistema para cumplir una función, para corresponder a una finalidad. [...] Lejos de funcionar sólo «en ne changeant pas» [...], la lengua cambia para seguir 7 funcionando como tal.” (Coseriu 1957:29–30)

This feature of functionality, as contended by Coseriu in (4) and by Paz in (1), is the basis and motivation for linguistic change: the speakers have to remodel their language in order for it to be able to function to their satisfaction to cope with a changing “world” (and possibly their changing conception of it). Language, is thus primarily and basically a ‘functional language’, belonging with the finalistic order of things. It is a speaker’s technical, or practical “knowledge” (an artifact), that is, know-how for speech (a saber hablar, a ‘competence to perform’). In this sense we can speak of an integrated functional theory of language and language change. Language change occurs in language (meta-)usage (meta-communication), as already claimed by Coseriu (1957), see (5): (5) “La lengua cambia justamente porque no está hecha sino que se hace continuamente por la actividad lingüística. En otros términos, cambia porque se habla: porque sólo existe como técnica y modalidad del hablar. El hablar es actividad creadora, libre y finalista, y es siempre nuevo, en cuanto se determina por una finalidad expresiva individual, actuál e inédita [...]. El hablante crea o estructura su expresión utilzando una técnica y una material anterior que le

7

Language does not pertain to the causal order but to the final order, to the facts that are determined by their function. If language is understood functionally, primarily as function, only secondarily as system — and it is in this way one has to understand it, in that it does not function because it is a system, but on the contrary, it is system in order to be able to perform a function, in order to correspond to a purpose [finalidad]. Rather than solely functioning “by not changing”, language changes in order to continue functioning as such. (Translation ONT)

312


proporciona su saber lingüístico. La lengua, pues, no se impone al hablante, sino que se le ofrece: el hablante dispone de ella para realizar su libertad expresiva.” (Coseriu 1957:69)8

However, even though language use, that is, speech, is individual, as stated in (5) (cf. ‘idio-lect’ above), it is irreducibly collective (cf. ‘dia-lect’ above), that is, social interaction/intercourse — communication (from lat. commune ‘common’; cf. de Saussure’s parole)/discourse. This essential feature of reciprocity (Coseriu’s ‘alterity’, ‘other-directedness’) is integrated with historicity, as stated in Coseriu (1957), see (6) (cf. also Paz in (1) above): (6) “[…] la historicidad del hombre coincide con la historicidad del lenguaje. […] la lengua pertence a su historicidad, a su ser tal y cual [...]. De esta manera, el hablar no deja de ser libertad expresiva y finalidad significativa individual, pero se realiza necesariamente en un marco de determinaciones históricas constituido por la lengua [...]. Por otra parte, el lenguaje tiene historicidad, y es fundamento mismo de la historicidad del hombre, porque es diálogo, hablar con otro: «una conciencia que significa presupone una conciencia que interprete, es decir, que acoja el signo y lo comprenda [...]. El hablar es siempre «comunicar». [...] mediante la comunicación «algo se convierte en común» […]; mejor dicho, la communicación existe porque quienes hablan ya tienen algo en común que se manifiesta en el hablar uno con otro [...]. En este sentido, el lenguaje es, al mismo tiempo, el primer fundamento y el primer modo de manifestarse de la intersubjectividad [...], del ser con otro, que coincide con el ser histórico del hombre.” (Coseriu 9 1957:69–70) 8

Language changes precisely because it is not made (done) but is being made continuously through linguistic activity. In other terms, it changes because it is being spoken: because it only exists as a technique and modality of speaking. Speaking is a creative activity, free and finalistic, and it is always new, as it is determined by an individual expressive finality, which is actual and “inedited”. The speaking subject creates and structures his expression by using a technique and an already existing material to which he applies his linguistic knowledge. The language thus does not impose itself on the language user, but it makes itself available to him: the speaker has it at his disposal to realize his expressive liberty. (Translation ONT) 9 Man’s historicity coincides with the historicity of language. Language pertains to his historicity, to his being such and such a person. In this way speaking is always expressive liberty and individual significative finality, but is necessarily realized in a framework of historical determination constituted by language. On the other hand, language possesses historicity, and is the very foundation of man’s historicity, because it is dialogue, the speaking with another: a consciousness which signifies presupposes a consciousness that interprets, that is, which grasps the sign and comprehends it. Speaking is always “communicating”. By way of communication “something becomes common”; or rather, communication exists because those who are speaking already have something in common that manifests itself in one person speaking with another person. In this sense, language is, at the same time, the primary foundation and the primary mode of manifestation of intersubjectivity, of being with another person, which coincides with man’s historical existence. (Translation ONT)


313

It is thus evident that Coseriu’s theory is “usage-based” (if not in the Langackerian sense, cf. Langacker 2000), and therefore we can describe his theory as pragmatic. Summing up Section 1, it is an integrated functionalpragmatic theory of language and language change. Towards understanding language and language change In approaching the understanding of language change, i.e. of the “changing” cultural instrument, we need to situate the historical, functional language of the single individual speaker, his communicative competence,10 in a wider perspective. Understood as ‘Pattern’ (dynamis), as practical (procedural) knowledge, his language is, on the one hand, taken for granted and used as an “instrument” of communication; on the other hand, the pattern itself is an issue which is subject to change, it is ever “emergent”: the instrument has to be shaped and re-shaped continuously in an adaptation to the changing conditions of use, etcetera. However, this instrumental historical pattern, this idiomatic knowledge, does not function in and of itself: it is integrated in a wider system of communicative competences (‘levels of knowledge’, cf. Coseriu 1952), as in (7): 2.

(7) “[…] el hablar es una actividad universal que se realiza por individuos particulares, en cuanto miembros de comunidades históricas. Por lo tanto, puede considerarse en sentido universal, en sentido particular y en sentido histórico. El hablar κατ αδύ ναμι νes el saber hablar, en el cual pueden distinguirse un escalón universal, otro particular, y otro histórico: este último es, precisamente, la «lengua» como acervo idiomático, o sea, como saber hablar según la tradición de una comunidad. El hablar κατ ’’ ε νέ ργ ε ι α es, en lo universal, el hablar simplemente: la actividad lingüística concreta, considerada en general; en el particular, es el discurso (el acto o la serie de actos) de tal individuo en tal oportunidad; y en lo histórico es la lengua concreta, o sea un modo de hablar peculiar de una comunidad, que se comprueba en la actividad lingüística como aspecto esencial de la misma. En cuanto al hablar κατ ’’ έ ργον , no puede haber un punto de vista propiamente universal, pues se trata siempre de «productos» particulares: a lo sumo, puede hablarse de la «totalidad de los textos». En lo particular, el hablar como «producto», es justamente, el texto; y en lo histórico se identifica nuevamente con la «lengua» como «acervo idiomático», pues el «producto histórico», en la medida en que se conserva (o sea, en la medida en que se acepta como modelo para actos ulteriores y se inserta en la tradición), se vuelve

10

Cf. Bynon (1977:172): “If [...] we assume that each speaker’s linguistic activities extend over a certain range which may be defined in social and geographical terms, we may in turn use this range to define his communicative competence as the amount of linguistic heterogeneity which is part of his experience and lies within his powers of perception and interpretation, though not necessarily of active manipulation.”

314

OLE NEDERGAARD THOMSEN hablar κατ αδ ύναμι ν, es decir, saber lingüístico. Ello significa que la «lengua» no es nunca propiamente ’ έ ργ ον .”(Coseriu 1957: 45–46)11

As is partly evident from the above quotation (7), there are three levels of linguistic knowledge/know-how for speech (cf. Coseriu 1988 [1987]): 1. Universal: elocutional knowledge: language-general pattern: technique of coherent speaking/understanding and thinking 2. General: idiomatic knowledge: language-specific, historical pattern, tradition of speaking 3. Particular: expressive knowledge: specific patterns of discourse production and reception Thus, when communicating the speakers perform activities of three kinds, applying the three levels of knowledge/know-how: (1) general speech acts in accordance with universal patterns of speaking/understanding, thinking, and reasoning (cf. Gricean maxims; Andersen’s 2001b general semiotic Principle of Markedness Agreement; Hawkins’ 2004 universal laws of processing), (2) language-specific, idiomatic, speech acts (“locutions”) in accordance with a specific tradition of speaking, and (3) discourse acts (manifesting general patterns of language use, e.g. telling a story, writing a condolence letter, etc.). That speech (discourse) is manifested according to these three different kinds of knowledge can be represented as in Diagram 1.

Speaking is a universal activity that is realized by particular individuals, as members of historical communities. Therefore, speaking can be considered in a universal, a particular, and a historical sense. Speaking, as a potentiality, is know-how for speech, linguistic knowledge, in which a universal, a particular, and a historical level can be distinguished: the last level is precisely the language as an idiomatic resource, that is, as know-how for speech in accordance with the traditions of a community. Speaking as a process is in universal terms speaking tout court: the concrete linguistic activity considered in general; in particular terms it is the discourse (the act or the series of acts) of a given individual in a given opportunity (situation); and in historical terms it is the concrete language, or, a mode of speaking peculiar to a given community, which is given in the linguistic activity as the essential aspect of it. Speaking as a product cannot be considered in a universal perspective proper, because it always concerns particular products, but only as a “totality of texts”. In particular terms, speaking as a product is, precisely, the text; and in historical terms it is (again) identified with the “language” as “idiomatic resource”, because the “historical product”, in the degree to which it is conserved (or, in the degree to which it is accepted as model for ulterior acts and inserts itself in the tradition) it again becomes speaking as pattern, that is to say, as linguistic knowledge. This means that “language” is never product in any proper sense. (Translation ONT) 11


315

Elocutional Idiomatic Expressive Speech

Diagram 1: Speech and the three kinds of linguistic knowledge.

When focusing on idiomatic knowledge — the historical functional language (know-how) of the speaker — the picture has to be slightly modified, because there is a level in between idiomatic and elocutional knowledge, a level of linguistic universals (for an extensive exposition of linguistic universals, cf. Coseriu 1974). Inside idiomatic knowledge there are three levels, or functional planes (Coseriu 1977 [1968]), again, a universal, a general, and a particular, so to speak, viz. the level of the functional type, the functional system, and the functional norm — all levels of the dialect of the individual speaker-listener, cf. Coseriu (1957) in (8a) and (8b), Coseriu (1977 [1968]) in (9), visualized in Diagram 2: (8a) “En las estructuras que constituyen la lengua es importante distinguir entre lo que es simplemente normal o común (norma) y lo que es oposicional o funcional (sistema).” (Coseriu 1957:53–54)12 (8b) “De un modo general, [...] una lengua funcional (lengua que puede hablarse) es un «sistema de oposiciones funcionales y realizaciones normales», o, mejor, sistema y norma. El sistema es «sistema de posibilidades, de coordenadas que indican los caminos abiertos y los caminos cerrados» de un hablar «comprensible» en una comunidad [...]; la norma, en cambio, es un «sistema de realizaciones obligadas» [...], consagradas social y culturalmente: no corresponde a lo que «puede decirse» sino a lo que ya «se ha dicho» y tradicionalmente «se dice» en la comunidad considerada [...]. El sistema abarca las formas ideales de realización de una lengua, es decir, la técnica y las pautas del correspondiente hacer lingüístico; la norma, los modelos ya realizados históricamente con esa técnica y según esa

12

In the structures that constitute the language it is important to distinguish between that which is quite simply normal or common (norm) and that which is oppositional or functional (system). (Coseriu 1957:53–54) (Translation ONT)

316


pautas. De esta manera, el sistema representa la dinamicidad de la lengua, su modo de hacerse, y, por lo tanto, su posibilidad de ir más allá de lo ya realizado; la norma, en cambio, corresponde a la fijación de la lengua en moldes tradicionales; y en este sentido, pecisamente, la norma representa en todo momento el equilibrio 13 sincrónico («externo» e «interno») del sistema.” (Coseriu 1957:55–56) (9) “2.1. Las lenguas representan el «saber hablar» en cuanto históricamente determinado (perteneciente a comunidades históricas). Una lengua es, por lo tanto, un saber que se manifiesta en una actividad, un sistema de procedimientos o modos de hacer, o sea, un saber técnico — precisamente, una técnica histórica del hablar —, y presenta los caracteres comunes de todos los saberes técnicos [...]. Ahora bien, en los modos técnicos que constituyen una lengua, se pueden distinguir tres estratos functionales [...]: norma, sistema y tipo lingüístico. La «norma» abarca lo que en el hablar de una comunidad lingüística es técnica históricamente realizada, lo que en ese hablar es realización común y tradicional, aun sin ser necesariamente funcional [...]. El «sistema» representa el conjunto de las oposiciones funcionales (distintivas) comprobables en el mismo habla r, las reglas distintivas según las cuales ese hablar se reliza y, por consiguiente, los límites funcionales de su variabilidad; como tal, el sistema va más allá de lo históricamente realizado, pues abarca también lo que sería realizable de acuerdo con las mismas reglas ya existentes (parcialmente aplicadas en la norma). [...] Finalmente, el «tipo lingüístico» abarca los principios funcionales, es decir, los tipos de procedimientos y categorías de oposiciones del sistema, y representa, por ello, la coherencia funcional comprobable entre las varias secciones del sistema mismo. Así interpretado, el tipo es una estructura lingüística objetiva, un plano funcional de la lengua: es, simplemente, el nivel de estructuración más alto de una 14 técnica lingüística [...].” (Coseriu 1977 [1968]:194–195) 13

In a general sense, a functional language (a language which can be spoken by an individual) is a system of functional oppositions and normal realizations, or, better, system and norm. The system is system of possibilities, of coordinates that indicate the open ways and the closed ways of a speaking which can be comprehended in a community; norm, on the other hand, is a system of obligatory realizations, socially and culturally dedicated: it does not correspond to that which ought to be said, but to that which has already been said and traditionally is said in a given community. The system comprises the ideal forms of realization in a language, that is, the technique of the corresponding linguistic activity; the norm is the models that have already been realized historically with this technique. In this way, the system represents the dynamicity of the language, its mode of activity, and thus, its possibility of transgressing the limits of what has already been said; the norm, on the other hand, corresponds to the fixation of the language in traditional moulds; and in this sense, precisely, the norm represents, in every moment, the synchronic equilibrium (external and internal) of the system. (Coseriu 1957:55–56) (Translation ONT) 14 The languages represent the “know-how for speech” as historically determined (pertaining to historical communities). A language is, accordingly, a know-how (knowledge) that is manifested in an activity, a system of procedures or modes of making, or, a technical knowledge, precisely, a historical technique of speaking, and presents the features common to all technical knowledges. Now, in the technical modes that constitute a language, it is possible to distinguish three functional levels: norm, system, and linguistic type. The “norm” comprises that


317

Universals Type System Norm

Diagram 2: Universals and the three levels of idiomatic knowledge.

In the current stage of human language, the supra-idiomatic universal cultural level does not change; but linguistic change occurs on the functional levels of the norm, the system, and the type (cf. Coseriu 1977 [1968]). The functional (idiomatic) language (as a cultural instrument of speaking) always tends to be such that in its application every act of communication is in accordance with the constraints of the elocutional knowledge (cf. the Gricean maximes of quality, quantity, manner, and relation (relevance)): the structure of the grammar will accordingly be prone to processes of grammaticalization to this effect — the functional language is adapted to its use. Before we can continue approaching the understanding of language change, the picture has to be enlarged — the scope has to be widened: the language of the speaker, his or her functional language, is not only a cultural instrument, it is also a biological, psycho-physical “organ” (as in the Chomskyan tradition of the language faculty, “the human language”, cf. Chomsky 2000, 2005; Hauser, Chomsky & Fitch 2002) — just as man himself belongs to which in the speech of a linguistic community is historically realized technique, what in this speech is common and traditional realization, although not necessarily functional […]. The “system” represents the totality of functional (distinctive) oppositions which can be verified in the same speech, the distinctive rules in conformity with which this speech is realized, and, consequently, the functional limits of its variability; as such, the system is more progressive than the historically realized, insofar as it also comprises what would be realizable in accordance with the same rules already existing, however only partially applied in the norm. […] Finally, the “linguistic type” comprises the functional principles, that is, the types of procedures and categories of oppositions of the system, and represents, through that, the functional coherence verifiable between the various sections of the system itself. Interpreted this way, the type is an objective linguistic structure, a functional plane of the language: it is, simply, the highest structural level of a linguistic technique […]. (Coseriu 1977 [1968]:194–195) (Translation ONT)

318


both the natural and the cultural order of the universe. The biological capacity (the universal, biological Laws of Language, fysis) is the foundation for the cultural, communicative competences, dealt with above (the set of specific “laws”, rules, norms, or prescripts, nomos, (to be) obeyed by the speaker). The biological foundation is the result of a phylogenetic evolution (cf. e.g. Croft 2000, this vol.), it is an inherited trait of the human species (an inalienable possession), and as such it is present in each member of the species (passed on from generation to generation, in the present stage unchanged). It has an ontogenetic “development” in the single individual speaker (i.e. the maturation of the language faculty in the child, his so-called “language growth”). It is competence to perform in the biological sense of psychophysical speech production (articulation) and reception (perception). (This is the basis of the “internal causes” of linguistic change, as against the “external”, sociological ones having to do with e.g. prestige/stigma, language planning, etc.) The picture of linguistic competences is now roughly as represented in Diagram 3. Biological Elocutional Idiomatic Expressive

Diagram 3: The biological foundation of linguistic knowledge.

Human language understood as a cultural instrument, as a “memetic” phenomenon, is embedded in the biological language faculty (Coseriu 1952, 1987). Being a cultural entity, it is primarily a semiotic, especially a symbolic instrument of verbal communication and cognition (exo- and endo-semiosis). Just as we speak of the biological language faculty as a genetically inherited property of the individual and an ‘essential’ characteristic of the species, we conceive of the cultural ‘functional language’ as an acquired property (an alienable possession) of each single individual in the given speech community, applied as a know-how in speaking, cf. (10) from Coseriu (1957):


319

(10) “Los modos lingüísticos que se comprueban en el hablar concreto manifiestan […] el «saber lingüístico» de los hablantes. Para cada sujeto hablante la lengua es un saber hablar, el saber cómo se habla en una determinada comunidad y según una determinada tradición. Sobre la base de este saber, el hablante crea su expresión que, en cuanto coincide con las de otros hablantes o se adopta por ellos, integra (o llega a integrar) la lengua comprobada en el hablar. En este sentido todo hablante es creador de lengua «para otros». Pero el hablante no crea sino excepcionalmente sus propios modelos: el saber lingüístico lo adquiere continuamente de 15 otros hablantes [...].” (Coseriu 1957:57)

From this quotation it is evident that ‘idiomatic’ knowledge is first and foremost a “norm” acquired from the surroundings, a norm for him to follow. The totality (population) of functional languages of a speech community is its ‘historical language’. However, a historical language may be a disparate set, as expressed in (11), from Coseriu (1957): (11) “[…] hay que subrayar que la lengua funcional no debe confundirse con la lengua histórica o idioma (como, por ej., la lengua española, la lengua francesa, etcétera). Una lengua histórica puede abarcar no sólo varias normas sino también varios sistemas. [...] El «español» es, por lo tanto, un «archi-sistema» dentro del cual quedan comprendidos varios sistemas funcionales [Esos sistemas pueden ser regionales y pueden también coexistir en la misma región (por ej., en distintos estratos sociales o culturales).] El equilibrio entre los sistemas abarcados por un archisistema puede llamarse norma histórica [...].” (Coseriu 1957: 56–57)16

The ‘historical language’ would then be a(n alienable) property of the speech community: the speech community is a population of speakers, each speaker possessing his own functional language(s). The historical language is 15

The modes of linguistic activity that are observed in concrete speaking manifest […] the linguistic knowledge of the speakers. For each speaking subject the language is know-how for speech, knowledge of how one speaks in a given community and according to a given tradition. On the basis of this knowledge, the speaker creates his expression, which, insofar as it coincides with the knowledge of the other speakers or is adopted by those other speakers, integrates (or comes to integrate) the language verified in speech. In this sense, each speaker is the creator of the language “for others”. However, the speaker only exceptionally creates (ex novo) his own moulds: he acquires his linguistic knowledge continuously from other speakers. (Translation ONT) 16 It has to be stressed that the functional language should not be confounded with the historical language or idiom (e.g. the Spanish language, the French language, etc.). A historical language may comprise not only various norms but also various systems. Thus, “Spanish” is an “archi-system” within which various functional systems are comprised. Those systems may be regional, or may even coexist in the same region (e.g. in distinct socio-cultural strata). The equilibrium between the systems comprising an archi-system may be termed the historical norm. (Translation ONT)

320


accordingly simply the population of functional languages, their union set, in Coseriu’s term a ‘dia-system’. In another sense, the historical language would be a separate, hypostasized entity, lying “out there” between the speakers, cf. Harder 2003. However, insofar as the speech community is not a macroindividual, only a set (communicational network) of individuals, not a de Saussurean masse parlante, the historical language could not be a monolithic de Saussurean langue. We are accordingly nominalists in this respect. Language change: Meta-grammar and meta-communication The primary and basic contention of the integrated functional–pragmatic theory of language and language change is that language change and acquisition occurs in (conversational) speech — in language usage, that is, in language as process —, but we should be more precise: language change (reanalysis) is manifested (actualized) in speech, the change (innovation) itself being a covert mental process, belonging at the level of the idiomatic functional language. The language user renews his functional language — reshapes his communicative instrument before applying it. The first claim is that “language change does not exist” (Coseriu 1988 [1982]) in the traditional sense of “change” (the traditional, de Saussurean sense implies that the essence of language is static ‘being’, not dynamic ‘becoming’). It does not exist because the essence of (a functional) language is its being processual rather than a static object (a result). The next claim — on the face of it, contradictory — is that a functional language (and thereby the derivative, historical language, idiom) is always and constantly changing: each occurrence of speech, each text/discourse, is in principle “new” — even when it is or contains ‘repeated speech’, that is, a repetition (replication) of previous speech — precisely because each situation of language use is a unique historical entity, uniquely located in time and space, and performed by particular language users — it is a token. In each situation of communication, the speakers “negotiate” their dialects, that is, their idiomatic knowledges, and thus after each communicative situation, each functional language is up-dated, and derivatively, the historical language (idiom) is renewed. Accordingly, change occurs in the mental grammars, but owing to speech, conditioned by the factors of speech. However, each (‘normal’) utterance is put forth with a claim that it belongs, or should belong, within the already sanctioned norms (in this sense only norm continuity really exists) — but with the proviso that these norms always be ‘functional’ (in the sense of “efficient”), or ‘optimal’: the best ones under the given circumstances in the given communicative situation. In that sense each occurrence of language use is a ‘renewal’ of the language, a 3.


321

speech that is (meant to be) ‘up-to-date’, that is, belonging within the valid current norms, in accordance with the idiom “in force”. That language use is constantly implying a ‘negotiation’ of the current norms of the language, of what are the norms “in force” — a “legislation” — means that we have to operate with a level of meta-usage (of meta-communication, or meta-dialogue — see below), and of meta-grammar, that is, the technique of how to negotiate language (here not Coseriu’s idiomatic metalinguistic ways of speaking about language in a given tradition of speaking). Meta-grammar must then be a component operative in language acquisition (“learning” and “teaching”; cf. the Chomskyan ‘language acquisition device’; Andersen’s (2001a:245) “innate principle of grammar formation, which is a kind of procedural knowledge”) and language change, both processes being specific kinds of ‘negotiating’ language. This gives the picture in Diagram 4: Meta-grammar Grammar Meta-usage Usage

Diagram 4: The object and meta-levels of language and speech.

The picture does not imply that there are two separated kinds of language use—use integrates meta-usage and “object-level” usage (‘normal’ linguistic conduct). A conception of linguistic change as a “form of communication” was already proposed by Labov (1980). Later, it was adopted and modified by Andersen (1989) who describes it as in (12): (12) “[...] [I]t is reasonable to recognize two kinds of metadialogue which take place concurrently with the use of language for its regular, communicative purposes. One of these concerns the constitution of the linguistic system. It consists of learners’ experimentation with their language [...]. Its messages are the substantive innovations a learner produces and the perceived feedback regarding the successfulness and acceptability of these innovations, which serves to confirm or disconfirm the learner’s hypotheses about the productive system and the norms of the language.

322


The other metadialogue concerns the norms of usage more narrowly, for it is focused on the social values of novel expressions. This is a dialogue in which all members of a speech community participate throughout their lives. In this metadialogue, the use of an innovation in a specific context is tantamount to a motion that it be accepted for general use in such contexts, a motion which the interlocutors may second, reject or, for the moment, leave undecided. As they select novel and traditional expressions in accordance with their individual hypotheses about their appropriateness — relative to the genres of discourse, styles of diction, categories of role and status and social class recognized in the culture — the speakers in effect negotiate the norms that they look upon as their community norms. [...] The [...] metadialogue [...] is part and parcel of the life of any language tradition. For no community values can be established except through the dialectic of usage, and no values can be maintained except through renewal. Thus it is by the temporary consensus mediated by this unspoken dialogue that the tacit conventions of a language are shaped and constantly reshaped as long as the language is spoken.” (Andersen 1989:24f)

As is transparent from the quotation in (12), not only the language ‘as such’ is negotiated, but also its social evaluation is subject to metalinguistic “discussion”. In addition to this socio-cultural evaluation, any linguistic entity is subject to a (re-)valuation in terms of markedness whereby a paradigmatic option may be either obsolete (‘archaic’, defunct, non-productive), marked, or unmarked. In the course of markedness valuation ‘markedness shifts’ may occur, as envisaged in e.g. Dik (1997, TFG 1; see also Becerra Bascuñán, this volume; Andersen 2001a, b, c). The meta-grammar has as its premise the language universals (the constitution of a possible human, cultural language as regards the panchronic, diachronic, and synchronic universals). Universals Meta-Grammar Grammar Use

Diagram 5: The Universals as premise.


323

The meta-grammar contains linguistically adapted ways of reasoning manifest in the elocutional competence mentioned above — Peircean Abduction, Induction, and Deduction (cf. Andersen 1973; Anttila 1975, 1992).17 In language acquisition the speaker uses his meta-grammar, the universals of language, and the linguistic output of the surroundings, the others’ use, as well as the non-linguistic context of speaking in the creation (‘abduction’) of his object-level grammar,18 just as the speaker-listener in the creation of objectlevel linguistic behavior — in production and reception — in addition to his object-level grammar must have access to the universal ways of reasoning of the elocutional knowledge in order to be able to formulate and interpret discourse (using peripherally the biological capacity in articulation and perception). It is important to stress that language ‘updating’ also concerns inductive, i.e. frequency and probability aspects of language use, as focused on in usage-based Cognitive Linguistics (cf. Langacker 2000; Tomasello 2000, 2003) and in Functional Phonology (Boersma 1998).19 For instance, frequent linguistic patterns behave differently diachronically from infrequent patterns (cf. the retention of system non-conforming forms like the past tense of the English copula, was; cf. system-conforming *beed). In general, the structure of the individual’s idiomatic, historical functional language is shaped by its psycholinguistic and phonetic use, that is, by processing (cf. Berg 1998; Boersma 1998; Hawkins 2004) — the cultural instrument has to be efficacious.20 This does not imply that language change is ‘teleological’ in the sense that the language users intend the change as a goal in final causation (they may have no conscious knowledge that they change their 17

For a critique of Andersen’s (and others’) (mis-)application of Peircean Abduction, see Deutscher (2002). For a revised schema of the Andersenian “abductive” model, see Janda & Joseph (2003). For Chomsky’s early recognition of abduction in language acquisition, see Chomsky (1974 [1972]). 18 For a detailed investigation of the learning of a production grammar in Functional Phonology, see Boersma (1998:269–346). 19 This is denied any relevance in Andersen (2001a:244) who claims that “[...] there is no reason to suppose that speakers keep running-frequency counts [...]”. By the way, this denial seems inconsistent with a further suggestion made by Andersen (2001a:235): “[...] one might suppose that all speakers have a memory of experienced usage, and that speakers generally feel most comfortable keeping their own usage within the bounds of what is usual in their experience.” This memory component is (part of) the functional norms of the speaker. 20 This is stated in Hawkins’ (2004:3) Performance-Grammar Correspondence Hypothesis: “Grammars have conventionalized syntactic structures in proportion to their degree of preference in performance, as evidenced by patterns of selection in corpora and by ease of processing in psycholinguistic experiments.” The hypothesis is robustly confirmed by Hawkins’ results.

324


language, cf. Coseriu 1988 [1982]), but change is indeed ‘teleological’ in the functional sense that the language users always and everywhere tend to locally optimize their functional languages (Boersma 1998), with different kinds of functionality arising due to competing motivations (cf. Berg 1998). As already stated, the existence of the meta-grammar is a reflection of the essential property of language of its ‘reflexivity’, that is, that it can be about, or directed at, itself — that messages can be about (the constitution of) the code, to use Jakobson’s (1960) terminology. Also stated above was the essential property of language of its ‘reciprocity’ (Coseriu’s ‘alterity’). This feature reflects man’s basic and irreducible ‘collective intentionality’ (Searle 1990b; Tomasello & al. 2005). It implies that the single individual’s functional language is a collective ‘dia-lect’ rather than a private ‘idio-lect’ (Coseriu 1980a), and that speech, or use, is ‘dia-logic’ rather than ‘mono-logic’. Thus, the meta-communication (negotiation) involved in (language acquisition and) change is also dialogic, cf. (13) from Coseriu (1957): (13) “3.1. El cambio lingüístico tiene su origen en el diálogo: en el paso de modos lingüísticos del hablar de un interlocutor al saber del otro. Todo aquello en que lo hablado por el hablante — en cuanto modo lingüístico — se aleja de los modelos existentes en la lengua por la que se establece el coloquio, puede llamarse innovación. Y la aceptación de una innovación, por parte del oyente, como modelo para ulteriores expresiones, puede llamarse adopción [Se trata, naturalmente, del diálogo reducido a su esquema mínimo. El diálogo real es mucho más complejo. El hablante real no sólo innova, sino que, al mismo tiempo, difunde innovaciones ajenas. Además, una «innovación» puede surgir también en el oyente, por ej., por imperfecciones de la percepción o por incomprensión de lo significado por el hablante. Por otra parte, cada uno de los dos interlocutores del diálogo es al mismo tiempo hablante y oyente, y cada hablante se oye también a sí mismo. Finalmente, el oyente no «aprende» del hablante sólo «innovaciones», sino también modos tradicionales que simplemente desconoce.] Esta distinción quizás parezca obvia y de poca importancia; sin embargo, es fundamental para la comprensión y el planteamiento correcto del problema teórico del cambio lingüístico. Muchos estudiosos parecen pensar que, habiendo explicado la «innovación», han explicado el «cambio»; mas éste es otro error que procede de tratar el problema en el plano de la lengua abstracta. En efecto, en la lengua abstracta cada modelo es único (un fonema, una palabra); pero a cada modelo de la lengua abstracta corresponden un gran número de modelos en los muchos saberes individuales, y no es pensable que éstos se modifiquen simultáneamente.” (Coseriu 1957:78f)21 21

Linguistic change has its origin in dialogue: in the transmission of linguistic modes of speaking from one interlocutor to the knowledge of the other. Everything in the speech of the language user, in terms of linguistic mode, which diverges from the existing models in the


325

Thus, according to Coseriu, language change is (the accomplishment of) the joint interaction of the initiating proposal by the initiator — her metaidiomatic speech act of ‘innovation’ — and the terminating acceptance by the respondent — his meta-idiomatic speech act of ‘adoption’ and ‘selection’ in use (i.e. ‘deductive’ change, Andersen 1973). Language change, by being a conversational dialogue, is subject to the Gricean principle of cooperation, which belongs with elocutional knowledge. The change of the ‘historical language’ of the speech community (‘idiom’) is established by the ‘diffusion’ of the innovative variants through the communicative networks of the speech community, until the final ‘mutation’ (of the historical language) in the speech community as a whole. For a precise definition of Coseriu’s original proposal of the types of speech act accomplished by the speakers in the diachronic metalinguistic dimension, see Coseriu (1988 [1982]:151,157). It is noticeable that ‘reinterpretation’ (e.g. reanalysis of the linguistic type) as well as ‘application’ (the synchronic functioning of a given functional level yielding change on the next lower level, as e.g. the application of the type determining a change of the system and actualized as ‘drift’, cf. Andersen 1990) are types of innovation. Andersen (2001a) adds ‘acquisition’ as a meta-act similar to adoption. A juridical, speech-act theoretical conception of language and language change 4.1 A multi-level theory of language change As contended in the previous section, usage is two-fold, as is grammar: (1) there is a primary object level of linguistic behavior and corresponding 4.

language in which the dialogue is established, may be termed innovation. And the acceptance of an innovation, by the hearer, as model for further expressions, may be termed adoption [Here we are speaking of a dialogue reduced to its minimal schema. A real dialogue is much more complex. The real speaker not only innovates, but at the same time also diffuses innovations made by other speakers. Furthermore, an “innovation” may also arise with the hearer, for instance owing to imperfections of perception or faulty comprehension of what the speaker means. In another way, each one of the two interlocutors of the dialogue is at one and the same time speaker and hearer, and each speaker also is his own hearer. Finally, the hearer not only “apprehends” “innovations” from the speaker, but also traditional modes of speaking which he quite simply does not know.] This distinction may seem obvious and of no importance; however, it is fundamental in order to correctly understand and pose the theoretical problem of linguistic change. Many researchers seem to believe that, having explained the innovation, they have eo ipso explained the “change”; there is an additional error which proceeds from treating the problem in terms of the abstract language. In effect, in the abstract language each model is unique (one phoneme, one word [i.e. types, ONT]); but to each model in the abstract language corresponds a great number of models in many individual knowledges [saberes], and it is not conceivable that those are modified simultaneously. (Translation ONT) [For a treatment of the last claim above, see Fortescue, this vol.]

326


patterns, or norms of this behavior (in a general sense of ‘norm’) — the functional language, and (2) a secondary meta-level of meta-communication and meta-grammar (the technique of ‘negotiating’ the object-level grammar in acquisition and change). The proposal of the present section, and the specific, and hopefully original and innovative, proposal of the overall contribution, is that this conception is of the same kind as a juridical system. However, rather than being a metaphor, this conception is supposed to be factual: language 22 belongs to the same social order as a juridical system, I shall claim. The object level is a level of ‘conduct’ where the language users follow the norms of their functional languages — or transgress them, as the case may 23 be. The meta-level concerns two sublevels, namely the ‘legislative’ and the ‘adjudicative’ sublevel.24 The legislative sublevel, as the term suggests, is the level of linguistic ‘legislation’ where (a) the norms of the language and their validity are negotiated, new rules being proposed and accepted (passed) or rejected; and where (b) these norms are ‘promulgated’ (made public/proclaimed) for adoption and acquisition, and, if successful, are transmitted (diffused: adopted and acquired), in the speech community. The role models in adoption and acquisition, the ‘promulgators’, either passively display the norms (by following them on the level of conduct), or more actively “teach” them (by being ‘instructors’). Correspondingly, the adopting and acquiring subjects more or less actively “learn” the norms from their models (primarily by way of Abduction, cf. above). Promulgation is an interactive meta-dialogue, as is perhaps especially evident in lexical acquisition (cf. E. Clark 1999). The promulgation meta-dialogue is intimately connected with adjudication on the next sub-level, in that the language learner normally presupposes that the linguistic behavior of the model, the ‘promulgator’, represents a ‘correct’ application of the linguistic norm (his functional language), and that the norm itself is ‘correct’ (that is, “identical” with the other functional languages of the (competent) members of the speech community). (It should be noticed that in speaking, and even within the same discourse, one and the same speaker may manifest both an innovative variant and the older conservative variant (cf. Andersen 2001a:234), on the face of it perhaps an enigmatic situation. This 22

For a treatment of a juridical system in the context of speech act theory, see the treatise (A. Ross 1968) by the Danish legal philosopher and philosopher of language, Alf Ross. 23 Notice that normal discourse production on the object level concerns not only the ‘referential’ function of language (the Searlean speech acts, directed at the ‘context’) but also an integrated “meta-discursive” function (non-Searlean speech acts directed at the discourse itself, the ‘contact’, and the speech act participants; cf. Jakobson 1960). 24 In Jakobson’s (1956) terms meta-communication is directed at the ‘code’, his “metalingual” function (cf. the classic Jakobson 1960).


327

could only be understood in the way that the speaker promulgates the innovation (on the legislative level), but obeys the older norm-sanctioned rule on the level of conduct — a given speaker indeed performs several roles in a given discourse.) The adjudicative sublevel (judicial–executive sublevels) concerns linguistic ’adjudication’, that is, the sanctioning of linguistic behavior (cf. also Coseriu 1987): when adjudicating, a given language user (functioning as ‘adjudicator’) makes a sanctioning decision as to whether his interlocutor (e.g. a language learner) by his linguistic conduct has violated a linguistic norm, this violation evidently being established by comparing his own (understanding of the) norms “in force” — his own dialect — with that of his interlocutor. If there has been a violation, the adjudicator ‘disapproves’, and negative sanctions may be ‘executed’, in the form of negative feedback, e.g. correction. If the norms are followed, the adjudicator ‘approves’, and, for instance in the case of child language acquisition, positive feedback may be issued (Christiansen 2002). The correction by the adjudicator is, in terms of illocutionary force, a personal, disinterested directive, a so-called ‘exhortation’ (A. Ross 1968): the adjudicator, solely from respect for the linguistic norms, appeals, not to the interlocutor’s own interests, but to his respect for these norms (which he takes should be common).25 The violation of the norms may be intentional (willful) or unintentional (not willful). If the violation is due to lack of knowledge of the ‘norms’ (‘norms’ now in Coseriu’s sense), and a correction is executed of a system-conforming form, the language learner may either accept the correction on the face of it, and patch up with an added, adaptive norm filter (cf. Andersen 1973), or he may de facto accept the correction and revise, i.e. update, his version of the norms. (It should not go unnoticed that the language learner may be his own adjudicator — this is seen in self-corrections: beedwas, or the like.) In addition to being concerned with judging whether linguistic conduct is norm-following (‘correct’) or norm-breaking (‘incorrect’) with respect to the idiomatic, language-specific norms, the adjudicator may also judge whether it is a ‘congruent’ (coherent) or ‘incongruent’ (incoherent) instance of universalpragmatic elocutional behavior (cf. elocutional knowledge above, Coseriu 1988 [1987]), and whether it is an ‘apt’ or ‘inept’ discourse — belonging or not within the discursive norms valid in his speech community (cf. expressive

25

This is not to deny the evident occurrence in the speech community of heated (‘interested’) discussions on orthography, orthoepy, etc., that is, negotiations of the prescriptive historical norms.

328


knowledge above, Coseriu 1988 [1987]; cf. also Andersen’s 2001a appropriateness norms). 26 Having sketched rather programmatically the multilevel conception of language acquisition and change in general, let me now focus more specifically on the ‘legislative’ aspects. ‘Legislation’, that is, negotiation of the linguistic norms and their validity, is a meta-dialogue, as claimed several times already. The initiating turns are ‘adhortative’ speech acts (cf. first person plural imperative) where the ‘legislator’ proposes some new norms (innovation). In doing so, he commits himself to also follow these proposed norms, and he simultaneously requires his interlocutor to also follow them (or at least suggests or advices him to do so). The reaction by the other party, his responding turn at speaking, is a ‘commissive’: the interlocutor either accepts the proposal of the legislator, thereby committing himself to comply with the obligations of the new norms on the level of conduct (the proposal is passed), or he rejects it. Acceptance is followed by adoption of the new norms into the interlocutor’s dialect, and their application (selection) in discourse on the object level. By the respondent’s accepting the initiator’s proposal the proposed norm becomes ‘in force’, becomes valid — a reciprocal obligation, a collective agreement. By conceiving of language acquisition and change as a kind of linguistic legislation, as just sketched, it becomes self-evident — and perhaps trivial — that a single language user’s functional language (an i-language in Chomskyan parlance) is a collective (social-psychological) ‘dialect’, rather than a (totally) private ‘idiolect’: automatically, owing to the “promising game” of adhortative innovation and commissive acceptance and adoption in linguistic legislation, the speakers’ functional languages are ‘dia-lects’, that is, collective ‘agreements’ (communal obligations), or ‘conventions’. In this sense, the functional languages are norms, that is, patterns with a ‘directive’ direction of fit, whereby the linguistic behavior (on the level of conduct) has to fit the conventions, has to be ‘correct’ manifestations of the norms. (It is, of course, this question of fit that is adjudicated on the adjudicative sublevel.) The conventional character of the functional languages, the linguistic norms, is again to be understood on the basis of the fundamental, irreducible ‘collective intentionality’ of human beings (defended by Searle 1990b, 1995, 1998, 2001, 26

Corrections, or supportive behavior, may also occur with respect to the faulty (‘abnormal’) application of the biological/psychophysical linguistic “knowledge” — e.g. helping a person stuttering (notice that here it is capacity which is impaired, rather than its application being incorrect). Evidently, ‘abnormal’ speech behavior does not represent innovations and thus does not trigger linguistic change. Cf. also Dressler (1995).


329

2002, 2004; cf. also Harder 2003, Tomasello & al. 2005; cf. additionally Coseriu’s essential concept of ‘alterity’, i.e. reciprocity). It is important to stress that the same kinds of speech acts are operative on the meta-communicative level as on the communicative, object level. This is because both levels are constituted or regulated by the same universalpragmatic ‘elocutional’ knowledge (cf. Section 2). By themselves, of course, the receptive acts of acquisition, acceptance (rejection), and adoption, and the productive acts of instruction, proposal, and correction are mono-logic metacommunicative speech acts, performed with an ’individual intentionality’ (Searle 1983). However, no monologue occurs on its own, but only by being an integral part of a dialogue, a turn-taking, a collective interactivity. In the legislative meta-dialogues, the communicators, in “negotiating” the constitutive and regulative rules of their language, act with ’collective intentionality’ (Searle 1990b, 1998; Tomasello & al. 2005), presupposing the universal principle of cooperation — they co-perform a collective ‘promising game’, as mentioned above. The individual contributions to this ongoing meta-dialogue are converted from illocutionary into perlocutionary acts with perlocutionary effects, which cumulatively form the social, institutional fact of the linguistic conventions of the functional languages of the speech community (and thereby, derivatively, of its ‘historical language’, in Coseriu’s sense, see above). These agreements between the individual speakers of the speech community are only unspoken, or tacit, insofar as the legislative meta-dialogues make use of the same functional languages as the object-level dialogues, in reflexive, selfreferential meta-communicative acts. In that the same functional language both functions on the object level of conduct and on the meta-level of acquisition and change, it is, on the former dimension, a ‘practice’ of speaking and, on the latter, a ‘tradition’ of speaking (cf. Andersen, this vol.). (Ultimately, a speaker’s functional language is part of culture and thus a ‘memetic’ phenomenon.) The conception of a functional language as a juridical norm Before concluding my contribution I shall have to be more specific about the conception of a functional language as a juridical, legislative norm. (Accordingly, on this level, the distinction within the idiomatic “knowledge” of the single language user between norm, system, and type has to be disregarded — the overall character of language being its ‘normativity’.) Thus, it should be emphasized that the ontological status of a functional language is primarily “deontic” (know-how), only secondarily “epistemic” (knowledge). This implies that discourse production and reception, in “intra-dialectal” communication, are (primarily) based on active rule-following — in “interdialectal” communication, production, and especially, reception may second4.2

330


arily involve the “passive” part of the speaker’s communicative competence, that is, his hypothesized rules of another dialect than his own (cf. fn. 8). This passive part is primarily knowledge (belief) rather than know-how. The linguistic norms (in the general, juridical sense) are social norms, that is, impersonal directives that correspond to certain social facts in such a way that the patterns of linguistic behavior of the norms are in general followed by the speakers and are felt by them as collectively binding, or valid — “in force”. The linguistic norms are thus mutual obligations/conventions/ agreements between the individual speakers comprising (a section of) a speech community. They are, of course, tacit in the sense that they are internal(ized) functional languages (however, verbalized in meta-communication, as mentioned above) — they are unspoken and unwritten, a practice of speaking. Only parasitically may they be externally registered, as for instance in the case of orthographical norms and other written prescriptive norms, which are agreed upon only by an elitarian legislative assembly. That the linguistic norms are tacit means that they are not directly available, only indirectly through their manifestation in object-level discourse, as ‘followed pattern’ (which then comes to function as ‘promulgation’, cf. above). The linguistic norms only exist as tokens in the minds/brains of single individual speakers, as their internal functional languages, their dialects, and their hypotheses about each other’s dialects (cf. the concept of communicative competence, fn. 8). A (higher-level) communal, “shared” Dialect (the historical community language, termed ‘idiom’ by Coseriu), i.e. a monolithic type behind the individual token dialects, which are internalized in the individual speakers’ minds/brains, ‘subsists’ solely in a ‘third realm’, outside the real mental and external worlds (in the manner of Frege, Popper, and Leech — see Leech 1983; cf. also Harder 2003), that is, it only exists if such a universe does indeed exist. The present proposal, subscribing to Searlean realism, biological naturalism, and methodological individualism, does not accept such a phantom world.27 Rather, it only recognizes the population of the language users’ functional languages and their respective theories about the others’ functional languages. To be more specific about the characteristics of a linguistic norm, we may try to cross-classify it within the taxonomical scheme of A. Ross (1968), as seen in Table 1: 27

This is not to deny the “causal” reality of external, prescriptive norms, because they may be ‘internalized’, that is, become ‘functional norms’ (“i-norms”, cf. Chomskyan i-language), and thus become operative in speech production and reception. On the other hand, the external, “enorms” are ‘externalizations’ of the (more or less shared) i-norms of an elitarian group. (For an extensive treatment of the norm concept in language and language change, see Bartsch 1987.)


Impersonal directives

Manifestation

Quasi-commands

Law and convention

Whose interests are served by the compliance with the directive? Society’s

Constitutive rules based on mutual agreement Directives of morality

Rules of games

No one’s

Moral principles and judgments

No one’s

331

Source of effectiveness

Fear of sanctions and respect for impersonal authority Agreement and mutual approval One’s own approval

Table 1: A taxonomy of impersonal directives, according to A. Ross (1968).

A. Ross (1968) recognizes a specific kind of quasi-command, namely, customs, that is, norms whose correspondence between the impersonal directives and the social facts arises not through legislation in the proper, concrete sense (i.e., not my generalized sense), but spontaneously as a product of an organic and unconscious “evolution”, a slow process of adaptation under the pressure of forces (customs thus being functional phenomena). With this kind of quasi-command there exists no authority whose (sole) function it is to enunciate the directives. Customs may be legal (in the narrow sense, judicial) or conventional (non-judicial). The functional language of the single individual (dialect, idiomatic knowledge) seems to be foremost of the conventional kind of custom, but it seems evident that a given language also possesses ingredients of the constitutive rules of games, and of the regulative rules of directives of morality. However, unfortunately, in this rather programmatic paper there is no room for further discussion of these matters, except for mentioning that the concept of the historical language (of the speech community, Coseriu’s ‘idiom’), with its distinction between standard, nonstandard, or substandard varieties, etc., shows that language in the general sense somehow covers the whole continuum of impersonal directives. By now, it should be evident that language and language change necessarily has to be viewed within a non-reductionist, “maximalist” functional–pragmatic theory of the Natural Language User, his (meta-)communicative and (meta-)grammatical competences, as well as his (meta-)communicative linguistic behavior. Language is a phenomenon of ‘collective intentionality’, and thus both a biological ‘organ’ (natura) and a cultural ‘instrument’ (ars). This integral know-how is an “emergent” conventional custom, a habit (usus) — a (synchronic) practice and a (diachronic) tradition.

332


Conclusion and perspectivization It has been contended in the present contribution, following the vein of Coseriu, that language is ‘Process’ (energeia), in that it is constantly being renewed, renovated, up-dated by the co-work of the particular, historical language users in historical communicative situations, in concrete circumstances, and in particular dialogues. Language change belongs to the social (sociopsychological) normative order of things and is of a similar kind as legislation, decision, and execution in a legal system. Linguistic change and language acquisition occur in meta-communication between actual speakers. It is on or through this level that a language can be said to be constantly transmitted and acquired, formed and re-formed, (re-)created. The present chapter is (also) meant as a contribution to the ongoing endeavors of constructing a functionally adequate theory of the so-called Natural Language User (cf. Dik 1997; Bussuyt 1983, Weigand 1990, Bakker 1998, Hengeveld 2004a,b, Hengeveld & McKenzie, forthcoming; Nedergaard Thomsen, forthcoming), a theory that should be capable of accounting for the fact that language and speech are the outcome of the co-work of single historical natural language users (operating in collectives) of a given historical speech community both individually and as a collective, on the basis of individual and collective intentionality. 5.

Acknowledgements I would like to express my gratitude to Silvia Becerra Bascuñán for her invaluable help with the translations from Spanish and, especially, for many exciting discussions on Coseriu’s theory of language and language change — and not the least for the marvellous quotation from Octavio Paz. REFERENCES Andersen, Henning. 1973. “Abductive and deductive change”. Language 49. 765–793. Andersen, Henning. 1989. “Understanding linguistic innovations”. Language Change: Contributions to the study of its causes ed. by Leiv Egil Breivik & Ernst Håkon Jahr, 5–25. Berlin & New York: Mouton de Gruyter. Andersen, Henning. 1990. “The structure of drift”. Historical Linguistics 1987: Papers from the 8th International Conference on Historical Linguistics ed. by Henning Andersen & Konrad Koerner, 1–20. (= Current Issues in Linguistic Theory, 66.) Amsterdam & Philadelphia. John Benjamins. Andersen, Henning. 2001a. “Actualization and the (uni)directionality of change”. Actualization: Linguistic change in progress ed. by Henning Andersen, 225–248. Amsterdam & Philadelphia: John Benjamins.


333

Andersen, Henning. 2001b. “Markedness and the theory of linguistic change”. Actualization: Linguistic change in progress ed. by Henning Andersen, 21–57. Amsterdam & Philadelphia: John Benjamins. Andersen, Henning. 2001c. “Introduction”. Actualization: Linguistic change in progress ed. by Henning Andersen, 1–20. Amsterdam & Philadelphia: John Benjamins. Andersen, Henning. This volume. “Synchrony, diachrony, and evolution”. Anttila, Raimo. 1975. “Generalization, abduction, evolution, and language”. The Transformational–Generative Paradigm and Modern Linguistic Theory ed. by E.F.K. Koerner, 263–296. (= Current Issues in Linguistic Theory, 1.) Amsterdam: Benjamins. Anttila, Raimo. 1988. “The type and the comparative method”. Energeia und Ergon. Vol. 2: Das sprachteoretische Denken Eugenio Coserius in der Diskussion ed. Harald Thun, 75–83. Tübingen: Gunter Narr. Anttila, Raimo. 1992. “Historical explanation and historical linguistics”. Explanation in Historical Linguistics ed. by Garry W. Davis & Gregory K. Iverson, 17–39. Amsterdam & Philadelphia: John Benjamins. Bakker, Dik. 1998. “Language change: Some implications for the theory of Functional Grammar”. Working Papers in Functional Grammar 66. Bartsch, Renate. 1987. Norms of Language: Theoretical and practical aspects. London & New York: Longman. Becerra Bascuñán. This volume. “Grammaticalization of Indirect Object Cross-Reference in Spanish as a case of Drift”. Berg, Thomas. 1998. Linguistic Structure and Change: An explanation from language processing. Oxford & New York: Oxford University Press. Boersma, Paul. 1998. Functional Phonology: Formalizing the interactions between articulatory and perceptual drives. LOT/Netherlands Graduate School of Linguistics & IFOTT. The Hague: Holland Academic Graphics. Bussuyt, Alain. 1983. “Historical Functional Grammar: An outline of an integrated theory of language change”. Advances in Functional Grammar ed. by Simon C. Dik, chap. 14:301–325. Dordrecht & Cinnaminson, N.J.: Foris. Bynon, Theodora. 1983 [1977]. Historical Linguistics. Cambridge: Cambridge University Press. Chomsky, Noam. 1974 [1972]. “Noam Chomsky”. Discussing Language: Dialogues with Wallace L. Chafe, Noam Chomsky, Algirads J. Greimas, M.A.K. Halliday, Peter Hartmann, George Lakoff, Sydney M. Lamb, André Martinet, James McCawley, Sebastian K. Šaumjan, and Jacques Bouveresse ed. by Herman Parret, 27–54. The Hague & Paris: Mouton. Chomsky, Noam. 2000. New Horizons in the Study of Language and Mind. Cambridge & New York: Cambridge University Press. Chomsky, Noam. 2005. “Three factors in language design”. Linguistic Inquiry 36.1–22. Christiansen, Morten H. 2002. “On the necessity of multiple-cue integration in language acquisition”. Talk given at the Department of General and Applied Linguistics, University of Copenhagen.

334


Clark, Eve V. 1999. “Acquisition in the course of conversation”. Studies in the Linguistic Sciences 29.1–18. Coseriu, Eugenio. 1952. “Sistema, norma y habla”. Revista de la Faculdad de Humanidades y Ciencias 9.113–181. (= Sistema, norma y habla. Montevideo: Universidad de la Republica, Faculdad de Humanidades y Ciencias, Investigaciones y estudios, 1952. Repr. in Teoría del lenguaje y lingüística general, 11–113. Madrid: Gredos, 1962. 3rd, rev. edn., 1973.) Coseriu, Eugenio. 1957. “Sincronía, diacronía e historia: El problema del cambio lingüístico.” Revista de la Faculdad de Humanidades y Ciencias 15. 201–355. (= Sincronía, diacronía e historia: El problema del cambio lingüístico. Montevideo: Universidad de la Republica, Faculdad de Humanidades y Ciencias, Investigaciones y estudios, 1958. 3rd ed. Madrid: Gredos, 1978.) Coseriu, Eugenio. 1977 [1968]. “Syncronía, diacronía y tipología”. El hombre y su lenguaje: Estudios de teoría y metodología lingüística ed. by Eugenio Coseriu, 186–200. Madrid: Gredos. (Orig. in Actas del XI Congreso Internacional de Lingüística y Filología Románicas, Madrid 1965, vol. 1:269–281.) Coseriu, Eugenio. 1974. “Les universaux linguistiques (et les autres)”. Proceedings of the 11th International Congress of Linguists [1972], vol. 1:47– 73. Bologna: Il Mulino. (English translation: “Linguistic (and other) universals”. Linguistics at the Crossroads ed. by Adam Makkai, Valerie Becker Makkai & Luigi Heilmann, 317–346. Padua & Lake Bluff: Liviana Editrice & Jupiter Press, 1977.) Coseriu, Eugenio. 1979. “Humanwissenschaften und Geschichte: Der Gesichtspunkt eines Linguisten”. Årbok 1978.118–130. Oslo: Det Norske Videnskaps-Akademi. Coseriu, Eugenio. 1980a. “ ‘Historische Sprache’ und ‘Dialekt’ ”. Dialekt und Dialektologie: Ergebnisse des Internationalen Symposions “Zur Theorie des Dialekts”. Marburg/Lahn, 5–10 Sept. 1977 ed. by J. Göschel, P. Ivic, K. Kehr, 106–122. Wiesbaden: Steiner. (Repr. in Energeia und Ergon: Sprachliche Variation — Sprachgeschichte — Sprachtypologie ed. by Jörn Albrecht, Jens Lüdtke & Harald Thun, vol 1:45–61. Tübingen: Gunter Narr, 1988.) Coseriu, Eugenio. 1980b. “Vom Primat der Geschichte”. Sprachwissenschaft 5.125–145. Coseriu, Eugenio. 1988 [1982]. “Linguistic change does not exist”. UCLA Conference on Causality and Linguistic Change, Los Angeles, May 1982. Energeia und Ergon: Sprachliche Variation — Sprachgeschichte — Sprachtypologie ed. by Jörn Albrect, Jens Lüdtke & Harald Thun, vol. 1:147–157. Tübingen: Gunter Narr. Coseriu, Eugenio. 1988 [1987]. ”Die Ebenen des sprachlichen Wissens. Der Ort des ”Korrekten” in der Bewertungsskala des Gesprochenen”. Energeia und Ergon: Sprachliche Variation — Sprachgeschichte — Sprachtypologie ed. by Jörn Albrecht, Jens Lüdtke & Harald Thun, vol. 1:327–364. Tübingen: Gunter Narr, 1988.)


335

Coseriu, Eugenio. 1988. “Humboldt und die moderne Sprachwissenschaft”. Energeia und Ergon: Sprachliche Variation — Sprachgeschichte — Sprachtypologie ed. by Jörn Albrecht, Jens Lüdtke & Harald Thun, vol. 1: 3–11. Tübingen: Gunter Narr, 1988. Croft, William. 2000. Explaining Language Change: An evolutionary approach. London & New York: Longman. Croft, William. This volume. “The relevance of an evolutionary model to historical linguistics”. Deutscher, Guy. 2002. “On the misuse of the notion of ‘abduction’ in linguistics”. Journal of Linguistics 38.469–485. Dik, Simon C. 1997. The Theory of Functional Grammar, Part 1 & 2 ed. by Kees Hengeveld. Amsterdam & Philadelphia: John Benjamins. Dressler, Wolfgang U. 1995. “Wealth and poverty of functional analyses, with special reference to functional deficiencies”. Form and Function in Language: Proceedings from the first Rasmus Rask Colloquium, Odense University, November 1992, ed. by Sharon Millar & Jacob Mey, 11–39. (= RASK Supplement, 2.) Odense: Odense University Press. Durie, Mark. 1995. “Towards an understanding of linguistic evolution and the notion ’X has a function Y’ ”. Discourse Grammar and Typology: Papers in honor of John W. M. Verhaar ed. by Werner Abraham, Talmy Givón & Sandra A. Thompson, 275–308. Amsterdam & Philadelphia: John Benjamins. Fortescue, Michael. 2000. “Grammar as Pattern and Process”. Simon Dik Lecture 1999. Universiteit van Amsterdam, Amsterdam: Vossiuspers. Fortescue, Michael. This volume. “The non-linear nature of diachronic change”. Harder, Peter. 2003. “The status of linguistic facts: Rethinking the relation between cognition, social institution and utterance from a functional point of view”. Mind & Language 18:1.52–76. Harris, Roy. 1998. Introduction to Integrational Linguistics. Oxford: Pergamon Hauser, Marc D., Noam Chomsky & W. Tecumseh Fitch. 2002. “The faculty of language”. Science 198.1569–1579. Hawkins, John A. 2004. Efficiency and Complexity in Grammars. Oxford & New York: Oxford University Press. Hengeveld, Kees. 2004a. “The architcture of a functional discourse grammar”. A New Architecture for Functional Grammar ed. by J. Lachlan Mackenzie & María de los Ángeles Gómez-González, 1–21. Berlin & New York: Mouton de Gruyter. Hengeveld, Kees. 2004b. “Epilogue”. A New Architecture for Functional Grammar ed. by J. Lachlan Mackenzie & María de los Ángeles GómezGonzález, 365–378. Berlin & New York: Mouton de Gruyter. Hengeveld, Kees & J. Lachlan Mackenzie. 2006. “Functional Discourse Grammar”. Encyclopedia of Language and Linguistics ed. by Keith Brown, vol.4, 668–676. Oxford: Elsevier. (2nd ed.)

336


Hengeveld, Kees & J. Lachlan Mackenzie. Forthcoming. Functional Discourse Grammar. Oxford & New York: Oxford University Press. Jakobson, Roman. 1956. “Metalanguage as a linguistic problem”. Selected Writings, vol. 7: 113–121. Berlin: Mouton. Jakobson, Roman. 1960. “Concluding statement: linguistics and poetics”. Style in Language ed. by Thomas A. Sebeok, 350–377. New York: Wiley. Jakobson, Roman. 1973. Main Trends in the Science of Language. London: Allen & Unwin. Janda, Richard D. & Brian D. Joseph. 2003. “On language, change, and language change — Or, Of history, linguistics, and historical linguistics”. The Handbook of Historical Linguistics ed. by Brian D. Joseph & Richard D. Janda, 3–180. Oxford & Malden, Mass.: Blackwell. Labov, William. 1980. “The social origins of sound change”. Locating Language in Time and Space ed. by William Labov, 251–266. New York: Academic Press. Lakoff, George & Mark Johnson. 1999. Philosophy in the Flesh: The embodied mind and its challenge to Western thought. New York: Basic Books. Langacker, Ronald W. 2000. “A dynamic, usage-based model”. Usage-based Models of Language ed. by Michael Barlow & Suzanne Kemmer, 1–63. Stanford, Calif.: CSLI Publications. (= Center for the Study of Language and Information.) Leech, Geoffrey. 1983. The Principles of Pragmatics. London & New York: Longman. Nedergaard Thomsen, Ole. Forthcoming. “Funktionel (Diskurs) Grammatik”. Ny-danske Studier og Almen Kommunikationsteori. Nedergaard Thomsen, Ole. In preparation. “Prolegomena to an integrated theory of language and language change: Language as process (’ ε ν έ ργ ε ι α): In commemoration of Eugenio Coseriu”. Roskilde University. Ross, Alf. 1968. Directives and Norms. London: Routledge. Searle, John R. 1965. “What is a Speech Act?”. Philosophy in America ed. M. Black, 221–239. New York: Cornell University Press & London: Allen & Unwin. Searle, John R. 1969. Speech Acts: An essay in the philosophy of language. Cambridge: Cambridge University Press. Searle, John R. 1979. Expression and Meaning: Studies in the theory of speech acts. Cambridge: Cambridge University Press. Searle, John R. 1983. Intentionality: An essay in the philosophy of mind. Cambridge: Cambridge University Press. Searle, John R. 1989. “How performatives work”. Linguistics and Philosophy 12.535–556. Searle, John R. 1990a. “Consciousness, explanatory inversion, and cognitive science”. Behavioral and Brain Sciences 13.585–596. Searle, John R. 1990b. “Collective intentions and actions”. Intentions in Communication, ed. by Philip R. Cohen, Jerry Morgan & Martha E.


337

Pollack, 401–416. Cambridge, Mass. & London: Bradford Books, The MIT Press. Searle, John R. 1992. The Rediscovery of the Mind. Cambridge, Mass.: The MIT Press. Searle, John R. 1995. The Construction of Social Reality. London: Penguin Books. Searle, John R. 1997. The Mystery of Consciousness. New York: New York Review of Books. Searle, John R. 1998. Mind, Language and Society: Philosophy in the real world. New York: Basic Books. Searle, John R. 2001. Rationality in Action. Cambridge, Mass.: A Bradford Book, The MIT Press. Searle, John R. 2002. Consciousness and Language. Cambridge: Cambridge University Press. Searle, John R. 2004. Mind: A brief introduction. Oxford & New York: Oxford University Press. Tomasello, Michael. 2000. “First steps toward a usage-based theory of language acquisition”. Cognitive Linguistics 11:1/2.61–82. Tomasello, Michael. 2003. Constructing a Language: A usage-based theory of language acquisition. Cambridge, Mass.: Harvard University Press. Tomasello, Michael, Malinda Carpenter, Josep Call, Tanya Behne & Henrike Moll. 2005. “Understanding and sharing intentions: The origins of cultural cognition”. Behavioral and Brain Sciences 28.675–691. Weigand, Hans. 1990. Linguistically Motivated Principles of Knowledge Base Systems. Dordrecht & Providence, R.I.: Foris.

INDEX

A abduction (abductive change) 5, 6, 77, 81– 83, 323, 326 ablaut (strong) verbs 272, 276, 281, 284– 285 abstract nouns 235–252 accusative 137, 170, 172, 174, 194, 215, 222–224, 229–234, 263, 265–267, 291–294, 302–303 acquisition 5, 13, 18, 64, 67, 71, 81, 108– 109, 130, 165, 187, 309, 320–329, 332 activation 109 actualization 9, 12, 20, 27, 68, 72, 78, 134, 140–141, 144–145, 150, 153, 157, 164–65, 302–304, 320, 325 actuation 40, 104, 179 adaptation (adaptive process, innovation, rule) 6, 9, 18, 60, 78–80, 83, 85, 94, 114–117, 134, 138, 148–153, 157, 163, 166, 168, 307–308, 313, 317, 323, 327, 331 addressee 12, 68–69, 84, 118, 149–152, 214, 242 adject 136 adjudication 326–328 adoption 7, 9, 65–80, 83, 117, 294, 309, 319, 325–329 Alemannic 170–175 Aleut 22–23 alterity (other-directedness) 312, 324, 329 analogy (analogical change, extension, levelling) 67, 69, 92, 96, 116, 122, 174, 184–185, 207 animacy (Animacy Hierarchy, paradigm) 12, 134, 142, 145, 153–56, 180, 191, 197, 289–305 aspect 11–12, 114, 214–216, 224–229, 232–233, 289–305

assimilation 21–22 automatic speech recognition 43 B balanced harmony 25 base, vs. profile 223, 230, 236–250 base system 72 Bavarian 170, 174, 175, 178 behavior (conduct) 3, 18, 31, 63, 91, 109– 114, 117–119, 143, 155, 168, 323–331 binary oppositions 43–45, 48–49, 52, 55, 273–274, 279, 283–286 (Old) Bulgarian 292, 295, 298 C Cantonese 44, 48, 52, 55 case 70, 74, 170–175, 183, 186, 192–198, 200, 206, 214–215, 222–224, 229–233, 262, 292–295, 303 causality, cause 20–21, 24—25, 35, 80–82, 94, 96, 104–111, 114–117, 123—124, 164, 167–169, 176, 181, 311, 318, 323, 330 chain movement 34–39 change grammatical (semantico-syntactic) 24, 110, 163–182, 270 lexical 68, 116 linear, vs. non-linear 3, 17–29 morphological 68, 197, 270 semantic (content) 11, 27, 68, 115– 116, 235–252, 270 phonological (sound) 24, 33–40, 43, 71, 74–75, 106–107, 116, 125, 184–187, 196, 207 typological 5, 8–9, 19, 24

340 chronolect 152, 156 Chukotko-Kamchatkan 22 clitic (placement) 11, 133–157, 172, 253– 267 code, codification 9, 12, 20, 27, 44, 65, 69–70, 73, 76, 84–85, 102, 120, 123, 137, 140, 157, 171, 175, 177, 215, 267, 277, 324, 326 cognitive (level, domain, etc.) 17, 21, 24, 26, 37, 75, 77, 83–84, 91, 127, 165, 181, 235–236, 246–247, 251 Cognitive Grammar, cognitive linguistics, ~ semantics 10–11, 101, 235–236, 323 cohort (of speakers, cf. chronolect) 67, 71 communication 17, 25, 65, 107, 111, 114– 115, 117–118, 148, 151, 164–165, 181, 242, 249, 292, 308–309, 312, 314, 317–321, 324–326, 329–332 communicative competence 313, 308, 311, 313, 318, 320, 330–331 consonant shift 75 construal (operation) 11, 215, 235–251 contact 18, 20, 23, 30, 72–73, 113, 123, 131, 327 content content reanalysis 11, 269–286, 303 convention 5, 8, 17, 100, 103, 111, 114– 118, 120, 123, 125, 165, 168–169, 180–181, 239, 248, 309, 322, 328–331 convergence, vs. divergence 21, 23, 24, 64, 70, 111, 113 cooperation, coordination (joint action) 8, 117–118, 169, 325, 329 creativity 1, 6, 81–84, 106, 308, 311, 319 cross-reference (pro-drop) 8, 133–157, 260, 262 cryptanalysis 121–122 D Danish 10–12, 33, 80, 211–214, 232, 269– 287 dative 8–9, 134–140, 144, 147, 150, 163– 181, 186, 194, 198, 223–224, 229–230, 265–267 DATR 183–206 deduction (deductive change) 323, 325 dependent marking 137

INDEX

dialect, dialectology 13, 18, 21–23, 26, 64–66, 69, 72–74, 98–99, 163–181, 187, 280, 284, 315, 320, 327–331 diffusion 9, 20, 24, 27, 72–73, 325 distinctive features 4, 43, 51, 57 drag chain, see pull chain drift 8–9, 17, 19, 21, 25, 35, 37, 79, 133– 157, 183, 191, 197, 199–200, 204, 207, 325 Dynamic Syntax 10–11, 253–267 E empathy (Emapthy Hierarchy) 9, 142, 152–153, 157 Eskimo 22–23 event, structure 212–214, 236, 250, 296– 297 evolutionary model (theory), scenario 1, 7, 59–64, 76–85, 91–125, 163–170, 172, 176–181 extension 6, 67–84, 122, 183, 187, 206, 241, 279 F finality 312 Finno-Permic 22 form–function reanalysis 120, 122, 167, 169 freedom (expressive liberty) 1, 79, 312 Functional Grammar 135, 138, 142, 332 functional language (Coseriu) 9, 310–311, 313, 316–331 functional load 4, 43–58 G Generalized Analysis of Selection (GAS) 5–8, 91–124 German 9, 51, 57, 75, 110, 163–181, 183– 208, 241, 246 grammaticalization 3, 6, 8–12, 19–20, 24– 25, 64, 74, 103, 106–107, 114, 116, 121, 133–157, 172, 269–286, 317 parallel 11–12, 301–304

341

INDEX H head marking 8, 133–157 historical language 3, 9, 310, 319–320, 325, 329, 331 hyperanalysis 120–121 hypoanalysis 121

meta-grammar 320–325 metalanguage, metalingustic (metalingual) 322, 326 metanalysis 121 modal character, factor, form, use 269–286 mood 12, 114, 216, 269–286

I implementation 4, 9, 164–165, 179–180 indirect object 8, 11, 133–157, 171–172, 261 induction, inductive 323 information (theory) 43, 46, 48 innovation 5, 7–8, 21, 23, 27, 40, 60, 64, 66–85, 104–24, 141–142, 144, 167– 170, 173, 176–179, 320–322, 325, 328 interaction, interactor, interactivity 5–8, 23, 27, 94–96, 104–109, 114, 117, 146, 151–155, 167–168, 179–181, 312, 325–326, 329 interference 113, 122–123 intraference 122

N Network Morphology 183–208 norm 5, 28, 65, 141, 315, 319, 320, 326– 328 Norwegian 33, 105, 241

L language shift 23 learning 23, 183, 187, 321, 323 lingueme 7, 105–110, 114, 125, 167 M Mandarin 23, 25, 51, 57 markedness Markedness Agreement 141–146, 157 Markedness Shift 134–35, 140, 143, 146–149, 157 Martinet Hypothesis 44 mediopassive (middle) 269, 274, 278, 280–281, 286 memetics, meme 2, 60–61, 92–94, 318, 329 Mental Grammar 10, 211–233 merger, merging 4, 38, 44–45, 48, 52, 70, 173, 281, 283, 285, 293–294, 302–303 meta-communication, meta-dialogue 13, 76, 308, 311, 320–332

O object-level 308, 325–329 Ob-Ugric 22–23 P Paleosiberian 21 passive 11–12, 224, 269–86 perlocutionary (act, effect) 13, 329 Permic 22 phonemic contrasts (oppositions) 4, 43, 49–52 Polish 224, 297–298 Prague School 4, 43 practice (of speaking) 6, 13, 59, 65–66, 76, 83, 85, 329–331 Prepositional Dative Marking 9, 137, 139, 163–181 promulgation 325–326, 330 propagation 5, 8–9, 104, 106, 109, 111– 112, 115–117, 120, 124, 168–169, 178–180 pull (drag), vs. push chain 3, 20, 25, 33–41 R Radical Construction Grammar 8, 122 reanalysis 6, 9, 11–12, 64, 67–84, 120– 122, 140–141, 143, 150, 157, 167, 169, 172–178, 269–286, 293–294, 301–304, 320, 325 reconstruction 1, 20–23, 113, 176

342 reduplication, double representation, clitic doubling 133–57 replication, replicator 5–8, 61, 76–78, 81– 84, 93–96, 104–119, 123, 167–168, 179, 320 rhetorical devaluation 143, 147, 152 Russian 23, 197, 211–233, 289–304 S Samoyedic 22–23 Scandinavian, Old 23, 33–39, 183, 186, 269, 274, 281, 286 selection 1, 4–9, 18, 76–83, 92–96, 103– 119, 163–170, 177–181, 323, 325, 328 Slavic 289–303 ‘s-mood’ (passive mood) 269–286 sociolect 69, 153–156 Spanish 8–9, 11, 133–157, 253–267, 319 structural gap 37 subjunctive 121, 187, 279–286 syntacticization 213–214, 233 Swedish 11, 33, 39, 235–251, 269–286 T teleology, teleological explanation, change 7–8, 17–18, 25, 37, 78, 119–122, 165, 217, 323–324 tense 114, 216, 224, 271–286, 295, 297– 303 Theory of Utterance Selection (TUS) 4, 7, 91–125

INDEX

tone 24, 43, 48, 51–52 tradition (of speaking) 6, 13, 59, 64–85, 91, 99, 150, 310, 314–317, 321–322, 331 Turkic, Turkish 23, 191–194 U Ugric 22–23 Uralic 21–24 usage, usage rules, usage-based 2, 5, 20, 28, 60, 65–83, 104, 108–12, 118, 122, 134, 141, 157, 164, 166, 293–94, 302– 4, 310, 312–13, 321–26, 338–39 V variability, vs. invariance 49, 98–101, 109–110, 120, 125, 168, 316–317 variation 5, 7–13, 18, 21, 27–28, 45, 60, 66–67, 71, 78–79, 94–110, 118–119, 123–125, 139, 150, 153, 185–188, 191, 194, 199, 235, 237, 250–251 vowel shift 33–41 W wave (model/theory) 21, 27 Z Zealandic 279–281

CONTRIBUTORS Henning Andersen Department of Slavic Languages and Literatures University of California, Los Angeles P.O.Box 951502 Los Angeles, CA 900095-1502, USA e-mail: [email protected]

Lena Ekberg Institutionen för nordiska språk Lunds Universitet Helgonabacken 12. Box 201 221 00 Lund, Sweden e-mail: [email protected]

Silvia Becerra Bascuñán Christianshavns Gymnasium Prinsessegade 35 DK-1422 København K, Denmark e-mail: [email protected]

Michael Fortescue Institut for Nordiske Studier og Sprogvidenskab Københavns Universitet Njalsgade 120 DK-2300 København S, Denmark e-mail: [email protected]

Miriam Bouzouita & Ruth Kempson Department of Philosophy King’s College London, Strand London WC2R 2LS, United Kingdom e-mail: [email protected] & [email protected]

Lars Heltoft Institut for Sprog og Kultur Roskilde Universitets Center Universitetsvej 1. Postboks 260 DK-4000 Roskilde, Denmark e-mail: [email protected]

William Croft Department of Linguistics University of New Mexico Albuquerque NM 87131, USA e-mail: [email protected]

Brit Mæhlum Institutt for nordistikk og litteraturvitenskap Universitetet i Trondheim N-7491 Trondheim, Norway e-mail: [email protected]

Per Durst-Andersen Institut for Fransk, Italiensk, Russisk, Spansk og Tysk Copenhagen Business School Dalgas Have 15 DK-2000 Frederiksberg, Denmark e-mail: [email protected]

Ole Nedergaard Thomsen Institut for Sprog og Kultur Roskilde Universitets Center Universitetsvej 1. Postboks 260 DK-4000 Roskilde, Denmark e-mail: [email protected]

344

CONTRIBUTORS

Jens Nørgård-Sørensen Institut for Tværkulturelle og Regionale Studier Københavns Universitet Leifsgade 33 DK-2300 København S, Denmark e-mail: [email protected]

Dinoj Surendran & Partha Niyogi Department of Computer Science University of Chicago 1100 E. 58th Street, Ryerson Hall Hyde Park, Chicago, IL 60637, USA e-mail: [email protected] & [email protected]

Guido Seiler Deutsches Seminar Universität Zürich Schönberggasse 9 CH-8001 Zürich, Switzerland e-mail: [email protected]

Miguel Vázquez-Larruscaín Modern Languages Department College of Arts & Sciences Northeastern University 360 Huntington Avenue Boston, MA 02115, USA [email protected]

CURRENT ISSUES IN LINGUISTIC THEORY

E. F. K. Koerner, Editor

Zentrum für Allgemeine Sprachwissenschaft, Typologie und Universalienforschung, Berlin [email protected] Current Issues in Linguistic Theory (CILT) is a theory-oriented series which welcomes contributions from scholars who have significant proposals to make towards the advancement of our understanding of language, its structure, functioning and development. CILT has been established in order to provide a forum for the presentation and discussion of linguistic opinions of scholars who do not necessarily accept the prevailing mode of thought in linguistic science. It offers an outlet for meaningful contributions to the current linguistic debate, and furnishes the diversity of opinion which a healthy discipline must have. A complete list of titles in this series can be found on the publishers’ website, www.benjamins.com 281 Vermeerbergen, Myriam, Lorraine Leeson and Onno Crasborn (eds.): Simultaneity in Signed Languages. Form and function. Expected January 2007 280 Hewson, John and Vít Bubeník: From Case to Adposition. The development of configurational syntax in Indo-European languages. Expected December 2006 279 Nedergaard Thomsen, Ole (ed.): Competing Models of Linguistic Change. Evolution and beyond. 2006. vi, 344 pp. 278 Doetjes, Jenny and Paz González (eds.): Romance Languages and Linguistic Theory 2004. Selected papers from ‘Going Romance’, Leiden, 9–11 December 2004. 2006. viii, 320 pp. 277 Helasvuo, Marja-Liisa and Lyle Campbell (eds.): Grammar from the Human Perspective. Case, space and person in Finnish. 2006. ix, 280 pp. 276 Montreuil, Jean-Pierre Y. (ed.): New Perspectives on Romance Linguistics. Vol. II: Phonetics, Phonology and Dialectology. Selected papers from the 35th Linguistic Symposium on Romance Languages (LSRL), Austin, Texas, February 2005. 2006. x, 213 pp. 275 Nishida, Chiyo and Jean-Pierre Y. Montreuil (eds.): New Perspectives on Romance Linguistics. Vol. I: Morphology, Syntax, Semantics, and Pragmatics. Selected papers from the 35th Linguistic Symposium on Romance Languages (LSRL), Austin, Texas, February 2005. 2006. xiv, 288 pp. 274 Gess, Randall S. and Deborah Arteaga (eds.): Historical Romance Linguistics. Retrospective and perspectives. 2006. viii, 393 pp. 273 Filppula, Markku, Juhani Klemola, Marjatta Palander and Esa Penttilä (eds.): Dialects Across Borders. Selected papers from the 11th International Conference on Methods in Dialectology (Methods XI), Joensuu, August 2002. 2005. xii, 291 pp. 272 Gess, Randall S. and Edward J. Rubin (eds.): Theoretical and Experimental Approaches to Romance Linguistics. Selected papers from the 34th Linguistic Symposium on Romance Languages (LSRL), Salt Lake City, March 2004. 2005. viii, 367 pp. 271 Branner, David Prager (ed.): The Chinese Rime Tables. Linguistic philosophy and historicalcomparative phonology. 2006. viii, 358 pp. 270 Geerts, Twan, Ivo van Ginneken and Haike Jacobs (eds.): Romance Languages and Linguistic Theory 2003. Selected papers from ‘Going Romance’ 2003, Nijmegen, 20–22 November. 2005. viii, 369 pp. 269 Hargus, Sharon and Keren Rice (eds.): Athabaskan Prosody. 2005. xii, 432 pp. 268 Cravens, Thomas D. (ed.): Variation and Reconstruction. 2006. viii, 223 pp. 267 Alhawary, Mohammad T. and Elabbas Benmamoun (eds.): Perspectives on Arabic Linguistics XVII–XVIII. Papers from the seventeenth and eighteenth annual symposia on Arabic linguistics. Volume XVII–XVIII: Alexandria, 2003 and Norman, Oklahoma 2004. 2005. xvi, 315 pp. 266 Boudelaa, Sami (ed.): Perspectives on Arabic Linguistics XVI. Papers from the sixteenth annual symposium on Arabic linguistics, Cambridge, March 2002. 2006. xii, 181 pp. 265 Cornips, Leonie and Karen P. Corrigan (eds.): Syntax and Variation. Reconciling the Biological and the Social. 2005. vi, 312 pp. 264 Dressler, Wolfgang U., Dieter Kastovsky, Oskar E. Pfeiffer and Franz Rainer (eds.): Morphology and its demarcations. Selected papers from the 11th Morphology meeting, Vienna, February 2004. With the assistance of Francesco Gardani and Markus A. Pöchtrager. 2005. xiv, 320 pp. 263 Branco, António, Tony McEnery and Ruslan Mitkov (eds.): Anaphora Processing. Linguistic, cognitive and computational modelling. 2005. x, 449 pp. 262 Vajda, Edward J. (ed.): Languages and Prehistory of Central Siberia. 2004. x, 275 pp. 261 Kay, Christian J. and Jeremy J. Smith (eds.): Categorization in the History of English. 2004. viii, 268 pp.

260 Nicolov, Nicolas, Kalina Bontcheva, Galia Angelova and Ruslan Mitkov (eds.): Recent Advances in Natural Language Processing III. Selected papers from RANLP 2003. 2004. xii, 402 pp. 259 Carr, Philip, Jacques Durand and Colin J. Ewen (eds.): Headhood, Elements, Specification and Contrastivity. Phonological papers in honour of John Anderson. 2005. xxviii, 405 pp. 258 Auger, Julie, J. Clancy Clements and Barbara Vance (eds.): Contemporary Approaches to Romance Linguistics. Selected Papers from the 33rd Linguistic Symposium on Romance Languages (LSRL), Bloomington, Indiana, April 2003. With the assistance of Rachel T. Anderson. 2004. viii, 404 pp. 257 Fortescue, Michael, Eva Skafte Jensen, Jens Erik Mogensen and Lene Schøsler (eds.): Historical Linguistics 2003. Selected papers from the 16th International Conference on Historical Linguistics, Copenhagen, 11–15 August 2003. 2005. x, 312 pp. 256 Bok-Bennema, Reineke, Bart Hollebrandse, Brigitte Kampers-Manhe and Petra Sleeman (eds.): Romance Languages and Linguistic Theory 2002. Selected papers from ‘Going Romance’, Groningen, 28–30 November 2002. 2004. viii, 273 pp. 255 Meulen, Alice ter and Werner Abraham (eds.): The Composition of Meaning. From lexeme to discourse. 2004. vi, 232 pp. 254 Baldi, Philip and Pietro U. Dini (eds.): Studies in Baltic and Indo-European Linguistics. In honor of William R. Schmalstieg. 2004. xlvi, 302 pp. 253 Caffarel, Alice, J.R. Martin and Christian M.I.M. Matthiessen (eds.): Language Typology. A functional perspective. 2004. xiv, 702 pp. 252 Kay, Christian J., Carole Hough and Irené Wotherspoon (eds.): New Perspectives on English Historical Linguistics. Selected papers from 12 ICEHL, Glasgow, 21–26 August 2002. Volume II: Lexis and Transmission. 2004. xii, 273 pp. 251 Kay, Christian J., Simon Horobin and Jeremy J. Smith (eds.): New Perspectives on English Historical Linguistics. Selected papers from 12 ICEHL, Glasgow, 21–26 August 2002. Volume I: Syntax and Morphology. 2004. x, 264 pp. 250 Jensen, John T.: Principles of Generative Phonology. An introduction. 2004. xii, 324 pp. 249 Bowern, Claire and Harold Koch (eds.): Australian Languages. Classification and the comparative method. 2004. xii, 377 pp. (incl. CD-Rom). 248 Weigand, Edda (ed.): Emotion in Dialogic Interaction. Advances in the complex. 2004. xii, 284 pp. 247 Parkinson, Dilworth B. and Samira Farwaneh (eds.): Perspectives on Arabic Linguistics XV. Papers from the Fifteenth Annual Symposium on Arabic Linguistics, Salt Lake City 2001. 2003. x, 214 pp. 246 Holisky, Dee Ann and Kevin Tuite (eds.): Current Trends in Caucasian, East European and Inner Asian Linguistics. Papers in honor of Howard I. Aronson. 2003. xxviii, 426 pp. 245 Quer, Josep, Jan Schroten, Mauro Scorretti, Petra Sleeman and Els Verheugd (eds.): Romance Languages and Linguistic Theory 2001. Selected papers from 'Going Romance', Amsterdam, 6–8 December 2001. 2003. viii, 355 pp. 244 Pérez-Leroux, Ana Teresa and Yves Roberge (eds.): Romance Linguistics. Theory and Acquisition. Selected papers from the 32nd Linguistic Symposium on Romance Languages (LSRL), Toronto, April 2002. 2003. viii, 388 pp. 243 Cuyckens, Hubert, Thomas Berg, René Dirven and Klaus-Uwe Panther (eds.): Motivation in Language. Studies in honor of Günter Radden. 2003. xxvi, 403 pp. 242 Seuren, Pieter A.M. and Gerard Kempen (eds.): Verb Constructions in German and Dutch. 2003. vi, 316 pp. 241 Lecarme, Jacqueline (ed.): Research in Afroasiatic Grammar II. Selected papers from the Fifth Conference on Afroasiatic Languages, Paris, 2000. 2003. viii, 550 pp. 240 Janse, Mark and Sijmen Tol (eds.): Language Death and Language Maintenance. Theoretical, practical and descriptive approaches. With the assistance of Vincent Hendriks. 2003. xviii, 244 pp. 239 Andersen, Henning (ed.): Language Contacts in Prehistory. Studies in Stratigraphy. Papers from the Workshop on Linguistic Stratigraphy and Prehistory at the Fifteenth International Conference on Historical Linguistics, Melbourne, 17 August 2001. 2003. viii, 292 pp. 238 Núñez-Cedeño, Rafael, Luis López and Richard Cameron (eds.): A Romance Perspective on Language Knowledge and Use. Selected papers from the 31st Linguistic Symposium on Romance Languages (LSRL), Chicago, 19–22 April 2001. 2003. xvi, 386 pp. 237 Blake, Barry J. and Kate Burridge (eds.): Historical Linguistics 2001. Selected papers from the 15th International Conference on Historical Linguistics, Melbourne, 13–17 August 2001. Editorial Assistant: Jo Taylor. 2003. x, 444 pp. 236 Simon-Vandenbergen, Anne-Marie, Miriam Taverniers and Louise J. Ravelli (eds.): Grammatical Metaphor. Views from systemic functional linguistics. 2003. vi, 453 pp.

Competing Models of Linguistic Change: Evolution and Beyond (Current Issues in Linguistic Theory)

Postvelar Harmony (Current Issues in Linguistic Theory)

Loan Phonology (Current Issues in Linguistic Theory)

Anaphora Processing (Current Issues in Linguistic Theory)

Loan Phonology (Current Issues in Linguistic Theory)

Germanic linguistics: syntactic and diachronic (Current Issues in Linguistic Theory)

The Transformational-Generative Paradigm and Modern Linguistic Theory (Current Issues in Linguistic Theory)

Ancient Scripts and Phonological Knowledge (Current Issues in Linguistic Theory)

Evolution and Revolution in Linguistic Theory

Evolution and Revolution in Linguistic Theory

Linguistic Variables: Towards a Unified Theory of Linguistic Variation (Current Issues in Linguistic Theory) issue 0304-0763

Linguistic Categorization: Prototypes in Linguistic Theory

Stability, Variation and Change of Word-order Patterns over Time (Current Issues in Linguistic Theory)

Current Issues in Bilingualism: Cognitive and Socio-linguistic Perspectives

Current Issues in Bilingualism: Cognitive and Socio-linguistic Perspectives

Clinical Linguistics: Theory and Applications in Speech Pathology and Therapy (Current Issues in Linguistic Theory)

Linguistic Universals and Language Change

Linguistic Universals and Language Change

Phonological Theory and the Dialects of Italy (Current Issues in Linguistic Theory)

The Paradox of Grammatical Change: Perspectives from Romance (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory, CILT 293)

Typological Models in Reconstruction (Amsterdam Studies in the Theory & History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Proto-Japanese; Issues and Prospects (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Ergativity: Emerging Issues (Studies in Natural Language and Linguistic Theory)

Foundational Issues in Linguistic Theory: Essays in Honor of Jean-Roger Vergnaud (Current Studies in Linguistics)

Foundational Issues in Linguistic Theory: Essays in Honor of Jean-Roger Vergnaud (Current Studies in Linguistics)

The Typology and Dialectology of Romani (Current Issues in Linguistic Theory)

Linguistic evolution through language acquisition: Formal and computational models

Historical Syntax and Linguistic Theory

Historical Syntax and Linguistic Theory

A Linguistic Theory of Translation

Linguistics and Philosophy: Essays in Honor of Rulon S. Wells (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Competing Models of Linguistic Change: Evolution and Beyond (Current Issues in Linguistic Theory)

Postvelar Harmony (Current Issues in Linguistic Theory)

Loan Phonology (Current Issues in Linguistic Theory)

Anaphora Processing (Current Issues in Linguistic Theory)

Loan Phonology (Current Issues in Linguistic Theory)

Germanic linguistics: syntactic and diachronic (Current Issues in Linguistic Theory)

The Transformational-Generative Paradigm and Modern Linguistic Theory (Current Issues in Linguistic Theory)

Ancient Scripts and Phonological Knowledge (Current Issues in Linguistic Theory)

Evolution and Revolution in Linguistic Theory

Evolution and Revolution in Linguistic Theory

Linguistic Variables: Towards a Unified Theory of Linguistic Variation (Current Issues in Linguistic Theory) issue 0304-0763

Linguistic Categorization: Prototypes in Linguistic Theory

Stability, Variation and Change of Word-order Patterns over Time (Current Issues in Linguistic Theory)

Current Issues in Bilingualism: Cognitive and Socio-linguistic Perspectives

Current Issues in Bilingualism: Cognitive and Socio-linguistic Perspectives

Clinical Linguistics: Theory and Applications in Speech Pathology and Therapy (Current Issues in Linguistic Theory)

Linguistic Universals and Language Change

Linguistic Universals and Language Change

Phonological Theory and the Dialects of Italy (Current Issues in Linguistic Theory)

The Paradox of Grammatical Change: Perspectives from Romance (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory, CILT 293)

Typological Models in Reconstruction (Amsterdam Studies in the Theory & History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Proto-Japanese; Issues and Prospects (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Ergativity: Emerging Issues (Studies in Natural Language and Linguistic Theory)

Foundational Issues in Linguistic Theory: Essays in Honor of Jean-Roger Vergnaud (Current Studies in Linguistics)

Foundational Issues in Linguistic Theory: Essays in Honor of Jean-Roger Vergnaud (Current Studies in Linguistics)

The Typology and Dialectology of Romani (Current Issues in Linguistic Theory)

Linguistic evolution through language acquisition: Formal and computational models

Historical Syntax and Linguistic Theory

Historical Syntax and Linguistic Theory

A Linguistic Theory of Translation

Linguistics and Philosophy: Essays in Honor of Rulon S. Wells (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Recommend Documents