Pragmatic Markers in Contrast (SiP 2), Volume 2 (Studies in Pragmatics) (Studies in Pragmatics)

PRAGMATIC MARKERS IN CONTRAST STUDIES IN PRAGMATICS General Editor: Bruce Fraser Associate Editors: Kerstin Fischer, ...

Author: Karin Aijmer | Anne-Marie Simon-Vandenbergen

184 downloads 1091 Views 15MB 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

PRAGMATIC MARKERS IN CONTRAST

STUDIES IN PRAGMATICS General Editor: Bruce Fraser Associate Editors: Kerstin Fischer, Maj-Britt Mosegaard Hansen

The Studies in Pragmatics series is dedicated to publishing innovative, authoritative monographs and edited collections from all micro-, macro- and metapragmatic linguistic perspectives. Rooted in the interdisciplinary spirit of the Journal of Pragmatics, it welcomes not only book proposals from linguistics proper but also pragmaticallyoriented proposals from neighboring disciplines such as interactional sociology, language philosophy, communication science, social psychology, cognitive science, and information science. The goal of the series is to provide a widely read and respected international forum for high quality theoretical, analytical, and applied pragmatic studies of all types. By publishing leading edge work on natural language practice, it seeks to extend our growing knowledge of the forms, functions, and foundations of human interaction.

Other titles in this series: FISCHER

Approaches to Discourse Particles

Forthcoming: HABERLAND, CAFFI, HIRAGA & JANNEY

Future Prospects for Pragmatics

FETZER & FISCHER

Lexical Markers of Common Grounds

Proposals for the series are welcome, please contact the General Editor, Bruce Fraser: [email protected]

PRAGMATIC MARKERS IN CONTRAST

EDITED BY

KARIN AIJMER Gothenburg University, Sweden

ANNE-MARIE SIMON-VANDENBERGEN Ghent University, Belgium

Amsterdam ● Boston ● Heidelberg ● London ● New York ● Oxford Paris ● San Diego ● San Francisco ● Singapore ● Sydney ● Tokyo

Elsevier The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands

First edition 2006 Copyright © 2006 Elsevier Ltd. All rights reserved No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: [email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at http://elsevier.com/locate/permissions, and selecting Obtaining permission to use Elsevier material Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress ISBN-13: 978-0-08-044676-9 ISBN-10: 0-08-044676-0 ISSN: 1750-368X For information on all Elsevier publications visit our website at books.elsevier.com Printed and bound in The Netherlands 06 07 08 09 10

10 9 8 7 6 5 4 3 2 1

Studies in Pragmatics (SiP) General Editor Bruce Fraser Boston University, USA Associate Editors Kerstin Fischer University of Bremen, Germany Maj-Britt Mosegaard Hansen University of Copenhagen, Denmark Consulting Editor Jacob L. Mey University of Southern Denmark, Denmark Editorial Board Kent Bach, San Francisco State University, USA Diane Blakemore, University of Salford, UK Shoshana Blum-Kulka, Hebrew University, Israel Claudia Caffi, University of Genoa, Italy Alessandro Duranti, UCLA, USA Marjorie Goodwin, UCLA, USA Hartmut Haberland, University of Roskilde, Denmark William F. Hanks, University of California, USA Sachiko Ide, Tokyo Women’s University, Japan Mikhail Ilyin, Moscow State Institute of International Relations of the Foreign Affairs Ministry of the Russian Federation (MGIMO), Russia Richard W. Janney, University of Munich, Germany Kasia Jaszczolt, University of Cambridge, UK Elizabeth Keating, University of Texas, USA Sotaro Kita, University of Bristol, UK Ron Kuzar, University of Haifa, Israel Alec McHoul, Murdoch University, Australia Brigitte Nerlich, Nottingham University, UK Etsuko Oishi, Fuji Women’s University, Japan Srikant Sarangi, Cardiff University, UK Marina Sbisà, University of Trieste, Italy Maxim Stamenov, Bulgarian Academy of Sciences, Bulgaria Deborah Tannen, Georgetown University, USA

This Page is Intentionally Left Blank

TABLE OF CONTENTS List of Contributors ................................................................................................................. ix Introduction ............................................................................................................................... 1 Karin Aijmer, Gothenburg University Anne-Marie Simon-Vandenbergen, Ghent University The Function of Adverbial Connectors in Second Initial Position in English and Swedish.... 11 Bengt Altenberg, University of Lund The English Pragmatic Marker surely and its Functional Counterparts in Spanish................. 39 Angela Downing, Universidad Complutense de Madrid Polysemy Patterns in Contrast: The Case of Dutch toch and German doch ........................... 59 Ad Foolen, Radboud University Nijmegen On the Universality of Discourse Markers .............................................................................. 73 Bruce Fraser, Boston University “Not now” – On Non-Correspondence between the Cognate Adverbs now and nå ............... 93 Hilde Hasselgård, University of Oslo How Well can well be Translated? On the English Discourse Particle well and its Correspondences in Norwegian and German ........................................................................ 115 Stig Johansson, University of Oslo Contrastive Analysis of Adversative Relational Markers, Using Comparable Corpora ....... 139 Diana M. Lewis, University of Lyon 2 The Spanish Discourse Markers o sea and pues and their English Correspondences .......... 155 Anna-Brita Stenström, Bergen University Aspectual Particles in some European Languages ................................................................ 173 Willy Vandeweghe, Hogeschool Gent Dialogistic Positions and Anticipated Audiences – a Framework for Stylistic Comparisons 189 Peter R.R White, University of Adelaide Motoki Sano, University of Wollongong Vraiment and really in Contrast. When Truth and Reality Meet .......................................... 215 Dominique Willems, Ghent University Annemie Demol, Ghent University Constructions in Cross-Language Research. Verbs as Pragmatic Particles in Solv ............. 237 Jan-Ola Östman, University of Helsinki


LIST OF CONTRIBUTORS Karin Aijmer, Gothenburg University Anne-Marie Simon-Vandenbergen, Ghent University Bengt Altenberg, Lund University Annemie Demol, Ghent University Angela Downing, Universidad Complutense de Madrid Ad Foolen, Radboud University Nijmegen Bruce Fraser, Boston University Hilde Hasselgård, University of Oslo Stig Johansson, University of Oslo Diana Lewis, University of Lyon 2 Jan-Ola Östman, University of Helsinki Motoki Sano, University of Wollongong Anna-Brita Stenström, University of Bergen Willy Vandeweghe, Hogeschool Gent Peter White, University of Adelaide Dominique Willems, Ghent University


Pragmatic Markers in Contrast edited by K. Aijmer and A.-M. Simon-Vandenbergen © 2006 Elsevier Ltd. All rights reserved.

INTRODUCTION Karin Aijmer, Göteborg University Anne-Marie Simon-Vandenbergen, Ghent University

1.

CHALLENGES IN THE FIELD

As a result of internationalisation and new developments in linguistics there has been an increasing interest in cross-linguistic studies, shedding light both on similarities and differences between languages. Moreover, during the last decades contrastive studies have moved into new areas of research such as pragmatics and discourse analysis. This new direction in linguistics has also led to more interest in spoken communication and the study of pragmatic markers. From a modest start, research on pragmatic markers has exploded in the last 20 years both from an empirical point-of-view and theoretically. There are various important books in this area including Schiffrin (1987), Jucker & Ziv (1998), Lenk (1998), Hansen (1998), Andersen & Fretheim (2000), Fischer (2000), Aijmer (2002). In addition, empirical research has yielded in-depth analyses of particular items in different languages, described in a wide range of articles. Nevertheless, there are still major theoretical and descriptive challenges in this field of research. On the theoretical side there is a clear need for a model of communication which is rich enough to account for the complexity of the functioning of pragmatic markers in different languages. Pragmatic markers have been analysed in different frameworks (discourse analysis, relevance theory, argumentation theory), but a general model is currently lacking, and research findings regarding various aspects of the functions of pragmatic markers have as yet not been accommodated within an overarching theoretical framework. In order to establish such a common framework we need, for example, to agree on what we mean by pragmatic markers and what dimensions of the context are referred to by the pragmatic markers. A second challenge for research in this area is to deepen our insight into the multifunctionality of the pragmatic markers in order to arrive at a functional typology and a satisfactory account of meaning relations on the semantic and pragmatic levels. Much work remains to be done to get a better idea of the functions and distributions of pragmatic markers both monolingually and in a contrastive perspective. Empirical studies of pragmatic markers from many different languages and language pairs are needed to show to what extent similar discourse functions are found in the languages of the world. The aim of this volume is to bring together empirical studies of pragmatic markers across languages and to discuss the larger theoretical and methodological issues which such studies give rise to. The volume is innovative in its coverage of many different pragmatic markers which have not been discussed before in a contrastive perspective. The individual articles explore the possibility of comparing pragmatic markers across languages in order to

2 Pragmatic Markers in Contrast establish correspondences and to find out more about similarities and differences between languages. Different approaches and different languages are represented. Translation serves a crucial role and many papers make use of translation corpora for exploring meanings and correspondences between pragmatic markers across languages. The empirical analysis includes pragmatic markers in different languages and can also shed new light on the analysis of the markers in the source and the target language, for example showing whether they are polysemous and if so how many meanings they have. We believe that the comparison of the distribution and function of markers in several languages may result in a more precise circumscription of the range of meanings that pragmatic markers can have and may show where there are specific meanings which are lexicalized (or grammaticalized) in the form of pragmatic markers. The articles are synchronic but the functional variation also opens up a grammaticalization perspective. We shall first discuss the lack of a generally accepted terminology and justify our own use of the term ’pragmatic marker’ (Section 2). Section 3 discusses the problem of multifunctionality and the lack of functional taxonomies and ways of dealing with it. Different methodologies for studying pragmatic markers cross-linguistically are discussed in Section 4. An overview of the different contributions is given in Section 5.

2.

TERMINOLOGY

The question of terminology is less crucial but needs to be tackled. There is a lack of a generally accepted terminology and useful taxonomies in this area. A variety of different terms are used to refer to the items discussed here which are sometimes difficult to distinguish from each other. The terms ‘pragmatic marker’, ‘discourse marker’, ‘pragmatic particle’, ‘discourse particle’ are used by some as synonyms, by others as partly overlapping categories. Some researchers prefer the term ‘connective’ (cf. Crystal & Davy, 1975; Blakemore,1987; Bazzanella, 1990 ‘phatic connective’). The diversity of terms may reflect different linguistic approaches of the markers. The connection between the term ‘discourse marker’ and textual function is very clear in Schiffrin’s analysis. Schiffrin (1987) establishes a discourse model consisting of a number of extralinguistic structures (ideational, action, exchange structure) and regards (discourse) markers as indicators of the location of utterances within the emerging structures (Schiffrin, 1987: 24). On the other hand, Jucker & Ziv (1998: 2) chose the term ‘discourse marker’ as a convenient cover term ‘because it seems to be the one with the widest currency and with the least restricted range of application; one that enables us to include a broad variety of elements under a single conceptual umbrella’. We have opted here for ‘pragmatic marker’ rather than ‘discourse marker’ as a broad term. Discourse marker is the term which we use when we want to describe how a particular marker signals coherence relations. Pragmatic markers as we see them are not only associated with discourse and textual functions but are also signals in the communication situation guiding the addressee’s interpretation. The term as we are using it can also be defined negatively: if a word or a construction in an utterance does not contribute to the propositional, truth-functional content, then we consider it a pragmatic marker. In this we follow Fraser (1996: 168) who distinguishes pragmatic markers from propositional content items and

Introduction 3 defines them as ‘the linguistically encoded clues which signal the speaker’s potential communicative intentions’. For us ‘pragmatic marker’ is also a broad enough term to cover items which can then be further subclassified according to more detailed functional and formal characteristics. (For a list of formal, functional, phonological, sociological and stylistic features see e.g. Jucker & Ziv, 1998: 3). For example, discourse markers, adverbial connectors and routines are subclasses of pragmatic markers.

3.

WHY A CROSS-LINGUISTIC STUDY OF PRAGMATIC MARKERS?

Cross-linguistic studies of pragmatic markers are needed to find out more both about what is universal and what is language-specific. In this work we want to shed light on this question by including more empirical analyses of particular markers in a cross-language perspective. By adding more languages and more pragmatic markers we can also establish if there are specific meanings which are sources of pragmatic markers and learn more both about universal processes involved in grammaticalization and about how these processes are affected by sociolinguistic and cultural factors. We expect the volume to contribute to elucidating both what elements in different languages have in common and how they differ. The investigation of cross-linguistic correspondences can help us to find answers to different questions. First, such work can provide a detailed picture of the correspondences of words which do not have a direct correspondence in another language. Much recent comparative work has been concerned with typological questions and finding discourse universals. However, when we compare pragmatic markers across languages complete equivalence between words in different languages is unusual. Pragmatic markers operate at several linguistic levels simultaneously and a function which is expressed by a lexical item in one language can be expressed grammatically or by another word class in another language. For example, on the formal level a pragmatic marker can be translated into an auxiliary or a verb and it can be placed in a different position in the target language. Thus a sentence with German sondern corresponds to a cleft or a pseudo-cleft (John didn’t go to London, it was Peter) as well as to but (Peter is not stupid but lazy) in the English translations (Fernandez 1994: 77). Another example is provided by the German ‘Abtönungspartikeln’ (mal, eben, auch, doch), which have a wide range of correspondences in various languages, including lexical items, structural variation, prosodic shifts and other rhetorical procedures (Fernandez 1994: 78). Well has been the subject of several cross-linguistic studies. Bazzanella & Morra (2000) compared well and its Italian cognate bene in the translation of an English novel and found only a small overlapping area of meaning. In a study investigating well and its translations into Swedish and Dutch in bilingual corpora, Aijmer & Simon-Vandenbergen (2003) found that the marker was translated in many different ways reflecting the fact that neither Swedish nor Dutch have an equivalent marker. Among other forms, Swedish also used response particles (ja ‘yes’), conjunctions, modal particles and routine phrases. In Dutch the most

4 Pragmatic Markers in Contrast frequent correspondence was the pragmatic marker nou (‘now’). There were also interesting quantitative differences between the languages shown by translations. Another striking result was the high frequency of omission, especially in Swedish, which could be explained as reflecting the lack of conceptual content of the marker. Secondly, the cross-linguistic method can be used to investigate a larger semantic field. Simon-Vandenbergen & Aijmer (2002/2003) have shown how we can go from form to function and from function to form and construct a semantic field of of course. For example, naturally and obviously, both found as back-translations of items which came up as translation equivalents of of course, may be shown to be close synonyms in the field because they share a correspondence in the translation while, for instance, naturally and after all (which also occurs as an English translation of Dutch and Swedish equivalents of of course) share little meaning since they are not translated back with the same item in Swedish or Dutch. Thirdly, the cross-linguistic method can be used to investigate etymological or semantic cognates. This is an open field of research. Diverging polysemy (Altenberg & Granger, 2002:23) is illustrated by the case where markers develop in different directions. On the other hand, there are also examples where we find that markers in two languages have roughly the same meaning extensions (‘overlapping polysemy’ Altenberg & Granger, ibid.). For example Suzanne Fleischman and Marina Yaguello have suggested (1999) that similar meanings or functions may be lexicalized in several languages, suggesting that we can discover discourse universals on the basis of cross-linguistic research. As a contribution to ‘cross-linguistic pragmatics’ they undertook a comparison of like in English and genre in French. It was shown that the markers had the same functions, although the lexical origin of the markers in the languages compared was not the same. This study is particularly interesting since it gives support to the hypothesis that grammatical and semantic changes follow certain paths which can explained in terms of a movement from propositional meaning to textual or interpersonal meaning or in terms of a development towards more subjectification and intersubjectivity (‘the unidirectionality’ hypothesis; Traugott & Dasher 2002). Another good example that elements which have the same source or a similar meaning develop in the same direction in different languages is a study by Fraser & MalamudMakowski (1996). A comparison of a group of discourse markers in English and Spanish with the same function showed that the functions of denial and contrast corresponded closely. The authors therefore concluded that the function of Spanish en real dad is exactly the same as that of English in reality, viz. ‘the utterance U2 denying the explicit proposition of U1, with the same constraints for mood and responsibility of the assertion being denied’ (Fraser & Malamud-Makowski 1996: 878). It is clear that we need many similar investigations. Cross-linguistic comparisons help us to identify both similarities and differences between the markers and what it is that marks the boundary between the acceptable and unacceptable translations. A marker may lack a correspondence in another language; if there is a cognate in the other language it may have a different semantic value. Cross-linguistic comparison may also reveal partially overlapping distributions between languages. The area of functional equivalence can range from complete overlap to cases where the area amounts to almost nothing. Bruti showed, for example, that

Introduction 5 there is very little equivalence between Italian infatti and in fact in spite of the etymological relationship between the two markers (Bruti 1999).

4.

METHODOLOGY

It goes without saying that the basic methodology for studying the function of pragmatic markers is the analysis of natural discourse and that the methods used are those common in discourse analysis. However, additional methods of a more experimental type can yield deeper insights or provide answers to questions that cannot easily be answered on the basis of conversational or discourse analysis alone. Fischer (2000), for instance, has shown that varying the communicative situation in an experimentally controlled way leads to differences in the occurrence and use of certain pragmatic markers. This seems a productive way to test hypotheses about the dimensions of the communicative situation that the markers relate to. Moreover, the use of translation corpora and corpus linguistic methodology have revolutionized contrastive linguistics and made it possible to make more detailed comparisons of usage in two or more languages. Several types of bilingual or multilingual corpora can be used. A distinction can be made between comparable corpora and translation corpora. Comparable corpora are matched in terms of text type, subject matter and function with a corresponding corpus in another language. An advantage of consulting comparable corpora is that one avoids the influence of the translation effects on the data which are used. The difficulty with comparable corpora is, however, that one first needs to establish what it is one wants to compare. Especially in the case of pragmatic markers, where the other language frequently offers a large number of alternatives, it is difficult to establish correspondences between the languages. A translation corpus, on the other hand, makes it possible to establish cross-linguistic paradigms showing how an element in Language A is rendered by alternative variants in Language B. It is also possible to go from translations to sources in order to establish equivalents. The drawback of using translation corpora is that the translations are affected by the source language or by the individual translator (Gellerstam 1996). Comparable corpora and translation corpora can also be used in combination. This is the case with the EnglishNorwegian and the English-Swedish Parallel Corpus which contain comparable texts in the two languages as well as translations into the other language. Bilingual and multilingual corpora are used for research in a large number of areas and new methodologies and tools are developed for using them for empirical research. An approach which uses translation equivalents in a corpus to get a better picture of the meanings of lexical items has been proposed by Dyvik (2004). Dyvik’s method implies that we look at the meanings of a lexical item as mirrored in another language. Translation equivalents are based both on translations and sources. By going back and forth between sources and translations we can calculate an intertranslatability index between elements in two languages (Altenberg & Granger, 2002: 17f). The translations make it possible to confront elements in the two languages with each other and thus provide a better picture of similarities, differences and overlapping meanings. The methodology has for instance been used in Johansson & Oksefjell (1998), Hasselgård & Oksefjell (1999), Altenberg & Granger

6 Pragmatic Markers in Contrast (2002), Simon-Vandenbergen & Aijmer (2002/2003). The advantage of using translation corpora is especially strong for pragmatic markers, precisely because of their underspecified core meaning and their polysemous nature. Translations can moreover serve as a heuristic device for discovering the relevant contextual dimensions or for making more fine-grained divisions in these dimensions. The reason is that the translations force one to account for the contextual factors that lead to particular choices.

5.

STRUCTURE OF THE BOOK

The book consists of twelve chapters. The papers originate from a colloquium ‘Pragmatic markers in contrast’, held in Brussels 22 – 23 May 2003, subsidised by the Royal Flemish Academy of Belgium for Science and the Arts, with invited speakers representing expertise in this area of research. What the contributions show is that the picture of similarities and differences is a complex one and that many different factors need to be taken into account to explain these similarities and differences. The articles have been arranged in alphabetical order. In our brief description of the contents below, however, we shall deviate from alphabetical order to group certain articles. There are parallels between some contributions on the basis of the types of markers contrasted, on the basis of the types of data used, and on the basis of the types of questions raised. For instance, some articles share the comparison of cognates, many share the reliance on translation data, most articles are primarily descriptively oriented, while others (Östman and White & Sano) tackle broader issues, and one paper (Vandeweghe) is typological rather than contrastive. Various studies focus on cognates. Stig Johansson’s study of well goes beyond earlier cross-linguistic work on well (e.g. Aijmer & Simon-Vandenbergen, 2003; Bazzanella & Morra, 2000) by extending the analysis to more language pairs (English-Norwegian, EnglishGerman). Johansson’ analysis is particularly interesting because Norwegian has vel which is a cognate of well. Although vel can have a large number of different functions it is shown to be more marginal than well, which covers a wider area of use. The zero correspondences in the data were fewer than in Swedish (no doubt because of the frequency of vel) but more frequent than in Dutch, where well frequently corresponds to nou. The translations into Norwegian and German represented a wide range of correspondences in both languages. Another interesting cognate pair is now and nå which is studied in an EnglishNorwegian perspective by Hilde Hasselgård. Now/nå has meanings (functions) on three different levels which can be described with the help of the three Hallidayan metafunctions. As a temporal adverb nå/now has an ideational function. However, it can also be described on the textual level as a continuative and on the interpersonal dimension as a modal particle. Translations are used to describe quantitative and qualitative differences between the languages. English and Norwegian were seen to differ most markedly when now/nå was a discourse marker with continuative meaning. The Norwegian modal particle was often omitted in translation, indicating that it has little meaning.

Introduction 7 Ad Foolen compares the Dutch toch with the cognate German doch on the basis of data from the translation of a Dutch novel into German. Foolen found a recurrent ‘schematic’ meaning in all the uses of toch and doch where the different steps in the scheme could be described as affirmation-negation- (re)affirmation. This was seen to correspond to different functional subsystems (such as connectives, modal particles, causal markers, dialogic particles) in the languages in which doch and toch function. Although all the meanings were plausible applications of the schematic meaning of the particle there were language-specific uses, for example in causal contexts and when toch and doch are modal particles. The topic of Diana Lewis’ contribution is the cognates au contraire in French and on the contrary in English. The methodology in this case involves using comparable corpora of French and English political speeches. The differences established were both quantitave and qualitative. Au contraire was for instance more frequent than on the contrary in the compared databases suggesting that it has a wider functional domain. Diana Lewis’ paper draws attention to the fact that the conceptual meaning of related connectives can differ depending on the language. Thus on the contrary is inherently negative as a result of semantic shift and therefore not appropriate in the same contexts as au contraire. Because of its meaning it seems to be restricted to argumentational contexts in which one idea is rejected as false and replaced by a preferred, ‘true’ idea. Anna-Brita Stenström’s paper shows that when we include several related markers in the comparison the result is a complicated network of contrastive correspondences. The author compares two common Spanish discourse marker o sea and pues with the equally frequent English cos and well. The investigation is based on frequencies of the different markers in comparable corpora of adolescent speech in Spanish and in English. The markers are multifunctional, which is reflected in the translations into the other language. For example when o sea is analysed as a connective it can be translated by well or that is. O sea can also be epistemic. In this case it can also be translated by well but other translations are preferred. Pues corresponds to ‘cos’ as a causal connector although it would probably be untranslated in many cases. As a pragmatic marker on the discourse level it would, however, be translated by well. The careful comparison of cognates such as Angela Downing’s comparison of surely and its Spanish correspondences is additional evidence that we find a complex intertwining between universal and language-specific features when we compare pragmatic markers. Despite the fact that there are a number of lexemes in Spanish which are etymological cognates of surely (seguramente, seguro, (que)) they do not have the same functions, as is illustrated by translational mismatches. In conclusion, it is shown that the Spanish phrases are only partial equivalents of English surely mapping only some of the variations expressed by the English word and that they then express a weaker epistemic stance. In the paper by Dominique Willems and Annemie Demol vraiment in spoken and written French is compared with different uses of really. Moreover vraiment and really were compared with their Dutch correspondences echt and werkelijk. When one compares vraiment with really in English, they seem to be almost perfect equivalents. In Dutch both echt and werkelijk can be translation equivalents of really and vraiment. However they did not seem to share the more pragmaticalized meaning extensions. A striking finding was that really was

8 Pragmatic Markers in Contrast often not translated into Dutch and that really was sometimes inserted in the translation of Dutch sources containing no lexical counterpart of really. The contrastive analysis can also be used to shed light on phenomena in a single language. Bengt Altenberg investigates the text strategic principles determining the variation between (a) However, my father didn’t worry and (b) My father, however, didn’t worry and examines how the functional difference between the (a) and the (b) versions is expressed contrastively by studying the translations into Swedish. Because of the V2-constraint the Swedish correspondences of English connectors in second position are only exceptionally placed in this position. The typological difference between the languages explains that initial English connectors tend to drift into the middle field in the Swedish translations although there were differences depending on the semantic type of connector. Another approach represented in this volume is the language-typological one. Willy Vandeweghe discusses aspectual particles in some of the languages in Europe (cf English already/not yet/still/not anymore). It is shown that a large number of factors interact with aspectuality and that some of the factors which distinguish between aspectual predicates can be described on a ‘duality square’. The sample of languages and the conclusions drawn from it in Vandeweghe’s study are compared with the results from the EUROTYP project led by van der Auwera (van der Auwera 1991). It is for instance shown that there are few languages with no aspectual particles at all. Bruce Fraser’s paper also deals with a large number of different languages in a typological perspective. It is concerned with contrastive discourse markers, in particular with but, and with the question whether there is an equivalent in other languages that functions as but does in English. In order to test this, different contrastive contexts are established. The methodology used to gather the data for the contrastive analysis is particularly interesting. In order to determine if other languages also had a primary contrastive discourse marker equivalent to but, and if so whether it patterned like but, a survey was sent to native speakers of a large number of languages which included both contexts that accepted but in English and those which did not. There are, as we suggested earlier in this introduction, a number of theoretical issues which need to be confronted in (contrastive) work on pragmatic markers, and Jan-Ola Östman’s contribution tackles some of these wider issues. One question dealt with is that of the so-called ‘multifunctionality’ of pragmatic markers. Another issue is how we can find an adequate tertium comparationis in the field of cross-cultural research on pragmatic markers. Östman argues for a constructional approach to pragmatic phenomena and to pragmatic particles in particular. Cross-linguistic comparisons are, in this view, possible at the level of constructions only. The usefulness of a Construction Grammar approach is illustrated with particles in Solv (a dialect of Finland Swedish), and it is argued that detailed descriptions of language-internal phenomena are a prerequisite for cross-linguistic comparisons. Peter White and M. Sano also go beyond the analysis of specific items and propose a broader framework for analysing pragmatic makers intra-linguistically and crosslinguistically. Pragmatic markers are dealt with as options in the system of ‘engagement’, from which speakers select to position themselves dialogistically. The focus in this contribution is on those markers which are traditionally called ‘epistemic’, including I think, obviously, of course. The functionality of such items is examined in public discourse. While

Introduction 9 the engagement meanings are primarily studied in English, the authors raise the question to what extent the same options would be available in other languages. A preliminary study of the dialogistic positioning of the British and Japanese Prime Ministers at press conferences suggests, however, that there are similarities. The cross-linguistic applicability of the framework is a fascinating topic for further research. In conclusion, we hope that the diversity of approaches, theoretical frameworks, methodologies and language data brought together in this volume will both contribute to answering some questions and raise new ones, thus stimulating further research in this area. Finally, we would like to thank the Royal Flemish Academy of Belgium for Science and the Arts for its generous support of the colloquium which was at the origin of this book. We also owe special thanks to the anonymous referees of Studies in Pragmatics, whose suggestions have been very helpful, and to Lars Malmsten for his expert help with the technical aspects of this volume.

REFERENCES Aijmer, K. (2002). English Discourse Particles. Evidence from a Corpus. John Benjamins Publishing Company, Amsterdam/Philadelphia. Aijmer, K. & A.- M. Simon-Vandenbergen. (2003). The discourse particle well and its equivalents in Swedish and Dutch. Linguistics, 41-6, 1123-1161. Altenberg, B. & S. Granger (eds) (2002). Lexis in Contrast. Corpus-based Approaches. John Benjamins Publishing Company, Amsterdam/Philadelphia. Andersen, G. and T. Fretheim (eds), (2000). Pragmatic Markers and Propositional Attitude. John Benjamins Publishing Company, Amsterdam/Philadelphia. Bazzanella, C. (1990). Phatic connectives as interactional cues in contemporary spoken Italian. Journal of Pragmatics, 14, 639-647. Bazzanella, C. and L. Morra (2000). Discourse markers and the indeterminacy of translation. In: Argomenti per una linguistica della traduzione, On linguistic aspects of translation, Notes pour une linguistique de la traduction (I. Korzen and C. Marello, eds), pp.149-157. Edizioni dell' Orso, Alessandria. Blakemore, D. (1987). Semantic Constraints on Relevance. Blackwell, Oxford. Bruti, S. (1999). In fact and infatti: The same, similar or different. Pragmatics 9(4), 519-533. Crystal, D. & D. Davy. (1975). Advanced Conversational English. Longman, London. Dyvik, H. (2004). Translations as semantic mirrors: from parallel corpus to wordnet. In: Advances in Corpus Linguistics. Papers from the 23rd International Conference on English Language Research on Computerized Corpora (ICAME 23) Göteborg 22-26 May 2002 (K. Aijmer and B. Altenberg, eds), pp. 311-326. Rodopi, Amsterdam. Fernandez, M. M. J. (1994). Les particules énonciatives dans la construction du discours, Presses universitaires de France, Paris. Fischer, K. (2000). From Cognitive Semantics to Lexical Pragmatics. The Functional Polysemy of Discourse Particles. Mouton de Gruyter, Berlin & New York.

10 Pragmatic Markers in Contrast Fleischman, S. and M. Yaguello. (1999). Discourse markers in comparative perspective: A contribution to cross-language pragmatics. Hand-out at the ICLA Conference, Stockholm. Fraser, B. (1990). An approach to discourse markers. Journal of Pragmatics 14, 383-395. Fraser, B. and M. Malmaud-Makowski (1996). English and Spanish contrastive markers. Language Sciences 18, 863-881. Gellerstam, M. (1996). Translations as a source for cross-linguistic studies. In: Languages in Contrast. Papers from a Symposium on Text-based Contrastive Studies (K. Aijmer, B. Altenberg and M. Johansson, eds), pp. 53-62. Lund University Press, Lund. Hansen Mosegaard, M.-B. (1998). The Function of Discourse Particles. A Study with Special Reference to Spoken Standard French. John Benjamins Publishing Company, Amsterdam/Philadelphia. Hasselgård, H. and S. Oksefjell (eds) (1999). Out of Corpora. Studies in Honour of Stig Johansson. Rodopi, Amsterdam & Atlanta, GA. Johanssson, S. and S. Oksefjell (eds) (1998). Corpora and Cross-linguistic Research: Theory, Method and Case Studies. Rodopi, Amsterdam & Atlanta, GA. Jucker, A. H. and Y. Ziv (eds) (1998). Discourse Markers. Descriptions and Theory. John Benjamins Publishing Company, Amsterdam/Philadelphia. Lenk, U. (1998). Marking Discourse Coherence. Functions of Discourse Markers in Spoken English. Gunter Narr Verlag, Tübingen. Schiffrin, D. (1987). Discourse Markers. Cambridge University Press, Cambridge. Simon-Vandenbergen, A.-M. and K. Aijmer (2002/2003). The expectation marker of course. Languages in Contrast 4(1), 13-43. Traugott, E. C. and R. B. Dasher (2002). Regularity in Semantic Change. Cambridge University Press, Cambridge. van der Auwera, J. (1991). Beyond duality. In: Adverbs and Particles of Change and Continuation (J. van der Auwera, ed.), Eurotyp Working Papers Vol. 2, pp. 131-159. Strasbourg, European Science Foundation,


THE FUNCTION OF ADVERBIAL CONNECTORS IN SECOND INITIAL POSITION IN ENGLISH AND SWEDISH Bengt Altenberg, University of Lund

1.

INTRODUCTION

Most adverbial connectors have a variable clause position in English, but their favoured position is clause-initial (see Quirk et al., 1985: 643, Biber et al., 1999: 891). This seems natural in view of their function: their role is to relate a unit of discourse to the preceding context, and to signal the nature of this relationship as early as possible is likely to facilitate the listeners’ or readers’ comprehension of the progression of discourse. However, the exact position of a connector at the beginning of a clause may vary: 1. (a) (b)

However, in Britain the situation was different. In Britain, however, the situation was different.

Both these examples have a ‘multiple theme’, to use Halliday’s (1994: 52 ff.) terminology. Two adverbials occur before the subject but their order is different: in (a) the connector is placed initially before a place adverbial; in (b) it is placed ‘parenthetically’ in second position after the place adverbial. Both elements are ‘thematic’ in Halliday’s sense (1994: 38), i.e. they indicate the point of departure or ‘framework’ of the message, but they do this in different ways (on various types of framework, see Downing, 1991). While the connector is a ‘textual’ theme signalling a contrastive relationship with the preceding discourse (not indicated here), the place adverbial is an ‘ideational’ theme setting up a spatial framework for the rest of the message. By varying the order of these, different thematic priorities are indicated and the examples can be seen to fit into different contexts. While (a) suggests a logical or rhetorical type of discourse in which the main point is to signal a contrast, (b) also draws attention to a spatial break or shift in the progression of discourse. The discourse-related order of complex sentence beginnings realized by two or more adverbials has been investigated by Aletta Smits (2002) on the basis of a corpus of native and learner English. She makes a distinction between (a) sentences that have what she calls a ‘stepwise orientation’, as in (1a), where an initial connector (or other ‘peripheral’ adverbial) takes in its scope a clause which is itself preceded by a frame-setting adverbial, and (b) those that have a ‘complex orientation’, as in (1b), where the second adverbial is syntactically and prosodically parenthetic and serves to highlight (and in some sense modify) the first.1 1 Smits also distinguishes other types of orientation (2002: 80 ff.), but only the ‘stepwise’ and ‘complex’ types are relevant here. Smits explains the textual function of elements in second position in terms of ‘grounding’

12 Pragmatic Markers in Contrast

Following Virtanen (1992), who in turn develops ideas put forward by Givón (1983) and Enkvist (e.g. 1987), Smits demonstrates convincingly that while stepwise orientations are especially suited to types of discourse that have a predominantly rhetorical progression or ‘text strategic continuity’ highlighting a logical development of ideas, complex orientations fit most naturally into discourse types dominated by shifts in a spatial, temporal or topical line of development (cf. also Altenberg, 1998). Another pair of examples illustrating the two types of orientation are: 2. (a) (b)

For instance, last week John’s neighbour complained about the noise. Last week, for instance, John’s neighbour complained about the noise.

Smits’ analysis goes a long way to explaining the discourse-driven order of initial adverbials in English. However, some aspects remain unclear. For one thing, her study is restricted to complex beginnings consisting of two (or more) adverbials but, as she herself points out (2002: 135 f.), connectors (and other higher-level adverbials) may also alternate with a subject to create different discourse-related orientations: 3. (a) (b)

However, my father didn’t worry. My father, however, didn’t worry.

It is consequently of interest to find out what text strategic principles determine this kind of alternation. Secondly, one intriguing problem in Smits’ investigation is that the distinction between stepwise and complex orientations is particularly clear in cases involving certain connectors like however and for instance but less so in examples containing, for example, an additive connector like moreover. Thus, in the following pair, the (b) version, although quite acceptable in the right context, seems less likely to occur than the (a) version: 4. (a) (b)

Moreover, the Belgians have their chocolate. The Belgians, moreover, have their chocolate.

(2002: 59 ff.), but I will simply talk of ‘highlighting’ since a parenthetical connector typically draws attention to a preceding element by separating it from the rest of the clause and putting special focus on the highlighted element. The prosodic treatment of parenthetical connectors supports this (cf. Smits, 2002: 87 f). In the LondonLund Corpus of Spoken English (LLC) parenthetical examples of however either have no nuclear tone at all or, more commonly, carry a rising tone; in either case the connector is also normally followed (but not preceded) by a tone-unit boundary. In other words, clause-internal however typically closes a tone unit and is prosodically integrated into it: (a) ^one "!m\/ight want to 'argue how'ever# […] – that it ^might a'waken the 'urges of a :psychopath whose :sadism was !still 'sub":l/iminal# (LLC 12.7: 593-595) (b) ^fr\equently how/ever# . the ex^p\onents of c/ensorship# "^cl\aim to 'base their obj/ections# ^on "!purely pru"!d\/ential 'grounds# (LLC 12.7: 82-85) Other parenthetical connectors have a similar prosody, a tendency that clearly reveals their ‘backward-focusing’ function.

Adverbial Connectors in Second Position 13 This suspicion is supported by one of the informant experiments Smits conducted to supplement her corpus investigation (see Smits, 2002: 239 ff.). In this experiment (an ‘unscrambling’ task, without a context guiding the production), the great majority of her test subjects preferred to place moreover in absolute initial position (stepwise orientation), as in (4a), while very few chose to place it in second initial position (complex orientation), as in (4b). Interestingly, this tendency was much stronger among her native informants than among the Dutch learners of English that were also tested (2002: 255). Smits suggests that these results would have been different if moreover had been replaced by a “more mobile” connector like furthermore (2002: 241). Although this is doubtful, the suggestion indicates a complicating factor in the placement of mobile connectors in English: individual connectors tend to have very different positional preferences and our knowledge of these is insufficient. Descriptive grammars give information about certain general tendencies (see e.g. Quirk et al., 1985: 643, Biber et al., 1999: 890 ff.), but little is known about the positional profiles of individual connectors, especially their ability to occur parenthetically in second initial position. It is therefore important to investigate these tendencies in greater detail. Finally, the highlighting function of parenthetical connectors is also of contrastive interest. Swedish is a verb-second language that normally only permits one clause element before the finite verb in declarative main clauses. This means that, theoretically speaking, none of the initial sequences illustrated in (1) – (4) should be possible in Swedish. The question then is: how is the functional difference conveyed by the (a) and (b) versions above handled in Swedish? Is thematic highlighting of the kind illustrated in the (b) versions quite impossible or can a similar effect be achieved by other means?

2.

AIMS AND MATERIAL

In the present study I will briefly examine the questions raised above on the basis of a bidirectional English-Swedish translation corpus. I will first explore the positional tendencies of some common English adverbial connectors, with special focus on their readiness to occur parenthetically in second initial position. Using Smits’ approach, I will then examine the discourse factors promoting the placement of certain types of connectors in the thematic field in English, focusing again on parenthetic second position but extending the analysis to include connectors following subjects as well as adverbials. Finally, I will briefly look at the treatment of the corresponding phenomena in Swedish original texts and translations. The study is based on the English-Swedish Parallel Corpus (ESPC) (see Altenberg and Aijmer, 2000). This corpus consists of a wide range of text samples from original English and Swedish sources and their translations into the other language. Both fiction and non-fiction texts are represented in the corpus. The original English and Swedish texts have been matched as far as possible in terms of text type and purpose and the corpus can therefore be used both as a comparable corpus and as a translation corpus (on the advantages of this, see Johansson, 1998). The total size of the corpus (including original texts and translations from both languages) is over 2,8 million words. For a detailed description of the corpus, including an explanation of the text codes used in the paper, see Altenberg et al. (2001).


The notation describing the clause positions of adverbial connectors in both languages is that of Quirk et al. (1985: 490 ff.). The symbols used for the various positional slots (indicated by arrows) are shown in Figure 1. I I2 iM M iE E By then the book had been placed on the shelf Figure 1. Adverbial positions in English

The parenthetic positions that are of special interest here are I2 (between an initial adverbial and the subject), iM (between the subject and an operator) and M (between the subject and the main verb). When no further specification is needed, these positions will be referred to as ‘second position’. Only connectors occurring in declarative main clauses have been examined. Connectors with narrow ‘phrasal’ scope of the following kind have also been disregarded:2 5. Document leaks concerning, for example, intended social security cuts were always a hazard. (FF1)

3.

POSITIONAL TENDENCIES OF ENGLISH CONNECTORS

To establish the positional preferences of the English connectors in the corpus the clause position of connectors occurring at least twenty times in the English original texts were examined. Connectors having a fixed clause-initial position (so, yet) were excluded. As it turned out that many interesting connectors were not frequent enough in the original texts, positional data from the English translations were added for these connectors.3 The results are shown in Table 1. The connectors are sorted according to their tendency to occur in initial position (I). Connectors whose distribution is based on both the English original texts and translations are marked by an asterix. The table shows that the majority of the connectors prefer initial position, some almost exclusively so (e.g. besides, in other words, in addition, even so, furthermore). However, with quite a few connectors – at the bottom of the list – other positions predominate. Though, for example, never occurs at I (where it functions as a subordinating conjunction) and therefore 2 It should be added that many adverbial connectors are of course frequently used parenthetically to highlight an element elsewhere in a clause, for example after the operator or a negator to emphasize the truth value or polarity of a proposition: I do, however, recall one moment of piercing sadness. (JB1) Though this is another interesting property of mobile connectors, it will be ignored here. 3 It may be objected that data from the English translations are unreliable because of possible influence from the Swedish source texts. However, since Swedish connectors are (as we shall see) extremely rare in second position, placement of connectors in this position in the English translations is unlikely to be triggered by the Swedish source texts. As a control measure, the positional tendencies of some connectors were also examined in the British ICE-GB corpus (see note 5).

Adverbial Connectors in Second Position 15

and after all occur at I in less than a third of the cases. On the whole, the table gives clear evidence of individual positional ‘profiles’ among the connectors. Some have a strong tendency to occur at E (though, anyway, then), others prefer M (accordingly, therefore, still), while others are never placed there in the corpus (furthermore, consequently, moreover). Some connectors are extremely mobile without any clear preference for any position (e.g. however, for example, for instance), while others are practically restricted to one position (e.g. besides at I). Of special interest here are connectors appearing at I2, i.e. between an initial adverbial and the subject. As shown in the table, comparatively few connectors are used in this position, some notable exceptions being however (21%), for example (16%), for instance (14%), moreover (10%) and though (10%). I will return to these below.

Table 1. Positional tendencies of English connectors (%)

Connector besides* in other words* in addition even so* furthermore* as a result* on the contrary* consequently* moreover* above all* thus in any case* on the other hand nevertheless instead by the way* anyway all the same* for example for instance then still accordingly* however therefore after all though Total

I 98 94 93 92 92 89 84 83 77 72 66 63 63 61 58 57 52 50 49 45 42 40 37 35 31 22 0 55

I2 0 0 0 0 2 3 0 0 10 3 0 0 0 3 0 0 0 3 16 14 8 0 0 21 3 2 10 6

iM 0 0 0 0 4 3 6 0 2 0 2 6 17 6 2 5 0 0 4 9 3 12 3 11 3 11 12 5

M 0 0 0 0 2 0 3 17 6 10 30 9 17 26 2 5 1 0 14 14 8 48 57 18 56 26 2 16

iE 0 0 3 0 0 0 3 2 4 7 2 3 4 3 29 0 0 3 8 9 2 0 3 6 4 4 7 4

E 2 6 3 8 0 5 3 0 0 7 2 20 0 0 10 33 47 43 10 9 37 0 0 9 2 35 68 14

Total (n) 46 32 30 24 49 37 32 48 48 29 61 35 24 31 52 21 79 30 51 22 91 42 35 190 93 46 41 1319

16 Pragmatic Markers in Contrast Despite the great positional diversity demonstrated in Table 1, it is of some interest to see if the position of the connectors is related to the semantic type of relation they express. To establish this, the connectors were grouped into semantic categories in accordance with the classification in Quirk et al. (1985: 634) and the (combined) positional tendency of each category was calculated.4 The distribution of the five most common categories is shown in Table 2. As the table demonstrates, the additive connectors strongly prefer initial position (88%), while the other categories have a more varied distribution. The I2 position is rare with all the categories except those containing exemplifying and contrastive connectors. Hence, what this semantic grouping of the connectors suggests is that additive connectors tend to take scope over the entire clause in which they occur, whereas exemplifying and contrastive connectors often occur in second initial position highlighting a preceding adverbial for special thematic attention.

Table 2. Positions of English connectors by semantic type (%)

Type Additive Resultive Exemplifying Concessive Contrastive Total

4.

I 88 53 48 48 46 56

I2 3 3 15 2 13 6

iM 1 2 5 5 9 5

M 3 29 14 12 14 16

iE 2 2 8 2 9 4

E 2 11 10 31 8 13

Total (n) 202 365 73 282 298 1220

ENGLISH CONNECTORS IN PARENTHETIC SECOND POSITION

While Tables 1 and 2 give a good indication of which connectors tend to occur at I2, i.e. after an initial adverbial, they tell us little about connectors placed parenthetically after the subject. To establish this, it is necessary to examine connectors occurring at iM and M and distinguish cases where the connector is parenthetic from those where it is not. Parenthetic connectors are normally, but not always, set off by commas in the corpus. However, even in cases without 4 The classification agrees with that of Quirk et al. (1985: 634) with the exception that a distinction has been made between ‘contrastive’ (adversative) and ‘concessive’ connectors. The borderline between these is not always clear, however. Though, for example, is functionally (and prosodically) similar to the contrastive connectors, but I have followed Quirk et al. (1985) and classified it as concessive (for a more detailed discussion, see Fraser 1998 and Altenberg 2002). The exemplifying category only contains two ‘appositive’ connectors, for example and for instance. Connectors that are not common enough in the corpus to form a category (e.g. summative or transitional connectors) are excluded in the table.


commas the ‘reading pronunciation’ of the connector as parenthetic or non-parenthetic (i.e. prosodically ‘integrated’) is usually clear. Table 3 shows the same connectors as in Table 1 listed in terms of their tendency to occur in parenthetic second position following either an adverbial or a subject.

Table 3. Tendency of English connectors to occur parenthetically in second position

Connector

however on the other hand for instance for example though after all moreover* furthermore* on the contrary* therefore as a result* by the way* then accordingly* in addition* all the same* anyway besides* consequently* even so* in other words* still nevertheless thus in any case* instead

Second position ParenIntethetic grated 78 0 8 0 7 0 16 0 10 0 7 0 6 0 3 0 2 0 6 14 2 0 1 0 4 7 1 8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 13 0 6 0 8 0 2 0 1

Total

% parenthetic

190 24 22 51 41 46 48 49 32 93 37 21 91 35 30 30 79 46 48 24 32 42 31 61 35 52

41 33 32 31 24 15 13 6 6 6 5 5 4 3 0 0 0 0 0 0 0 0 0 0 0 0

As the table shows, the connectors describe a cline from those having a relatively strong tendency to occur parenthetically in second position to those that never do so in the corpus. At the top of the scale we recognize the contrastive connectors however and on the other hand, concessive though and the exemplifying connectors for instance and for example. Connectors with a weaker parenthetical tendency (< 20%) include concessive after all, the additive connectors moreover and furthermore and a mixture of resultive and other contrastive connectors. The table also shows that some connectors that do occur very often in second position are commonly or exclusively prosodically integrated into their clause. These include


therefore, still, accordingly, thus and then. Some of these can be either parenthetical or integrated, as illustrated by therefore in the following examples: 6. Columbus had been an itinerant peddler of old maps and an assiduous reader of the books by and about the ancient geographers […] But for the Enterprise of the Indies to work, for ships and crews to survive the long voyage, the Earth had to be smaller than Eratosthenes had said. Columbus therefore cheated on his calculations, as the examining faculty of the University of Salamanca quite correctly pointed out. (CSA1) 7. Of course they sentenced him [Socrates] to death. He'd mentioned also that friends of his wished him to recommend a fine of thirty minae, which they would guarantee, and this, therefore, was what he consented in the end to do. (JH1)

When the subject presents given information, as in (6), connectors like therefore are usually prosodically integrated in middle position in the clause. However, the subject in (7) is also given, so a simple given-new distinction is not sufficient to explain the difference. What seems to be decisive for the parenthetic treatment of therefore in (7) is its strong deictic character. The pronominal subject of reversed wh-clefts is normally anaphorically emphatic and it is this quality that is highlighted by therefore. Hence, unlike therefore in (6), which focuses on the predicate, parenthetical therefore in (7) has a clear backward-focusing effect. If we look at the five semantic types distinguished above, we find two divergent tendencies (see Table 4). While exemplifying and contrastive connectors are used parenthetically in second position in about a third of the cases, additive, concessive and resultive connectors are relatively seldom used in this way.

Table 4. Parenthetical connectors in second position by semantic type

Types Exemplifying Contrastive Additive Concessive Resultive

Parenthetical 23 95 10 15 13

% 32 30 5 4 3

Total 73 322 218 368 373

In other words, the position and functional use of the connectors seems to be closely related to their meaning. Exemplifying and contrastive connectors appear to be especially suited for highlighting a preceding adverbial or subject, while other types of connectors either tend to be placed initially (especially the additives) or be prosodically integrated within the clause. This difference will be examined more closely in the following section.


5.

THE DISCOURSE FUNCTION OF ENGLISH CONNECTORS IN SECOND POSITION

Let us now examine the function of some English connectors occurring in parenthetic second position in the corpus. The scrutiny will be limited to three types of connectors, those expressing contrast, exemplification and addition. The focus will be on the textual role played by the clause element preceding the connector. The examples are drawn from the English original texts; the Swedish translations will also be presented but comments on these will be postponed to section 6. 5.1

Contrastive connectors

The contrastive connectors occurring in second position in the corpus are however, on the other hand, by/in contrast and on the contrary. When a contrastive connector is placed parenthetically after an initial adverbial the latter is often a phrase or clause indicating a temporal or spatial shift in a dominant text strategic continuity (TSC), as illustrated in the following examples: 8.

By March 1905 the confidence of French investors had been so shaken by both the abortive Russian revolution and Russian reverses in the war against Japan that with the support of Delcassé, the French Foreign Minister, Raffalovich was distributing bribes to the tune of 200,000 francs a month. As usual in the case of agents of influence, it is difficult to assess the importance of the press support purchased in this way. In March 1905 even Raffalovich's largesse failed to prevent French banks breaking off negotiations for a further loan. By 1914, however, 25 per cent of France's foreign investment was in Russia (four-fifths of it in government loans) — as compared with only 9 per cent in the vast French Empire. (CAOG1)

I mars 1905 var de franska investerarnas förtroende så rubbat av både den felslagna ryska revolutionen och Rysslands motgångar i kriget mot Japan att Raffalovitj med stöd av den franske utrikesministern, Delcassé, delade ut mutor till ett belopp av 200 000 franc i månaden. Som vanligt när det gäller inflytelseagenter är det svårt att uppskatta betydelsen av det stöd från pressen ryssarna köpte sig på detta sätt. I mars 1905 kunde inte ens Raffalovitjs frikostighet förhindra att de franska bankerna avbröt förhandlingarna om ytterligare ett lån. År 1914 däremot låg tjugofem procent av Frankrikes utrikesinvesteringar i Ryssland (fyra femtedelar av dem i lån till regeringen) — vilket kan jämföras med bara nio procent i det väldiga franska imperiet.

9.

In Western Europe the development of the cabinets noirs was disrupted in varying degrees during the nineteenth century by public and parliamentary protests at interference with the post. In Britain, for example, the Decyphering Branch was

I Västeuropa avbröts utvecklingen av cabinets noirs i varierande grad under 1800talet på grund av allmänhetens och parlamentens protester mot att posten lästes. I till exempel Storbritannien lades Decyphering Branch (forceringsavdelningen)

20 Pragmatic Markers in Contrast abolished in 1844 after a Commons row over the opening of the correspondence of the exiled Italian nationalist, Giuseppe Mazzini. British sigint did not resume until the First World War. In autocratic Russia, however, the development of sigint was undisturbed by parliamentary protests. (CAOG1)

ned 1844 sedan gräl utbrutit i underhuset beträffande det faktum att den landsförvisade italienske nationalisten Giuseppe Mazzinis post öppnades. Den brittiska signalspaningen återupptogs inte förrän under första världskriget. I det autokratiska Ryssland kunde signalspaningen emellertid utvecklas utan att störas av parlamentariska protester.

In (8) there is a dominant temporal TSC, as indicated by the underlined adverbials in clauseinitial position. By 1914 marks a clear shift in this continuity at the same time as a contrast is expressed in the rest of the clause (the size of French investment in Russia before and after this year). In a sense, the clause can be said to express a double contrast, one signalled by the temporal adverbial and one in the rest of the clause. Prosodically, however neatly divides this contrastive information into two information (and intonation) units, each with a focus of its own. The effect is that the temporal shift is highlighted. In (9) the dominant TSC is spatial. The underlined initial adverbials signal the shifts and the final one is combined with a contrast (the ‘undisturbed’ state of affairs in contrast to the the ‘disrupted’ situation in Western Europe and Britain). Again, the position of the connector sets off the initial adverbial, highlighting the spatial shift and the double contrast created by the prosodic division. However, an initial adverbial does not have to take part in a dominant TSC to be highlighted. As long as it signals a clear temporal or spatial shift in the progression of the discourse, it can be set off by a parenthetic contrastive connector: 10. We have thus long had the people specially fitted to participate in such an expedition, but what we have lacked until very recently is the man with the qualifications to be the leader. Now, however, I believe, that we have such a man: he, who has had his Arctic baptism with an enterprise which has aroused attention all over the civilised world. (RH1)

Vi har således länge haft folk som är speciellt lämpade för att delta i en sådan expedition, men till helt nyligen har vi saknat en man som är kvalificerad att leda den. Nu har vi emellertid enligt min mening en sådan man: han som har fått sitt arktiska dop med ett företag som har väckt uppmärksamhet i hela den civiliserade världen.

Here there is no dominant TSC, yet the temporal shift is clear: while the first paragraph describes a situation in the past, the second moves the time frame to the present at the same time as a contrastive state of affairs is presented (the existence of a leader). Temporal shifts of this kind are in fact more common than shifts in a dominant TSC in the corpus. Often they indicate a ‘global’ time shift changing the temporal frame backwards or forwards across a considerable stretch of discourse. In such cases the change often coincides with a paragraph shift: 11. It was not until years later that Kate, looking back, began to think that Miss Simmons had

Inte förrän många år senare började Kate, när hon tänkte tillbaka, få en känsla av att miss


been physically frightened of her pupils, that the sight of them sitting or standing before her, sullen, bored, ominously fertile, potentially rebellious, must have been as unnerving to her as the sight of an angry mob to a Victorian millowner, as the sight of a mutinous hungry crew to a captain far from land. Kate even began, posthumously, to feel sympathy, for on the few occasions in adult life when she found herself before an audience of teenage girls, she could see that it took a good deal of nerve to face such a mass of young restless and hopeful bodies with confident equanimity, and poor Miss Simmons had not had much. At the time , however, Kate, like the other girls, chafed against restrictions, talked in class, picked up one or two O levels with not very good grades […] and found herself, by the age of sixteen, with not much idea of what she wanted to do next. (MD1)

Simmons hade varit rädd för sina elever, rent fysiskt; att åsynen av dem där de satt eller stod framför henne, buttra, uttråkade, olycksbådande könsmogna, potentiellt upproriska, måste ha varit lika skrämmande för henne som åsynen av en retad arbetarskara för en viktoriansk fabriksägare eller åsynen av en hungrig om myteri mumlande besättning för en kapten långt från land. Kate började till och med i efterhand känna en viss identifiering med henne, för vid de få tillfällen i sitt vuxenliv då hon talade till en publik av tonårsflickor fann hon att det krävdes en hel del mod för att stå inför en sådan massa rastlösa och förhoppningsfulla unga kroppar med bibehållet sinneslugn och självförtroende, och stackars miss Simmons hade inte haft mycket av någondera. Men då gjorde Kate som de andra flickorna, trotsade reglerna, pratade under lektionerna, fick dåliga betyg i det mesta […] och fann sig vid sexton års ålder stå där utan någon egentlig uppfattning om vad hon hade lust att göra.

Although the most common function of highlighted initial adverbials in the corpus is to signal temporal or spatial shifts in discourse, other reasons are not uncommon. Various modal and circumstantial adverbials can act as thematic frameworks such as speaker attitude, viewpoints, reasons, conditions, purpose etc (cf. Downing, 1991). Unlike temporal and spatial adverbials they rarely take part in a TSC, but in contrastive contexts they can sometimes be seen to signal significant shifts in the progression of discourse. Two examples will suffice to illustrate this: 12. Securing titles to the land was theoretically possible, because the government operates various programmes as a means of providing land for the landless. In reality, however, the process is a nightmare of applications, appeals and procedures which can take years. (LT1)

13. The Bolsheviks saw the Civil War from the beginning as part of a great Allied plot. In reality, the revolt of the Czechoslovak

Det var teoretiskt möjligt att säkra jorden, eftersom staten har olika program för att de jordlösa ska få mark. I praktiken är det dock en mardrömslik uppgift med en rad ansökningar och papperskvarnar som kan ta år.

Redan från början betraktade bolsjevikerna inbördeskriget som en del av en stor allierad komplott. I själva verket var det inte de

22 Pragmatic Markers in Contrast Legion had been prompted not by the Allies but by fears for its own survival after attempts by Leon Trotsky, now Commissar for War, to disarm it. To Lenin and Sovnarkom, however, it seemed evident that the Czechs were the tools of 'the AngloFrench stockbrokers'. (CAOG1)

allierade som hade framkallat den tjeckiska legionens revolt, utan dess oro för den egna överlevnaden sedan Lev Trotskij, nu folkkommissarie för krigsärenden, försökt avväpna legionen. För Lenin och sovnarkom tycktes det emellertid uppenbart att tjeckerna var redskap åt "de engelsk-franska börsmäklarna".

In (12) In reality contrasts with theoretically in the preceding sentence. It thus creates a ‘double’ contrast which is emphasized by the parenthetic connector. In (13) the initial adverbial expresses a viewpoint which is linked to, and partly contrasts with, other perspectives expressed in the preceding text (The Bolsheviks, In reality). The examples presented so far have all contained an initial adverbial. However, equally common in the corpus are cases in which a subject is set off by a following contrastive connector. In most of these cases the highlighting of the subject is clearly discourse-related. Sometimes the subject signals a break in a dominant ‘topic’ continuity, i.e. a line of development marked by a sequence of animate or inanimate participants in initial clause position. Two examples illustrating this are: 14. Mars and Venus both had atmospheres dominated by carbon dioxide, with only small proportions of oxygen and nitrogen. More important, both had atmospheres close to the chemical equilibrium state; if you took a volume of air from either of those planets, heated it to incandescence in the presence of a representative sample of rocks from the surface, and then allowed it to cool slowly, there would be little or no change in composition after the experiment. The Earth, by contrast, has an atmosphere dominated by nitrogen and oxygen. (JL1)

Både Mars och Venus atmosfärer dominerades av koldioxid. Bara en liten del av deras atmosfärer utgjordes av syre och kväve. Vad som var ännu viktigare var att de båda planeternas atmosfärer bestod av en gasblandning vars sammansättning var nära kemisk jämvikt. Om man tog en del av planeternas atmosfärer tillsammans med en typisk del av det material dess yta bestod av och hettade upp alltsammans till glödgning, då skulle proportionerna mellan de olika gaserna vara ungefär densamma som från början när det hela hade svalnat av. Jorden å andra sidan har en atmosfär som domineras av kvävgas och syrgas.

15. There are two human types, basically, people I grund och botten finns det två who bottle their emotions up and people who let it all come roaring out. Introverts and extraverts if you prefer. Introverts, as is well known, tend to internalise their emotions, their rage and their self-contempt, and this internalisation, it is equally well known, produces cancer. Extraverts, on the other hand, let joyous rip, rage at the world, divert

människotyper, folk som lägger lock på sina känslor och folk som låter allting välla ut. Introverta och extroverta, om ni så vill. Som bekant lutar introverta åt att stänga inne sina känslor, sin vrede och sitt självförakt, och just den där internaliseringen – återigen ett välbekant faktum – framkallar cancer. Extroverta å andra sidan öser på för fullt,

Adverbial Connectors in Second Position 23 their self-contempt on to others, and this over-exertion, by logical process, causes heart attacks. (JB1)

tömmer sin vredes skålar över omvärlden, låter sitt självförakt gå ut över andra; det är bara logiskt att den sortens överspel leder till hjärtattacker.

As these examples show, the highlighted subjects are part of a dominant TSC and separately contrasted with a preceding member of this continuity. The result is a double contrast which is emphasized by the parenthetical connector. The introduction of a new contrastive topical theme is usually highlighted in this way, whether it participates in a dominant topic continuity or not. An example of a contrast that is not part of a continuity is (16) where Few contrast with everyone: 16. He took out his cheque book. His credit card Hans id-kort skulle inte behövas här för att would not be needed here to back the cheque, for everyone knew him, this was where he had his account; he had already caught the eye of one of the cashiers and said good morning. Few, however, knew his Christian name. (RR1)

styrka checken, eftersom alla kände honom; det var här han hade sitt konto. Han hade redan mött blicken från en av kassörskorna och hälsat godmorgon. Men det var inte många som kände till hans förnamn.

Although the highlighted subject normally presents new or contrastive information, this is not always the case: 17. For this reason, the genre of myth has never been entirely abandoned, although we are inclined to think of it as archaic. If it dwindled to the children's bedtime tale in some societies, in parts of the world protected by forests or deserts from international megaculture it has continued, alive, to offer art as a system of mediation between the individual and being. And it has made a whirling comeback out of Space, an Icarus in the avatar of Batman and his kind, who never fall into the ocean of failure to deal with the gravity forces of life. These new myths, however, do not seek so much to enlighten and provide some sort of answers as to distract, to provide a fantasy escape route for people who no longer want to face even the hazard of answers to the terrors of their existence. (NG2)

Av detta skäl har myten som genre aldrig helt övergivits, trots att vi har en benägenhet att betrakta den som arkaisk. Om den i en del samhällen har reducerats till godnattsagor för barn, har den i andra delar av världen, vilka skyddas ifrån den internationella megakulturen av skogar och öknar, fortsatt att leva och erbjuda konsten som ett sätt att förbinda individen och existensen. Och den har dessutom gjort en virvlande come-back utifrån Rymden, en Ikaros inkarnerad i Läderlappen och hans gelikar som aldrig faller ner i misslyckandets hav för att ta itu med livets tyngdlagar. Dessa nya myter försöker emellertid inte så mycket upplysa och komma med något slags svar som de vill roa, erbjuda en fantasins flyktväg för människor som inte längre ens vill utsätta sig för risken att få svar på sin tillvaros fasor.

24 Pragmatic Markers in Contrast Here the subject is clearly given (referring back to myths that have made a ‘whirling comeback’) but it nevertheless represents a referential shift in the topic continuity initiated by the genre of myth and continued by it in the intervening text. Occasionally even a personal pronoun may be highlighted: 18. Progress on Agenda 2000 I do not need to remind the President-inOffice of how essential it is to meet the timetable, not only to secure the necessary democratic legitimacy for the Agenda 2000 package but also to allow programmes and policies to be implemented in a timely and effective manner. We welcome his efforts in this respect. We, however, need to make progress on our priorities. Parliament's priorities on the general framework regulation are, in fact, clear. (EMCC1)

Framstegen för Agenda 2000 Jag behöver inte påminna rådets ordförande om hur viktigt det är att tidsplanen kan hållas, inte bara för att säkerställa det nödvändiga demokratiska berättigandet av Agenda 2000paketet utan även för att låta alla program och politiska planer genomföras i tid och på ett effektivt sätt. Vi välkomnar hans insatser i det avseendet. Vi behöver dock göra framsteg i våra prioriteringar. Parlamentets prioriteringar med avseende på de allmänna rambestämmelserna är faktiskt tydliga.

In this extract from a speech in the European Parliament there is a clear contrast between the priorities of the Council (represented by the President-in-Office and his efforts) and those of the Parliament (we). We is a given participant (and can even be regarded as being part of a topic continuity represented by I … we … we) but is nevertheless highlighted because of the double contrast. In other words, the subject does not have to convey new information to be highlighted. An example of a slightly different kind is the following: 19. These improvements in safety have been achieved particularly by the use of on-site assistance teams, supported by the PHARE and TACIS programmes. These have resulted in a radical change in safety culture. I would note that this change in safety culture and the increased operational safety cannot be measured in direct monetary terms. The fact that these benefits cannot be readily measured, however, does not make them any less real. (EADA1)

Dessa säkerhetsförbättringar har särskilt uppnåtts genom det arbete som har utförts av assisterande grupper på plats, stödda av Phare- och Tacis-programmen. Detta har resulterat i en radikal förändring i säkerhetskulturen. Jag måste säga att denna förändring i säkerhetskulturen och den förbättrade driftsäkerheten inte direkt kan mätas i pengar. Det faktum att dessa säkerhetsvinster inte låter sig mätas på ett enkelt sätt, gör dem dock inte mindre verkliga.

Here the subject is not contrastive; instead it sums up the information in the previous sentence and so provides a cohesive link with the preceding text. This anaphoric ‘chaining’ strategy (see Smits, 2002: 129) is not very common in the corpus. Occasionally a fronted complement is highlighted by a contrastive connector:

Adverbial Connectors in Second Position 25 20.

Stora Enso is, by capacity, the world's second largest forest products company, holding a leading position, in many of its core businesses. More important than the size of the company, however, are our strategic objectives. (STO1)

I fråga om tillverkningskapacitet är Stora Enso världens näst största skogsindustrikoncern med en ledande ställning inom många av sina kärnverksamheter. Viktigare än koncernens storlek är dock dess strategiska inriktning.

Here a comparative attribute of the subject has been fronted as a marked theme. The construction achieves two things: the comparative link with the preceding sentence is emphasized and the subject is given end-focus as the most important part of the message. 5.2

Exemplifying connectors

Two exemplifying connectors have a tendency to highlight a clause-initial element in the corpus: for example and for instance. The initial element is either a subject or an adverbial. The adverbial is usually a temporal or spatial phrase or clause. The highlighted element, especially when it functions as subject, typically represents an instance (hyponym) of a larger (superordinate) set mentioned in the preceding context: 21. The Scots have a well deserved reputation for Skottarna har ett välförtjänt gott rykte för innovation and inventiveness. In my particular part of Scotland, the North-East, we have many businesses, enterprises and academic and research institutions working on projects and ideas that are very worthy of dissemination both throughout Europe and in the wider world. Dundee, for example, one of the two cities in my area, is well and truly becoming a centre of excellence in life sciences. Aberdeen is acknowledged as the energy capital of Europe. (EHUD1)

22. When Plato died, he was working on the Laws, his bleak, misanthropic proposal for a totalitarian society in which all members were prisoners of a rigid orthodoxy that would not tolerate thinkers like himself and Socrates and in which the penalties for varieties of common transgressions were unmerciful. For a first offense of impiety, for example, the punishment was five years in prison. For a second, it was death without burial. (JH1)

uppfinningsrikedom och nyskapande. I den del av Skottland jag kommer ifrån, den nordöstra delen, har vi mycket affärsverksamhet och många företag och akademiska och forskningsinstitutioner som arbetar med projekt och idéer som är väl värda att spridas både över Europa och i övriga världen. Dundee t.ex., en av de två städerna i mitt område, håller verkligen på att bli ett utmärkt center för biovetenskap. Aberdeen är känt för att vara Europas energihuvudstad. När Platon dog arbetade han på "Lagarna", sitt ödsligt misantropiska förslag till ett totalitärt samhälle där alla medborgare var fångar under en sträng ortodoxi som inte skulle tolerera tänkare sådana som han själv och Sokrates och där straffen för olika slag av vanliga överträdelser var skoningslösa. För en första ogudaktighetssynd var straffet till exempel fem års fängelse. För en andra var det döden utan begravning.

26 Pragmatic Markers in Contrast In both these examples the highlighted element can also be seen as initiating a TSC (for another example, see the function of in Britain in (9)). In most cases, however, the highlighted element is not part of a dominant TSC. What triggers the presentation of an example with its highlighted initial element is often a general statement in the preceding context – a claim that raises expectations of a clarification or illustration. An example is presented, prefaced by a new topical theme (the subject) or a circumstantial framework (an adverbial) specifying the conditions under which the example is valid. Whether the highlighted element is part of a dominant TSC or not is immaterial, but it typically indicates a thematic shift of some kind: 23. How did I get on to that? Oh yes, everyone else Hur kom jag in på det här? Javisst ja: det around here has changed their name. It’s true, and it’s quite a thought, isn’t it? Gillian, for example, she changed her name when she married me. Her maiden name was Wyatt, but now she’s called Hughes. I don’t flatter myself that she was eager to take my name. I think it was more that she wanted to get rid of Wyatt. (JB1)

var ju detta med att man byter namn, lite till mans! Det gör "man" minsann, vilket kan få en att fundera, inte sant? Gillian, till exempel, hon bytte namn när hon gifte sig med mig. Hennes flicknamn var Wyatt, men nu heter hon Hughes. Jag inbillar mig inte att hon hyste någon större iver att få mitt namn. Jag tror snarare att hon ville bli av med Wyatt.

24. What now? He considered going to his sister's, Vad skulle han nu ta sig till? Han but she probably wouldn’t want Edward either. To tell the truth, this wasn’t the first time there had been complaints. Last week, for instance, Macon's brother Charles had stopped by to borrow a router, and Edward had darted in a complete circle around his feet, taking furious little nibbles of his trouser cuffs. Charles was so astonished that he just turned his head slowly, gaping down. "What’s got into him?" he asked. "He never used to do this." Then when Macon grabbed his collar, Edward had snarled. (AT1)

funderade på att fara till sin syster, men hon ville antagligen inte heller ha Edward. Sanningen att säga var det inte första gången någon klagade. Förra veckan till exempel, när Macons bror Charles tittade in för att låna en grep, for Edward runt i en cirkel runt fötterna på honom och nafsade ilsket i hans byxben. Charles blev så häpen att han bara stirrade. "Vad har tagit åt honom?" frågade han. "Så här brukar han aldrig göra." När Macon högg tag i halsbandet på Edward morrade han. Han drog upp överläppen och morrade.

In (23) the left-dislocated subject Gillian is heralded by the superordinate term everyone else in the preceding clause and the main reason for highlighting it is that it introduces a new theme which does two things: it specifies a subtype of the superordinate term and it introduces a new topical theme. (It should be observed that Gillian does not present new information here – she is one of the main characters of the story – but rather a referential shift and a new topical theme.) In (24), which describes the increasingly alarming behaviour of a dog (called Edward), the ‘example’ is heralded by the immediately preceding context (this wasn’t the first time …). The highlighted time adverbial satisfies the expectations raised by

Adverbial Connectors in Second Position 27 the context and emphasizes the temporal shift (and perhaps the recent nature of the event: ‘only last week’). The most important reason for highlighting the initial element thus seems to be that it either introduces a new topical theme (in the case of subjects) or a new circumstantial theme (in the case of adverbials). As a result the initial element normally conveys new information. However, there are exceptions. One example of this has already been given in (23); another is: 25. Macon went over to his study window. […] Up in one of the oak trees, a bird sang what sounded like the first three notes of "My Little Gypsy Sweetheart." "Slum... ber... on..." it sang. Macon wondered if even this moment would become, one day, something he looked back upon wistfully. He couldn’t imagine it; he couldn’t think of any period bleaker than this in all his life, but he'd noticed how time had a way of coloring things. That bird, for instance, had such a pure, sweet, piercing voice. (AT1)

Macon gick fram till fönstret i sitt arbetsrum. […] Högt uppe i en av ekarna sjöng en fågel någonting som påminde om de tre första tonerna i "My Little Gypsy Sweetheart". Macon undrade om även detta ögonblick en dag skulle bli någonting som han såg tillbaka på med längtan i sinnet. Det trodde han inte; han kunde inte minnas att han någonsin hade genomlevt en lika dyster period, men han hade ju lagt märke till att tiden sätter färg på saker och ting . Den där fågeln hade till exempel en så ren, ljuv, genomträngande stämma.

Here That bird does not introduce a new referent, nor does it present a subtype of a previously mentioned superordinate term. Instead, it is anaphoric, referring back to a bird four sentences earlier in the text. What it does is to reactivate this referent after several intervening sentences containing another topical theme. This reintroduction of an earlier participant as the clausal theme (which also represents a break after a local topical continuity) is sufficient reason for highlighting it. 5.3

Additive connectors

As we have seen, unlike contrastive and exemplifying connectors, additive connectors such as moreover, furthermore, in addition and what is more are rarely placed parenthetically in second initial position (3%).5 In fact, only one example occurs in the original English texts in the corpus: 26. The character of the environmental debate has changed dramatically in the past 30 5

Miljödebatten har avsevärt ändrat karaktär under de senaste 30 åren. I näringslivet har

The strong tendency to place additive connectors in clause-initial position in the ESPC is confirmed in the spoken and written material of the British ICE-GB corpus. An examination of the three connectors furthermore, moreover and in addition shows that they are all overwhelmingly placed in initial position: in addition in 97% of the cases (36 out of 37), furthermore in 96% (22 out of 23) and moreover in 86% (18 out of 21). In addition is never used in second position, furthermore only once (after a subject) and moreover three times (once after an adverbial, twice after a subject).


years. Business has moved from seeing environmental issues as a nuisance to a recognition that our common future depends on our ability to achieve sustainability, in environmental, economic, and social terms. At ABB, furthermore, we understand that companies with our geographic, business and technology scope can and must contribute to achieving real progress on sustainable development. (ABB1)

man övergått från att betrakta miljöfrågorna som ett nödvändigt ont till insikt om att vår gemensamma framtid beror på vår förmåga att uppnå hållbar utveckling miljömässigt, ekonomiskt och socialt. Inom ABB är vi dessutom medvetna om att företag med vår geografiska räckvidd och affärsmässiga och tekniska omfattning kan och måste bidra till att uppnå verkliga framsteg för denna utveckling.

Here the connector highlights an initial spatial adverbial. This adverbial, At ABB (a multinational company), can be regarded as a subset (though in adverbial form) of Business, the topical theme of the previous sentence. It announces that the information added in the rest of the clause does not apply to business in general but to a specific company. By highlighting the adverbial in initial position this referential shift from a general (topical) theme to a more specific (spatial) theme is emphasized. Hence, the parenthetic connector is textually justified. This becomes apparent if we reverse the order of the first two elements: 26a. Furthermore, at ABB we understand that companies with our geographic, business and technology scope can and must contribute to achieving real progress on sustainable

development. Now at ABB is within the scope of Furthermore, i.e. the entire proposition following the connector is presented as an addition to the information in the previous sentence. The narrowed relevance of the addition is lost and there is a risk that the entire clause is interpreted as just another piece of information about Business. However, it is the order presented in (26a), with the connector preceding the adverbial (stepwise orientation), that is by far the most common in the corpus. Out of ten examples with furthermore in pre-subject position, nine begin with the connector. An example is: 27. When attacked by a leopard, for instance, all the dominant males [i.e. baboons] rush towards the attacker and together are able to intimidate it. Singly, the leopard would defeat them, but together they can drive it away. They can also help one another to defeat an attack by rival male baboons that are attempting to take over their shared females. Furthermore, if one of the females becomes too dominant they can pool their male resources to put her in her place. (DM1)

När babianer anfalls av t.ex. en leopard, störtar alla de dominerande hanarna mot angriparen och tillsammans förmår de avskräcka den. Leoparden skulle klara en ensam babian, men gruppen av babianer kan köra bort leoparden. De kan också hjälpa varandra att avvärja en attack från en rivaliserande babianhane som försöker ta över deras gemensamma honor. Om någon av honorna blir för dominerande, kan de dessutom med gemensamma hanliga resurser sätta henne på plats.

Adverbial Connectors in Second Position 29 Here it is equally difficult to imagine a reversal of the two initial elements. The conditional clause acts as a circumstantial framework limiting the relevance of the following matrix clause and it is this combined information that is added to the account of the cooperative abilities of male baboons. In other words, the initial sequence in (27) is a good example of stepwise orientation, in which the initial connector takes scope over the rest of the sentence. If the adverbial had been highlighted in initial position this would have falsely suggested some textual relevance of the condition, e.g. a significant shift from another type of condition or from another type of framework. With the exception of (26), all the additive examples containing an adverbial in presubject position have the connector in initial position. This tendency is so strong that it even counteracts the otherwise strong tendency of temporal shifts to be highlighted by a parenthetical connector: 28. Since autumn we have been implementing our redefined strategy. As outlined in the strategy, we have continued to sell non-core assets. In 1999 we released capital in total more than EUR 200 million. In addition in January 2000 a letter of intent was signed to sell the main part of the Group's power assets outside mills for EUR 1,850 million. (STO1)

Sedan hösten 1999 har den modifierade strategin implementerats. I enlighet med strategin har vi fortsatt att avyttra ickekärnverksamheter. Under 1999 frigjordes kapital motsvarande mer än 200 MEUR. Vidare undertecknades i januari 2000 en avsiktsförklaring avseende försäljning av huvuddelen av koncernens krafttillgångar utanför tillverkningsenheterna för 1.850 MEUR.

Here the temporal shift is presented as textually insignificant; what is of greater importance is the enumeration of the effects of the company’s strategy to ‘sell non-core assets’. There is no example in which an additive connector sets off a preceding subject, but 62 examples of the reverse order, with an initial connector. The latter order is preferred even when the subject conveys new information or represents a shift in a TSC. The following example from another company report contains an enumeration of the successes of the company during the past year. The enumeration takes the form of a TSC with varying topical themes. Despite these shifts each clause is added as a unit to the list, and the last item introduced by In addition is no exception: 29. 1999 was the first year for the new Stora Enso. The start of the new company could hardly have been more successful. The difficult merger process went quite smoothly, the productivity programmes continued successfully and the expected synergies were greatly surpassed. In addition strong market conditions contributed to a very strong result for Stora Enso in its first year as a unified company. (STO1)

1999 var första året för det nya Stora Enso. Den nya koncernens start kunde knappast ha blivit mera fram gångsrik. Den svåra samgåendeprocessen genomfördes tämligen smidigt, produktivitetsprogrammen fortlöpte lyckosamt och förväntade synergier överträffades betydligt. Därtill bidrog gynnsamma marknadsvillkor till ett mycket starkt resultat för Stora Enso det första året som ett enat företag.

30 Pragmatic Markers in Contrast We can conclude, then, that unlike contrastive and exemplifying connectors, additive connectors have a strong tendency to be placed in clause-initial position irrespective of the context. Whether this tendency to take the entire following clause in their scope is due to their ‘additive’ function or to a register-specific property of the connectors examined in the corpus is difficult to say without a detailed examination of various additive alternatives (e.g. the parenthetical use of too). The parallel distribution in the ICE-GB corpus (see note 5) indicates that the tendency is a general feature of the kind of additive connectors examined here.

6.

THE POSITION OF SWEDISH CONNECTORS

Like English connectors, most Swedish adverbial connectors have a variable position in the clause but, unlike their English counterparts, they are generally placed in the ‘middle field’ after the finite verb in declarative main clauses (cf. Teleman et al., 1999: 124).

Table 5. Positional tendencies of Swedish connectors (%)

Connector därtill å andra sidan tvärtom dessutom vidare för övrigt däremot framför allt i så fall därför sålunda ändå förresten dock trots allt alltså således exempelvis i alla fall emellertid till exempel (t.ex.) nämligen Total

I 87 81 66 56 54 50 48 48 46 45 41 34 30 18 14 12 8 8 6 3 3 0 29

I2 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

iM 0 0 0 0 0 0 4 0 0 0 0 0 0 0 4 0 0 0 0 0 1 0 0

M 4 15 13 30 31 30 31 31 42 41 36 38 32 67 54 63 69 60 36 67 63 44 47

iE 9 4 6 14 15 10 14 21 4 13 18 19 17 26 11 22 21 32 23 29 25 19 19

E 0 0 16 1 0 10 0 0 8 0 5 9 21 2 18 2 3 0 34 1 8 3 4

Total (n) 23 27 32 181 71 20 99 58 24 422 22 234 53 323 28 169 39 25 96 202 75 66 2289

This difference is seen clearly in Table 5, which shows the positional tendencies of the most common connectors in the Swedish original texts, sorted according to initial (I) position.6 6 The list is restricted to connectors occurring at least 20 times in the Swedish original texts. It should be added that when a Swedish clause begins with an element other than the subject, subject-verb inversion takes place, i.e.


Only a handful of connectors favour clause-initial position, and the most common position is medial (M). The difference between the two languages is displayed graphically in Figure 2.

Figure 2. Positions of English and Swedish connectors (%)

The difference between the two languages is also clearly noticeable in the Swedish translations where initial English connectors tend to drift into the middle field in the Swedish versions. Many illustrations of this tendency are given in section 5 above, e.g. in examples (9), (10) and (27). A more general picture of this drift is presented in Figure 3, which shows the distribution of initial clause elements (adverbials, subjects and connectors) in the English original texts (EO) and in the Swedish translations of these (ST). As clause-initial English connectors are moved to middle position in the Swedish translations, the proportion of initial adverbials and subjects increases in the Swedish texts.

Figure 3. Initial elements in EO and ST (%)

the subject is placed after the finite verb. In such cases a connector may be placed between these two elements, as in: Idag är emellertid vädret bättre (lit. ‘Today is however the weather better’). This position has been classified as medial (M).

32 Pragmatic Markers in Contrast It is also instructive to look at the positional distribution of the most common semantic types of connectors in the Swedish original texts. As shown in Table 6, the tendency of the Swedish types to occur in clause-initial position is broadly similar to that of the English types (cf. Table 2). Additive and resultive connectors have the strongest tendency to be placed initially, while that of contrastive and, especially, exemplifying connectors is much weaker. The main difference between the English and Swedish types is that the latter are generally less common initially and that exemplifying connectors (till exempel, exempelvis) are strikingly rare in that position. However, the scalar pattern again indicates that there is a connection between initial position and semantic type of linkage. There are also certain similarities between the two languages, especially in their tendency to signal ‘addition’ early in a clause.

Table 6. Position of Swedish connectors by semantic type (%)

I

I2

iM

M

iE

E

Additive Resultive Concessive Contrastive Exemplifying

Type

56 35 25 17 4

0 0 0 0 0

0 0 0 0 1

29 48 39 57 62

15 16 20 23 27

1 1 16 2 6

Total (n) 353 676 358 683 100

Total

29

0

0

47

19

4

2289

What is of particular interest here, however, are the Swedish correspondences of English connectors in second position. Because of the V2-constraint, this position is normally occupied by the finite verb in Swedish declarative main clauses and, theoretically speaking, connectors should not be possible here. As demonstrated in Table 5 and Figure 1, this is also largely true. Swedish connectors are only exceptionally placed at I2 (after an initial adverbial) or iM (after an initial subject or a fronted object or complement). However, the exceptions are interesting. As shown in Table 7, a few connectors occur occasionally in second position in the Swedish original texts and translations. Interestingly, these are of the same semantic types – contrastive and exemplifying – as the connectors that tend to occupy the same position in English. Some examples from the Swedish original texts are:7

7

Däremot is seldom set off by commas in the Swedish texts but prosodically it has the same characteristics as parenthetical English connectors.

Adverbial Connectors in Second Position 33 Table 7. Parenthetical connectors in second position in the Swedish texts

Connector

Swedish originals

Swedish translations

9 1 0 0 0

4 5 2 2 1

10

14

däremot ‘however’ till exempel ‘for example’ å andra sidan ‘on the other hand’ å sin sida ‘on the other hand’ i sin tur ‘in turn’

Total

30. Med hakorna djupt nedborrade mot brösten, med tårade svidande ögon kämpade sig några påpälsade fotgängare fram i den ylande motvinden. Steg för steg med blicken tjurigt fästad på den glattslipade gatan, bara en skolängd framför tåspetsarna, strävade de vidare och förbannade att de hade tvingats lämna den goda husvärmen. I medvinden däremot flög människorna fram och var tvungna att ta spjärn med klackarna för att bromsa upp farten och inte förlora herraväldet över sig själva. (KOB1)

31. Det var en mycket svår period för den internationella sjöfarten. Åtskilliga rederier försvann, andra rekonstruerades medan somliga inte bara lyckades överleva utan också växa. Den svenska rederinäringen stod under perioden 1870 fram till 1890 i stort sett stilla, den norska rederinäringen däremot expanderade. (TR1)

A few pedestrians in fur coats were struggling along against the wind, their heads bowed and tears streaming from their eyes. Step by step, their eyes glued stubbornly to a spot on the slippery pavement barely a shoe's length in front of their toes, they pressed on, cursing the moment they had been forced to leave their nice, warm homes. Those lucky enough to have the wind behind them, on the other hand, were having to dig in their heels in order to slow down and keep control of themselves.

This was a very difficult time for the international shipping industry. Many shipowners disappeared and others were restructured, but some not only managed to survive but also expanded. Sweden's shipping sector largely stood still during the period from 1870 to 1890, while the Norwegian shipping industry expanded.

32. Samtidigt som svensken är korkad lever han i At the same time as the Swede is 'corked,' he ett av de tekniskt mest avancerade länderna i världen och redan på trettiotalet skrevs den berömda boken om Sveriges fiffiga ekonomi "Sweden — the middle way". Detta går inte ihop. Men vilka nationalklichéer går ihop? Tysken, till exempel, han är både nacksvullen, preussisk officer och virrig professor på en gång. Fransmannen är en

lives in one of the most technically advanced countries in the world; a renowned book, Sweden — The Middle Way, on the country's already cleverly-managed economy, appeared in the 1930s. All this doesn’t hold together. But what national clichés actually do? The German , for example, is at once the swollen-necked Prussian officer and the

34 Pragmatic Markers in Contrast amorös person med tio älskarinnor, samtidigt som han ägnar sig åt ett förkrossande intensivt familjeliv och vars hem är hans borg antagligen i högre grad än engelsmannens. (IU1)

confused professor. The Frenchman is an amorous person with ten mistresses, who also devotes himself to a crushingly intense family life; his home is his castle probably to a higher degree than the Englishman's.

The textual factors determining the highlighting of the initial elements in these examples are the same as in the English texts. In (30) and (31) a contrastive connector sets off an initial element that is in separate contrast with an element in the preceding text (underlined). The subject in (31) also indicates a shift in a topical TSC. The subject in (32) specifies the idea of ‘national character’ suggested by nationalklichéer in the preceding sentence at the same time as it initiates a TSC. In other words, the placement of the connectors in second position is governed by the same textual principles as we have seen operating in English contrastive and exemplifying examples. Examples of connectors placed in second position in the Swedish translations have already been given in section 5 and will not be repeated here. Contrastive connectors (däremot, å andra sidan) in second position are illustrated in (8), (14) and (15), and exemplifying connectors (t.ex. ‘e.g.’, till exempel) in (21), (23) and (24). In all these cases the parenthetical Swedish connector has the same highlighting function as the connector in the English originals. The fact that the Swedish V2-constraint appears to be overruled in these cases suggests that the connector has a special grammatical status when it is placed parenthetically after another element. Teleman et al. (1999: 137, 152 f.) regard connectors in this position as focusing adverbials that are closely associated with, and ‘subordinate’ to, the preceding element, syntactically and prosodically. Because of their subordinate status they do not violate the V2-constraint. This focusing use of certain Swedish connectors is very similar to that of parenthetical English connectors and the Swedish analysis could just as well be applied to the English connectors (cf. Quirk et al., 1985: 632).8 Occasionally the highlighting effect of an English connector in second position can be achieved by placing the corresponding Swedish connector before the textually relevant constituent inside the highlighted element. This narrow focusing is possible with a very mobile connector like till exempel. One example has already been given in (9), where In Britain, for example, is rendered as I till exempel Storbritannien in the Swedish translation. Another similar example is the following:

8 Another case when a Swedish connector is used in apparent violation of the V2-constraint is when it is placed clause-initially without causing inversion: – Ja, det är självklart, men jag träffar ju farsan 'Yes, of course, but I see Dad, too. You do that. I också. Det gör man, jag menar dom är ju inte elaka mean, they 're not nasty to me. On the contrary, mot en, tvärtom, min mamma och jag har faktiskt Mum and I get along well together. rätt kul ihop. (MR1) In such cases, however, the connector can be analysed as a ‘free annex’ placed in the grammatical ‘forefield’ outside the syntactic structure of the clause (see Teleman et al., 1999: 452 ff.).


33. Young adult males are driven out of this group and form separate bachelor parties that must bide their time until they can steal young females and set up an oligarchy of their own. The advantage of this arrangement is that it provides an efficient means of defence against predators. When attacked by a leopard, for instance, all the dominant males rush towards the attacker and together are able to intimidate it. (DM1)

Unga vuxna hanar körs bort ur denna grupp och bildar separata ungkarlsflockar som måste bida sin tid tills de kan stjäla unga honor och själva skapa en oligarki. Fördelen med denna modell är att den ger ett effektivt försvar mot rovdjur. När babianer anfalls av t.ex. en leopard, störtar alla de dominerande hanarna mot angriparen och tillsammans förmår de avskräcka den.

The fact that the Swedish translators try to reproduce the highlighting effect of parenthetical connectors in the English original texts in one way or another, and that a similar use is occasionally found in the Swedish original texts, underlines the textual importance of the highlighted elements and the need to make them prominent in the progression of discourse. However, it should be emphasized that parenthetical connectors in second position are very rare in the Swedish texts: in the Swedish originals they represent no more than 0.4% of all the examples containing an adverbial connector. Even if we regard these exceptional cases as subordinate focusing devices that do not violate the V2-constraint, they are far less common than the corresponding expressions in the English texts. This indicates that Swedish lacks – or is at least far less ready to use – an important syntactic device for highlighting an initial element in contexts where this is textually effective. Whether this is due to the V2-constraint or not, the fact remains that Swedish normally relies on initial position alone as a means of signalling thematic prominence. This in turn means that the starting-point of the message acquires special importance. In the great majority of the examples in the Swedish original texts and translations this is also sufficient to create a textually appropriate discourse.

7.

CONCLUSION

Inspired by Smits’ (2002) investigation of complex sentence beginnings in English, this study has examined the textual function of adverbial connectors in second position in English and Swedish. The starting-point has been a corpus-based survey of the positional tendencies of frequent connectors in both languages, but the focus has been on the discourse function of the initial element (typically a subject or an adverbial) and the similarities and differences revealed between the two languages. The corpus shows that the connectors of both languages vary greatly in their positional preferences, but also that this variation is patterned and tends to reflect the type of linkage they perform. In both languages additive connectors have a strong tendency to be placed initially, whereas second position is most characteristic of contrastive and exemplifying connectors, especially in English but to some extent also in Swedish. This cross-linguistic similarity is to a large extent discourse-driven: while additive connectors are mainly associated with a ‘rhetorical’ progression of discourse linking entire propositions to the preceding text, contrastive and exemplifying connectors often serve to highlight the initial

36 Pragmatic Markers in Contrast element, syntactically and prosodically, thereby indicating a significant break or shift in the development of discourse. Hence, despite the syntactic differences between English and Swedish, there are interesting parallels between the two languages in the way that connectors adjust to, and signal, the progression of discourse. It hardly needs to be emphasized that the present study has been limited in several respects. For example, it has been restricted to written texts and has focused on certain types of adverbial connectors in second initial position. What is needed is an extensive investigation of the relationship between clause position and discourse function of a wide range of expressions (including interpersonal adverbials like in fact, of course and actually). As Smits has shown, corpora (supplemented by informant experiments of various kinds) can contribute a great deal to our understanding of the discourse-driven function of elements with a variable clause position. Contrastive studies can tell us what is language-specific and what is language-independent in this respect and are therefore especially revealing. Bilingual and multilingual corpora, in particular, open up new and exciting possibilities for such research in the future.

REFERENCES Altenberg, B. (1998). Connectors and sentence openings in English and Swedish. In: Corpora and Cross-linguistic Research: Theory, Method, and Case Studies (S. Johansson and S. Oksefjell, eds), pp. 115–143. Rodopi, Amsterdam. Altenberg, B. (2002). Concessive connectors in English and Swedish. In: Information Structure in a Cross-linguistic Perspective (H. Hasselgård, S. Johansson, B. Behrens and C. Fabricius-Hansen, eds), pp. 21–43. Rodopi, Amsterdam. Altenberg, B. and K. Aijmer (2000). The English-Swedish Parallel Corpus: a resource for contrastive research and translation studies. In: Corpus Linguistics and Linguistic Theory (C. Mair & M. Hundt, eds), pp. 15–33. Rodopi, Amsterdam. Altenberg, B., K. Aijmer and M. Svensson (2001). The English-Swedish Parallel Corpus (ESPC): Manual. Department of English, Lund University. Available at: www.englund.lu.se/research/corpus/corpus/espc.html. Biber, D., S. Johansson, G. Leech, S. Conrad and E. Finegan (1999). Longman Grammar of Spoken and Written English. Longman, Harlow. Downing, A. (1991). An alternative approach to theme: a systemic-functional perspective. Word, 42, 119–143. Enkvist, N. E. (1987). A note towards the definition of text strategy. Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung, 40, 19–27. Fraser, B. (1998). Contrastive discourse markers in English. In: Discourse Markers. Descriptions and Theory (A.H. Jucker and Y. Ziv, eds), pp. 301–326. John Benjamins, Amsterdam. Givón, T. (1983). Topic continuity in discourse: an introduction. In: Topic Continuity in Discourse: a Quantitative Cross-language Study (T. Givón, ed.), pp. 1–42. John Benjamins, Amsterdam. Halliday, M.A.K. (1994). An Introduction to Functional Grammar. 2nd ed. Arnold, London.

Adverbial Connectors in Second Position 37 Johansson, S. (1998). On the role of corpora in cross-linguistic research. In: Corpora and Cross-linguistic Research (S. Johansson and S. Oksefjell, eds), pp. 3–24. Rodopi, Amsterdam. Quirk, R., S. Greenbaum, G. Leech and J. Svartvik (1985). A Comprehensive Grammar of the English Language. Longman, London. Smits, A. (2002). How Writers Begin their Sentences. Complex Beginnings in Native and Learner English. LOT, Utrecht. Teleman, U., S. Hellberg and E. Andersson. (1999). Svenska Akademiens grammatik. Del 4. Norstedt, Stockholm. Virtanen, T. (1992). Discourse Functions of Adverbial Placements in English. Clause-initial Adverbials of Time and Place in Narratives and Procedural Place Descriptions. Åbo Akademi University Press, Åbo.



THE ENGLISH PRAGMATIC MARKER SURELY AND ITS FUNCTIONAL COUNTERPARTS IN SPANISH Angela Downing, Universidad Complutense de Madrid

1.

INTRODUCTION

In its most common use in British English, surely is a pragmatic marker of epistemic stance which indexes the speaker’s attitude towards the proposition in which it occurs, and also to the state of knowledge or intentions of the addressee, as illustrated in examples (1) and (2): 1. Surely you knew? BNC FS8 413-414 2. Surely it is time to make an effort to reduce the drug bills. BNC B7K In cases in which the hearer is directly involved, the speaker’s attitude often implies challenge, disapproval or reproach, which are inferred by the addressee, as can usually be seen from ensuing reponses. Surely functions as a positioning device for redefining common ground between individuals, and by veiling the impositive force of certain speech acts, allows the intended meaning to be conveyed covertly while still conforming with the standards of conventionally polite behaviour. The aim of this empirical study was to ascertain whether there is in Peninsular Spanish a similar short, one-word marker which can be used to convey meta-messages of a social or moral kind, as surely does. My hypothesis was that a marker with all or most of the characteristics and functions of surely is not available in present-day Peninsular Spanish, despite the fact that a number of lexemes exist in Spanish which are cognates of surely. Differences in the semantic evolution of the cognate words account for the mismatches that will be discussed in later sections of this chapter.

2.

DATA AND METHOD

Data for surely were taken from the British National Corpus (BNC). A first general search yielded 6345 occurrences. For an earlier paper (Downing, 2001), using Wordsmith Tools, I made a sub-corpus of concordances of surely co-occurring with each of the personal pronouns, and with unstressed there, all of these functioning as subject of the clause. In each case three sets of concordances were made according to the position of surely within the clause or utterance as initial, medial and final. A statistical analysis of the results can be found in Downing (2001). Stances were found to vary in type according to the position of surely together with the type of subject pronoun. In addition, other contextual factors were important


in suggesting speakers’ rhetorical goals. A representative selection of these concordances has been used for the present study. For the Spanish data I have used the on-line facility of the Corpus de Referencia del Español Actual (CREA) of the Real Academia Española to carry out a qualitative analysis of the meanings and uses of certain Spanish cognates of surely, comparing these with the English marker. Secondly, I discuss two syntactic constructions which partially correspond to some of the nuances expressed by surely. All the texts used were taken from radio discussions, interviews and other interactive programmes recorded between 1991 and 1996.1

3.

A CHARACTERISATION OF SURELY

In comparison with other stance adverbials pertaining to the domain of certainty such as certainly, definitely, obviously, indisputably, assuredly, evidently, among others listed by Biber and Finegan (1989), the meaning of surely is opaque. It is not what it might seem at first sight. While other adverbs in the group such as clearly and obviously are frequently interchangeable within a certain stretch of discourse, surely is hardly ever interchangeable with any of them (but see section 3.2). Nor is it usually paraphraseable by I am sure that, in the way that certainly is paraphraseable by I am certain that.., it is certain that. ‘It is sure that’ is uncommon in British English, presumably because of the subjective nature of ‘sure’ as opposed to ‘certain’. The raised construction as in ‘they are sure to win’ will, however, be taken into account in later discussion. Surely operates covertly in discourse, and as a result, its attitudinal meaning is not recovered directly, but by inference. This contrast between surely and other adverbs of certainty corresponds to what Palmer (1986:64) describes as two different though related types of epistemic modality, one concerned with inference, the other with confidence. Comparing epistemic must with the adverb certainly, Palmer finds that the modal verb indicates that the speaker is inferring from available information, whereas the adverb indicates the degree of confidence the speaker has in what s/he is saying. Similarly, modal will is different from probably. Epistemic will often expresses a judgement or prediction based on known facts or evidence, and in particular, on what is usually the case, as in Coates’ corpus-based example Her mother will know her age, the implication being that mothers generally know their children’s age. The term Assumptive suggested by Palmer (1986:62) captures this kind of epistemic meaning appropriately, in contrast with the Deductive meaning expressed by must. These differences are relevant in attempting to specify the meaning of surely. While certainly and definitely are markers of considerable epistemic strength, surely, I suggest, is like must in that it expresses deduction or inference from known facts or evidence. And as with must, despite the strength of commitment traditionally associated with it, surely may contain an element of doubt. A basis of assumption rather than deduction is also relevant in certain uses of surely, as in example (5) below. Both must and will co-occur with surely as 1

I wish to express my thanks to Geoffrey Leech, Diana Lewis and the participants and organisers of Contact Forum 2003, for their useful comments and suggestions on earlier versions of this study. Also to Raquel Hidalgo and Laura Hidalgo for their help and discussion of the Spanish illustrations. I am solely responsible for any deficiencies in the final version.

Surely and its Functional Counterparts in Spanish 41 ‘modally harmonic’ forms (Lyons, 1977:807) which are mutually reinforcing (Halliday, 1970:331) or express the same degree of modality (Coates, 1983:45), as illustrated by the following examples from the British National Corpus: 3. ‘Surely you must realise that I couldn’t accept the responsibility…’ BNC CJX 2102 4. Surely you can’t believe that Jane would want to see you?’ BNC FSG (can’t as negation of epistemic must)’ 5. Producing a play without an interval these days is asking for trouble. This they will surely know. BNC K52 5392 6. And surely I’ll hear the lift if it brings my inquisitor up? She knew it was a lie. BNC FPO 1649 The explanation for this opaqueness of surely, which permits a greater degree of dissimulation than is possible with the other ‘certainty’ adverbs, lies in the historical development of surely. In accordance with The Shorter Oxford English Dictionary on Historical Principles, fifth edition (2002), surely is seen to have evolved from its more literal and concrete meanings of ‘safely’, ‘securely’, ‘firmly’, ‘without danger’, through a gradual process of desemanticisation to a more psychological meaning in which it was almost interchangeable with ‘certainly’, ‘assuredly’, ‘confidently’, especially as an intensifier. Finally, through subjectivisation and automatisation, it has evolved towards its present stance and evidential meanings. Subjectivisation is defined by Traugott as a semantic-pragmatic process whereby ‘meanings become increasingly based in the speaker’s subjective belief/state/attitude towards the proposition’ (1995:31). It usually happens in such cases of semantic bleaching that a former meaning may persist alongside the predominant meaning, depending on such factors as position in the clause and contextual environment. This appears to be the case with some uses of medial surely. I will return to these factors presently. In undergoing a process of automatisation (Aijmer, 1986), surely is automatically triggered as a pragmatic response to something said, heard or otherwise perceived in the contextual situation. It is the best possible form available and can express a number of stances ranging from the confrontational to self-questioning, mitigation and agreement-seeking. In a previous paper, (Downing, 2001) I considered surely as an evidential marker in British English in so far as it indexes states of knowledge, not only the speaker’s, but also the addressee’s assumed state of knowledge, at a particular point in a particular discourse. With regard to speaker’s knowledge, the implication conveyed by surely is that the speaker’s opinion is the only true, reasonable or reliable one, an opinion based on the ‘mode of knowing’ of belief (Chafe, 1986), the sources of this belief being experience, sensory perception or moral right. As regards the addressee, surely encodes, adopting Haviland’s (1987:343) view of the moral as well as epistemological character of evidentiality, “what the interlocutor is taken to know, or should know or apparently (perhaps culpably) “fails to know”. To this we can add “what the addressee did, intended, or failed to do”, for utterances with surely also address deeds and intentions.

42 Pragmatic Markers in Contrast This view of evidentiality is close to epistemic and deontic modality. In fact, surely can signal a deontic as well as an epistemic bias, as will be explained shortly. 3.1

Surely as a Biased Question

Surely occurs in declarative clauses, but like negative-interrogative questions (in its main use in British English) it is conducive (Quirk et al, 1985:808; Biber et al, 1999:1113) or biased (Huddleston and Pullum, 2002:1.879). Declaratives which are intended as questions rather than statements have been called ‘queclaratives’ by Geluykens (1987) following Sadock (1974). Concluding his experimental study on rising intonation contours in queclaratives, Geluykens finds that the role of rising intonation in such utterances is overestimated, and is not able to overrule pragmatic cues such as the use of different personal pronouns. Not mentioned or used as pragmatic cues in the test utterances are the lexical ‘confidence markers’ (Huddleston and Pullum, 2002) such as surely, so, then and of course, which lie outside the propositional content of the question. That these undoubtedly reinforce the interpretation of the declarative as a question is confirmed by the frequent presence of a question-mark in printed materials, and the fact they they elicit a response, which is almost always forthcoming. In addition, they are explicit metapragmatic markers (Verschueren, 2000:447), which like other pragmatic markers and discourse markers, index the speaker’s metapragmatic awareness “of how their utterances fit in with preceding or following discourse” (Aijmer and Simon-Vandenbergen, 2004:1783). Among the confidence markers surely is the most heavily biased. A biased question is one “where the speaker is predisposed to accept one particular answer as the right one” (Huddleston and Pullum, 2002:879). We can compare surely in queclaratives with positive and negative-interrogative questions, in order to pin down yet another feature of surely which is not shared by other ‘certainty’ adverbs. Let us take examples (7a) and (7b): 7. (a) (b)

Surely I have the right to know that at least? BNC H7W 077 Surely he doesn’t have a right to know that?

While positive interrogative questions are neutral (Do I have a right to know that? ), that is to say not biased to either a positive or negative answer, positive queclaratives expect an answer with the same propositional content as the question. For instance, a positive queclarative such as (7a) Surely I have a right to know that? expects the positive answer You have a right to know that’. A negative queclarative such as (7b) Surely he doesn’t have a right to know that expects a negative answer (No, he doesn’t have a right to know that.) An attested example of a negative queclarative with initial surely is (8), which expects a negative answer having the illocutionary force of acquiescence such as ‘No, I’m not’: 8. You ain’t putting tomato on lasagne, surely BNC KCT 7057 However, surely, as a pragmatic confidence marker is more complex than so or of course. A positive queclarative with initial surely can be paraphrased by a plain negative-interrogative question, or a plain declarative followed by a question tag, as in:

Surely and its Functional Counterparts in Spanish 43 Surely I have a right to know that = Don’t I have a right to know that? = I have a right to know that, don’t I? A negative queclarative with initial surely can be paraphrased by a negative declarative followed by a question tag: Surely he doesn’t have a right to know that? = He doesn’t have a right to know that, does he? Another example (9) from the BNC illustrates a further nuance of this marker: 9. Surely we can do better for people with mental problems and their families? BNC EA4 895 Here the plain negative-interrogative question Can’t we do better? could serve as a grammatical equivalent to (9). However, with an ‘inclusive we’ pronoun, surely in certain types of context invites collaboration in a joint action or attitude and can function as a politeness marker, at the same time as it indexes the speaker’s dominant stance. Accordingly, a more appropriate equivalent would, I suggest, include an elicitation of the hearer’s opinion, expressed by Don’t you think / agree etc. We can compare: Surely we can do better

= Can’t we do better = Don’t you think/agree that we can do better?

In all these semi-equivalents, the nuance of dominant stance is much less evident than when surely is present. Like negative-interrogative questions, surely, particularly when initial, is based on contradictory assumptions, a fact which explains the element of doubt referred to previously. These assumptions are distinguished by Quirk et al (1985:808) as the ‘old expectation’ and the ‘new expectation’ respectively. That is to say, a positive declarative marked by surely has a positive bias but also a negative bias. So, for example, in Surely I have a right to know that ‘I have a right to know that’ is the old assumption and is positive, but evidence now suggests a negative assumption that I do not have a right to know that. According to this view, the old expectation tends to be associated with the speaker’s hopes and wishes, while the new expectation disowns this. Consequently, such negatively oriented questions often express disappointment or annoyance (Quirk et al,1985:808); together with, I would add, an implicit criticism or reproach directed covertly to the addressee if s/he is involved. Biber et al explicitly claim that negative questions disown a negative expectation and embrace a positive one (1999:1114). With negative queclaratives reinforced by surely, however, a negative response is not disowned, since it aligns with the speaker’s expectation, as seen in discussing (7b). We shall see in section 7.2 an instance of this taking place in the Spanish data, related to the illocutionary force of the utterance and at the same time to the speaker’s rhetorical goals. Surely can illustrate the three types of bias distinguished by Huddleston and Pullum (2002:879):

44 Pragmatic Markers in Contrast a) Epistemic bias: With this type the speaker knows or expects that one answer is the right one, as in (10): 10. ..and indeed there is only a metre between the dwellings and the adjoining boundaries. Surely there is a need for screening, it is in a very very prominent position. BNC JTD 279 The expected answer would be ‘yes, there is a need for screening.’ b) Deontic bias. With this the speaker judges that one answer ought to be right, as in (11): 11. You remember her surely? I know I do. BNC HOM 984-985 (= You ought to remember her). c) Desiderative bias occurs when the speaker wants one answer to be the right one, as in Surely she liked it?, when referring to a present chosen by the speaker. However, as surely is based on contradictory assumptions, one bias overlays another. In (12), for instance, a desiderative bias, wishing to believe he is alive, prevails for a moment over epistemic bias, here the knowledge or strong expectation, from the evidence of the letter, that he is dead: d) Desiderative-epistemic 12. Yet I think I knew, at that moment, that he was dead. That fatal happiness! But hope, desperate and defiant, sprang up at once. He might be a prisoner of war. Yes, surely he was a prisoner of war! I read the telegram in disbelief. BNC AMC 1341 Similarly, a deontic bias expecting the answer ‘Yes, we should toughen this up’ is at odds with an epistemic bias ‘it appears that you have a different idea’ in (13): e) Deontic-epistemic 13. I mean, it’s acceptable to whom, for heaven’s sake? Surely we should toughen this up a bit and say, I mean….BNC KN3 239 Contradictory assumptions are inherent to surely and are not overt but covert. They mark stance surreptitiously. The other ‘certainty’ adverbs such as certainly, evidently and indubitably are not based on contradictory assumptions and for this reason are not interchangeable with initial surely. 3.2

Medial Surely

The interactive meaning of medial surely is more difficult to pin down than that of initial and final surely, not least because in medial position it functions as an intensifier. As is pointed out by Quirk et al (1985:593), referring to intensifiers in general, ‘When positioned next to part of the communication, without being separated intonationally or by punctuation, their effect is often to emphasise that part alone, although there may be ambivalence as to whether

Surely and its Functional Counterparts in Spanish 45 the emphasis is on the part or the whole.’ Whereas surely in initial position is tonic and may be slightly detached from the main utterance, occupying a whole tone unit, in medial position it is attested in the London-Lund Corpus as carrying heavy stress but forming part of the tone unit in which it occurs (Downing, 2001). Examples from the BNC, which is not transcribed for prosodic information, suggest that the same pattern could apply to the whole or part immediately following surely, namely surely be back, surely know and surely wrong, respectively: 14. He is now undergoing a recouping, re-generating process. He will surely be back. BNC AAN 73-74 15. Producing a play without an interval these days is asking for trouble. This they will surely know. BNC K52 5392 16. It is surely wrong to draw a line between the right of another to live or die. BNC ANA 899-900 More importantly for this study, the clause in which medial surely occurs is not functioning as a queclarative. If we transfer surely to initial position in each case, the difference immediately becomes clear. Surely he will be back, surely they will know this and surely it is wrong are queclaratives implying contradictory assumptions, if only for the sake of argument, for surely can be used disingenuouosly, and the speaker’s real belief may be less confident than that overtly expressed. Medial surely is different. These attested utterances are not paraphraseable as ‘Won’t he be back? Won’t they know? Isn’t it wrong?’ Instead they make a fairly confident prediction regarding the future, a fairly confident assumption regarding the present, and a reinforced deontic statement, respectively. They would still do so, however, without the addition of surely; in fact the prediction and statement would be stronger without the adverb. The function of surely, then, in such cases is to heighten the rhetorical strength of the assertion, for surely is used as a rhetorical strategy. Medial surely does this, I claim, because it represents the penultimate stage of the development of surely, namely as a rough equivalent of certainly (I would suggest ‘almost certainly’ in most cases). This meaning has persisted in medial position. It can be paraphrased as ‘I am sure’, in the case of the first two examples, and is replaceable by undoubtedly in the third (‘undoubtedly wrong’). It is this meaning and use of surely that appears to be the most common use in present-day American English. In final position surely is an epistemic parenthetical prosodically often in the ‘tail’ of the previous intonation unit and may be reduced phonologically. Like initial surely, final surely makes for a biased question, based on contradictory assumptions, as in (17), 17. ‘Oh come on, you’ve got time for a cup of tea, surely?’BNC H9V 1797 where the assumptions are ‘I think you have /should have time for a cup of tea, but it appears that you haven’t.’


4.

CHARACTERISTICS OF PRAGMATIC MARKERS

In considering surely and its cognates as pragmatic markers it is useful to characterise pragmatic markers by a number of common features. (All conditions are numbered consecutively here for ease of reference.) Schiffrin (1987: 328) makes the following tentative suggestions as to what conditions allow an expression to be used as a marker (she refers to items such as oh, well, and, but, or, so, because, now, then, y’know and I mean as discourse markers, but recognises that they gain pragmatic effects in discourse) (ibid.). 1. it has to be syntactically detachable from a sentence 2. it has to be commonly used in initial position of an utterance 3. it has to have a range of prosodic contours (e.g. tonic stress and followed by a pause, phonological reduction) 4. it has to be able to operate at both local and global levels of discourse, and on different planes of discourse 5. this means that it has to have no meaning, a vague meaning, or to be reflexive (of the language of the speaker) Jucker and Ziv (1998:3) cite the four conditions for pragmatic markers suggested by Hölker (1991:78): 6. they do not affect the truth conditions of an utterance 7. they do not add anything to the propositional content of an utterance 8. they are related to the speech situation and not to the situation talked about 9. they have an emotive, expressive function rather than a referential, denotative, or cognitive function. Further characteristics adapted from Brinton’s (1996) longer list reordered and abbreviated by Jucker and Ziv (1998:3) and Andersen (2000:21) include the following, among others which are similar to those listed by Schiffrin: 10. they occur outside the syntactic structure or are only loosely attached to it 11. they are optional rather than obligatory features 12. they are features of oral rather than written discourse 13. they are short and (often) phonologically reduced 14. they are stylistically stigmatised 15. they constitute a heterogeneous set of forms which are difficult to place within a traditional word class (including items like ah, actually, and, just, like, now, really, well, I mean, I think and you know) 16. they have little or no propositional meaning. Andersen (2000:40) argues against the importance traditionally given to non-propositionality as an essential property of pragmatic markers, preferring to consider this feature as usual but not essential. More pertinently, “the label ‘pragmatic’ is meant to suggest a relatively low degree of lexical specificity and a high degree of context-sensitivity.” Furthermore, pragmatic markers may indicate speaker attitudes, carry out speech act functions or serve to increase politeness and solidarity among speakers (ibid.).

Surely and its Functional Counterparts in Spanish 47 Surely fulfills the majority of the conditions specified above. It is not, however, stylistically or socially stigmatised as are I mean, like and innit (by some speakers). Nor is surely difficult to assign to a word class. In its main current use surely has been bleached out semantically but it is nevertheless clearly an adverb of stance functioning as an adjunct.

5.

SURELY AS A MARKER OF STANCE

Stance is taken to be, in the words of Du Bois (2000) “the dialogically enacted positioning of a socal agent with respect to alignment, power, knowledge, belief, affect, and other socially salient categories”. To these must be added the stances of a moral nature referred to in the previous section. Surely operates in interactive discourse and is triggered at a particular point in the discourse by the immediate epistemic deduction that, despite assumed common ground, assumptions differ between speaker and hearer. Fracturing of the common ground through dissent with the speaker’s assumptions or beliefs is perceived or imagined as imminent, and surely is triggered as a marker of self-validation. The speaker adopts the position of dominant stance (Brazil, 1995), namely, that s/he knows better than the hearer, or rather, s/he presents things as if s/he believed this, for surely can be used as an argumentative strategy, or in a teasing, jokey way. Surprisal is often a feature of utterances with surely, especially when it occurs in initial position and with appropriate intonation (a high fall-rise or a high rise- fall). Surprisal in the evidentiality literature refers, under the term ‘mirativity’, to the linguistic marking of an utterance as conveying information which is new or unexpected (Delancey, 2001). More exactly, to the linguistic marking of “knowledge based on the immediate personal perceptions of events” (Lazard, 2001:361, emphasis in the original). It is not only verbal exchanges that alert the speaker to present or imagined conflict, but also visual, sensory or inferred stimuli. Surely is an automatic response which manifests itself in stances ranging from the confrontational and challenging to self-questioning, teasing and persuasive, according to context. ‘Stance is therefore not just something we think or feel, an interior state of our being. Stance is something we do: we enact our stance in the public space of dialogic interaction’ (Du Bois, 2000). Stances in interactive discourse articulate alignment and opposition in the positioning of selves relative to others (ibid.). Given the characteristics of surely specified above, opposition is likely to be more common than alignment, with alignment on the part of the hearer being sought by means of persuasion rather than enacted as speaker‘s positioning vis-à-vis a previous speaker. Surely is one of what Haviland calls ‘fighting words’, albeit with a veneer of politeness which enables it to function interactively as a quite subtle weapon in argumentation. It is the contextual factors which upgrade or downgrade the degree of confidence expressed by surely, signalling the speaker’s doubt, hope or expectation in addition to the above-mentioned stances. A wide variety of stances and rhetorical uses are discussed in Downing (2001).


6.

SPANISH COGNATES OF SURELY

The first set of data investigated in the Spanish corpus consisted of lexical items or constructions based on the adjective seguro, a cognate of English sure. a) Seguramente This is the adverb which corresponds morphologically to surely, but which appears to be undergoing epistemic weakening. It is placed as approximating certainty on a scale of doubt in the Gramática Descriptiva de la Lengua Española. Both the Diccionario de la Lengua Española, 22nd edition (2001) (DRAE) and The Oxford Spanish Dictionary (2001) list two senses, the first related to certainty, the second ‘probably’. English equivalents suggested by the latter are ‘I am sure that’ and ‘probably’. A sample from the CREA showed that either equivalent would be valid in many cases, but with the greater relevance of ‘(most) probably’, since an element of doubt seems to be almost always present. This view is reinforced by comparing seguramente with seguro in the same set of data. Seguro invariably expresses greater confidence and rhetorical strength. An example of an informal use of seguramente in the data is provided by (18), a message on an answer-phone: 18. Quedaremos en seguramente el viernes a las cuatro de la tarde, ¿vale? [We’ll meet most probably on Friday at four. Right?] In more formal or less interactive contexts, medial surely is a possible counterpart of seguramente: 19. Pero se debería desde luego seguramente legislar [But it should of course surely be legislated] and, conversely, seguramente is a possible counterpart to medial surely in formal contexts as in He will surely be back and This they will surely know (14) [Volverá seguramente] and (15) [Esto seguramente lo sabrán]. b) The adjective seguro Seguro is used both as a modifier (un lugar seguro, una fecha segura) (a safe place and a definite date, respectively) and predicatively after the verb ser: Ese aeropuerto no es seguro (That airport is not safe) and estar: Esa escalera no está segura (That ladder is not safe/secure) (OSD), in all cases with marking for number, gender and person. It has partially followed a similar trajectory to that of surely. From meaning ‘free from danger or risk’, it too has taken on the psychological meaning of ‘certain’, as in El éxito está lejos de ser seguro (Success is far from certain). Estar seguro/a (de) que…., expresses subjective certainty as in Estoy seguro (de) que ganará ( I’m sure he’ll win). c) The invariant form seguro (que)… This expression is not listed in the 22nd edition of the DRAE (2001), despite its great frequency of use in everyday spoken Peninsular Spanish. One informal equivalent is I bet: Seguro que llegan tarde,’ I bet they’ll be late’. The Spanish expression is not restricted to such a colloquial use, however, and as a result this equivalent is only partially valid. It appears to be evolving as an epistemic marker from ‘es seguro que’ (It is certain that..) as in … y ‘nos dijeron eso: “Llevarse ropa porque es seguro que os la compran” (CREA) [Take

Surely and its Functional Counterparts in Spanish 49 clothes with you as it’s certain that people will buy them], and as such, it warrants further consideration here. Seguro (que) is not, however, opaque, its meaning being that of a confident assertion or prediction, as illustrated in the following examples from CREA. The raised form be sure to and, more informally, I bet are suggested as equivalents: 20. Las imágenes que van a ver ahora, seguro que les llaman la atención. [The pictures you are about to see now are sure to attract your attention] 21. …y tu eras un chico con pelo, y ahora seguro que gastas todo en maquillaje aquí en… [and you used to have plenty of hair, and now I bet you spend everything on cosmetics here at..] 6.1

A Characterisation of Seguro (que) as a Marker of Epistemic Stance

1. Formal aspects a. Its form is invariant, not being marked for person, number or gender. Consequently, it is not analysable in its present day use as an ellipted form of estar seguro (as described in 6). It is more likely to be derived from ‘es seguro que..’. b. Seguro introduces a content clause, with the complementizer que but it is not integrated into any ‘main’ clause structure, since ‘seguro que’ has no subject or verb. c. It appears to be rarely modified by intensifiers, the only instance in the data being that of the attenuative casi (almost), as in ¡Casi seguro que no! (Almost certainly not!) d. Seguro que is commonly used in initial position, in which case its scope extends to the end of the sentence. e. It is syntactically detachable from the sentence. It can occur in final position, without the complementizer, as an epistemic parenthetical, even co-occurring with an initial use, as in (22): 22. …y es un espectáculo para Madrid. Seguro que vamos a triunfar, seguro. [and it’s a show for Madrid. It’s sure to be a success. Sure to.] f. It can occur in elliptical verbless clauses (¡Seguro que sí! ¡Seguro que no!). g. Its grammatical status is therefore not clear. 2. Functional and social aspects a. It has an emotive or expressive function, rather than a referential function. b. It is a feature of oral rather than of written discourse. c. It is not socially stigmatised. d. It is related to the speech situation. 3.Semantic and pragmatic aspects a. Seguro (que) is based on the speaker’s strong subjective belief in the truth of the statement it introduces. It makes a confident assertion, which may be somewhat confrontational in an upfront way, as in English I bet. It does not challenge by implicature, as surely does. It is a

50 Pragmatic Markers in Contrast confidence marker. When what is stated is a prediction, seguro (que) indexes strong confidence in the predicted outcome. Data seem to favour positive rather than negative predicted outcomes, although the latter is attested in one example ‘seguro que fallamos’ [We are sure to fail], referring to a series of international football matches. b. Seguro (que) does not appear to function regularly as a queclarative (biased declarative). That is, it does not appear to signal a contradiction of speaker’s expectations, as surely does, but rather, reinforces the speaker’s expectations. c. This basic meaning can be modified, however, by means of a question tag (¿no? or ¿verdad?) which typically requests confirmation of the content expressed in the clause, as in (23): 23. Porque la semana que viene, seguro que tenemos concurso otra vez, ¿no? Sí, tenemos concurso. [Because next week we are sure to have the quiz show again, right? Yes, we have the quiz show.] In such cases, the epistemic stance is that of momentary doubt, perhaps feigned. The same effect may be achieved simply by intonation, signalled in writing by punctuation. There is only one instance of this in the data: Pero, ¿seguro que no funciona? [But it’s not working for sure?] d. Its use does not appear to be triggered by an anticipation of imaginary or possible dissent on the part of the addressee, as surely is. e. Nor does it appear to be triggered by judgements of what is right, of upholding a moral stance, or of the addressee failing to know or do what is right. In other words, it has no deontic bias. f. It only rarely requests confirmation from the addressee. It does not draw the addressee into agreeing, confessing etc. as surely very often does. g. Nor does it invite collaboration in a joint action, as surely often does, especially with ‘we’ as subject. Rather, it affirms the outcome of the action. h. Seguro que appears to express a confident, often optimistic assertion, professing certainty and reassurance in a situation which is not certain, or is still unresolved, as in (24): 24. ¡hombre!, que tenga un poquito de confianza, que seguro que se recupera pronto. [Come on! Have a bit of faith, you’re sure to get well soon.] Glossing this utterance with initial surely (Surely you’ll get well soon) would be highly inappropriate in the context, as it would indicate contradictory assumptions regarding the possibility of the addressee´s recovery. Even medial surely would not be entirely free of a negative implication, in British English at least: (You’ll surely get well soon), although it may be acceptable in American English. In Spanish, seguramente would be inappropriate in this context because of its implication of probability rather than certainty. As illustrated in (24), be sure to, a form which is available for the expression of predictions and assumptions is usually a good equivalent of seguro que. Unless used as a question, or with a tag, seguro (que) does not admit a trace of doubt, as surely does. As a counterpart to initial surely, when expressing a non-aligning, challenging stance, therefore, seguro que functions best when followed by a tag, by which it becomes a queclarative, as in

Surely and its Functional Counterparts in Spanish 51 Surely you knew [Seguro que lo sabías, ¿verdad?]. Similarly, with final surely:You remember her surely? I know I do, which approximates to ‘Seguro que la recuerdas, ¿no?’, both markers express a less confrontational stance. By contrast, Seguro que la recuerdas without the tag is equivalent to ‘You’re sure to remember her’, a confident prediction. A large number of features present or potential in the use of surely are absent from the uses of seguro, as revealed by the data examined. They are complemented however, by the negative-interrogative questions and queclaratives. This fact appears to confirm the hypothesis that the stances expressible by surely are not shared by one single Spanish pragmatic marker cognate with surely. Rather, many stances expressed in Spanish will be coded as negative-interrogative questions.

7.

BIASED QUESTIONS AND QUECLARATIVES IN THE SPANISH DATA: ¿NO CREE QUE…? AND NO ME NEGARÁ QUE

The second set of data investigated in the Spanish corpus consisted of negative-interrogative questions of the kind ¿No cree(s) que..?.’[Don’t you think that…] and negative declaratives of the kind No me negarás que… [you won’t deny that…]. Negative-interrogative questions and certain negative queclaratives in the Spanish data are able to express a range of stances and rhetorical uses which correspond approximately to some of those expressed by surely in a biased declarative. Interestingly, they approximate to surely in precisely those uses and stances for which seguro que is an inadequate counterpart. I will take ¿No cree(s) que…? [Don’t you think/believe that…] as a typical negativeinterrogative in the data, and No me negará(s) que [You won’t deny that…] as a typical queclarative. The bracketed morpheme concords with the familiar second person tú, the plain form (cree, negará) with the third person polite pronoun usted. The following examples illustrate the use of each of these structures, together with suggested glosses and stances: A Stances of opposition: challenge, confrontation The first two extracts are taken from an interview with the author of a supposedly fictional work in which the characters and details bear a strong resemblance to the members of the royal family of Monaco. The speaker is the interviewer in both cases. 1. ¿No cree que..? 25. – Usted siempre ha dicho que es capaz de reventar a alguien en tres palabras. Con la familia real monegasca se ha ensañado. ¿No cree que ha sido demasiado cruel? [You have always said you can ruin someone in three words. You have tormented the royal family of Monaco. Don’t you think you have been too cruel? Surely you have been too cruel? ] This negative-interrogative in Spanish is biased towards a positive response (Yes, I have been too cruel)

52 Pragmatic Markers in Contrast 2. No me negará que 26. No me negará que son personajes identificables y muy parecidos a la familia real monegasca. [You won’t deny that the characters are identifiable persons and very similar to the members of the royal family of Monaco.] [Surely you won’t /can’t deny that the characters are identifiable persons….] This utterance is a conducive negative declarative with the force of a question, but without lexical reinforcement in Spanish. Like surely, it elicits a response and is biased towards a negative answer (No, I won’t [Eng can’t] deny..), as does a negative declarative question containing surely (Surely you don’t deny..). Both the Spanish version and the English one with surely are biased declarative questions or ‘queclaratives’. B Stances of alignment: Inviting collaboration in a joint action Extract (27) comes from a popular television programme, Un, dos, tres: 27. Jordi, Jordi, un momentito un momentito. ¿Tu no crees que antes de empezar deberíamos dar un aplauso muy fuerte todos para atraer la suerte? [Jordi, Jordi, just a moment. Don’t you think that before starting we should all give a big round of applause to bring us luck?] [Shouldn’t we all give a big round of applause..? Surely we should all give a big round of applause…] 7.1

Features of ¿No cree(s) que and No me negará(s) que…

a. They are both syntactic structures with a complementizer which, in a traditional approach, would be analysed as ‘matrix’ clauses. The complement clause which they introduce, however, carries the propositional content of the utterance. b. They occur predominantly in initial position, where they ‘frame’ the following content. c. They are both syntactically detachable from the sentence. They can occur in final position, without the complementizer, as an epistemic parenthetical (No me lo negarás requiring an object pronoun lo [‘it’] in this position). d. They do not affect the propositional content of the complement clause. e. They occur in face-to-face situations in spoken rather than written discourse (with the exception of fictional dialogue). f. They are more common in speech than in writing. g. The meaning expressed by the stance-carrying element is not its normal literal meaning. The preceding excerpts illustrate all these features except final position, which is illustrated in (28), which is about supposed beings from other worlds: 28. Llamarlos ángeles a todos quizá no sea lo más prudente, ¿no crees? [To call them all angels might not be the most prudent, surely? / lit. don’t you think?] The listed features are those which fulfil many of the criteria for pragmatic markers. Even syntactic status is shared by pragmatic markers such as I mean, I suppose. It would seem possible then, I suggest, to consider ¿no cree(s) que? and no me negará(s) as formulas which are on the way to being grammaticised as epistemic fragments or phrases similar to I think,

Surely and its Functional Counterparts in Spanish 53 (Thompson and Mulac, 1991; Kärkkäinen, 2004), and functioning as epistemic, evidential or evaluative markers (Thompson, 2002). In the latter work Thompson argues that ‘what conversationalists are engaged in doing with their talk crucially involves the complement; in the majority of cases the complement “overrides” the “main clause”, and the “main clause” is there to provide speaker stance towards the assessments, claims, counterclaims and proposals’ (p. 134). It is true that subjects other than first person tend not to be considered in discussions on markers. Yet findings on the surely corpus (Downing, 2001) showed that the most common subject after (mainly anticipatory) it collocating with initial surely was you. Furthermore, when interrogatives, and particularly biased interrogatives are used, the you and the I are inseparable participants in the communicative event. As Thompson and Mulac point out, it has often been noticed that markers of evidentiality and epistemicity are skewed towards first person singular declaratives and second person questions (1991:322). 7.2

Rhetorical Uses of ¿No cree que and No me negará que as Epistemic Phrases

In claiming that ¿No cree que.? and No me negará can be considered as epistemic phrases which in appropriate contexts can carry out pragmatic functions similar to those of surely, the question is how to distinguish a pragmatic use from their use as normal questions and statements, respectively. The first reply to this question is that negative-interrogative questions and conducive declaratives (queclaratives) are not ‘normal’ questions in the sense of being neutral. They are always biased, as surely is when in initial and final positions. It is worth mentioning that initial position of surely is overwhelmingly more frequent than final position. The same occurs with ¿no cree (que)?. No me lo negará in final position is not attested in the present data. Secondly, to see what a marker is doing, not only the propositional content or representation but the procedural meaning has to be considered, as Aijmer and SimonVandenbergen (2004:1790) point out with reference to in fact. The procedural function treats the speaker as instructing the hearer to interpret an utterance in a certain way, by means of inference. In the examples seen in section 7 the meaning of the stance-carrying element ¿No cree que..? and No me negará is inferred as not its normal, literal meaning. Thus, in the extract about the supposedly fictional work whose characters bear a strong resemblance to the royal family of Monaco, the interviewer is not asking the writer’s opinion. Rather, he is challenging the writer’s judgement in writing the book at all in its actual form. There is a clearly deontic bias in the formulating of the negative questions: the writer should not have made his fictional characters so similar to the members of the Monaco royal family. These negative questions index an implied contrast between what is the case and the speaker’s judgement of what should have been the case. Similarly, in the second extract from the same interview the future form negará of Spanish negar (‘deny’) is not functioning as a tense marker making reference to future events. Rather, it is an expression of strong subjective epìstemic assumption, as is ‘surely you will not deny’. Although with strict regard to polarity (25) expects a positive answer (Yes I have been too cruel), while (26) expects a negative answer (No, I can’t deny that the characters are

54 Pragmatic Markers in Contrast identifiable persons and very similar to the royal family of Monaco), the illocutionary force expected in the answer is the same in both: that of admission on the part of the writer that the interviewer is right. The attaining of such an admission is the rhetorical goal that the interviewer aims for in using conducive questions and queclaratives. The speaker is the interviewer, by convention the dominant speaker, the one who is licensed to ask provoking, probing and confrontational questions.The addressee is the agent of the reprehensible action in (25) and (26) and is the creator of the characters which meet with the interviewer’s disapproval. As a result, a reproof or reproach is implied and the addressee is invited to admit that he acted wrongly. In example (27), the function of the epistemic fragment ¿Tu no crees que..? is likewise not that of requesting a factual opinion about the suitability of giving applause, much less permission, from the unseen director of the television programme. The presenter and moderator of this highly popular television programme is not obliged to consult the director on matters such as giving a round of applause. What she is doing by means of the negativeinterrogative phrase is to create involvement by including the director, herself and the invited audience in giving a round of applause. This strategy is more effective than the hortative ‘Let’s give a round of applause’ and much more collaborative than the marked imperative ‘You all give a round of applause’, which would be unthinkable in this context. The addition of the optional second person pronoun tú (the familiar form) reinforces this interpretation.

8.

THE POLITENESS FACTOR IN THE USE OF SURELY AND ITS SPANISH COUNTERPARTS

Linguistic politeness is the factor that draws together the uses of surely, seguro (que), no me negará and ¿no crees (que)?. In all the examples cited, a plain assertion, one containing none of these markers, would be more direct and impositive. By prefacing the face-threatening utterance with one of these epistemic phrases, speakers distance themselves somewhat from the propositional content expressed in the rest of the clause. At the same time they establish either closer or more distant interpersonal space between themselves and their addressee, according to the context. It is this closeness or distance which in part enables the adoption of a variety of different stances on the part of the speaker, reflecting alignment or opposition, which are interpreted as the illocutionary force of the clausal utterance as a whole. With surely and no crees que the implication of the speaker’s superior status is usually strong when the addressee is the subject-agent of the content clause. When an inclusive ‘we’ is subjectagent, this stance is different, usually invitating collaboration, often with a deontic bias. The pragmatic implicature is therefore that of a more or less superior stance, but polite, within the bounds of acceptable (public) behaviour. Procedural meaning in context is the all-important factor which “permits reinterpretation of a construction, leading to a polysemic use of a given construction” (Chorodowska-Pilch, 1999:345). Surely has traversed a long period of polysemy and has all but reached the status of an opaque pragmatic marker whose principal force is recovered by inference. However, a trace still remains in the use of surely as an intensifier, used in medial position, in which case it is close to ‘certainly’. Without the benefit of a fuller context, including intonation and visual

Surely and its Functional Counterparts in Spanish 55 evidence, there may be a certain ambiguity in medial surely. Seguro has advanced less far in the grammaticisation process, but nevertheless has reached the status of epistemic marker in one use, that illustrated in this study, while still retaining some of its polysemous uses in other contexts. No crees que, like surely, is based on contradictory assumptions which are implied rather than stated. The conducive nature of these constructions accounts for the fact that a speaker can put across some quite damaging remarks while still conforming to the standards of conventionally polite behaviour. Conversely, the speaker can turn what might have otherwise been an impositive directive into an invitation to cooperate. 8.1.

Summary of the Spanish Counterparts of Surely

The predominant opaque meaning and use of surely in British English is marked as surely1. In this use surely is typically initial, but also occurs in final position. The secondary meaning and use of surely is marked as surely2. It occurs in medial position. Surely1 _______________________ (initial, final) Surely1 _______________________

¿No cree(s) que? / other negativeinterrogatives No me negará que

Surely1

Seguro que….+ tag ¿no?/ ¿verdad?

_______________________

Surely2 _______________________ (medial = undoubtedly, almost certainly)

Seguramente

Be sure to ______________________ I bet _______________________

Seguro que Seguro que

9.

CONCLUSIONS

The data revealed that there is no single pragmatic marker in Peninsular Spanish which serves as a counterpart to surely in all its nuances. This is not surprising if we realise that in British English at least surely functions somewhat differently when in initial and final positions as against medial. In initial and final positions surely functions as a queclarative (a declarative which is a biased question, based on contradictory assumptions), with greater rhetorical strength when initial than when final. In the above diagram this meaning of surely is labelled as surely1, as opposed to surely2, that of medial position. The nearest equivalents of surely1 in the Spanish data are negative –interrogatives such as ¿no cree(s) que? or negative declaratives used with an ‘assumptive’ meaning such as no me negará(s). Both, however, lack the strength typical of initial surely, although they do provide the touch of politeness that also characterises surely. Seguro que followed by a negative tag (¿no? or ¿verdad?) may serve as a marker similar to surely1 when the speaker’s confidence is of a high degree, hardly admitting

56 Pragmatic Markers in Contrast a contradictory assumption, giving rise to a strong positive bias and challenging stance, as in Surely you knew (Seguro que lo sabías, ¿verdad?). A negative-interrogative question captures some of the nuances of surely when a) indignation or surprise are a factor, as in (25) and (26); b) when the speaker anticipates real or imaginary dissent, as in the same examples. The use of ¿No cree que deberíamos…? (Don’t you think we should /Surely we should…) in (27) is a polite invitation to agree on a course of action. For surely2, in medial position, seguramente is the nearest equivalent when a fairly confident assertion is implied politely, especially in formal contexts, as in This they will surely know / Esto seguramente lo sabrán. (Cf. Eduardo Verdú, un joven escritor que seguramente conocerás.. [Eduardo Verdú, a young writer whom you will surely know].) For a stronger assertion the semantic field can be enlarged to include the constructions be sure to to account for the usual use of seguro que (You are sure to recover/ Seguro que te recuperas) and I bet, the latter a very informal variant for seguro. I conclude by suggesting that the Spanish constructions examined in the study have acquired pragmatic meanings alongside their literal meanings and in their pragmatic uses have almost become formulas. With the exception of seguramente, which occurs in all positions, they occur typically in spoken Spanish in initial position, they are related to the speech situation and fulfil several of the criteria, including a certain mobility, proposed for the identification of pragmatic markers. I suggest that the syntactic fragments seguro (que), no crees (que) and no me negará (que) may be considered as epistemic phrases whose function is that of indexing a stance taken by the speaker with relation to the propositional content, while repositioning the speaker vis-à-vis the addressee, as surely does. No cree que and no me negará can make for either opposition to or empathy with the addressee according to the context, with a tendency towards the confrontational. Other expressions share some of the features of these epistemic phrases, (No me querrás decir que.. /surely you don’t mean..), for instance, but they are not detachable from the sentence. Seguro que, by contrast, most typically projects a reassuring and optimistic stance. In all, these phrases serve as partial equivalents of English surely, mapping some of the variations expressed by the English marker, although with a weaker dominant stance.

REFERENCES Aijmer, K. (1986). Discourse variation and hedging. In: Corpus Linguistics II. New Studies in the Analysis and Exploitation of Computer Corpora (J. Aarts and W. Meijs, eds). Rodopi, Amsterdam. Aijmer, K. and A-M. Simon-Vandenbergen (2004). A model and a methodology for the study of pragmatic markers: the semantic field of expectation, Journal of Pragmatics 36, 1781–1805. Andersen, G. (2000). Pragmatic Markers and Sociolinguistic Variation. John Benjamins, Amsterdam. Biber, D. and E. Finegan (1989). Styles of stance in English: Lexical and grammatical marking of evidentiality and affect. Text 9 (1), 93–124.

Surely and its Functional Counterparts in Spanish 57 Biber, D., S. Johansson, G. Leech, S. Conrad and E. Finegan (1999). Longman Grammar of Spoken and Written English. Longman, London. Bosque, I. and V. Demonte (1999). Gramática Descriptiva de la Lengua Española. Real Academia Española. Espasa Calpe, Madrid. Brazil, D. (1995). A Grammar of Speech. Oxford University Press, Oxford. Brinton, L. J. (1996). Pragmatic Markers in English. Grammaticalization and Discourse Functions. Mouton de Gruyter, Berlin. Chafe, W. (1986). Evidentiality in English conversation and academic writing. In: Evidentiality: The Linguistic Coding of Epistemology (W. Chafe and J. Nichols, eds.), pp. 261–272. Volume XX in the series Advances in Discourse Processes. Norwood, N,J.: Ablex. Chodorowska-Pilch, M. (1999). On the polite use of ‘vamos’ in Peninsular Spanish. Pragmatics 9 (3), 343–355. Coates, J. (1983). The Semantics of the Modal Auxiliaries. Croom Helm, London. Delancey, S. (2001). The mirative and evidentiality. Journal of Pragmatics 33, 339–382. Diccionario de la Lengua Española (DRAE). 22nd edition, (2001). Academia de la Lengua, Madrid. Downing, A. (2001). “Surely you knew!” Surely as a marker of evidentiality and stance. Functions of Language 8, 253-285. Du Bois, J. (2000). Taking a stance. Constituting the stance differential in dialogic interaction. Paper presented at the Annual Meeting of the American Anthropological Association, San Francisco, November 18. Geluykens, R. (1987). Intonation and speech act type. An Experimental approach to rising intonation in queclaratives. Journal of Pragmatics, 11, 483–494. Halliday, M.A.K. (1970). ‘Functional diversity in language as seen from a consideration of modality and mood in English’. Foundations of Language 4, 1–5. Haviland, J.B. (1987). Fighting words: Evidential particles, affect and argument. In: Proceedings of the Thirteenth Annual Meeting of the Berkeley Linguistics Society on Grammar and Cognition, February 14–16 1987, Berkeley, California. (J. Aske, N. Berry, L. Michaelis and H. Filip, eds), pp. 343.354. Berkeley: California. Hölker, K. (1988). Zur Analyse von Markern. Korrektur-und Schlussmarker des Französischen. (Zeitshcrift für Französische Sprache und Literatur – Beiheft 15. Steiner, Stuttgart. Huddleston, R. and G. K. Pullum (2002). The Cambridge Grammar of the English Language. Cambridge University Press, Cambridge. Jucker, A. H. and Y. Ziv (1998). Discourse markers: Introduction. In: Discourse Markers. Description and Theory (A. H. Jucker and Y. Ziv eds). John Benjamins, Amsterdam. Kärkkäinen, E. (2004). Epistemic Stance in English Conversation. A description of its interactional functions with a focus on I think. John Benjamins, Amsterdam /Philadelphia. Lazard, G. (2001). On the grammaticalization of evidentiality. Journal of Pragmatics, 33, 359–367. Lyons, J. (1977) Semantics, vol.2. Cambridge University Press, Cambridge. Palmer, F.R. (1986). Mood and Modality. Cambridge University Press, Cambridge.

58 Pragmatic Markers in Contrast Quirk, R. , S. Greenbaum, G. Leech, J. Svartvik (1985) A Comprehensive Grammar of the English Language. Longman, London. Sadock, J. M. (1974). Toward a Linguistic Theory of Speech Acts. Academic Press, New York. Schiffrin, D. (1987). Discourse markers. Cambridge University Press, Cambridge. Shorter Oxford English Dictionary on Historical Principles, fifth edition (2002). Oxford University Press, Oxford. The Oxford Spanish Dictionary. Spanish-English, English-Spanish (2001). Oxford University Press, Oxford. Thompson, S. and A. Mulac (1991). A quantitative perspective on the grammaticization of epistemic parentheticals in English. In: Approaches to Grammaticalization (E. C. Traugott and B.Heine, eds) Vol.II, pp. 313–329. John Benjamins, Amsterdam/ Philadelphia. Thompson, S. (2002). “Object complements” and conversation: towards a realistic account. Studies in Language, 26 (1), 129–164. Traugott, E. C. (1995). Subjectification in grammaticalization. In Subjectivity and Subjectivisation. Linguistic Perspectives (D. Stein and S. Wright, eds), pp. 31–54. Cambridge University Press, Cambridge. Verschueren, J. (2000). Notes on the role of metapragmatic awareness in language use. Pragmatics 10 (4), 439–456.


POLYSEMY PATTERNS IN CONTRAST: THE CASE OF DUTCH TOCH AND GERMAN DOCH Ad Foolen, Radboud University Nijmegen

1.

INTRODUCTION

Polysemy is a pervasive property of natural languages. Language users apply existing linguistic means to new communicative problems. If the new use gets established and if it differs sufficiently from the already existing use(s), a polysemous pattern is the result. Every meaningful linguistic form can undergo such a process of ‘polysemization’: morphemes, content words, function words and syntactic constructions. Despite the intensive research on polysemy in the past 25 years, it would be an illusion to think that linguistic theory has reached a full and final understanding of the processes that lead to polysemy patterns and that further research could thus only consist of an application of existing models to new data. Research can still contribute to more insight in constraining conditions under which polysemous extensions can take place and possible patterns that result from such extensions. And what the exact criteria should be for distinguishing separate senses in a polysemous item (instead of considering the item as monosemous) is still a matter of dispute too (cf. Croft & Cruse, 2004: ch. 5). New methods of research can help throw light on such theoretical questions. In recent years, the traditional grammatical approach (explication of native speaker intuitions, supported by substitution and permutation operations on constructed examples) has been supplemented by analysis of electronic corpus data. More recently, we see a growing interest in typological research in which the semantic map approach is used for the analysis of ‘polyfunctionality’ (cf. Haspelmath, 2003). Finally, there is the detailed analysis of contrastive corpus data, cf. Aijmer and Simon-Vandenbergen (2003), who showed that the polysemy pattern of a much studied pragmatic marker like English well can be understood better by looking at the translation of such an item in a restricted number of other languages (Swedish and Dutch in their study). The research presented here made use of contrastive data, supplemented, where necessary, with monolingual data. The focus is on Dutch toch, with German doch coming into the picture through the translation data. I hope that the polysemy patterns of toch and doch contribute insights that are relevant for polysemy theory in general. The translation data come from the translation of a novel by Maarten ‘t Hart, De kroongetuige (1983), which was rendered into German in 2001, under the title Die schwarzen Vögel. The Dutch version contained 270 occurrences of toch and the German version had 231 uses of doch. In 179 cases, there was a congruent translation: a Dutch toch was translated with doch; in 143 cases,

60 Pragmatic Markers in Contrast however, there was a non-congruence: 91 occurrences of toch were translated as zero or in an other way, whereas 52 occurrences of doch did not correspond with a toch in the original version. The examles below are all from the novel by Maarten ‘t Hart, unless otherwise stated.

2.

TOCH AND DOCH

Dutch toch and German doch are cognates. The same item developed into dock in Swedish (Stig Johansson, p.c.), and though in English. The particle resulted from a deictic pronoun (‘that’), to which two emphatic particles were added (cf. Hentschel, 1986: 43). For many centuries, Dutch used toch and doch as variants, but in the last two centuries toch became the unmarked form. The use of doch in Dutch is now restricted to the function of an adversative conjunction in formal style. Dutch toch and German doch belong to the particles with the highest frequency (cf. Hentschel, 1986: 247). Their high frequency corresponds with their multifunctionality. They participate in different word classes (adverb, modal particle, conjunction, answering particle), so that their polysemy is at the same time a case of heterosemy (Lichtenberk, 1991). Just as with other pragmatic markers, it is not always clear when we should distinguish a separate use of toch or doch. In the literature, differences in constructional context, accentuation, substitution and paraphrase possibilities and systematic variation in translation are used as criteria for assuming a separate use. In this paper, I will take a more functional approach in distinguishing different uses. Toch and doch participate in a variety of functional subsystems in the language, and it is this variety that will be the basis for distinguishing different uses of these particles.

3. 3.1

THE PROBLEM OF ‘CORE MEANING’ Schematic meaning

An analysis of the uses of a pragmatic marker typically takes the form of a polysemy pattern. Family resemblances characterize the relation between the members of the pattern, and one of the members may be prototypical, i.e. occupy a central place, but need not (cf. Hansen 1998: 87). Such an analysis does not require a second, abstract, level of analysis, on which one assumes something that is variably called the abstract meaning, core meaning, basic meaning, ‘Grundbedeutung’, etc. This ‘monostratal’ view does not exclude the possibility that certain semantic features remain constant across the members of the polysemy pattern. What I found in the present study was indeed such a stable meaning that recurred in the different uses of toch and doch. This ‘schematic meaning’ can be characterized as a scheme with ‘three steps’. The utterance containing toch/doch is ‘affirmative’: The speaker believes the proposition expressed in the utterance. But this belief is affirmed in reaction to a negation (of that utterance). This negation may be explicitly present in the preceding utterance, or it may be imaginary, in the mind of the speaker, by way of an internal, fictional dialogue. A negation, in its turn, only makes sense against the background of the supposed relevance of an affirmation (Givón 1979: 103 ff). Again, this affirmation may precede in the ongoing

The Case of Dutch Toch and German Doch 61 discourse or have imaginary status. This results in a scheme of three (sequential) steps: affirmation – negation – (re)affirmation. For German doch, the three step analysis has already been explicitly proposed by the Japanese Germanist Sekiguchi (1939/1977). His view has not been generally adopted in the literature, but recently, Graefen (2000) made a similar proposal for German doch, independently of Sekiguchi, which I take as an indication that the intuition that toch and doch involve three steps has some plausibility. An additional supporting argument could be the observation that there are other particles in natural languages that involve three sequential steps. An example is German immerhin, which indicates that the assertion does not fulfil an initial ‘high’ expectation, but is higher than the low expectation that replaced the initial high expectation (Weydt, 1979). The sequence here is: high – low – middle. And such a three step structure is not only relevant for particles: Couper-Kuhlen (2000) and Barth-Weingarten (2003) assume that concessivity in dialogue always involves a minimum of three steps, which they call the ‘Cardinal Concessive Schema’: a statement by a first speaker, followed by a concession of a second speaker, followed by a ‘counterstatement’ by the same second speaker. As noted already, the presence of such a persistent schematic meaning does not force us to assume that this meaning has psychological relevance on a level of analysis that is distinct from the level on which the different family resemblance members are situated. Apparently, it is possible for the schematic meaning to be such a strong part of the meaning of toch/doch that it is reproduced every time this marker finds a new application. This is what I will assume here, as it is the simplest assumption. This view does not forbid a language user or a linguist to ‘abstract’ the schematic meaning out of the network of uses, but a language user can very well be a perfect native speaker without making such an abstraction. Here, I am following the view that Langacker has formulated many times: “I am suggesting that these contextdependent variants may be more fundamental than the context-neutral schematization we tend to regard as primary.” (Langacker, 2000: 125). 3.2

Heteroglossia

The 3-step schematic meaning of toch/doch makes it a perfect case for the concept of heteroglossia. Aijmer and Simon-Vandenbergen (2003) take this notion from Bakhtin (1981) and apply it to their analysis of well. The notion implies that language use always involves different ‘positions’, different viewpoints, which can cause a tension in discourse, in particular when the viewpoints are incompatible. When the incompatible viewpoints are taken up by different participants in the discourse, a speaker has to handle the situation with care, for example by acknowledging that he perceives the other position and respects it, but at the same time indicating that he nevertheless maintains his own position. The different positions do not have to be explicitly present in the discourse as positions taken by different participants. The speaker can prevent tensions if he indicates that he is aware that an alternative position could be taken. In my view, it is precisely the different ways in which the alternative position is present in the discourse (explicitly formulated by a different participant in the discourse, explicitly formulated as a possibility by the speaker himself, or implicitly hinted at by the speaker) that provides a natural basis for distinguishing

62 Pragmatic Markers in Contrast different uses of the particle. These different possibilities correspond with the different functional subsystems in the language in which toch and doch participate. In what follows, I will present different uses of toch and doch, give arguments why each of the proposed uses deserves being distinguished as a separate use and discuss the question which additional meaning, besides the constant three step schematic meaning, seems to be relevant for each of the uses. I will not try to present a full picture of all possible uses of toch and doch, as my primary aim is not descriptive exhaustiveness but possible theoretical relevance of the approach.

4.

4.1

TOCH AND DOCH IN THE SYSTEM OF DIALOGUE PARTICLES German doch 1. G. D. E.

2. G. D. E.

S: Ist das nicht erlaubt? H: Doch, natürlich … (p. 151) Mag dat niet? Natuurlijk wel … (p. 103) Isn’t that allowed? Of course it is. Aber ich glaube niemals, dass er hier Pop gehört hat. Doch, ganz bestimmt, … (p. 152) Maar ik kan nooit geloven dat hij hier naar pop heeft geluisterd. O, jawel hoor, ik weet het zeker … (p. 104) But I can never believe that he listened to pop music here. Oh, yes, he did, I am certain of it.

These examples show that in reaction to a negative question (1) or statement (2), German can use doch, whereas Dutch uses wel or jawel in this context. So, in the Dutch system of dialogue particles, wel seems to be the item that fills the position of ‘third step marker’. In the examples, only the negative second and the reaffirmative third step are present, but the positive first step is implicitly present in the preceding fragments of the text. One can easily construct dialogues in which the first step is explicitly present: That is a duck – No it isn’t – It is (German: doch, Dutch: jawel). German doch thus participates in the functional field of ‘affirmation markers’, where we find, in the first place, yes and no and their counterparts in other languages. A language can restrict itself to these two elements, or add a third marker, which is used when an affirmation is a ‘reaffirmation’ after a negative counterposition. The three positions show themselves most clearly in a ‘yes-no-yes’-dialogue. In this use, the marker can take different forms in different languages: English uses emphasis (I do like it), French uses the marker si (KerbratOrecchioni, 2001), etc. In other syntactic positions, a different form can be used, for example the English conjunct still.

The Case of Dutch Toch and German Doch 63 According to Hentschel (1986: 44) and Molnár (2002: 117), German doch established itself as a dialogue particle rather late, in the 18th century. The question why Dutch toch did not take on this use, can only be answered by diachronic research. Apparently, wel was around as a competitor. But German had a similar item, wohl, the counterpart of wel. Wohl is used emphatically when the reaffirmation takes the form of a full sentence: It ís a duck, Dutch: Het is wél een eend, German: Es ist wóhl eine Ente. This shows that the language specific distribution of the available forms looks rather unpredictable. Particular contextual aspects must be taken into consideration, aspects which do not directly pertain to the function of the marker as such and which are often of a rather superficial syntactic character. It would be too general to state contrastively that German doch is used as the dialogic reaffirmation particle and Dutch toch is not. In German, doch can be replaced by wohl if the context is a full sentence, and in Dutch one can add toch to wel, although this is an infrequent option. 4.2

Flemish toch

Dutch exists in different variants, one of them being the Dutch as it is spoken in the Flemish part of Belgium. In this variant, a particular use of toch has developed, which, like German doch, occurs primarily in dialogue, which is why it is analyzed in this section. A typical example is the following: 3.

‘Is er wat’? vroeg hij. ‘Is something wrong’ he asked. ‘Jij wou me spreken’, zei Eva. ‘You wanted to talk to me,’ Eva said. ‘O ja?’zei hij. ‘Did I?’ he said. ‘Vanmorgen toch,’zei Eva. ‘This morning at least,’ Eva said. ‘Nu niet meer,’ zei Vincent. ‘Not any more,’ Vincent said. (Patricia de Martelaere, Littekens. Novel. Amsterdam: Meulenhoff, Leuven: Kritak, 1990)

In this dialogue, Eva answers the initial question with a statement: ‘you wanted to talk to me’; then the hearer questions that answer, after which, in the third turn, Eva returns to her initial claim, but now in a weakened version ‘you wanted to talk to me this morning’. The speaker steps down, makes her claim more modest, and it is this discourse function that in Flemish can be marked by a sentence final toch. ‘At least’ is often an appropriate translation of this use in English; in Netherlandish Dutch, ‘tenminste’, or ‘althans’ would be used. What we see here, is the schematic meaning, enriched with the feature ‘partial return’. This feature involves a concessive aspect: the hearer might be right that the initial answer was too strong, so the return is only partial. The hearer’s viewpoint is not only acknowledged as a possible viewpoint, it is given some real validity. This use seems a natural semantic development of the heteroglossic meaning of toch. A similar ‘weakened return’ to first position is present in German immerhin (see 3.1). This use of toch does not only occur in real dialogue, where the positions are taken by different participants. A speaker can use it in a monologic discourse which, through the presence of this toch, gets a strong dialogic flavour, see the following examples:

64 Pragmatic Markers in Contrast 4.

D. E.

Ik denk dat je dat gebouw nog wel herkent. Dat hoop ik toch. (CGN: Belgisch) I think you recognize that building. At least I hope so (Corpus Spoken Dutch, Belgian part)

5.

D.

We laten even terzijde dat Amerikaanse persbureau’s – en andere zijn er niet meer, toch geen internationale – de neiging hebben om Amerika gelijk te stellen met de wereld. (De Standaard 20/02/2001) We disregard for the moment that American news agencies – and other agencies don’t exist anymore, at least no international ones – tend to consider America and the world as identical.

E.

In (4), ‘think’ is weakened to ‘hope’, in (5), the parenthetic claim ‘there are no others’ is followed immediately by the weakened claim ‘at least no international ones’, to pre-empt a possible protest by the imagined hearer. Note, that toch is not final here, but from the context, it is clear that the intended use is the one we have been discussing in this section.

5.

TOCH AND DOCH IN THE SYSTEM OF ADVERSATIVE CONNECTIVES

In this use, the function of toch/doch is primarily connective. Here, the second, negative, position in the three step scheme is an implication of the preceding sentence, in the same way as when a denial of expectation but is used. In small but brave, the first conjunct implicates not brave, and it is this implication that the second conjunct reacts against. It is no surprise that in English translations of the connective use of toch/doch we find but, nevertheless, (and) yet, even so or still. Just as with dialogic doch, there are language specific syntactic conventionalizations of adversative toch/doch. In German, doch can easily be used in the position of a coordinating conjunction, cf. (6): 6.

G. D. E.

Doch von der Garderobe her rief er fröhlich: Jozef Lambert. (p. 63). Maar bij de kapstok riep hij vrolijk: Jozef Lambert. (p. 43) But at the hatstand he called joyfully: Jozef Lambert.

In contrast, Dutch toch normally does not occur in the conjunction position. Instead, the conjunct position or the first sentence position are used (see below, examples 8 and 9). But there is one context, where toch occurs as a conjunction, namely when two adjectival phrases are coordinated: met die diepe, toch beweeglijke stem, ‘with that deep but flexible voice’ (novel, 2004). It is a remarkable fact, that in this context toch can be substituted by doch: Een klein doch zeer belangrijk verschil, ‘a small but very important difference’ (magazine, 2004). This doch gives the text a formal flavour. More remarkable still, this formal style doch can, though very rarely, be found as a conjunction between two full sentences:

The Case of Dutch Toch and German Doch 65 7.

D.

E.

Tegenover de overeenkomsten tussen (1) en (2) staan evidente verschillen, doch die behoeven niet op de tegenstelling tussen object en perifere bepaling te wijzen. (Linguistic manuscript, 2003). Besides the similarities between (1) and (2), there are evident differences, but they do not necessarily point to the difference between object and peripheral adjunct.

The presence of the verb behoeven instead of hoeven ‘need’ also indicates that the author aims at a formal style here. Dutch toch/doch in conjunction position seems to be a rest of an older system (see section 2). The ‘normal’ uses of adversative connective toch are not found in the conjunction position. In Dutch, there are two alternative syntactic positions available. On the one hand, there is the conjunct position, separated by a comma from the main clause, as in (8): 8.

D.

E.

Ik maak nota’s. Waarvan je vertraagd of soms helemaal niet hoort of ze gewerkt hebben. Toch, voor mij moet wat ik doe met onderwijs te maken hebben. (Magazine, June 2003). I produce reports. With much delay, or never, do you hear whether they had some effect. Still, what I do must, for me, have something to do with education.

The other possibility is provided by the first sentence position that precedes the V2 position in Dutch and German. The subject is the unmarked choice for this position, but most other constituents are eligible as well. In Dutch, toch is a possible choice, cf. (9) and (10): 9.

D. G. E.

Toch zou ik daarmee nog maar even wachten. (p. 94) Dennoch würde ich damit noch etwas warten. (p. 139) And yet, I would wait a bit before doing so.

10.

D. G. E.

Toch wilde ik de cruciale vraag nog stellen. (p. 93) Und doch wollte ich die entscheidende Frage noch stellen. (p. 138) Even so, I still wanted to ask the crucial question.

These examples and their translations show again language specific peculiarities. Example (10) shows that German doch can occur in the first sentence position, but the translator has decided to let a coordinating conjunction precede. And in (9) the translation of toch is dennoch. Apparently, German does not ‘like’ a bare doch in first sentence position. This might have to do with the fact that initial doch in German is a conjunction. If it could be, at the same time, a sentence adverb in first sentence position, then this would be a case of syntactic ambiguity at the beginning of the sentence, which is avoided by the system. In this section, I proposed to assume a textual adversative function for toch and doch, which is realized in varying positions at the beginning of the sentence. Note that I have given the functional similarity more weight here than the syntactic differences. German and Dutch differ in their preferences where to put the marker syntactically: German easily puts doch in the coordinative conjunction position, Dutch prefers the conjunct and the first sentence position. In the conjunction position, Dutch has preserved doch as an option, with formal style connotation.

66 Pragmatic Markers in Contrast Up till now, the analysis suggested that there is no principled difference between the unmarked adversative conjunction but and adversative connective toch/doch in sentenceinitial position. That would mean that, in this contex, the schematic meaning of toch/doch has been reduced to the same two step scheme that is assumed for but (but q = q, and not r, which could be deduced from the foregoing), the first positive step of the original schematic meaning of toch/doch being cancelled. But a look at the broader textual context reveals that the distribution of these adversative connectives is not random. Doherty (2001: 238) took a closer look at the distribution of the different adversative markers and found that “the examples with doch suggest that it returns to an earlier assumption in relation to the new or contrasted element.” Doherty proposes to call doch and dennoch ‘resumptive adversative connectors’. This fits in, of course, very well with the three step schematic meaning I have proposed for all uses of doch and toch.

6.

TOCH AND DOCH AS MODAL PARTICLES

When doch/toch occurs unstressed in the middle field of a sentence, they are to be considered as modal particles. The modal particle use of toch and doch is a very frequent one. Of the 270 occurrences of toch in ‘t Hart’s novel, 178 belonged to this category, most of which were translated as doch. What is the function of modal particles, i.e. in which functional subsystem of language do they fit? Much theorizing has been devoted to this question, and no generally accepted answer seems to have resulted from it. It has been observed that the distribution of modal particles is strongly related to sentence type, which could be seen as an indication that their functioning is related to the speech act dimension of language use. This is the approach taken in Waltereit (2001: 1414): “Their purpose is essentially to accommodate at minimal linguistic expense the preparatory conditions of the speech act they occur in.” This approach seems fruitful for explicating the effect of adding the modal particle toch/doch to the various sentence types. This is not the place to do so in full detail, but to give an idea, let us briefly look at at few speech acts and sentence types in which toch/doch occur. A preparatory condition for a request is, that H does not know that S wants him to do X. A request with doch/toch indicates: you already know that I want you to do X. The schematic meaning of toch/doch leads back to the ground for this deviation from the normal preparatory condition: the request had been made before. However, non-compliance on the part of the hearer has led to a repetition of the request. The end-effect is a request with some urgency, often irritation, cf. (11): 11.

Imperatives (18 cases) D. Ach, hou toch op met dat gezeur. (p. 8) G. Ach, hör doch endlich damit auf. (p. 11) E. Oh, stop nagging, for goodness’ sake!

A declarative sentence typically contains new information. The accompanying preparatory condition thus is: H does not know p. Toch/doch indicates that this condition does not hold,

The Case of Dutch Toch and German Doch 67 i.e., it indicates that S assumed that H did know p and shared the same attitude to p. This leads to the special effect that the bringing up of p is understood as a ‘check’, cf. (12): 12.

Declarative sentences (140 cases in the original Dutch version of the novel) D. Maar ik kan toch thuis koffie zetten. (p. 8) G. Aber ich kann doch zu Hause Kaffee machen. (p. 10) E. But I can make coffee at home, can’t I?

The ‘check’ function can be paraphrased as follows: the speaker wants to make sure that he and the hearer share the same attitude to the proposition expressed by the sentence. In this context the three steps are: We share positions – There is reason to be uncertain about your attitude – So, I, speaker, indicate that, on my part, the proposition still holds, and I would like to know how you see this. A slightly rising intonation at the end is part of this construction. There is a distributional difference to be observed between Dutch and German: In Dutch, ‘checking’ toch has the additional possibility of occurring in sentence-final tag position. This ‘end shift’ is rather new, and seems especially popular in informal style, more in Netherlandish Dutch then in Flemish Dutch. From a functional perspective, there is no reason to assume a ‘different’ use of toch here, however. The distributional innovation seems a very natural, iconic one: The orientation of the sentence is ‘forward’, it invites a (confirming) reaction from the hearer. Placing the particle at the end of the utterance instead of the middle field, is a natural shift, and it has the extra benefit that it can function as a turntaking marker. Here are some examples: 13.

D. E.

Wat is nou consequent? Het kan alle kanten opvliegen, toch? (Magazine, 2003) What is consistency? It can develop in any directions, can’t it?

14.

D.

Mijn zoon was altijd goed met geld. Als je dan ook nog eens de bank achter je hebt staan dan kan het niet stuk, toch? (Newspaper, 2003). My son always had a feeling for money. And when your bank is, on top of that, fully supportive, then things can’t go wrong, can they?

E.

Besides the directive and declarative context, there is a third context in which toch and doch are frequently used as modal particles, namely exclamatives, cf. (15 and (16): 15.

Exclamatives (9 cases) D. O, mevrouw, wat kent u de misdaad toch slecht. (p. 134) G. Oh, Mevrouw, wie schlecht kennen Sie sich doch im Verbrechen aus. (p. 194) E. Oh, madam, how unfamiliar you are with crime!

16.

Exclamative if-sentence D. Ach, als het zo toch eens gegaan kon zijn. (p. 145) G. Ach, wenn es sich doch so abgespielt hätte. (p. 209) E. Oh, if only it could have happened that way!

68 Pragmatic Markers in Contrast As (15) and (16) illustrate, Dutch and German are parallel in this respect, but there is an extra distributional possibility in Dutch that German does not have. Dutch can realize an exclamation by only expressing the addressee’s name: Jan! Usually, the name is preceded by maar or followed by toch, or both: (maar) Jan (toch)! This extended use of Dutch toch in exclamative contexts is, like in the case of the toch that can be shifted to the end of the declarative sentence, a peculiarity that is not part of the German pattern. Again, I see no reason to assume a separate use for this Name+toch-exclamative. Just as in full sentence exclamatives, the meaning is ‘surprise’, usually with a negative flavour. In the exclamative contexts, the schematic meaning of toch and doch applies in the subjective sphere of the speaker’s belief. In the first stage there is the observation that p, then follows the disbelief, but in the end there is no escape: the speaker has to accept p, because the observed reality offers no escape. Indication of an up-down-up movement of the speaker’s feeling or attitude is what exclamative toch/doch contributes. There is one context of modal particle use of toch/doch, where Dutch and German are remarkably different, namely in wh-questions, cf. (17): 17.

Wh-questions (11 cases) D. Wie heeft toch je pad gekruist, toen ik bij m’n moeder was? (p. 49) G. Wer ist dir nur über den Weg gelaufen, als ich bei meiner Mutter war. (p. 72) E. Who on earth crossed your path when I was visiting my mother?

None of the 11 occurrences of toch in the context of a wh-question was translated by doch. Instead, other particles were chosen, like nur in (17). The situation is, however, different again in case a German wh-question has the special function of ‘memory search’. Then doch is a typical choice, whereas toch is not in Dutch. As this use did not occur in the corpus, the following example was constructed: 18.

Memory search D. Wie hiess er doch? G. Hoe heette hij nog maar? E. What was his name again?

For this special function of memory search, Dutch has a variety of possible markers, typically consisting of a cluster of particles: nog maar, ook al weer, etc. Why is it that the modal particle uses in Dutch and German run parallel in the other sentence types and not in whquestions? I am afraid, a synchronic, structural answer will be hard to find. A diachronic study might reveal certain contextual factors that led to this contrastive picture. Toch and doch are embedded in a different linguistic system and the diachronic processes of extension of other particles have undoubtedly played a role in the present outcome for toch and doch.

7.

TOCH AND DOCH IN GROUNDING CONTEXTS

Dutch toch and German doch occur in contexts that are often characterized as ‘causal’ (for example Borst, 1985: 74 ff). The contexts are, however, not the same in the two languages. In

The Case of Dutch Toch and German Doch 69 Dutch, the context is a sentence in which an action or a proposal for a possibly controversial action X is formulated. In a sentence that can follow or precede, this action is justified. This justificatory sentence contains a toch in the middle field. The intonation contour of this sentence is special: strong accent on toch, reduction of accents in the rest of the sentence. The two sentences together can be paraphrased as follows: ‘now that what is stated in the tochsentence is the case, action X can be carried out as well. In other circumstances, objections to doing X would be reasonable, but against the background of what is stated in the tochsentence as an undeniable fact, objections make no sense’. If the undeniable fact in itself is evaluated as ‘bad’ the whole construction gets a touch of resignation: X might as well be done, it does not make any difference, facts are what they are. In the novel by Maarten ‘t Hart, 7 examples of this use of toch were found, see example (19): 19.

Toch in declarative sentences, meaning ‘anyway’ (7 cases) D. Nu ik hier toch ben, zou ik graag eens wat rondkijken, kan dat? (p. 54) G. Nun, wo ich einmal hier bin, würde ich mich gern ein wenig umsehen, Geht das? (p. 79) E. Now that I am here anyway, I would like to look around a bit, is that OK?

In the German translation of (19) a different particle (einmal) is chosen. This is already an indication that Dutch and German differ in the way they use toch and doch in a grounding context. German ‘causal’ doch is used in a slightly different context. The following example is again from the translation of ‘t Hart’s novel. Note that there is no toch in the Dutch original: 20. G. D. E.

Fast ohne hinzusehen, konnte ich die Rille finden, bei der Cassio mit seiner Geschichte beginnt; hatte ich doch schon so oft diese Rille gesucht. (p. 46) Bijna blindelings kon ik de groef vinden waar Cassio zijn verhaal begint, had die groef al zo vaak gezocht. (p. 32) I could find the groove almost without looking, since I had, after all, looked for that groove so often before.

As has been pointed out in the literature (see for example Brauße, 2001), this use of doch occurs in a construction with strict syntactic properties. The doch-sentence has to follow the one it relates to (in Dutch it can also precede) and starts with the finite verb (which is very marked for the declarative sentence type, cf. Önnerfors, 1997: 155ff), and doch is unstressed and situated in the middle field. The construction as a whole seems to have been grammaticalized much more strongly than is the case with the Dutch justificatory toch. The paraphrase of (20) is as follows: Because the speaker had searched for the specific groove on the record so often before, he could easily find it again this time, almost without looking. The function of the doch-sentence is explanatory; the speaker explains why it was so easy for him to find the groove. With doch, the speaker indicates that the content of the sentence that serves as the explanation, is known to the hearer, so there was no urgent need to state it, but it is stated nevertheless, just in case the hearer might have forgotten. In this example, the reader does indeed know from the previous pages, that the main character had listened to this passage of Othello again and again.

70 Pragmatic Markers in Contrast In the grounding uses discussed here, the three steps of the schematic meaning pertain to different contextual dimensions in Dutch and German. In Dutch, the first step is the stated fact (which is explanatory for action X), the second one involves the idea that this fact might not be the case. In the third step this negative step is acknowledged as a relevant possibility, but not as reality. This results in marking the statement as an undeniable fact (‘I am here, whether you like it or not …’). In the German case, the target of the particle is the knowledge state of the hearer. Step 1: you know X. Step 2: you might temporarily not be thinking of X. Step 3: that is why I am mentioning it here, because it is relevant in this context, it explains what was said in the previous sentence.

8.

CONCLUSION

The above analysis of different uses of toch and doch is not meant to be exhaustive. There is, for example, the possibility of having toch/doch strongly stressed in the middle field as the main focus of the sentence, comparable with a stressed negative particle (He is nót present). Emphatic negation can be used to counter an opposite claim or thought. With toch/doch it is thoughts or observed facts, not verbalized claims (then one uses wóhl in German and wél in Dutch, see section 4), that the speaker returns to. The context is thus typically one where a speaker is ‘reflecting’ on his own thoughts and observations: Je bent er tóch/Du bist dóch da (‘you are present!’), implying that I thought or saw before that you were present, then I thought that this was not so, but now I see that you are really present. The Dutch-German contrastive analysis made clear, more than a monolingual study could have done, that the language specific uses of ‘the same’ item are indeed rather specific, but at the same time that in all these applications a single schematic meaning was preserved. Despite the varied uses that are made of the same item, the schematic meaning turned out to be robust. Each use is embedded in a particular subsystem of the language, that of dialogic particles, adversative connectives, modal particles and causal markers, and in each of these contexts, additional meaning can become part of the full meaning of the particle in that context. Theories which advocate sparsity in linguistic description would prefer to leave such extra meaning to the context (implicatures, generated by the interaction between the item and specific contexts), but the findings in this study rather support the assumption of richer conventionalized meanings for specific members of the polysemy pattern. This does not mean that there is no room for contextual implicatures: ‘resignation’ can occur as an implicature of Dutch ‘justificatory toch’ if it occurs in contexts where actions are justified by arguing that doing otherwise would not help change a negatively evaluated situation (‘I am having another piece of cake, I am toch fat’), ‘irritation’ can be an implicature when the modal particle is used in a directive context where the hearer seems unwilling to follow the directive, etc. A proper linguistic description should find the right balance between the schematic meaning that holds for all members of the polysemy pattern, extra conventionalized meanings that hold for specific members, and, finally, possible implicatures that occur when specific members are used in certain interactional constellations. I hope to have shown that a contrastive approach can contribute to finding such a balance.

The Case of Dutch Toch and German Doch 71

REFERENCES Aijmer, K. and A. Simon-Vandenbergen (2003). The discourse particle well and its equivalents in Swedish and Dutch. Linguistics, 41, 1123–1161. Bakhtin, M. (1981). The Dialogic Imagination. University of Texas Press, Austin. Barth-Weingarten, D. (2003). Concession in Spoken English. On the Realisation of a Discourse-pragmatic Relation. Narr, Tübingen. Borst, D. (1985). Die affirmativen Modalpartikeln doch, ja und schon. Niemeyer, Tübingen. Brauße, U. (2001). Die kontextuellen Varianten des Konnektors doch. Ein Ausdruck von Relationen zwischen Widerspruch und Begründung. In: Wort und (Kon)text (P. Kocsány & A. Molnár, eds.), pp.151–171. Peter Lang, Frankfurt am Main. Couper-Kuhlen, E. (2000). Concessive patterns in conversation. In: Cause, Condition, Concession, Contrast. Cognitive and Discourse Perspectives (E. Couper-Kuhlen & B. Kortmann, eds.), pp. 381–410. Mouton de Gruyter, Berlin. Croft, W. and D. A. Cruse (2004). Cognitive Linguistics. Cambridge University Press, Cambridge. Doherty, M. (2001). Discourse relators and the beginnings of sentences in English and German. Languages in Contrast, 3, 223–251. Givón, T. (1979). On Understanding Grammar. Academic Press, New York. Graefen, G. (2000). Ein Beitrag zur Partikelanalyse. Beispiel: doch. Linguistik online, 6. http://viadrina.euv-frankfurt-o.de/~wjournal/2_00/graefen.html Hansen, M.-B. Mosegaard (1998). The Function of Discourse Particles. Benjamins, Amsterdam. ‘t Hart, M. (1983). De kroongetuige. De Arbeiderspers, Amsterdam. German translation by Marianne Holberg, 2001, Die schwarzen Vögel. Piper, München. Haspelmath, M. (2003). The geometry of grammatical meaning: Semantic maps and crosslinguistic comparison. In: The New Psychology of Language (M. Tomasello, ed.), Vol. 2, pp. 211–242. Lawrence Erlbaum, Mahwah, N.J. Hentschel, E. (1986). Funktion und Geschichte deutscher Partikeln. Ja, doch, halt und eben. Niemeyer, Tübingen. Kerbrat-Orecchioni, C. (2001). Oui, Non, Si : un trio célèbre et méconnu. Marges Linguistiques, 2, 95–119. http://www.marges-linguistiques.com Langacker, R. (2000). Grammar and Conceptualization. Mouton de Gruyter, Berlin. Lichtenberk, F. (1991). Semantic change and heterosemy in grammaticalization. Language, 67, 475–509. Molnár, A. (2002). Die Grammatikalisierung deutscher Modalpartikeln. Peter Lang, Frankfurt am Main. Önnerfors, O. (1997). Verb-erst-Deklarativsätze. Grammatik und Pragmatik. Almqvist & Wiksell International, Stockholm. Sekiguchi, T. (1939/1977). Was heißt doch? In: Aspekte der Modalpartikeln (H. Weydt, ed.), pp. 3–9. Niemeyer, Tübingen. Waltereit, R. (2001). Modal particles and their functional equivalents: A speech act-theoretic approach. Journal of Pragmatics, 33, 1391–1417.

72 Pragmatic Markers in Contrast Weydt, H. (1979). Immerhin. In: Die Partikeln der deutschen Sprache (H. Weydt, ed.), pp. 335–348. Walter de Gruyter, Berlin.


ON THE UNIVERSALITY OF DISCOURSE MARKERS Bruce Fraser, Boston University

1.

INTRODUCTION

In an earlier paper (Fraser, 1996) I proposed the functional class of Pragmatic Markers, linguistic forms that have meaning but do not contribute to the propositional content of the sentence of which they are a part. I suggested that there are four subclasses: (1) Basic Markers, such as please and I promise, which signal aspects of the direct literal basic message intended by the speaker; (2) Commentary Markers, such as frankly and certainly, which signal a message commenting on the basic message; (3) Parallel Markers, such as Sir, damned, and hey, which signal a message accompanying the basic message; and (4) Discourse Markers, such as and, but, and so, which signal how the discourse segment, of which they are a part, is semantically related to the previous segment. In this paper I am concerned with Contrastive Discourse Markers (CDMs), a sub-class of Discourse Markers, represented in bold type in the following examples, 1. a. b. c. d. e.

He started late. But he arrived on time, David rarely goes out. Instead, he just stays home and drinks. Susan is 6 foot tall. In contrast, her sister is only 4 feet, 11 inches. I’m a nurse. However, my husband won’t permit me to work. Give up? On the contrary, I’ve only begun to fight.

where the second segment, S2, has a type of contrastive relationship with the prior segment, S1. In particular, I am concerned with the Contrastive Discourse Marker but and whether it has an equivalent in other languages that functions as but does in English. I take the following as working hypotheses: 1: There is a single, primary Contrastive Discourse Markers in every language (in English, it is but). HYPOTHESIS

HYPOTHESIS

2: The uses of these primary CDMs are the same across

languages. The first hypothesis is presumably true although to my knowledge it has not been explored. The second hypothesis is worthy of study as research on DMs goes beyond individual terms


in a single language. If Hypothesis 2 is shown to be strongly supported, then, for example, but in English, aber in German, mais in French, and pero in Spanish, would all enjoy the same uses when they function as DMs. It would mean that the primary CDM for each language could occur in the following linguistic contexts: 2. a. b. c. d. e. f. g.

John is fat, ____ Mary is thin. I’m a nurse. ____ my husband won’t let me work. The flower was beautiful, ____ it was plastic. A: All the boys left. B: ____ there were only two boys to start with. A: I realize that John is sick. B: ____ John isn’t sick. John died yesterday, ____ he had been ill. A: John is home. B: ____ I just saw him at the store.

The choice of but as the first DM to examine as it occurs across languages seems reasonable, since I would predict that if there is a universality across languages with any DMs, markers like but, so, and and would be the prime candidates, in contrast to other more specific DMs such as rather, consequently, and furthermore. The purpose of this paper is threefold. First, I will present a statement of the different possible contrastive contexts in English and explore to what extent the DM but occurs in each of them.1 Second, I will describe the methodology used to gain the data that bears on these hypotheses. And third, I will present the progress to date on testing the two above hypotheses, showing the extent to which each are borne out and suggest how to account for apparent anomalies.

2.

THE CONTRASTIVE CONTEXTS OF ENGLISH AND THE OCCURRENCE OF BUT

In order to establish the distinctive contrastive contexts within English discourse, I considered the following set of CDM of English,2 3. all the same, (al)though, but, contrary to expectations, conversely, despite, even so, except, however, in spite of, in comparison (with), in contrast (to), instead (of), nevertheless, nonetheless, only, on the other hand, on the contrary, rather (than), or, regardless, still, to the contrary, whereas, while, yet and determined, when each CDM was present in discourse, what relationship between the segments is signaled. For example, in (4a),

1

I am not considering the use of but in connection with any other CDMs, for example, “We started late but arrived on time, nevertheless” or with DMs from another class, for example, “He restricted her to home for two weeks, but, in addition, took away her computer.” See Fraser (2005b) 2 This inventory of English CDMs is intended to be exhaustive although it excludes archaic cases like contrariwise, and notwithstanding.

On the Universality of Discourse Markers 75

4. a. I could go to the movies tonite. On the other hand, I could also stay home and watch TV. b. John doesn’t plays the piano anymore. Instead, he listens to opera on his radio. c. A: I regret John is sick. B: Except John isn’t sick. He just looks that way. the CDM on the other hand signals that there are two contrasting alternatives under consideration, one conveyed by S1, the other by S2; in (4b), instead signals that the segment following, S2, conveys what actually occurred (or will occur), in contrast to the information provided in S1, which conveys a potential action which did not (will not) occur. And in (4c), except signals that the segment following contradicts and rejects a presupposition derived from S1, namely, that “It is true that John is sick.” For these three examples, although certainly not for all cases, the contrastive contexts are mutually exclusive: on the other hand cannot occur in the context created in (4b) or (4c), and so forth. As a result of this examination of contrastive contexts in English, I have found that there is a major distinction – direct vs. indirect contrastive contexts – and within each type of contrast context, several subtypes. I treat them in turn. A. Direct contrastive context A direct contrastive context (DCC) occurs when, for a sequence S1 – CDM – S2, it is possible to compare the explicit interpretation of S2 with the explicit interpretation of S1 and thereby derive what can be construed as a meaningful semantic contrast of the type signaled by the particular CDM. For example, in sequences (5) containing the CDM on the other hand, a DCC exists if the hearer is able to find a semantic relationship between S2 and S1 which can be construed as an alternative. In (5a) the alternative relationship is obvious, but in (5b) I am unable to find an alternative contrast, in fact, any contrast at all. 5. a. I could go to the movies tonite. On the other hand, I could also stay home and watch TV. b. Susan worked for a long time on the puzzle. On the other hand, the pizza burned up. There are six different DCC in English, determined by the one or more CDMs with similar meanings which define them: (1) SIMPLE CONTRAST (but; not only…but also); (2) ALTERNATIVE CONTRAST (on the other hand; or); (3) COMPARATIVE CONTRAST (in comparison; in contrast); (4) PAIRED-OPPOSITE CONTRAST (conversely); (5) REPLACEMENT CONTRAST (instead; rather); and (6) CORRECTIVE CONTRAST (on the contrary). In each case, the resulting interpretation of the S1-CDM-S2 sequence is a contrast between S2 and S1 that the speaker considers to be worth noting.3

3 Some CDMs which signal the type of direct contrast are optionally present, some obligatorily. I will not address this here. See Fraser (2005a).


Simple direct contrast The simple type of DCC occurs when but is the CDM, where the CDM signals that the relationship between the segments constitutes a semantic contrast of some sort or another between S2 and S1, and where there is no other, more specific relationship intended. The examples in (6) are illustrative. 6. a. b. c. d. e.

Three is a prime number, but four is not. Enter quietly, but leave noisily. Harry plays soccer but not football. Jim is very smart, but Henry is very athletic. He gave toys to the boys but she gave dolls to the girls.

In each of these examples, when the two segments are contrasted a meaningful contrast is revealed. For example, in (6a) it’s prime numbers; in (6c) it’s sports Harry plays. This not the case with (7), 7. There is food on the table. But don’t you dare touch it. where there is no semantic contrast between the explicit meaning of the two segments.4 It is interesting to note that with a simple DCC, either order of the segments is permissible with no change in interpretation, and that only declarative and imperative sequences are acceptable. There is a variation of simple direct contrast, illustrated in (8). 8. a. b. c. d. e.

John is (not only) a cop but he is a lawyer also. Not only does Sam play baseball, but Mary plays too. (Not only) Take an apple, but take a pear also. (Not only) Don’t stand in the hall. But don’t sit down either. A: John left early. B: But didn’t Mary leave early also.

These sequences contain a compound DM, not only…but also, where the initial not only is often optional.5 This compound CDM, in contrast to the CDM but just discussed, is signaled by the presence of also or another form such as too, in addition, or either (when S1 is negative).6 Just as above, there is direct contrast here, between the two segments, although the not only may be only implied. Here, also, the segments are balanced, occurring in either order. Again, only declarative and imperative sequences are permissible.

4

This sort of contrastive context is discussed below. Note that in (8e) that the “not only” is obligatorily deleted when S2 is an interrogative. The also marker occurs optionally when the segments are conjoined with and, but with one salient difference: the sense of contrast is absent, reflected by the unacceptability of the not only. a) John is a cop and/but he is also a carpenter. b) John is a cop and/*but he is a carpenter. c) (Not only) is John a cop *and/but he is also a carpenter. 5 6


Alternative direct contrast The next three cases are refinements on simple DCC. The first is defined as occurring when the CDM on the other hand, on the one hand…on the other hand, alternatively, or or is present, where the CDM signals that the relationship between the segments constitutes alternative states or actions that are in contrast. This context is reflected in (9).7 9. a. I could go to the movies tonight. On the other hand I could (also) stay home and watch TV. b. John is very smart. On the other hand, Jim is very athletic. These options may be inherent in the meaning of the two segments, as in (9a). In this case, the meaning signaled by the CDM is redundant in which case it is deletable. Or the alternatives relationship may have to be provided by the discourse context. For example, the contrast between the two segments in sequence (9b) as it stands does not appear to provide alternatives. However, if the prior discourse were about needing a member for a role in a movie who was both smart and athletic, and if John were smart but non-athletic and Jim were athletic but not smart, then the sequence (9b) meets the requirements of on the other hand and is acceptable. The CDM but may occur in the alternative DCC. When the discourse segments S2 and S1 specifically reflect the alternative options in their explicit interpretations, as in (9a), the sequence with but receives the same interpretation as if on the other hand were present, namely, the but provides the contrast, the segments provide the alternatives. If, however, as in (9b), the segments do not provide the alterative relationship, the sequence with but may be interpreted as only simple contrast unless the discourse context fills in the details. Comparative direct contrast The third contrastive context is defined as occurring when in comparison (with/to this/that), by comparison, compared to, in contrast (with/to this/that), contrasted to/with is present, where the CDM signals that the relationship between the segments constitutes a comparison or a contrast. The sequences in (10) are illustrative. 10. a. b. c. d.

John weighs 180 pounds. In comparison/In contrast, Jim weighs 150. He isn’t even 5 feet tall. In comparison/In contrast, she is well over 6 feet. Iraq is a dictatorship. *In comparison/In contrast, the U.S. is a democracy. John can’t see very well. *In comparison/In contrast, he can hear perfectly.

Although the distinction between the CDMs in comparison and in contrast is hard to illustrate, I suggest (though with some hesitation) that the former occurs only when the value of the aspect is continuous (weight, height, intelligence, etc.) whereas the latter can occur 7 As in most of the cases, all members of a defining group of CDMs do not occur in every variation of the particular DCC at issue.


both with continuous and discrete variable (type of government, car, etc.). In this DCC, the sequences may be declarative only. The CDM but may occur in this DCC, subject to the same restrictions as were enumerated above for but occurring in the on the other hand DCC. Likewise, on the other hand may occur, subject to the same conditions. Paired-opposite direct contrast The fourth case of direct contrast is defined as occurring when alternatively or conversely is present, where the CDM signals that the relationship between S2 and S1 constitutes a paired opposition contrast. The sequences in (11) are illustrative. 11. a. What we gain in speed, we lose in sensitivity. Conversely, what we gain in sensitivity, we lose in speed. b. All athletes are not intellectuals. Alternatively all intellectuals are not athletes. Again, only declarative sequences are permitted. The CDMs but and on the other hand may occur in this context, subject to the conditions discussed above. However, in comparison/in contrast may not occur. Likewise, conversely may not occur in the comparison DCC. Replacement direct contrast The fifth direct DCC is defined by the CDMs instead, instead of (this/that), rather, rather than(this/that), where the CDM signals that S1 must represent an action or a state which did not/will not occur, in contrast with S2, which represents an action or a state which did/will occur. Looking at instead first, there are four ways in which the speaker can linguistically indicate that S1 did not/will not occur. These are illustrated in (12). 12. a. b. c. d.

John didn’t agree with her. Instead, he took exactly the opposite position. Mary rarely goes shopping. Instead, she stays at home and knits. She should have tried to open the door. Instead, she simply ignored the boys. He expected to win easily. Instead, he lost by a large margin.

In the first context, (12a), S1 contains an explicit, unincorporated negation, which is metalinguistic negation (Horn, 1985), where it is not the entire proposition that is rejected but rather an aspect of S1.8 In the second, (12b), S1 contains a negative adverb of time (never, rarely, seldom, infrequently,…). In the third, (12c), S1 contains a conditional modal (could, should or might) followed by have, implying that the action or state represented did not occur, while in the fourth, (12d), S1 contains a higher-order verb which implies the negative.9 8 If the verb contains incorporated negation, e.g., disagree, the sequence is not acceptable, presumably because to disagree, for example, is considered an action performed, while to not agree is considered the action of agreeing but not performed. 9 The modal would is not acceptable, as in “I would have done it. Instead, I just sat there and watched.”

On the Universality of Discourse Markers 79 Interestingly, instead does not occur when S1 represents a state, as in (13), which is a case of metalinguistics negation (Horn, 1985). 13. a. *Harry is not a policeman. Instead, he’s a security guard. b. *He wasn’t exactly crazy. Instead, he was just a little drunk. c. *John was not encouraged. Instead, he was seriously discouraged. A variation of instead, instead of (doing that), may be used when S1 contains an explicit negative with no modal, as illustrated in (14). 14. a. I didn’t participate. Instead, I just observed. b. Instead of participating, I just observed. The explicit negation is removed from S1 when the instead of form is used. Since this is not possible in the other three cases, (12b-d), this presumably accounts for the fact that they do not occur with instead of. While rather and instead are generally considered to be roughly synonymous, the context for rather is slightly different. Whereas the first two contexts of instead (explicit negation and negative adverbial, illustrated in (12a) and (12b), respectively) are acceptable for rather, the latter two contexts are not. However, when S1 represents a state, as in (13), rather is acceptable. There is a variation of rather, rather than (doing that), that may be used analogous to instead of only when S1 is an explicit negative without a modal. 15. a. I didn’t go. Rather, I just lounged around at home. b. I didn’t go. Rather than going, I just lounged around at home. c. Rather than going, I just lounged around at home. The CDM but occurs in three of the four direct replacement contexts illustrated in (12), but the interpretation of the sequences seems to be that of direct contrast rather than the sense of instead. Only in the third and fourth instead sequences, where the interpretation of S1 is interpretable as carrying a negative implication, does but occur, as in (16). 16. a. She should have tried to open the door. But she simply ignored the boys. b. He expected to win easily. But he lost by a large margin. Yet notice that the but in (16) is involved in simple direct contrast and the sequences do not carry the sense of instead. Corrective direct contrast This final DCC is defined by on the contrary/to the contrary, where the CDM signals that S1 constitutes the an action/state which is in correct/inaccurate, in contrast with S2 which constitutes the version of an action/state which is correct/accurate. There are two separate


cases: (1) where S1 and S2 are spoken by a single speaker; and (2) where S1 and S2 are spoken by different speakers. In the one-speaker sequences, as in (17), 17. a. George didn’t fight for his country. On the contrary, he never showed up at all. b. Mary didn’t make a trivial mistake. On the contrary, she made a horrendous error. c. Mary didn’t make a horrendous error. On the contrary, she made only a trivial mistake. there are two variations of contrast, albeit different from those contrasts already discussed. The first, as in (17a), requires the contrast to be the opposite value, or nearly so, of an S1 term (e.g., fighting v. not showing up).10 The second, as in (17b), requires the correction to be a different value along the same continuum. (e.g., trivial mistake v. horrendous error). The correction may go in either direction, as 17c) shows.11 For the one-speaker case, the speaker can signal S1 as being inaccurate in one of three ways: (1) by using metalinguistic negation, similar to the instead case above, where it is not the entire proposition that is rejected but rather an aspect of S1, as in (18a); 2) by asking a rhetorical question, as in (18b); and 3) by assigning the value of rejected term (gorgeous) as the responsibility of someone else, a view which can then be corrected by the speaker, as in (18c). 18. a. I'm not happy. On the contrary, I'm ecstatic! b. Throw the game? On the contrary, I’m going to go all out to win. c. You think she is gorgeous. On the contrary, I find her rather ordinary looking. The CDM but does not occur in this one-speaker DCC with the same interpretation as on the contrary. Where it does occur, for example (18c), the interpretation is contrary to expectation, discussed below. The two-speaker corrective contrast context is very different. Segment S1 consists of the first speaker’s contribution setting forth one message, while S2 consists of the second speaker’s message, contradicting the accuracy of the prior message. If S1 is negative, it is propositional not metalinguistic negation. As in the one-speaker context, on the contrary signals that the S2 segment is in the way of a correction to the facts asserted, advice given, implication of a question asked, as the sequences in (19) illustrate. 19. a. A: He has arrived. B: On the contrary, he won’t arrive for 45 minutes. b. A: Arrive on time. B: On the contrary, arrive a little late to avoid the rush. c. A: Don’t/Do you agree with me? B: On the contrary, I emphatically disagree with you. If but does occur in the corrective contexts, as it could in (19a), it signals a simple contrastive context, not a corrective one. If the sequence is declarative, as in (20), 10

Here, also, an incorporated negation is not possible. Interestingly, when S1 represents an assertion of a positive state attributed to another, a denial such as “That’s incorrect” or something analogous may replace on the contrary. You said that she made a trivial mistake. That’s incorrect/On the contrary, She made a horrendous error.

11


20. A: John is at home. B: On the contrary/You are wrong. I just saw him at the mall. the CDM can be replaced by an assertion of incorrectness. B. Indirect contrastive context An indirect contrastive context (ICC) occurs when, for a sequence S1 – CDM – S2, no meaningful direct contrast can be made between the explicit interpretations of S2 and S1. In this second type of contrastive context, a contradiction, one type of contrast, must be found between the explicit interpretation of S2 and an inference derived from S1. For example, sequence (21a), 21. a. I’m a certified nurse. But my husband won’t let me work. b. I am willing to work. does not permit a meaningful direct contrast. However, the second segment S2 (My husband won’t let me work) directly contradicts the inference (21b), which is one of many inferences which can be derived from S1. The interpretation such sequences, where the explicit interpretation of S2 contradicts with an inference of S1, is that the speaker intends to reject the inference as being irrelevant in this context.12 There are four semantic relationships of indirect contrast: (1) S2 contradicts a contextual inference of S1; (2) S2 contradicts a presupposition of S1; (3) S2 contradicts an entailment of S1; and (4) S2 contradicts an aspect of an illocutionary act conveyed by S1. I will treat these cases in turn. S2 Contradicts a contextual inference By far the most robust of these four types is the ICC which is defined by the presence of although, but, contrary to expectations, despite, except, however, in spite of, nevertheless, nonetheless, only, still, whereas, while or yet, where a contextual inference is derived from S1 and contradicted by the explicit interpretation of S2, with the result that the inference from S1 is rejected as irrelevant in this context. This process where there is indirect contrast is, of course, different from that of direct contrast, where the sequences is interpreted as a contrast of messages with the interpretation that the particular contrast stands. There are several sub-cases of contextual inference. The first of these sub-cases is defined by the CDMs but, contrary to expectation, except, only, and however, where the meaning of the CDM is roughly “contrary to expectation,” and the inference derived from S1 reflects this CDM meaning. These are illustrated in (22)13. 22. a. I’m a certified nurse. However/Only I can’t work since I have a cold. b. It’s raining. But/However/*Only/*Except I want to go out since I need some fresh air. 12 13

Bell called this “cancellation” and Blakemore refers to the process as “contradiction and elimination.” The CDMs although, whereas, and while are subordinate conjunctions and are not considered here.


c. d. e. f. g.

We arrived late to the party. Contrary to expectations, no one seemed to mind. Take off all your clothes. But/Except/Only don’t do it right now. John left early. But/However/Except didn’t Mary leave then too? A: Where did he go? B: But/*Only/*However, why do you want to know? A: Stop the car right now. B: But/*However/*Only I’ll skid if I try to stop.

The above sequences show that there is great flexibility for the types of sequences (e.g., declarative – interrogative) but except for but, these CDMs are not uniformly acceptable in all contextual inferential contexts. For example, only or except don’t occur in imperativedeclarative sequences. Note also that neither on the other hand or any of the other direct contrast CDMs occur in these cases. (See Fraser, 2005b for details.) When but is used, there may be a direct contrast possible, as in (22d), thus rendering the sequence pragmatically ambiguous until the discourse context is brought to bear. In some of these sequence, for example (22a), segment S2 consists of the contrasting message (“I can’t work”) but, in addition, specifies the reason for the message (“I have a cold”), which may also constitute the entire S2: “But I have a cold.” Such sequences illustrate that, in contrast to the earlier variations, two distinct S2 contents are generally possible: (1) content which contradict the contextual inference; (2) content which provides a reason for the contradiction. A more restrictive-variation of indirect contrast involving contextual inference is defined by the presence of the CDMs all the same, but, despite, even so, in spite of, nevertheless, still, and yet, where S1 is always a factual statement, S2 is also a declarative segment, and where the interpretation of the CDMs is “despite S1.”14 Like the first sub-type, S2 contradicts a contextual inference of S1 and reject it. This is illustrated in (23). 23. a. b. c. d.

We left late. But/Nevertheless, we arrived on time. Yes I love you. In spite of that/Nevertheless, I won’t clean up your messes. It’s a terrible thing to say. Yet/All the same/Nevertheless, I wish he were dead. She isn’t very intelligent. Nevertheless/Still/Despite that, I like her.

When S1 is not factual, the sequence is typically unacceptable. Thus, in (24), where S1 is an opinion, the sequence is typically unacceptable. 24. John is probably right. But/*Nevertheless, not quite right. In some cases of a contextual inference, the S1 may be contextual, not verbal (See Blakemore, 2002; Fraser, 2001). For example, suppose a large teenage bully is taking a bicycle from the bike rack at school. His non-verbal message is “I am taking this bike whether you like it or not.” When the bike’s owner, a smaller teenaged boy cries, 25. But you can’t do that. That’s MY BIKE!

14

However and other CDM discussed above may occur in this ICC, but they signals a “contrary to expectation” interpretation.


he is rejecting the contextual inference of the bully’s actions which can be assumed to be “I can take this bike.” It is interesting that however nor any of the other CDMs in this group do not occur in this context. Still another variation, which uses predominantly but and however, is reflected in (26). 26. a. A: John was chosen to be a MacArthur fellow. B: But/However (this is under standable, since) he is very smart. b. A: John died. But/However, (after all) (this was expected, since) he had been very ill. c. A: John died. B: But/However, why is this surprising? He had been very ill. d. She speaks flawless German. But/However, this is not surprising, given she is German. where the inference of S1 is that the content of the message is surprising/unsurprising (reasonable/understandable), and the primary message of S2, perhaps elided, is that is was not so unsurprising/surprising. There is often an anaphoric this/that referring back to S1 and there is a marked intonation pattern to signify the interpretation. Only declarative-declarative and declarative-interrogatory sequences occur with this interpretation. Of course, the reason has to justify the (un)surprise, as shown in (27). 27. *John died. But this is understandable since he was in good health. S2 Contradicts an entailment of S1 This variation is defined by the presence of the CDM but, except, and only (the latter two being more restricted), where segment S2 contradicts an entailment derived from S1, with the interpretation that the entailment is rejected. 28. a. A: Some of the boys left. B: But/Except only one of the boys left. [“Some” entails “more than two.”] b. A: I dried the dishes tonight. B: But how is that possible? You didn’t move a muscle. [“Drying dishes” entails “physical activity.”] c. A: Why did John murder Harry? B: But/Except Harry’s not dead, only sleeping. [“Murder” entails “killing someone.”] These sequences require two speakers, there is often a reference to S1 with a this/that, and, like the however sequences discussed above, S2 may include the reason or justification for the rejection. S2 is typically spoken with a marked, final-rising intonation with the interpretation of “I don’t believe you.”


S2 Contradicts a presupposition of S1 This variation is defined by the presence of but, except, and only, where S2 contradicts a presupposition derived from S1 with the interpretation that the presupposition is rejected.15 As above, the interpretation is “You are wrong/incorrect/misinformed/etc.” 29. a. A: I regret John is sick. B: But/Only John isn’t sick. He just looks that way. [“Regret” presupposes the truth of the propositional complement embedded under it.] b. A: When did he die? B: But/Except he didn’t die. He just left town. [When-questions presuppose that truth of the sentence proposition.] c. A: He is the King of France. B: But there’s no King of France. [Reference to a specific entity presupposes the existence of that entity.] As above, two speakers are required and the reason for the contradiction is often present. S2 Contradicts a felicity condition on a speech act convey by S1 This variation is also defined by the CDM but, except, only, and, for representative speech acts (e.g., claims, assertions, stipulations, etc.) on the contrary, where S2 contradicts a felicity conditions on a speech (illocutionary) act conveyed by S1.16 For the sequences in (30), 30. a. A: I apologize for knocking over that vase. B: But/Except you weren’t responsible, since you weren’t even here. [One who apologizes is responsible for the offensive act] b. A: I order you to sit down. B: But, you don’t have the right to order me. [One who orders has the right to do so.] c. A: I will give you roses for your birthday. B: But/Only you know you can’t do that. [One who promises is capable of carrying out the promise.] d. A: John is home. B: But that’s impossible, since I just say him at the mall [One who claims is assumed to be speaking truthfully.] In all these sequences, the rejections must be of a felicity condition associated with a direct rather than an implicated (indirect) illocutionary act. Thus in (31a), 31. a. A: It’s warm in here. B: But you don’t have any basis to assert that. b. A: It’s warm in here. ( => Turn down the heat. ) B: *But you know I can’t do that. the direct claim with its felicity condition of truthfulness can be rejected whereas in (31b), the implied request to turn down the heat, which has a speaker felicity condition of hearer ability to do so, cannot be rejected.

15

I am referring to pragmatic not logical presupposition. The defining conditions on an Illocutionary Act must be satisfied if the act is to be successful. The Felicity Conditions must be satisfied if the act is both successful and without defect. Cf. Bach & Harnish (1979).

16


C. Summary of contrastive contexts and the occurrence of but In all, there are 10 distinct contrastive contexts in English. Of the DCC, the first four contexts (simple contrast, alternative contrast, comparative context, and pair-opposition contrast) all permit but to occur, with the interpretation of the sequences subject to the restrictions stated. The interpretation of the sequence is that the segments stand in contrast in accordance with the individual CDM meaning. The fifth (replacement contrast) context permits but, but without the instead sense, while the sixth (corrective contrast) context does not permit but to occur at all, although some sequences appear that way. Of the ICC (arising from S2 contradicting with a contextual inference, an entailment, a presupposition, or a felicity condition derived from S1), but may occur in all, signaling contradiction with the inference of S1 while rejecting the inference which is rejected as irrelevant for this discourse. I note in passing that the sequencing of DMs is fairly involved, including the sequencing of but with other CDMs, as illustrated in (32). 32. a. I could go to the movies tonite. But, on the other hand, I could also stay home and watch TV. b. John weighs 180 pounds, but, in contrast, Joe weighs 130. c. We arrived late. But no one seemed to mind, however. d. All athletes are not intellectual. But, conversely, all intellectuals are not athletes. e. John rarely goes out, but instead, stays home and drinks. But may usually occur as the first of two CDMs in sequence, although in some cases the sequence sound more acceptable if the second CDM is placed finally in S2, as in (32c), and others combinations are downright unacceptable, e.g., but – except and but—on the contrary. In any event, the sequencing of but is not considered in this paper.

3.

NON-DM USES OF BUT

Among the non-DM uses of but there are several systematic uses and then a variety of other, idiomatic uses. I will briefly examine them. A. Systematic uses of but as a non-dm The first non-DM use of but is that of a topic change pragmatic marker (cf. Fraser, 1996). This group of markers, which includes (but, by the way, incidentally, while I think of it, lest I forget), has the function of signaling a reorientation of the conversational topic. As such, it does not signal a semantic relationship between S1 and S2 and therefore the members of this group of lexical items are not DMs. The following sequences are illustrative. Note that but is not always acceptable as the marker. 33. a. A: I had dinner with George last night. B: But/Incidentally did you get paid the money? b. I promise to go. But/By the way, leave me alone for the time being.


c. A: The dinner was delicious. B: But/Lest I forget, where was Tim this evening? d. This is a typical July day in New England. (*But) Cathy Fuller, by the way, is going to sit in for Ron Della Chiesa as your host this afternoon. The second non-DM use occurs when but is a preposition, roughly equivalent to except, with the exception of, apart from, aside from, excepting, excluding, save, etc. For this use, the head of the PP must contain a definite NP or a determinate or indefinite pronoun of absolute meaning such as not all, nothing, no one, everyone, every, any, each, everything as opposed to some, many, a couple, a few, etc. The interpretation is that the object of the preposition but is the sole exception to the information represented in the other segment.17 34. a. b. c. d. e. f.

Everyone/nobody/*somebody but John was released to go home. Who but/except our old from Tom should turn up. Except for/*but taking her to the circus, nothing would satisfy her. He can’t tolerate anything but/except classical music. Come anytime but/except now You are anything/nothing but a liar.

The third use of non-DM but is that of an adverb, where it is synonymous with only, simply, just, and sometimes merely, as shown in the following examples. 35. a. b. c. d. e.

I have but/only/just a moment. It seems but/just yesterday. I saw Jack but/only once. A: will you win. B: I can but try. He is but/merely a child.

The fourth systematic use of non-DM but occurs with the combination all but (nearly/almost/ practicably/almost) where the interpretation is just less than the scope content. 36. a. b. c. d.

He has all but/nearly clinched the championship. He all but/nearly strangled me. The paper money in Russia is all but/practicably worthless. They are finding it all but/almost impossible to make a living.

The fifth and final systematic use of non-DM but is illustrated by the following sequences, where S1 must be an apology and carry a deferential tone. 37. a. b. c. d. e. f). 17

I may be wrong, but… I apologize for saying this, but don’t you think you should drive more slowly? I don’t mean to interfere, but isn’t that a little bit dangerous? Pardon me, but where is the john? I don’t want you to think that I’m biased, but this is the way I was brought up. I’m not sure if this is relevant, but isn’t that bag leaking.

Note that except for but not but can occur initially


I note that and is not permitted in these sequences, 38. a. I don’t mean to interfere, *and isn’t that a little dangerous? b. Pardon me for saying so, *and you have bad breath.

B. Non-systematic uses of but The other uses of non-DM but are must less systematic, some being unique idioms. They are listed below without comment. 39. a. but good. I’ll get you but good. b. If I could (would/had) but Verb… If I could but explain. c. Rhetorical question What belief is so foolish but some will embrace it? d. can (could) not help but wonder if … I can’t/couldn’t help but wonder if we did the right thing yesterday. e. but of course A: Is it done? B: (But) of course it’s done/*it’s not done. f. but the thing that/what pleases him/her most/least He is happy with his work, with his marriage, and with his children. But what/ the thing that pleases him most is the fact that he has finally stopped drinking. g. but above all Thanks are due to John and Mary. But above all, I want to thank Harry. h. but for = except for But for the grace of God, that was I.

4.

THE SURVEY

In order the determine if other languages have a primary CDM equivalent to but and, if so, if they pattern similar to but in English, I constructed a survey which was sent to native speakers of Arabic, Catalan, Chinese, Danish, Dutch, Finnish, French, German, Greek, Hebrew, Japanese, Korean, Norwegian, Spanish, and Swedish. At least two people responded from each language. The survey was constructed to include: 40. Those contexts that accept the DM use of but in English John is not fat ___ is very thin. We started late ___ we arrived on time.


41. Those contexts that reject the DM use of but in English I rarely should have tried to open the door. _____ I simply ignored them boys. Mary didn’t make a trivial mistake. _____ she made a horrendous error. 42. Those use of but that are not DMs in English All left ___ one. He is ___ a child. I’ll get him ___ good. I used mostly declarative sequences both because of space limitation on the survey, but also on the assumption that if the “but” term in another language was viable, it would have a greater chance with declarative forms. There were a number of deficiencies in the initial survey. (See the Appendix at the end of this paper for the specific survey questions.) Foremost among them was the fact that I used the contrastive contexts of English (as presented above), and didn’t consider if the contrastive contexts in other languages were different. Second, there was no guarantee that the respondents were in fact truly bilingual, which might have biased their judgments. Third, after subsequent work on the general problem, I realized that I had left certain critical areas uncovered. And fourth, the level of linguistic sophistication varied greatly among the respondents. Dispite this, the initial results were very encouraging. All speakers identified a primary CDM equivalent to but in English. With a few exceptions, all the uses of but as a DM in English were judged to be similarly used by at least one of the individuals from each language. In many cases, all agreed. Where there was disagreement between languagespeakers, it was noted and subsequent surveys will attempt to tease out the reason(s) for this difference in judgement. The most consistent disagreement was with sequence such as 43. John is not tall but short. where speakers from a variety of languages indicated that an alternative form (Spanish: pero/sino; German: aber/sondern; Hebrew: ela/aval) was required in this context. This was to be expected. There was no other outstanding exception. The sequences of non-DM uses of but provided some curious results. Three of the sequences were surprisingly judged acceptable across many of the languages. 44. a. You may think I’m crazy, ____ where is the dog? b. Thanks go to John and Mary. ____ above all, I want to thank Harry. c. You have to get up early. ____ after all, how much sleep do you need? with (44a) acceptable to nearly everyone. These results were unexpected and will require further research.


5.

CONCLUSION

As has been obvious, this was an initial effort to ascertain if there is a universal CDM in all languages equivalent to but in English and, if so, whether the uses are the same or nearly so. While the results are promising, considerable revision in terms of the deficiencies stated above and re-surveying of English-Languagei bilinguals as well as examination of Languagei vs. Languagej, before any definite conclusions can be drawn.

REFERENCES Bach, K & R Harnish (1979). Linguistic Communication and Speech Acts. MIT Press, Cambridge, MA Bell, D. (1998). Cancellative discourse markers: a core/periphery approach. Pragmatics 8(4), 515-541. Blakemore, D. (2002). Relevance and Linguistic Meaning: The Semantics and Pragmatics of Discourse Markers. Cambridge Studies in Linguistics 99. Cambridge University Press, Cambridge Fraser, B. (1996). Pragmatic markers. Pragmatics 6(1), 167-190. Fraser, B. (2001). The case of the empty S1. J of Pragmatics 33(10), 1625-1630. Fraser, B. (2005a). To appear. Towards a theory of discourse markers. In Approaches to Discourse Markers (K. Fischer, ed.), Studies in Pragmatics Series, Elsevier Press, Amsterdam. Fraser, B. (2005b). To appear. Sequencing of discourse markers. Horn, L. (1985). Metalinguistic negation and pragmatic ambiguity. Language 61, 121-174.


APPENDIX I am soliciting your help in learning the extent to which there is an equivalent of the English term but in your language (e.g. Spanish: pero, German: aber) and if so, if it enjoys the same possibility of uses. For example, will the but of your language fit into the contexts below: John is fat,____ Mary is thin. A: John is here. B: ____that can’t be right; I just saw him at the mall. I’m a nurse, ____ my husband won’t let me work. Lunch is on the table. ____don’t touch anything yet. At this time, I am only interested in whether or not but occurs in various linguistic environments, not whether there are other alternatives that are possible. In what follows, I have set down some simple sequences in which the but from your language might be used. Please pick the term in your language which corresponds to but, and then test whether or not it could occur in the space provided in the sequence. A simple “N” or “Y” will suffice. If the term for but does not occur in a particular sequence (e.g., in Spanish, pero does not occur in “John is not tall ___ short” while sino does), please mark this with a N. Please note: I have included some examples sequences in which but does not occur acceptably in English. Your equivalent of but may or may not occur in this context. Don’t consider English acceptability when you mark your answer. Don’t worry about punctuation, e.g., the absence or presence of a comma or a period, or whether the sequence sounds better if the segments are combined into one sentence. Thanks for your help.


Questionnaire: Name: ________________________ First Language: __________ but Term: _________ DM Use of but in English 1. John is fat, ____ Mary is thin. 2. John is not tall ____ short. 3. I suggest that you take some cookies, ____ leave the cake alone. 4. Take an orange, ____ leave the apples alone. 5. You promise to help me. ____ you let me down. 6. John is a cop, ____ he’s also a carpenter. 7. Take a peach, ____ take an apple too. 8. Don’t move your hand, ____ don’t talk either. 9. Take one, ____ don’t take more than one. 10. I’d take more, ____ I’m full. 11. I could give you this book, ____ frankly, I don’t want to. 12. I’m a nurse. ____ my husband won’t let me work. 13. I know this bus goes to Count Hall. ____ does it also go to the gym 14. The lunch is on the table. ____ don’t touch anything. 15. A: It’s warm in here. B: ____ turn up the heat anyway. 16. A: What time is it? B: ____ don’t you have a watch? 17. A: What time is it? B: ____ why do you want to know? 18. Take a drink, ____ be careful you don’t spill any. 19. I cut a finger yesterday, ____ it wasn’t mine. 20. John died yesterday, ____ he had been ill. 21. A: John speaks perfect German. B: ____ he was raised in Germany. 22. A: John was assassinated in Dakar. B: ____ I always thought he died of natural causes. 23. A: Now you know all the facts. B: ____ I’m still not convinced he is guilty. 24. Say what you will, ____ you’re not going to make me get upset. 25. The flower was beautiful, ____ it was plastic. 26. A: Consider this triangle. B: ____ this figure has four sides. 27. A: All the boys left. B: ____ there were only two boys to start with. 28. A: I realize that John is sick. B: ____ John isn’t sick. 29. A: When did he die? ____ he didn’t die. he only left town. 30. A: I apologize for disturbing you. B: ____ you have nothing to apologize for. 31. A: John is home. B: ____ I just say him at the store. 32. I should have tried to open the door. _____ I simply ignored the boys. 33. Mary didn’t make a trivial mistake. _____ she made a horrendous error.


Not Systematic Uses of but that might occur in your language 1. Everyone ____ John was here. 2. Nothing ____ soda would satisfy her. 3. He is nothing ____ a child. 4. I have ____ a moment. 5. If you could ____ explain, I’d be very grateful. 6. You may think I’m crazy, ____ where is the dog? 7. I’m not sure if this is relevant, ____ isn’t that bag leaking? 8. A: Is it finished. B: ____ of course it’s done. 9. A: John didn’t leave. B: ____ of course he didn’t leave. 10. He has all ____ clinched the championship. 11. I can’t help ____ obey her. 12. Thanks are due to John and Mary. ____ above all, I want to thank Harry. 13. ____ for the grace of God, there go I. 14. I will get you ____ good. 15. You have to get up early. ____ after all, how much sleep do you need.


“NOT NOW” – ON NONCORRESPONDENCE BETWEEN THE COGNATE ADVERBS NOW AND NÅ Hilde Hasselgård, University of Oslo

1.

INTRODUCTION

The Norwegian nå and the English now are cognate adverbs that have the same etymology and the same basic meaning; i.e. reference to present time.1 Both words are frequent, but as shown in Table 1, below, the figures for nå are much higher than those for now. Such a discrepancy cannot be due to chance; it must reflect a difference between the languages as regards the meaning and the use of these seemingly equivalent words. A parallel corpus such as the English-Norwegian Parallel Corpus (ENPC) is ideal for researching such differences as it contains equal amounts of original texts in both languages with translations into the other. There are 50 source text extracts of about 10,000–15,000 words for both languages, of which 30 are fictional and 20 are non-fictional. Both original and translated texts have been culled from published sources.2 As is becoming standard practice in lexical studies based on parallel corpora (cf. Johansson, 1998), I refer to translations of a word as correspondences, which also includes ‘zero’ correspondences; i.e. the word has been omitted in the translation. It may be noted that one can also talk about the correspondences of a word in going from translations to original text. The concern of the present study, however, is to compare usage in original English and Norwegian in order to establish any differences in meaning between nå and now and to examine what happens to the words in translation into the other language. The main focus of the study is on non-correspondences; i.e. cases in which nå has not been translated with now and vice versa. The material for the study thus consists of all the instances in the ENPC where the two words do not correspond to each other. The paper has been organized according to the different meanings and discourse functions of nå and now.

2.

THE OCCURRENCE OF NOW AND NÅ IN THE ENPC

Searches in the ENPC show a surprisingly high degree of non-correspondence between the cognate adverbs now and nå. The words differ in the extent to which they get a congruent translation (i.e. nå translated into now and vice versa); Norwegian nå corresponds to 1 2

See e.g. The Oxford English Dictionary and Bokmålsordboka. See further Johansson (1998) and http://www.hf.uio.no/iba/omc.


something other than now in almost 40% of the cases, while now corresponds to something other than nå in 26% of the cases. (See Table 1.)

Table 1: The occurrence of nå and now in the English-Norwegian Parallel Corpus.3

nå/now total nå now N N % Fiction 684 175 English 893 235 26.3 Non-fiction 209 60 Fiction 982 363 Norwegian 1245 493 39.6 Non-fiction 263 130 There is also a certain difference between the genres: the overall frequency of nå/now is higher in fiction, and the degree of correspondence is lower in non-fiction, particularly going from Norwegian to English. Genre differences will not, however, be a major concern of the present study. The following factors can be hypothesized as possible explanations of the divergence between nå and now: • Nå/now with time reference may be considered redundant because present time reference has been established by the verb phrase. The languages may differ in the extent to which they allow this redundancy. Translators may be less tolerant of redundancy than source text writers and thus omit the redundant word from the translation (zero correspondence). • Both nå and now can be discourse markers, but their functions seem to be different; i.e. now as a discourse marker is generally a continuative, while nå can be a modal particle (definitions are given below). This may have consequences for the use of nå/now as a discourse marker in general, and certainly for the way in which a discourse marker is rendered in translation. • Nå/now may be part of an idiomatic collocation which is translated as a unit. The corresponding unit need not include nå/now (e.g. nå for tiden – these days; right now – med en gang). The present study will consider these and (possibly) other factors in an attempt to describe the different uses of now and nå in English and Norwegian.

3

The Norwegian adverb/particle nå (‘now’) has been separated from the homonymous verb nå (‘reach’). This has been done automatically using the tagged ENPC for the originals and manually for the translations. The variant forms no and nu have been included with nå. The figures in Table 1 thus cannot be replicated using a standard lexical search in the ENPC.

Non-correspondence Between the Cognate Adverbs now and nå 95

3.

THE MEANINGS AND FUNCTIONS OF NÅ/NOW

Clearly, the core meaning of nå and now is reference to present time. The temporal use can safely be considered primary to other uses, and is listed as the first/primary meaning in the Oxford English Dictionary (OED) as well as Norwegian dictionaries. It may be noted that both nå and now occur with tenses other than the present. Thus, the potential for referring to time can be extended beyond the present, especially in narrative texts, where nå/now refers to reference time rather than to the time of speaking. The temporal meaning of nå and now is illustrated by (1) and (2), respectively.4 We may note the past tense in (2). 1. Begge konstablene står nå på huk og tar av seg luene. (LSC1) Lit: ‘Both constables stand now crouched…’ Both of the policemen crouch down and take off their caps. 2. Naturally, he knew who owned them now, everyone in Kingsmarkham knew that. (RR1) Han visste selvfølgelig hvem den nåværende eieren var. Det visste alle i Kingsmarkham. Lit: ‘He knew of.course who the present owner was.’ Both now and nå also have other uses, where the temporal meaning is either bleached or completely absent. The ‘non-temporal’ now is described by Schiffrin (1987) as a discourse marker, used for marking a shift in ‘idea structure, orientation and participant footing’ (1987: 244). Along similar lines, Aijmer (2002) describes the discourse particle now as a ‘topic changer’. The same use is described by Halliday (1994: 53) as continuative, signalling that ‘a new move is beginning’, or ‘a move to the next point if the same speaker is continuing’. This use is illustrated in example (3). According to Bokmålsordboka, nå can have a similar function though its use seems to be more limited; i.e. the word is classified as an interjection, expressing surprise, impatience, irritation, etc; cf. (4). 3. Now don’t shout, or do anything stupid. (RDA1) Vær så snill å ikke begynne å skrike eller gjøre noe dumt. Lit: ‘Please don’t begin to shout or do anything stupid.’ 4. “Nå, ja, så får dere komme, da.” (TTH1) Lit: ‘Now, yes, then may you come, then.’ “Well, you can all come then.” A further meaning of nå, not shared by now, is epistemic modality, as noted by e.g. DavidsenNielsen (1996: 303) and Faarlund et al (1997: 824). Characteristics of modal particles, according to Faarlund et al is that they express the speaker’s attitude to the proposition, the relationship between the proposition and the real world, and the speaker’s relationship with the hearer (ibid.). This ties in with the description of disjunct adverbials given by e.g. Quirk et 4

All examples are from the ENPC unless otherwise stated. The original is given first, followed by the translation. Examples in Norwegian – whether original or translated – are followed directly by a word-for-word translation, marked as ‘lit[eral]’.


al (1985: 620 ff). In brief, nå can function as a modal particle, somewhat similar to jo (Swedish ju, German ja), basically signalling that the information is well known by the speaker, but not necessarily shared by the hearer. 5. Og da ville misunnelige sjeler med skadefryd kunne konstatere at geniet Leonardo, ja han kunne nå ikke skrive selv de enkleste ord riktig! (ANR1) Lit: ‘…the genius Leonardo, well he could [nå] not write even the simplest words correctly’ Envious souls would then be able to state with malice that the genius, Leonardo, could not even write the simplest word correctly! In a systemic-functional perspective, the textual functions of nå/now can be described along the three metafunctions: as a time adverb, nå/now functions as a marker of temporal circumstance and is thus part of the ideational (transitivity) system. Both nå and now have extended functions as textual elements, fulfilling the role of continuative (Halliday 1994: 53). Furthermore, the modal particle use of Norwegian now can be seen as an extension along the interpersonal dimension, where modality belongs (Halliday 1994: 88ff).

Table 2: Distribution of discourse functions of now (not ‘nå’) and nå (not ‘now’) (ENPC fiction and non-fiction)

now nå Discourse functions of now and nå N % N % Time reference 173 73.6 387 78.5 Discourse marker 54 23.0 98 19.9 Ambiguous 8 3.4 8 1.6 Total 235 100 493 100 In the present study, the textual and interpersonal meanings of nå/now will be subsumed under the term ‘discourse marker’, as opposed to its more literal function of giving a time reference. The distribution of the discourse functions of now and nå is shown in Table 2. Both nå and now can occur as part of more or less fixed collocations, e.g. in a nominal use as complement of a preposition, or other types of fixed or semi-fixed phrases (now and then, right now, nå for tiden, nå igjen). In the analysis, the whole expression in which nå/now is a part has been classified as temporal or as a discourse marker. In both languages the temporal function dominates. As noted by both Schiffrin (1987) and Aijmer (2002), the distinction between the temporal now and the discourse marker now is not always clear-cut, not even when considering the wider context; thus a handful of examples have been categorized as ‘ambiguous’. Although the absolute frequencies differ between the languages, the proportional distribution of discourse functions is relatively similar. Syntactically, nå and now function at clause level as time adverbials, continuatives, and modal particles; and at phrase level as complements of prepositions (a nominal use), and less commonly, as modifier of an adjective or noun (e.g. the now empty covers of the book) and as

Non-correspondence Between the Cognate Adverbs now and nå 97 (part of) a conjunction (now that). To some extent, the syntactic function correlates with discourse function, i.e. the temporal nå/now typically functions as (part of) a time adverbial, while the textual nå/now functions as continuative and the interpersonal nå as modal particle. The distribution of syntactic functions is shown in Table 3. In both languages the time adverbial function is dominant, followed by continuative in English and by modal particle in Norwegian. The languages thus differ most markedly in the use of nå/now as a discourse marker.

Table 3: Syntactic functions of now/nå

now nå N % N % time adverbial 153 65.1 383 77.7 continuative 55 23.4 8 1.6 modal particle 0 0 87 17.6 complement of preposition 19 8.1 12 2.4 modifier in NP/AdjP 3 1.3 1 0.2 conjunction 2 0.9 0 0 other 3 1.3 2 0.4 TOTAL 235 100.1 493 99.9

4.

NÅ/NOW AND TENSE

Table 4 shows the form of the verb phrase in clauses with nå/now. Temporal and nontemporal meanings of nå/now have been listed separately, as this distinction was believed to be significant for the co-occurrence with different verb forms. It was noted above that both nå and now can occur with tenses other than the present. As shown in Table 4, only about a third of nå/now co-occur with the present tense. With the temporal meaning of nå/now, the distribution of tenses is fairly similar in the two languages. Surprisingly, the past tense is most frequent in both English and Norwegian. This must be related to the fact that the material consists of written texts, mainly narrative, where the past tense is dominant.5 As regards the non-temporal uses of nå/now, the languages differ. In English, where the non-temporal use is mainly textual/continuative, now occurs most frequently in imperative clauses, and not particularly frequently in clauses with the past tense. In Norwegian, however, the interpersonal nå occurs frequently with both the past and the present tense, but not so frequently with imperatives. Note, however, that the imperative seems to be over-represented with the non-temporal nå/now in both languages, as compared to the temporal use.

5 By contrast, in 100 random examples of now in the British National Corpus, 53 occurred in clauses with the present tense, and only 20 with the past tense.


Table 4: Verb forms in clauses with now/nå

temporal now non-temporal now temporal nå non-temporal nå N % N % N % N % present tense 55 31.8 18 29.0 154 39.8 33 31.1 past tense 80 46.2 9 14.5 164 42.4 42 39.6 modal 18 10.4 6 9.7 52 13.4 15 14.2 imperative 2 1.2 22 35.5 4 1.0 9 8.5 non-finite 9 5.2 0 0 2 0.5 0 0 n.a. 9 5.2 7 11.3 11 2.8 7 6.6 Total 173 100.0 62 100.0 387 99.9 106 100.0 A note should be made of the categories in Table 4. Both present and past tense include perfective and progressive verb phrases. ‘Modal’ comprises all verb phrases containing a modal auxiliary, including future-referring expressions in both languages. ‘Non-finite’ includes participles (present and past) and infinitives. The fact that Norwegian lacks a participle form corresponding to the English –ing form probably explains the lower frequency of the non-finite category in Norwegian. ‘N.a’ (not applicable) indicates that nå/now is not part of a clause structure.

Table 5: Non-congruent correspondences of nå and now.

zero untranslated clause/sentence6 other time expression discourse marker other TOTAL

5.

Correspondences of nå (not ‘now’) N % 328 66.5

Correspondences of now (not ‘nå’) N % 128 54.5

34

6.9

8

3.4

83 22 26 493

16.8 4.5 5.3 100.0

68 19 12 235

28.9 8.1 5.1 100.0

CORRESPONDENCES OF NÅ AND NOW

Table 5 shows the overall distribution of (non-)correspondences of nå and now in the EnglishNorwegian Parallel Corpus. Apart from the striking difference in frequency between the two words, it may be noted that zero is a more frequent option as a correspondence of the Norwegian nå than of the English now, while ‘other time expression’ occurs more frequently as a correspondence of the English now. Otherwise the relative distribution of correspondences is rather similar. A full list of correspondences is given in Appendix 1.

6

I.e. the clause or sentence in which nå/now occurs is not found in the translated text.


Since both nå and now have distinct meanings and uses depending on discourse function, the following sections will discuss the time adverb and the discourse marker separately.

6.

TEMPORAL USES

A brief study of the ENPC material where nå and now are each other’s translations indicates that most of these instances contain the temporal nå/now. The two words thus seem to be well matched in function and meaning when used as time adverbs. Still, a substantial number of temporal nå/now also have non-congruent correspondences. These are shown in Table 6.

Table 6: Correspondences of temporal nå/now

Correspondences of nå Correspondences of now N % N % zero 253 65.4 90 52.0 untranslated clause/sentence 28 7.2 8 4.6 other time expression 82 21.2 67 38.7 discourse marker 4 1.0 0 0 other 20 5.2 8 4.6 TOTAL 387 100.0 173 99.9 The high proportion of zero correspondence with the temporal nå/now in both directions of translation indicates that translators have considered the word redundant for various reasons. In translation from Norwegian to English, the zero option will bring the target text closer to the frequency norms for English; cf. Table 1. However, this is not likely to be the main reason for zero correspondence, since the same tendency is seen in the opposite direction of translation, where the principle of ‘normalization’ would lead to a number of added nå. It is thus more likely that the avoidance of perceived redundancy is simply a feature of the translation process. Having said that, there seems to be a connection between the aspect and tense of the verb phrase and zero correspondence of the temporal nå/now. Fretheim (1983: 99) notes that nå will often seem redundant in a clause where the verb phrase is perfective. According to Fretheim this is due to nå encompassing the time of the utterance in all cases where the finite verb is in the present tense. However, this seems to be the case also when nå is used in combination with the past tense, as in (6), where the time adverb signals simultaneity with reference time rather than with utterance time. The shift to perfective aspect in the translation makes the time adverb redundant. 6. Nå var Nord-Amerikas østkyst konstatert av oppdagerne. (KP1) Lit: ‘Now was North-America’s east.coast determined by the explorers. The explorers had determined the extent of the American coast.


7. It was late afternoon now. (MM1) Det var blitt sent på ettermiddagen. Lit: ‘It had become late on the.afternoon.’ The same phenomenon is seen in translations going from English into Norwegian, as shown by (7). In addition to being perfective, the verb phrase in the translation of (7) is inchoative. The inchoative seems to have the same effect as the perfective as regards the redundancy of nå/now, thus giving a double reason why the adverb has been left out in the translation. While English and Norwegian have similar resources for expressing inchoation and perfectivity, only English has a grammaticalized progressive aspect. This, too, seems to have a bearing on the use and non-use of the temporal nå/now. Both adverbs can be used to mark stages in a sequence of actions, as in (8). However, this usage seems to be more common in Norwegian than in English. A possible reason is that since English has the progressive aspect as an option, a series of non-progressive verb forms automatically imply consecutive events (cf. Quirk et al, 1985: 187), and so the adverb is not needed to mark the sequence. 8. Berit hadde satt fram en skål potetgull, og nå knaste vi dem mellom tennene og sendte urolige blikk rundt i rommet. (LSC2) Lit: ‘Berit had set out a bowl [of] crisps, and now crunched we them…’ Berit set out a bowl of potato chips, and we crunched them between our teeth and sent restless looks around the room. 9. Han hadde steget ned fra hesten, og nå leide han den mot bekken. (SH1) Lit: ‘He had dismounted from the.horse, and now led he it towards the.stream.’ He had dismounted and was going towards the stream. By contrast, in (9) the use of nå in the Norwegian original seems to have triggered the progressive aspect in the translation, since it marks an action as ongoing and incomplete. Evidently, this is precisely the meaning conveyed by the progressive verb phrase in the translation, and thus the time adverb is redundant. The temporal sequence between the clauses is clearly marked by the perfective aspect in the first clause in both the original and the translation. The tense of the verb phrase seems to be relevant to the extent to which nå/now is regarded as redundant. The two directions of translation show slightly different tendencies; nå is more often omitted in translation into English in combination with the present tense than now is in translation into Norwegian (98 times vs. 22, i.e. 38% vs. 26%). In combination with the past tense, however, the proportion of omitted nå/now is greater going from English to Norwegian than vice versa (55 vs. 107 instances, i.e. 65% vs. 40%). As shown in (8) and (9) this may also be linked to aspectual differences, particularly the progressive/non-progressive contrast. The second most frequent correspondence of temporal nå/now is a lexical equivalent other than nå/now. Example (10) is a case in point, showing then as a correspondence of now.

Non-correspondence Between the Cognate Adverbs now and nå 101 10. Nå forlangte han å bli vist alterbildet som han hadde hørt om. (JW1) Lit: ‘Now demanded he to be shown …’ He then demanded to be shown the altarpiece that he had heard about. In clauses with the past tense, the meaning of now approximates that of then, marking a stage in a sequence of events. One should thus perhaps not be surprised to find then as a correspondence of nå 17 times in the corpus. Similarly, there are 13 examples of English now being translated by Norwegian så or da (‘then’). This type of correspondence may be seen as a kind of normalization, i.e. the translator chooses an adverb which seems to harmonize better with a past tense verb. There are also other examples of nå/now being replaced by another general time adverb that seems to be in harmony with the verb phrase, as in (11), where the Norwegian adverb lenger (‘any longer’) seems to fit better with the negative verb phrase, although nå would have been possible. 11. At one time, he lived on Stanley Place, but he’s not there now. (SG1) En gang i tiden bodde han i Stanley Place, men der holder han ikke til lenger. Lit: ‘… but there lives he not [any] longer.’ Many of the lexical correspondences give a more precise time reference than the deictic nå/now and may thus be regarded as cases of explicitation. Examples are expressions such as this time (example (12)), at the moment and in a minute instead of now, and i øyeblikket (‘at the moment’), med en gang (‘immediately’), and i dag (‘today’) instead of nå. Example (13) illustrates the most frequent lexical correspondence of now, namely i dag, which occurs in 19 out of the 67 instances where now corresponds to ‘other time expression’ (cf. Table 6). There are also a couple of examples in which nå is translated by today, showing the obvious relationship between these two pairs of words. Generally, the languages do not seem to differ markedly as regards the type or the relative frequency of lexical correspondences. A full list of the lexical correspondences of nå/now is given in Appendix 1. 12. Nå kikket hun rundt seg, men hun så ingen. (JG1) Lit: ‘Now looked she around herself, but she saw nobody.’ This time she looked all around but there was nobody in sight. 13. When life started on Earth some 3.8 billion years ago, the Sun was about 30 percent less luminous than now. (JL1) Da livet på Jorden begynte for omtrent 3,8 milliarder år siden, var solen omtrent 30 prosent mindre lyssterk enn i dag. Lit: ‘… was the.sun about 30 percent less luminous than today.’ When nå/now occurs as part of an idiom or collocation, its meaning tends to be temporal. Typical examples are prepositional phrases with nå/now (e.g. by now, til nå), as shown in (14), nå/now modified by an adverb (e.g. even now, akkurat nå); cf. example (15), now coordinated with another time expression (e.g. now and then), and nå modified by a prepositional phrase (e.g. nå for tiden). Most of these idioms/collocations correspond to a temporal expression in the translations. The second most frequent correspondence is zero,

102 Pragmatic Markers in Contrast though zero correspondences are much less frequent with the collocations than with nå/now on its own. Collocations involving nå have a zero correspondence in 23% of the cases of noncongruent translations (11 out of 47) while those involving now have a zero correspondence in 17% of the cases (7 out of 41). The most likely reason for the low proportion of zero correspondences is that the collocations have more ideational content than nå/now on its own and thus are more likely to be preserved in translation. 14. Up till now she had felt only an overwhelming relief that all would be well. (PDJ3) Hittil hadde hun bare følt stor lettelse over at alt ville gå godt. Lit: ‘Hitherto had she only felt great relief…’ 15. Først nå hadde postbudet vært der med dagens post. (JG1) Lit: ‘only now had the.mailman been there with the.day’s mail.’ The mailman had just delivered the day’s mail.

Table 7: Correspondences of nå/now as discourse markers

zero untranslated clause/sentence time expression other discourse marker7 other TOTAL

7.

Correspondences of Correspondences of Correspondences of continuative now continuative nå modal particle nå N % N % N % 34 61.8 1 12.5 68 78.2 0

0

2

25.0

3

3.4

0

0

0

0

0

0

18

32.7

5

62.5

11

12.6

3 55

5.5 100.0

0 8

0 100.0

5 87

5.7 99.9

NÅ AND NOW AS DISCOURSE MARKERS

When nå/now functions as a discourse marker, direct translation is difficult because the discourse meanings of nå and now are quite different. As shown in Table 3 above, the continuative function is common with the English now, but quite rare with Norwegian nå. On the other hand, Norwegian nå has a frequent use as a modal particle, which is not matched in English.8 These differences are reflected in Table 7, showing the correspondences of the 7

The term ‘discourse marker’ has been used liberally in the classification of translation correspondences and includes disjunct and conjunct adverbials. For original material the term is used more restrictively, covering continuative and modal uses only. 8 Aijmer (2002: 93-95) describes a modal particle use of the discourse marker now, typically occurring with imperatives (e.g. “now, wait a moment”), and typically conveying a subjective / affective meaning. This is of course close to my analysis of nå as a modal particle with imperatives. However, I have analysed this use as


discourse markers nå and now in translations. (Only Norwegian has a column for the modal particle function for the obvious reason that this function is not found in English.) When nå/now functions as a discourse marker function it only rarely occurs as part of a collocation. The only examples are there now, nå vel, nå engang and nå igjen. (See Appendix 2 for an overview.) 7.1

Nå/now as continuative

As continuatives, both nå and now typically occur in clause-initial position. The temporal meaning of the adverb is bleached or altogether absent, i.e. nå/now does not primarily refer to a point of time, but rather to a stage in the discourse (cf. Hasselgård 1997: 177). The Cambridge Advanced Learner’s Dictionary prefaces the entry for the discourse marker now by ‘in speech’, and explains that the word is ‘used in statements and questions to introduce or give emphasis to what you are saying’, ‘said when you want to make someone feel better or give them a gentle warning’, or ‘said to attract attention to what you are going to ask or suggest’. In the present material, continuative nå/now is found most frequently in dialogue sequences in fiction and can safely be described as interactive. (Only five of the 55 examples of continuative now come from non-fiction; three of these are from rather fiction-like texts, while two come from argumentative texts.) The low number of Norwegian nå with continuative function seems to reflect the general situation in Norwegian: a search in Norwegian original texts in the ENPC for initially placed instances of nå corresponding to now yielded only a small handful of examples, which indicates that this use is rare in this type of material. A further check of 200 random examples of clause-initial nå in the Oslo Corpus of tagged Norwegian texts gave no examples at all of the continuative use. 9 Moreover, the ‘Big Brother Corpus’ (of informal spoken Norwegian)10 contained only three out of 121 examples of clause-initial nå that could be classified as continuative, and two of these occurred in combination with ja. It is probably impossible to ascertain why nå is so rare as a continuative. However, the fact that time adverbs – among them nå – occur frequently and unmarkedly in initial position in Norwegian (Hasselgård 1997: 184) may contribute to blocking the discourse marker reading of a clause-initial nå. As indicated by the Norwegian dictionaries (Bokmålsordboka and Aschehoug og Gyldendals store norske ordbok), the continuative use of nå is often coupled with another marker, typically vel or ja, written as two words or one, as shown in (4) above and (16) below. 16. Nåvel, men da burde jo du være eksperten her ved bordet. (GS1) Lit: ‘Now.well, but then should [particle] you be the.expert here at the.table’ Well, then – you ought to be the expert around here.

continuative. This is due to clause-initial position, and the fact that its use does not seem to be much different from the use of now as a continuative in other contexts, i.e. it typically marks a change of topic or a new turn. 9 The Oslo Corpus contains 18.5 words of written Norwegian. See http://www.tekstlab.uio.no/norsk/bokmaal/. 10 The Big Brother Corpus contains about 150,000 words transcribed from the Norwegian version of the reality show ‘Big Brother’. See further http://www.tekstlab.uio.no/talespraak/bigbrother/.


It may be noted that both nåvel and nåja are twice as frequent in translations from English as in original material, though the numbers in both types of material are very low. Both words are used as translations of English continuatives, particularly well, in addition to now. Given the scarcity of the Norwegian continuative nå it is perhaps only to be expected that zero correspondences would be frequent in translations from English. It is all the more noteworthy that the few examples of continuative nå that do occur in original Norwegian should be given translations other than now. This indicates that nå and now differ in meaning when they function as continuatives. According to Aijmer (2002: 95) the discourse marker now typically signals a boundary in the discourse and/or a change of topic. The Norwegian continuative nå, on the other hand, seems to have an affective meaning (impatience, irritation, surprise) when it occurs on its own, typically as part of a response or as a marker of agreement in combination with vel or ja.11 The most frequent correspondence of the continuative nå is in fact well. The continuative well is described by Schiffrin (1987: 126– 127) as a ‘response marker’ – a way of negotiating e.g. disagreement, non-compliance with a request, self-correction, or simply to accomplish coherence, thus doing other types of discourse work than now. In English, the continuative function of now is quite common, cf. Table 3. Stenström (1994: 85) subsumes the discourse marker now under ‘framing’, frames being a device for signalling ‘that a message is on the way or that there will be a change of topics’. Example (17) illustrates this use. In (17), the continuative initiates the utterance and marks the start of a new topic, though not the termination of a previous one. It has been rendered in Norwegian by the continuative ja (‘yes’). Other correspondences of the continuative now include nei (‘no’), vel (‘well’), nåvel (lit: ‘now well’). 17. He was eating a nice pork pie, and when the waitress offered him dessert he said, “Oh, now, let me see, maybe I will try some at that,” … (AT1) Han spiste en lekker svinepai, og da oppvartningspiken kom med dessert til ham, sa han: “Ja, la meg se, kanskje jeg skal prøve litt av det der,” … Lit: ‘Yes, let me see …’ Some examples seem to combine the temporal meaning and the discourse marker function of now, i.e. the word may be ambiguous between a time adverb and a continuative. I would still interpret this meaning of now primarily as ‘at this stage in the discourse’, rather than ‘at this point of time’. As described by Schiffrin (1987: 244) and Aijmer (2002: 70), it signals a topic shift, or a new move in the discourse. Example (18), which occurs paragraph-initially, is a case in point. In the English original the topic shift (or more precisely, the return to a previous topic) is signalled by the discourse marker now, while the Norwegian translator has used the contrastive connector men (‘but’) to signal this.

11 Cf Bokmålsordboka on nåvel: brukt i konstaterende el. oppsummerende uttr: (‘ used in concluding or summarizing expressions’), and Aschehoug og Gyldendals store norske ordbok on nåja: uttrykker betinget enighet, tvil, innrømmende konstatering (‘expresses conditional agreement, doubt, reluctant conclusion’).


18. Now, Rembrandt had debts that he could not meet. (JH1) Men Rembrandt hadde altså gjeld som han ikke kunne gjøre opp. Lit: ‘but Rembrandt had as.we.know debts…’ Twenty out of the 55 cases of continuative now occur in imperative clauses, as in (19). In such contexts, the continuative seems to be more of an attention-getter than a topic-changer. As noted by Aijmer (2002: 93) it acts as an intensifier, expressing ‘the speaker’s involvement with the hearer’. The temporal meaning of now is not completely absent; the speaker of (19) obviously wants the requested action to take place immediately. Furthermore, now with imperatives can sometimes be taken to signal a stage in a series of actions. The most conspicuous function of now in (19), however, is that of the continuative: drawing attention to the ensuing message. 19. Now keep your nasty mouth shut so we can all watch this programme in peace. (RD1) Lukk den ekle lille munnen din så vi kan få se på dette TV-programmet i fred og ro. Lit: ‘Shut that nasty little mouth [of] yours …’ Another use of the affective now (Aijmer 2002: 93) is illustrated by (20). It occurs here in combination with there, though according to the Cambridge Advanced Learner’s English Dictionary it can also occur on its own or with a repeated now to ‘make someone feel better or give them a gentle warning’. The soothing effect is most likely achieved because now expresses the speaker’s personal involvement. 20. “There now, it’s nothing to fret about. (MW1) “Så, så, det er ikke noe å plage seg med. 21. “Did he now?” (RDO1) – Gjorde han det? Lit: ‘Did he that?’ Example (21) has a clearly non-temporal now in end position. Despite its non-initial position, it has been analysed as a continuative, because it marks a cohesive link to preceding context. However, this use of now approximates the meaning of modal particles, as it seems to add a subjective flavour to the utterance (see below for features of the modal particle nå). For instance one might substitute really without much difference in meaning. (21) is the only example of its kind in the material, which makes it hard to establish this use as a separate category of meaning. A search in the spoken part of the British National Corpus, however, (for ‘you’ + ‘now’) gave a few similar examples. All the cases where a clause-final now was clearly non-temporal were questions, such as (22). 22. Raymond: Jean:

I’d a love a vodka and coke! I’d give al-- almost anything. Would you now? (BNC: KDN 2570)


23. “Well, okay now, thanks a lot,” I called, easing toward the door. (SG1) “Javel, greit da, mange takk,” ropte jeg og trakk meg mot døren. Lit: ‘“yes well, okay then, many thanks”… A similar use is found in examples like (23), where now seems to be relatively devoid of meaning, but functions as a filler (Stenström, 1994: 69), and perhaps adds some affective overtones. The Norwegian correspondence in this case is da (‘then’), a particle which according to Solberg (1990: 79) can function as an emphasizer showing the speaker’s sincerity. It should be noted that the non-initial now with discourse marker function is very rare in the English material for this study, with only five examples (as against 49 initial ones), four of which occur in sentence fragments, as in (23). To sum up, the Norwegian nå is rare as a continuative. When it does occur, it is often coupled with another particle, especially vel and ja. Thus the meaning of the continuative will depend on the combination of particles; nåvel is summative or conclusive, while nåja expresses reluctant agreement or doubt. The English continuative now, on the other hand, is quite frequent, and also polysemous. Its main use is that of a topic changer (Aijmer 2002), as in (18). Further it may be used as an attention-getter, or to establish ‘control of the conversational floor’ (ibid.: 93). The discourse marker now can also sound reassuring or soothing, as in (20), and possibly in (21)–(23), where it clearly resembles modal particles. 7.2

Nå as modal particle

The modal particle nå is said to signal speaker orientation. It is described as a booster, “emphasizing what the speaker knows and that s/he will insist on and convey to the hearer” (Faarlund et al, 1997: 824, echoing Fretheim, 1981 and Solberg, 1990).12 Thus in (24), the speaker confidently presents a proposition as a given fact. As Solberg (1990: 57) points out, the speaker does not imply that the information is known to the hearer (in which case the particle da (‘then’) would have been used instead). 24. Men graut er nå kost for trell og småbonde både hverdag og helgedag. (TTH1) Lit: ‘But porridge is now food for slave and small.farmer …’ But still, porridge is the only food slaves and small farmers have every single day. However, there are uses of the modal nå in the material that do not quite fit this description. Quite frequently, modal nå occurs in a non-factive environment, i.e. in direct and indirect questions (25)–(26), conditionals (27), and imperatives (28). 25. “Hvor gjorde jeg nå av den boken?” (KF1) Lit: ‘Where did I now [put] that book?’ “Where did I put that book?”

12 My translation. The original is: “Forsterkaren no er ego-orientert. Han framhevar det som sendaren veit, og som ho/han vil insistere på og formidle til mottakaren.”


26. Han fór rundt og studerte fugler, hva nå det kunne være godt for. (HW2) Lit: ‘He ran around and studied birds, what now that could be good for’ He had scurried around studying birds, for whatever good that might do. 27. Hvis jeg nå hadde vært der, hva så? (KA1) Lit: ‘If I now had been there, what then?’ Assuming I was there, what about it? 28. Mor deg nå riktig godt, og skriv når du får lyst. (EHA1) Lit: ‘Enjoy yourself now really well …’ Enjoy yourself a lot and write when you feel like it. In combination with a question word (a hv-word, corresponding to the English wh-words) the particle nå is probably still ego-oriented, but rather than signalling speaker’s knowledge, it signals that the speaker thinks s/he ought to know. This is illustrated in (25), where the speaker seems to ask himself as much as the hearer, and in (26), where the speaker tries to imagine a possible reason for running around studying birds. It may be noted that the combination of a hv-word and nå often corresponds to whatever, whoever etc. The use of modal nå in conditionals seems to establish a counter-factual scenario, which is well represented by the translation of (27). I.e. the speaker considers for a moment that something may be the case, while at the same time rejecting the possibility. This may have to do with what Solberg (1990: 60) refers to as the argumentative or defiant meaning of the particle, i.e. the particle can mark a contrast to a known fact. In combination with the imperative the particle nå seems to signal insistence and personal involvement on the part of the speaker. It is thus, as the Norwegian reference grammar indicates, a booster. The meaning contributed by nå is something like ‘you must know that I really mean this’. While the message given in (28) is a pleasant one, the booster nå can also be used in other contexts, as in (29), where the ‘personal involvement’ consists of irritation and impatience. This use is not completely dissociated from the temporal meaning of nå; most likely the use of nå in (29) also indicates that the mother wants the discussion to stop immediately, though if reference to present time were the only meaning of nå, it would have been placed clause-finally (‘Hold opp nå’). 29. “Hold nå opp!” roper moren og reiser seg med et brak. (LSC1) Lit: ‘Hold now up’ “Stop it!” yells Mother, rising with a start. It is interesting that both now and nå as discourse markers frequently co-occur with imperatives (cf. above). However, their functions are not identical. Most importantly, the clause-medial position of the Norwegian particle excludes the function of attention-getter, which was established as a function of the continuative now. Furthermore, like all modal particles, the clause-medial nå with imperatives is most likely to be unaccented; cf. Faarlund et al (1997: 824). Thus the continuative now emphasizes the whole message, or turn, while the modal particle nå gives the message a subjective flavour, which can be linked to the ego-


orientation mentioned above; here it emphasizes what the speaker knows that s/he wants and insists on conveying to the hearer (ibid: 824). A further use of the particle nå seems to have a concessive meaning. This use differs from the typical modal particle in that it occurs initially rather than medially in the clause. It often collocates with other modal particles such as jo or vel, or with modal disjuncts such as selvfølgelig (‘of course’), but it also occurs on its own, as shown in (30). It may be noted that the wider context excludes a temporal reading of nå in (30). Rather, the particle implies that the speaker regards the proposition as given, and wants to show the hearer that he is aware of this. 30. Nå kan man ikke snakke om forgangen skjønnhet i vraket av en borerigg. (KT1) Lit: ‘Now can one not talk about faded beauty…’ One cannot really talk about faded beauty in the wreck of a drilling rig. It may be argued that this use of the particle nå should be seen as a continuative rather than a modal particle. However, it implies the kind of defiance or contrast typical of the modal nå, as described by Solberg (1990: 60), thus bringing in a subjective perspective and contributing more to the interpersonal than to the textual domain. We may note that the concessive nå triggers subject-operator inversion due to the verb-second constraint, thus behaving syntactically as an adverbial rather than as a continuative.13 Lexical correspondences occur in only 16 of the 87 examples of modal nå. These are: well, anyway, certainly, what in the world, just, still, obviously, actually, in fact, so, after all, in any case, even if. While (30) has really as a correspondence of the concessive nå, the translation of (31) makes the epistemic meaning of the modal nå explicit by means of the modal adverb certainly. 31. Jeg for min del synes nå det er ufredstider hverdag som helgedag her i huset. (TTH1) Lit: ‘I for my part think [nå] there are unpeaceful times … For my part there certainly is enough fighting and disturbances in this house every single day, and that includes the holy days. The central meaning of the modal nå is associated with ‘ego-orientation’, i.e. the speaker’s involvement with the message. The exact meaning seems to be dependent on the communicative function of the message. In the use shown in (24) as well as (30), nå marks the message – a statement – as well-known or self-evident to the speaker, who then insists on imparting this knowledge to the hearer. Nå in questions seems to direct the question to the speaker as much as to the hearer. In imperatives the modal nå marks the message as subjective, and perhaps to some extent predictable or evident from the situational context. Finally, a clause-initial nå marks concession – the speaker’s acknowledgement that the proposition is counter-factual.

13 In Norwegian, as in all Germanic languages other than English, the finite verb has to be the second constituent in main declarative sentences. Thus, subject-operator inversion occurs after fronted non-subject constituents except conjunctions, continuatives and other peripheral clause elements such as vocatives and interjections.


8.

CONCLUDING REMARKS

The main conclusion to be drawn from this study is that discourse function is crucial to the translation of nå/now. Interestingly, the time adverb has developed different discourse functions in the two languages, and works chiefly as a textual marker (continuative) in English and as an interpersonal, modal particle in Norwegian. Table 8 is an attempt to sum up the meanings and uses of now and nå. The categories of meaning are not absolute, and the boundaries between them are not watertight. In particular, the primary meaning of ‘present time’ is evident in many uses of nå/now. The meaning of the English continuative can be glossed as ‘now listen’ or ‘now I’ll turn to this’, while the modal nå, though primarily an emphasizer, also indicates ‘do this now’ when occurring in imperatives. Furthermore, when the continuative now indicates empathy or speaker sincerity, it is reminiscent of the egoorientation of the modal particle. On the other hand, the concessive nå clearly has textual functions although syntactically it is not a continuative, and it is speaker oriented.

Table 8: Meanings of now and nå

IDEATIONAL

(TEMPORAL)

NOW

NÅ

utterance time reference time current stage in sequence of events

utterance time reference time current stage in sequence of events

INTERPERSONAL (MODAL)

‘ego-orientation’ personal involvement marker of ‘given fact’ (speaker’s knowledge) emphasizer / booster counter-factual scenario concession

TEXTUAL empathy, speaker sincerity (CONTINUATIVE) launching new turn topic shift / change calling attention to message/turn

emotional response-marker in combination with vel, ja: summative / signal of (reluctant) agreement

Among the non-correspondences of nå/now identified in this study, zero was the most frequent choice for all the discourse functions. However, the reason for omitting nå/now will vary, since the continuative now is not easily matched in Norwegian, and the modal particle nå has no direct counterpart in English. The temporal nå/now, however, is more easily matched, and thus dominates the material (not comprised by this study) where nå and now correspond. The study also revealed that most of the lexical correspondences of nå/now (other than now/nå) are time expressions. Some such translations (e.g nå ‘at that moment’) can be


regarded as explicitation, while others may be chosen to create better harmony between the time adverb and the verb phrase. The following hypotheses were presented at the outset of this study: • Temporal nå/now may be considered redundant because present time reference has been established by the verb phrase. The languages may differ in the extent to which they allow this redundancy. Translators may be less tolerant of redundancy than source text writers and thus omit the redundant word from the translation. • The different discourse functions of now and nå (continuative vs. modal particle) may have consequences for the use of nå/now as a discourse marker in general, and certainly for the way in which a discourse marker is rendered in translation. • If nå/now is part of a collocation which is translated as a unit, the corresponding unit need not include nå/now. The study has shown all of these points to be relevant. However, some refinement is needed. Obviously, a temporal nå/now may seem redundant, as it can generally be worked out from the context. However, some factors seem to contribute to the redundancy of nå/now, particularly the presence of the perfective aspect. We have also seen that temporal nå may have a zero correspondence in connection with a progressive verb phrase in English. Similarly, an inchoative verb phrase seems to make nå/now redundant. Since the temporal now is rather more frequent in Norwegian than in English, it is likely that this type of redundancy is more acceptable to Norwegians. I would still claim that the noncorrespondence of the temporal nå/now is mainly due to translation phenomena, i.e. the translators omit redundant features or opt for a more explicit alternative. It seems plausible that the translators should be less tolerant of redundancy than source text writers; in most cases nå/now is just as redundant in the source text as it would have been in the translation. The discourse marker functions of nå/now were even more different in the two languages than expected, so that nå and now can hardly ever be each other’s translation. The continuative function differs markedly in frequency between the languages, and seems to be more clearly lexicalized – or pragmaticalized – in English. This may be due to the fact that time adverbials occur more frequently in initial position in Norwegian than in English, and are thus less marked. Thus an initially placed nå, even if unstressed, will more easily be interpreted as temporal, unless it is accompanied by another continuative marker, such as ja or vel. Furthermore, to the extent that nå occurs as a continuative in Norwegian, it seems to be closer to the ‘response-marker’ well (cf. Schiffrin, 1987: 127) than to now, especially in the compounds nåvel and nåja. The modal particle function of nå is not paralleled in English, with the possible exception of the clause-final now illustrated in (21)–(23). As with the related modal particles nok and vel, the typical correspondence is zero (cf. Johansson & Løken, 1997: 104). To the extent that lexical correspondences of the modal particle occur in the English translations, they are mostly disjunct or conjunct adverbials, such as really or anyway. As regards nå/now as part of an idiomatic expression, I thought originally that this factor would account for a greater share of the examples than it actually did. In fact, such expressions made up only 17% of the English material and 10% of the Norwegian. As


mentioned earlier, most of these expressions were temporal, and generally translated by another temporal expression. The proportion of zero translations was lower for this group than for nå/now on its own, probably due to the fact that they have more lexical content. In sum, the temporal uses of now/nå have been found to be relatively similar in English and Norwegian, while the same words as discourse markers are rather different, although related uses can be found among both the continuative and the modal meanings. The cross-linguistic corpus method used in the present study has proved useful in discovering meanings and uses of nå and now. The frequent non-congruent translations of these innocent-looking words indicate a discrepancy between the words and throw the meanings of nå/now into relief. As a translation represents an interpretation of the source text, non-congruent correspondences reveal a lot about the meaning of nå/now, thus providing an excellent basis for establishing the range of meanings and uses of these polysemous words as they are used in their proper context to shape messages according to the needs of speakers and hearers.

REFERENCES Aijmer, K. (2002). English Discourse Particles: Evidence from a Corpus. John Benjamins, Amsterdam / Philadelphia. Aschehoug og Gyldendals store norske ordbok. (1991). Kunnskapsforlaget, Oslo. Baker, M. (1992). In other Words. A Coursebook on Translation. Routledge, London/New York. Bokmålsordboka (1993). Universitetsforlaget, Oslo. Cambridge Advanced Learner’s Dictionary on CD -ROM, Version 1.0. (2003). Cambridge University Press, Cambridge. Davidsen-Nielsen, N. (1996). Discourse Particles in Danish. In: Content, Expression and Structure: Studies in Danish Functional Grammar (E. Engberg Pedersen et al, eds.), 283–314. John Benjamins, Amsterdam / Philadelphia. Faarlund, J. T., S. Lie and K. I. Vannebo (1997). Norsk Referansegrammatikk. Universitetsforlaget, Oslo. Fretheim, T. (1981). ‘Ego’-dempere og ‘alter’-dempere. Maal og Minne, 86–100. Fretheim, T. (1983). Perfektum og det temporale ‘DA’ og ‘NÅ’. Norsk Lingvistisk Tidsskrift 2, 97–113. Halliday, M.A.K. (1994). An Introduction to Functional Grammar. 2nd ed. Edward Arnold, London. Hasselgård, H. (1997). Time and space adverbials in English and in Norwegian, with special reference to initial position. Norsk Lingvistisk Tidsskrift 15, 165–189. Johansson, S. (1998). On the role of corpora in cross-linguistic research. In: Corpora and Cross-linguistic Research (S. Johansson and S. Oksefjell, eds.), 3–24. Rodopi, Amsterdam.


Johansson, S. and B. Løken. (1997). Some Norwegian discourse particles and their English correspondences. In: Sounds, Structures and Senses. Essays Presented to Niels Davidsen-Nielsen on the Occasion of His Sixtieth Birthday (C. Bache and A. Klinge, eds.), 149–170. Odense University Press, Odense. Oxford English Dictionary, OED Online . Quirk, R, S. Greenbaum, G. Leech and J. Svartvik. (1985). A Comprehensive Grammar of the English Language. Longman, London. Schiffrin, D. (1987). Discourse Markers. Cambridge University Press, Cambridge. Solberg, T. K. (1990). Modalpartikler i norsk. Hovedoppgave (MA thesis), University of Oslo. Stenström, A.-B. (1994). An Introduction to Spoken Interaction. Longman, London.

Sources of material: The English-Norwegian Parallel Corpus: http://www.hf.uio.no/iba/prosjekt/ The British National Corpus: http://www.natcorp.ox.ac.uk/ The Oslo Corpus of Tagged Norwegian Texts: http://www.tekstlab.uio.no/norsk/bokmaal/english.html The Big Brother Corpus: http://www.tekstlab.uio.no/talespraak/bigbrother/


Appendix 1: List of lexical correspondences of nå/now found in translations. Nontemporal correspondences are marked by a #. Lexical correspondences of NÅ (not ‘now’) – 62 types, 131 tokens

Lexical correspondences of NOW (not ‘nå’) – 48 types, 99 tokens

actually # after all # again already and # any longer any more anyway # at that juncture at that moment at that time at the moment at this moment at this point at this stage certainly # current even if # every once in a while from then on from then onwards here (#) hitherto in a minute in a way # in any case # in fact # in the event (#) it’s time for just late

allerede alt altså # av i dag av og til bare så vidt # da dette # en gang i blant en gang imellom etter hvert for øyeblikket fra tid til annen før først hele tiden her (#) her og der (#) heretter hittil i dag i øyeblikket igjen innimellom ja # lenger med det samme med en gang men # nei # når

later much longer next no longer not until then nowadays obviously # oh # present since then so (#) so far still (#) suddenly tag question # temporal subclause the next day the big moment then these days this # this early this first this moment this time thus far today up to then well # what in the world # when

nåvel # nåværende og # på dette tidspunkt se så # ser du # slik # straks så (#) så lenge så, så # sånn # sånn ja # til dags dato temporal subclause ved nærmere ettertanke # vel #

Appendix 2: List of collocations including nå/now in the original texts. Non-temporal expressions are marked by a #. NOW as part of collocation in original texts (40 tokens, 15 types)

NÅ as part of collocation in original texts (47 tokens, 16 types)

before now by now even now every now and then for now from now on now and again now and then

akkurat nå allerede nå fra nå av først nå i neste nå i samme nå ikke nå lenger inntil nå

now that only now right now there now # until now up till now up to now

nå avdøde nå engang # nå for tida/tiden nå i det siste nå igjen nå ja # nå om dagen til nå



HOW WELL CAN WELL BE TRANSLATED? ON THE ENGLISH DISCOURSE PARTICLE WELL AND ITS CORRESPONDENCES IN NORWEGIAN AND GERMAN1 Stig Johansson, University of Oslo

1.

INTRODUCTION

The discourse particle well is an enigmatic word which has attracted the attention of a great many scholars. What does it actually mean? As shown by Diane Blakemore (2002: 128ff.), interpretations range from ‘all is not well’ to ‘all is well’. In a paper from 1979 Jan-Ola Östman suggests that a cross-linguistic perspective may facilitate the description of what he calls textual particles in “a) widening our views about the phenomenon itself; and (b) aiding both language-specific and contrastive-linguistic description” (Östman, 1979: 177). To what extent can the meaning of well be illuminated by a cross-linguistic study? To what extent can its meaning be conveyed in other languages? These are the questions addressed in the present paper. In a recent cross-linguistic study Karin Aijmer and Anne-Marie Simon-Vandenbergen (2003) investigate the English discourse particle well and its translations into Swedish and Dutch. Focusing on the role of translations in revealing the function of well, they claim that well is “a heteroglossic option, accommodating the utterance to the context, in particular the hearer’s expectations” (p. 1128). In another recent paper, which also deals with well in a translational perspective, Carla Bazzanella and Lucia Morra (2000) focus on the indeterminacy of translation. To what extent can discourse markers be handled in translation? While Aijmer and Simon-Vandenbergen base their study on corpora of English fiction texts and their Swedish and Dutch translations, the material examined by Bazzanella and Morra is a single literary text and its translation into Italian. These two papers are the background for my own study. I go beyond them in the following respects: •

the study is extended to other language pairs: English-Norwegian and, to some extent, English-German;

•

although the main focus is on translations of well, the study also includes sources which give rise to well in English translations.

1

For comments on an earlier version of the paper, I am indebted to Karin Aijmer, Gothenburg University, and Anne-Marie Simon-Vandenbergen, Ghent University.


The investigation is based on the English-Norwegian Parallel Corpus and the Oslo Multilingual Corpus.2

2.

MATERIAL AND METHOD

The English-Norwegian Parallel Corpus (ENPC) is a bi-directional translation corpus consisting of English original texts and their translations into Norwegian as well as Norwegian original texts and their English translations. Because it is structured in this way, it makes possible a range of studies comparing original texts in the two languages as well as original and translated texts, as shown in Figure 1 (see further Johansson, 1998). The ENPC contains 50 extracts of 10,000-15,000 words for each language and their translations into the other language, in all about 2.6 million words. The texts are divided into two main types: fiction (30 original texts for each language) and non-fiction (20 original texts for each language). All the texts were published in the course of the last couple of decades. A wide range of authors and translators are represented. The translations were made by professional translators and have presumably gone through an editing process before publication.

ENGLISH ORIGINALS

NORWEGIAN TRANSLATIONS

ENGLISH TRANSLATIONS

NORWEGIAN ORIGINALS

Figure 1 The structure of the English-Norwegian Parallel Corpus

In the larger Oslo Multilingual Corpus (OMC) the material is extended to other languages. In particular, the aim has been to adapt the model for the ENPC to the language triple EnglishNorwegian-German, making it possible to compare original texts and translations across three languages; see Figure 2.3

2 3

See: http://www.hf.uio.no/iba/prosjekt and http://www.hf.uio.no/german/sprik The figure was originally drawn in this way by my colleague Cathrine Fabricius-Hansen.

Well and its correspondences in Norwegian and German 117

GERM ORIG

ENG TRANS

NOR TRANS

NOR ORIG

ENG ORIG

GERM TRANS Figure 2 The Oslo Multilingual Corpus: English-Norwegian-German

All the texts in the two corpora were aligned at the sentence level with the aid of the Translation Corpus Aligner, a program developed by Knut Hofland, University of Bergen. Searches are made using the Translation Corpus Explorer originally developed by one of the members of the Oslo team, Jarle Ebeling. Using the program we can retrieve all instances of well together with the corresponding portions of texts in the other languages included in the corpora. As discourse particles are not tagged in the corpus, there is necessarily a manual sifting stage eliminating irrelevant occurrences (e.g. well used in an adverbial function). The study of the ENPC material includes both translations and sources. The questions asked are: how is well rendered in translations into Norwegian, and what are the Norwegian forms which give rise to well in translations into English? As the OMC has not been built up to the same extent, the study of the OMC material is limited to the way well is translated into Norwegian and German. Before I present my results, it is important to mention a serious shortcoming of the material. The discourse marker well is characteristic of conversational interaction, where the speaker and the addressee are in direct contact and where prosody is crucial for the interpretation, but all our material is written. The closest we get to real conversation is in fictional dialogue, and this is naturally where we find the majority of the instances. The material has the advantage, however, that it has been interpreted independently by translators in the process of translation. What we are studying is the result of this interpretation (and recreation) process, which simultaneously illuminates the function of well and shows how and to what extent it can be conveyed in other languages.


3.

WELL AND ITS NORWEGIAN CORRESPONDENCES

To begin with, I will give an overview of Norwegian correspondences, based on the ENPC material. All instances of well in sentence-initial position (e.g. Well, he had earned it) and following a comma (e.g. His eyes were blue, well, maybe more a blue-green) were retrieved from the fiction texts of the ENPC. The search produced the majority of the instances of the discourse marker well and a handful of instances of other uses. After sifting the material, I was left with 231 examples from the original texts and 98 examples from the texts translated from Norwegian; see Figure 3. There is a wide range of correspondences, but the figure focuses on the Norwegian discourse particle vel and on zero correspondences, i.e. cases where there is no lexical expression in Norwegian corresponding to well.

Figure 3 English well and its correspondences in the fiction texts of the ENPC

As the subcorpora of original and translated texts are equal in size (number of texts and approximate number of words), we can compare raw frequencies. We note that well is more than twice as common in the original texts as in the translations. Such a difference is by no means uncommon in the case of expressions of this kind.4 It suggests that there is no clear counterpart in Norwegian corresponding to English well. As translators are typically influenced by the linguistic choices in the source text, they will be less inclined to choose this form than are authors of English original texts. 3.1

The discourse particle vel

The ENPC material does, however, contain the Norwegian discourse particle vel, which is in fact the most common correspondence, accounting for about 22% of the translations and about 17% of the sources). Vel is etymologically related to English well, Swedish väl, Dutch 4

See, for example, the observations on Norwegian modal particles in Johansson and Løken (1997).


wel, and German wohl, but the words have developed in partly different ways in these languages. In Norwegian, Swedish, Dutch, and German they are commonly used as unstressed modal particles. A use corresponding to the English discourse particle well is found only in Norwegian and Dutch (for Dutch, see Abraham 1984: 35f., Heinemann, 1985: 50f., and Westheide, 1985: 193f.), but in the Dutch material considered in Aijmer and SimonVandenbergen’s study there are only a couple of instances translating the English discourse particle well, and it is noted that “Its infrequency in the corpus may be due to the fact that it has been superseded by nou in northern Dutch, which is the variety used in the corpus translations” (Aijmer and Simon-Vandenbergen, 2003: 1149). The English-Norwegian material suggests that Norwegian vel can often function as an equivalent of well, as in these examples:5 Preceding a declarative sentence: 1. "I've been to the Vaucluse." "Well, that's where they'll escape to. …" (ABR1) "Jeg har vært i Vaucluse." "Vel det er dit de skal rømme. …" Preceding an interrogative sentence: 2. "Well, what did you think," said Jack. (ST1) "Vel, hva synes du?" sa Jack. Preceding an imperative sentence: 3. Anthony's full, measured, sexy voice reminded her of the enemies that she confronted, and she arrived outside their door wearing, as she knew, her look. "Well, Alice, come in," said Anthony, heartily but falsely, for it was Theresa who was her friend. (DL2) "Anthonys dype, avmålte, sexy stemme minnet henne på at det var fiender hun kom til å stå overfor, og da hun sto utenfor døra til leiligheten deres, visste hun godt at hun hadde fjeset sitt på igjen. "Vel, Alice, kom inn da," sa Anthony, hjertelig, men falskt, for det var Theresa Alice var venn med. Preceding a sentence fragment: 4. We had this tremendous row with Oliver. Well, an argument, anyway. (JB1) Vi hadde nemlig en voldsom krangel med Oliver. Vel, iallfall en diskusjon. In medial position: 5. "But these carpets are priceless. It would be an act of er... well, sheer vandalism...." (ST1) "Men disse teppene er uvurderlige. Det ville jo være den reneste... æh... vel, vandalisme..."

5

In the examples from the corpus material, the English version is always given first. The original text is identified by a text code. For an identification of the codes, see the web addresses (cf. note 2).

120 Pragmatic Markers in Contrast Unattached: 6. "Don't you curse at me, Macon Leary!" They paused. Macon said, "Well." Sarah said, "Well, anyhow." (AT1) "Hold opp å banne til meg, Macon Leary!" De tidde. Macon sa: "Vel." Sarah sa: "Likevel — " See also (9) below. Like well, vel can be used both to express agreement (7) and disagreement or qualified agreement (8): 7. This conversation ended with "I think you should come and see for yourself." "Very well, I will," said he, agreeing as much to what had not been said as to what had: his difficulty in keeping up with his wife's unspoken languages was the main reason he had been pleased to leave her. (DL1) Denne samtalen endte med: "Jeg synes du burde komme og se selv." "Vel, så gjør jeg det," sa han, like enig i det som var blitt sagt som i det som ikke var nevnt: hans vanskeligheter med å holde tritt med konens uuttalte meninger var hovedårsaken til at han hadde forlatt henne med glede. 8. "You're not married?" "Well, I am, but she's... living elsewhere." (AT1) "Å," sa hun, "du er ikke gift?" "Vel, jeg er det, men hun... bor et annet sted. Note the sequence very well in (7). By itself, well more typically indicates qualified agreement, as in (8). Well and vel can both be used to respond to a move from a previous speaker (as in the last two examples) and to take a new initiative, as in (9): 9. Quietly they went down the stairs of their house where they had so thoroughly taken possession, and into their great family room, and let themselves out into the garden that was mysterious and hidden from them, not yet theirs. "Well?" said Harriet humorously as they got into his car to return to London. "And how are we going to pay for it all if I am pregnant?" (DL1) Stille gikk de ned trappen i huset sitt, som de hadde tatt så fullstendig i besittelse, og inn i det store familierommet og derfra ut i haven, som var mystisk og skjult for dem, ikke deres ennå. "Vel?" sa Harriet spøkefullt da de satte seg i bilen hans for å kjøre tilbake til London. "Og hvordan skal vi klare å betale for dette hvis jeg er gravid?" Here, as well as in (3) above, well is discourse-initial. Although generally discussed in the context of dialogue, well is also found in monologue. The same is true of vel. Both are for example found in the representation of a person’s thought, as in: 10. He gazed down at the sink, and the warmth from the dishes drifted gently up into his face. Well, you have to carry on. (AT1)

Well and its correspondences in Norwegian and German 121 Han stirret ned i oppvaskkummen, og varmen fra serviset drev mildt opp i ansiktet hans. Vel, livet må gå videre. Arguably, we can think of this as a kind of internal dialogue. In spite of the correspondence in many cases, Norwegian vel is clearly more marginal than well. A study of the frequency patterns of the two discourse particles is instructive; compare Figure 3 above and Figure 4 below. Vel is far less common than well, with a total of 31 instances in Norwegian original texts and 52 in translations from English, as against 231 examples for well in English original texts and 98 in translations from Norwegian. In the case of vel the number goes up in the translations, due to the similarity to the English discourse particle well, which is frequent in the English source texts; for English well, we find the opposite relationship (cf. the comment on Figure 3 above). Most clearly the difference is shown by the common occurrence of zero correspondences in the case of well and by the fact that other forms are selected in close to two thirds of the cases (cf. Figure 3).

Figure 4 Norwegian vel and its correspondences in the fiction texts of the ENPC

To sum up, we can say that the two discourse particles definitely overlap, but well covers a wider area of use. It is by far the most frequent correspondence in relation to vel (other notable correspondences are 4 instances of all right, 2 of good, and 1 of right). Zero correspondences hardly occur in the case of vel. Apparently, it is regularly equated with well, while the opposite is true only in a minority of cases. 3.2

The discourse particles nå and tja

Two correspondences which occur repeatedly are nå and tja. Nå is identical to the Norwegian word for ‘now’, but it can also be used as a discourse particle (and as a modal particle). Tja, which dictionaries describe as originating from an extension of ja (‘yes’), is only used as a discourse particle. Typical meanings assigned by dictionaries to these discourse particles are:

122 Pragmatic Markers in Contrast nå: tja:

used to express surprise or irritation used to express doubt or deliberation

I presented the examples from the corpus material to four Norwegian informants, who were asked individually whether it would have been possible to use vel instead of nå and tja. The options given were: + (yes), - (no), and ? (undecided). In most cases, one or more of the four informants indicated that vel would have been a possible alternative. Although there was a good deal of disagreement, vel was more commonly accepted as an alternative to tja. Such examples included: 11. "That was a war," I said. "Well, it wasn't really — I suppose it was." (RDO1) Det var en krig, sa jeg. Tja, ikke egentlig — det var kanskje det. 12. "… You have a perfect life. You're famous and everything." "Famous? Well —" (ABR1) "Du har et perfekt liv. Du er berømt og allting." "Berømt? Tja —" With some of the nå examples vel was unanimously rejected: 13. "Talk of the devil," says Frank. "Well, here you are: Paul — Andrea." (ABR1) "Når du snakker om sola," sier Frank. "Nå, her er du: Paul Andrea." 14. "You've had too much to drink again," says the blonde girl, with the smallest hint of a slur in her voice, shaking back her Lippizaner mane. "Well, come on, it's almost day. We must go." (ABR1) "Du har fått for mye å drikke igjen," sier den lyse kvinnen, med en anelse av snøvling i stemmen, mens hun slenger den store lippizaner-manken bakover. "Nå, kom igjen, det er nesten blitt dag." 15. "Well now, Jim, what have you been up to?" (FF1) "Nå, Jim, hva har du foretatt deg?" The contrast between the two groups of examples is striking. In (11) and (12) the discourse particle responds to the previous utterance. In the other examples it is discourse-initial (13) or introduces a change in the discourse (14, 15). Apparently, well is compatible with both situations, while vel, tja, and nå are more limited in their use. No hard-and-fast rules can be laid down, however, as regards vel vs. well; cf. the comment below. The elicitation test was not ideal, as only a limited context was provided, and a test of this kind is always more or less artificial. The following comment from one of the informants is worth noting (given in translation): Difficult (almost impossible) to reject vel, as it has so little content (?), so that it can be used almost everywhere (well, at least in written Norwegian). In speech, well, it often seems a bit artificial.


Here is an additional clue to the difference between well and vel. The two discourse particles are sufficiently similar to be equated in a great number of cases (and almost always if we use vel as a starting-point), but different conditions of naturalness seem to apply. This requires further investigation, preferably using real conversations recorded in comparable contexts.6 3.3

The response particles ja and nei

Next to vel and zero, the most common correspondence is the affirmative response particle ja (‘yes’), which was the most common overt translation in the English-Swedish material examined by Aijmer and Simon-Vandenbergen (2003: 1134, 1151). Swedish and Norwegian ja are in fact multifunctional and are used not just for an affirmative response, but also for regulating conversational interaction. Examples (16) and (17) are fairly straightforward cases where ja (ja) expresses affirmation: 16. "… I can see you're a doc." "Well, you're right, I am. …" (OS1) Jeg kan se at du er doktor." "Ja ja. Du har forresten rett. Jeg er det. …" 17. William said, "Bridget has to be told by you that she is welcome. Well — we all do, from time to time," he added, in his facetious way, and could not help sending a look at his wife. (DL1) William sa: "Bridget må få høre av dere at hun er velkommen her. Ja — det ønsker vi jo alle fra tid til annen," tilføyde han på sin spøkefulle måte og kunne ikke la være å se på sin kone. Note the co-occurrence in (17) with the modal particle jo, which can be roughly glossed as ‘as you know’.7 In (18) and (19) we see cases of qualified agreement: 18. Dorothy could not get out of her what could possibly be wrong: she said, "Very well, I suppose I'll have to come, then." (DL1) Dorothy klarte ikke å få ut av henne hva som var så galt, så hun sa: "Ja, ja, jeg får vel komme, da." 19. "… You are only just thirty, David. You two go on as if you believe if you don't grab everything, then you'll lose it. Well, that's the impression I get, listening to you talk." (DL1) "… Du er bare såvidt tredve, David. Dere to driver på som om dere tror at hvis dere 6

In the absence of a representative Norwegian spoken language corpus, I have consulted the Norwegian version of the TV programme ‘Big Brother’, transcribed at the Text Laboratory, University of Oslo: http://www.tekstlab.uio.no/talespraak/bigbrother/. In the material that had been transcribed at the time of writing (approximately 75,000 words) there were 93 instances of vel. Only a few of these represented the discourse particle, and they were all found in sequences: ja vel (2), vel jamen (1), nei vel (1), vel vel (1). 7 A similar Swedish example is discussed in Aijmer and Simon-Vandenbergen (2003: 1140), where it is pointed out that “ja … ju is used to acknowledge a conflict of interests and to mark the speaker’s willingness to defend a particular position against an accusation”.


ikke får alt, nå med én gang, vil dere miste det. Ja, det er i hvert fall det inntrykket jeg får når jeg hører på dere." The qualification is made explicit by the modal particles vel (corresponding to I suppose) and da (corresponding to then) in the first example and by the addition of i hvert fall (‘in any case’) in the last example. A further use is illustrated in: 20. "… when a woman brushes her coat after you've bumped into her by accident, well, that's when you say enough's enough." (MW1) "… når en dame børster av kåpen sin etter at en tilfeldigvis har støtt borti henne, ja, da sier en at nok er nok." In this case ja is used to emphasize a point. Less commonly, well is translated by the negative response particle nei. In this case it often co-occurs with a negative statement in the context. Note the co-occurrence with the modal particle jo in (22). 21. Nothing the least unusual. Well, I've thought over every moment of that party time and time again, sifting for pointers, signals, ways of knowing how to do things differently from the way they got done. (JSM1) Ikke det minste uvanlig i det. Nei, jeg har tenkt gjennom hvert minutt av den grillfesten om og om igjen, forsøkt å skjelne varseltegn, en pekepinn om noe vi burde ha skjønt og som kunne ha forhindret at det gikk som det gikk. 22. "Well, Sam," Tom whispered, "I don't know nothin' about children, but I do know enuff not to beat 'em and make 'em that scared." (MM1) Nei, Sam, jeg vet jo ingenting om unger, men jeg skjønner nok til ikke å slå dem eller skremme dem. 23. Beverley thought, well, I'm not bleedin' shavin' yer. (ST1) Nei, nå får det være nok, tenkte Beverley; jeg har fanken ikke tenkt til å barrbererei! Another use is illustrated in: 24. "Well, look a here," Etta marveled as she stroked the cracked red fist of the baby. (GN1) "Nei, har du sett," sa Etta beundrende mens hun strøk den røde, rynkete barneneven. This is a special use of nei found in exclamatory utterances.8 It may seem like a contradiction that well is found to correspond both to ja and nei. It is a sign of the multifunctionality of the response particles as well as of the elusiveness of well. In a way well can be regarded as covering the area between yes and no, and it actually occurs fairly often in sequences together with these response particles, suggesting that there is some

8

The same use is found with Swedish nej. In Dutch as well the negative response particle is used in exclamations (Anne-Marie Simon-Vandenbergen, private communication): nee maar (‘no, but!’).


doubt about which option to choose.9 Where such sequences are found in the material, the response particle tends to be translated while well has no formal counterpart (see Section 3.4). The following example is of particular interest: 25. I sensed that little hesitation when I floated the subject of his double chin. You mean you didn't notice it? Yes, well, in the dark with the light behind him... (JB1) Jeg ante en liten stussing da jeg sjøsatte bemerkningen om dobbelthaken hans. Mener du at du ikke la merke til den? Ja, nei, i mørket med lyset bakfra... The doubt expressed in the original text is here marked by a combination of the two Norwegian response particles. This is the only instance in the corpus material, but it is a use which I have noticed in Norwegian texts, as in: 26. Det faktum at …, kan det ha påvirket gangsynet deres? Ja, nei, sikkert. (from an interview in the newspaper Dagens Næringsliv) (‘The fact that …, could it have influenced your judgement’ [gangsyn = approx. ‘good enough sight to walk alone’]?’ – ‘Yes, no, certainly.’) 27. Det er sagt om henne at hun var veldig glad da hun var glad, og tilsvarende langt nede da hun var nede? Ja. Nei. Jeg kan ikke si imot det. (from an interview in the newspaper Dagbladet) (‘It has been said about her that she was very happy when she was happy, and just as depressed when she was depressed?’ – ‘Yes. No. I can’t deny it.’) Although such examples are hard to find in printed texts, I have frequently noticed ja nei in conversation. These are a couple of examples I have noted down from recent interviews with Norwegian authors: 28. Hvorfor [valgte du å skrive en bok om] London? – Ja nei jeg har alltid likt London. (‘Why [did you choose to write a book about] London?’ – ‘Yes no I have always liked London.’) 29. Tenker du best når du skriver eller når du snakker? – Ja nei når jeg skriver antakelig ja nei JO. (‘Do you think best when you write or when you speak?’ – ‘Yes no when I write probably yes no YES.’) It seems to me that the sequence yes no (spoken as one prosodic unit) may be developing as a new routine covering part of the area of English well.10

9

A search in the British National Corpus gave these results: well yes 439 instances, yes well 281 instances, well no 732 instances, no well 307 instances. Apparently, it is most common to open with well and come down with a negative choice, perhaps because a negative response requires more hedging. 10 Another compromise between ‘yes’ and ‘no’ is nja, recorded among the Swedish translations in Aijmer and Simon-Vandenbergen (2003: 1134). This form occurs in Norwegian as well, though no instances were found in the ENPC. In the Norwegian Big Brother transcriptions (cf. note 6 above) there were three instances of ja nei and one of nja.

126 Pragmatic Markers in Contrast The last example includes the response particle jo, which is sometimes found to correspond to well. This form is generally used to affirm the positive content of a previous negative utterance, as in: Er du ikke fra Sverige? – Jo det er jeg (‘Aren’t you from Sweden?’ – ‘Yes, I am’). In the examples from the ENPC material it occurs in responses to positive questions, as in: 30. "Was he handsome? Well, I will tell you. …" (TH1) "Om han var pen? Jo, nå skal du høre. …" 31. "… What would you do for money?" "Well, with you gone I could go to the DHSS, couldn't I? …" (PDJ3) "… Hvordan skal du få tak i penger?" "Jo, når du er reist, kan jeg gå på sosialen, ikke sant? …" A similar Swedish example is taken up by Aijmer and Simon-Vandenbergen (2003: 1141), who make the point that the response particle “does not simply mean that something new is coming but marks the start of a lengthy explanation”. This agrees well with the two Norwegian examples just quoted. 3.4

Zero correspondences

Zero correspondences, or cases of omission, were noted by Aijmer and Simon-Vandenbergen (2003: 1153): 21% in the English-Swedish translations and 7% in the English-Dutch translations. In the English-Italian text examined by Bazzanella and Morra the omission rate was as high as 39%. Apparently, Italian differs greatly from English as regards the means of conveying the meaning of the discourse particle well (or the relevance of marking this type of meaning). The rate of omission is somewhat lower in Norwegian (16%) than in Swedish, no doubt due to the availability of the discourse particle vel. Of the languages mentioned here, Dutch seems to be closest to English, with a low rate of zero correspondence and a favoured marker, nou (‘now’), which occurs in about 30% of the cases. One reason why well can be omitted is what Aijmer and Simon-Vandenbergen call doubling of function, i.e. cases where the meaning of well is supported by other means. Such examples are found in the ENPC material as well, as in: 32. "But you have at some point to understand that your father is not prepared any longer to share his ill-gotten gains with Jasper and all his friends." "Well, at least he is prepared to see they are ill-gotten," said Alice earnestly. (DL2) "Men før eller senere vil du likevel bli nødt til å innse at din far ikke lenger er innstilt på å dele sine urettmessig ervervede rikdommer med Jasper og alle vennene hans." "Han er i det minste i stand til å innrømme at de er urettmessig ervervet," sa Alice alvorlig. 33. Sonia wasn't daft. Well, not then, anyway, unless that's what the abandoning of your own life for your children can be called. (FW1) Sonia var ingen tosk. Ikke da i hvert fall, bortsett fra hva det at du kaster bort livet ditt for barnas skyld, kan bli kalt.

Well and its correspondences in Norwegian and German 127 Here the qualification is also conveyed by i det minste (‘at least’) and i hvert fall (‘anyway’), and well can be dispensed with in the translation. Where well co-occurs with another response word, it may be left untranslated, as in: 34. "Problematic?" "Well, yes. …" (SG1) "Problematisk?" "Ja. …" 35. "I meant, don't you get any nice personal post?" "Well, no, not today, actually." (MD1) "Jeg mente bare — får du ingen hyggelige, personlige brev?" "Nei, ikke idag i hvert fall." Note also i hvert fall in the last example. In other cases it seems to be sufficient that the translation contains a connecting element of some kind, as in: 36. "… I'm just wondering." "Well, then stop wondering." (TH1) "Jeg bekymrer meg ikke, jeg bare lurer på det." "Så får du slutte å lure." 37. "Nobody told me that." "Well, they should have." (AT1) "Det har ingen fortalt meg." "Det skulle de gjort." [lit. ‘that should they done’] The connection is established here by så (‘then’) in (36) and by the fronted object pronoun in (37). Often it is impossible, however, to identify a plausible reason for the omission. Aijmer and Simon-Vandenbergen (2003: 1154) point out that there are cases where the translation of well is usually not omitted, viz. when “it is focused on by syntactic means or explicitly oriented to by the participants in the discourse or by the author”. Examples from the ENPC material: 38. There was a silence. "Well. …" (AT1) Det ble en pause. "Vel. …" 39. Jasper's tight grip on her upper arm took her out of the room into the hall, where Alice said, "This place smells." "Well," said Jasper, in the clumsily placating way she knew was meant as love, "I suppose you'll put a stop to that." (DL2) "Det stinker her," sa Alice. "Nå ja," sa Jasper, med det klossete, forsonlige tonefallet som skulle uttrykke kjærlighet, "det får vel du satt en stopper for, om jeg kjenner deg rett."


In (38) well is unattached and forms a separate unit, in (39) it is separated from the rest of the utterance by the reporting clause, where there is in addition an indication of the way the utterance is spoken.11 In these cases there is without exception an overt corresponding form. Zero correspondences go both ways, i.e. they are found both as translations and as sources. While we may suggest some reasons for the omission of a translation, it may seem more difficult to account for the addition of well where there is no formal counterpart in the Norwegian source text. About every fifth occurrence of well in the English translations (more exactly 21%) have no clearly identifiable source.12 In a number of cases we note a parallel between the omission and the addition of well, as in: 40. Well, what is he called, then? Mogens Klint, said the man who called himself Joseph Frost. "Hva heter han så?" "Mogens Klint," sa han som kalte seg Joseph Frost. (FC1) 41. If she had to take from it, well, then she had to and she didn't worry about it. Måtte hun bruke av den, så måtte hun, og sørget ikke over det. (HW1) 42. "No, I've been in other places all day, so I haven't heard the news," he answered. "Well, you can be glad of that. "Nei, jeg har vært andre steder i hele dag, så jeg har ikke hørt alle ryktene," svarte han. "Det kan du være glad for. [lit. ‘that can you be glad for’] (THA1) In (40) and (41) the addition of well parallels the omission in (36) above. In (42) there is a fronted pronoun, as in (37) above. In other cases well corresponds to some kind of graphic marking, such as the dash in (43) and the full stops in (44): 43. You see, it wasn't always easy, either for her or for us. She was only a young girl and, well, we're not exactly young any more, you know...." "Du vet, det var jo ikke lett bestandig, hverken for henne eller oss. Hun var jo bare ungpiken, og vi — vi [lit. ‘we — we’] er jo ikke nettopp unge lenger..." (KF2) 44. "Yes, it does sound a bit, well you know — a bit far-fetched," ventured Karin. "Det høres svært... usannsynlig," sa hun forsiktig. [lit. ‘it sounds very… implausible, said she cautiously’] (EG1) In (43) well expresses the hesitation marked by the dash and the doubling of the subject pronoun. The multiple addition in (44) is remarkable: yes, emphatic do, well, you know. It is a reflection of the wording of the reporting clause, which is presumably also responsible for the rendering of the intensifier svært (‘very’) by the weaker form a bit, which is made even more tentative by the repetition. The last example neatly illustrates that lexical correspondence, or word-for-word translation, is not the rule in translation. Translators respond to the context in their re-creation 11 A similar example containing the adverb placatingly is given by Aijmer and Simon-Vandenbergen (2003: 1154). 12 Again we can point to a parallel with Norwegian modal particles; see Johansson and Løken (1997).

Well and its correspondences in Norwegian and German 129 of the meaning of the original text. A couple of examples will suffice to illustrate the role of contextual interpretation. In both cases well has been added by the English translator. 45. "Reported missing from an old people's home? Who reported it? Oh? Well, we'll talk about that later." "Jaså, forsvunnet fra et aldershjem? Hvem meldte henne savnet? Javel. Vi snakkes ved." (EG1) 46. "Have you been at the crisp-bread again?" Hildegun rolled her eyes in suffering towards the ceiling and answered with brazen politeness. "No, mother, we haven't. We're just drinking rose-hip tea." "Well that costs money too. You've been using electricity." "Har dere tatt knekkebrød igjen?" Hildegun himlet lidende mot taket og svarte med uforskammet høflighet: "Neida, mor. Vi drikker bare nypete." "Den koster også penger. Dere har brukt strøm." (BV1) These are good examples showing how well can serve to smooth the interaction. In (45) well is used to round off a telephone conversation. Example (46) opens with a reproach in the form of a question; the child tries to take the brunt of the reproach by using just (corresponding to bare in the Norwegian original); the mother continues with another critical remark, opening with a mitigating well. The need for such grease in the interaction apparently varies in different languages, as do the means of expression, as we shall now see in a comparison of well and its translations into Norwegian and German. 3.5

Other correspondences

Apart from the cases discussed above, there were many other correspondences in the ENPC material, most of them occurring just once or twice. The range of forms used to translate well will be documented further in the next section.

4.

SOME ASPECTS OF NORWEGIAN AND GERMAN CORRESPONDENCES

As explained in Section 2, the study of the OMC material is limited to the way well is translated. At the time of the investigation, the OMC contained 32 English original texts – with a couple of exceptions they were all fiction texts – with translations into both Norwegian and German. Of these, 26 contained examples of the discourse particle well, in all 226

130 Pragmatic Markers in Contrast instances. A survey of correspondences, broadly grouped by category, is given in Table 1.13 As far as possible, comparable Norwegian and German forms have been aligned. The most frequent correspondences are highlighted.

Table 1 Well in English original texts: German and Norwegian correspondences German Discourse particles also (DP) also gut also hör mal also schön ja ja also

na na ja (naja) na gut na schön na und? nun nun ja nun gut nun sag mal tja tja, sannelig (‘truly’) vel vel ... i hvert fall (‘in any case’)

38 3 1 1 6 1

12 24 1 6 1 17 17 1 1 11 1 57 1

Norwegian altså

1

ja ja ja (ja, ja) ja … i hvert fall (‘in any case’) jaså javel jo jo… kan du skjønne (‘you see’) nei (‘no’) nei … men (‘but’) nei, nå får det være nok (‘enough!’) nå nåja (nå ja) nåvel (nå vel)

24 4 1 2 3 3 1 6 1 1 7 11 7

tja

13

Modal particles (MP)

eben ja

1 3

da jo

7 3

DP + MP

na … ja na ja … doch na ja ... eben na ja ... schon nun ja ... eben tja ... eben

2 2 1 2 2 1

ja ... jo ... da javel, greit da (‘OK then’) ja ... jo nei … jo vel ... da

1 1 2 2 1

Conjunctions

aber

7

men men … nå (MP)

oder und und nun

1 3 1

13

og ... egentlig (‘really’)

10 1 1

The table includes all instances of the discourse particle, including sequences like: oh well, very well, well now, well then, well well. It is quite likely that correspondence patterns would differ depending upon the sequence, but instances were too few to allow any safe conclusions.


Adjectives

gut sehr gut schön sicher

2 1 1 1

fint utmerket (‘excellent’) greit nok (lit. ‘OK enough’)

1 1 1

Adverb(ial)s

auf jeden Fall jedenfalls bloss da dann trotzdem

1 1 1 1 1 1

iallfall i hvert fall (hvertfall) i og for seg (‘an und für sich’) sannelig (‘truly’) så (‘then’)

1 4 1 1 2

Interjections or exclamations

ach ach wirklich aha grosser Gott hm

2 1 1 1 1

du verden (lit. ‘you world’) å

2 1

Other

mag sein nicht direkt zugegeben

1 1 1

hør her (lit. ‘listen here’) jeg vet ikke (‘I don’t know’) hun har jo rett (‘she is MP right’) tro? (lit. ‘believe?’)

1 1 1 1

Omission Gap or other problem Total

36 3 226

30 4 226

As shown in the table, there is a wide range of correspondences in both languages. Similar categories of forms are used in Norwegian and German. Most often we find discourse particles. There are also: modal particles (note Norwegian jo and da);14 combinations of discourse particles and modal particles (again, jo and da in Norwegian); conjunctions, particularly the German and Norwegian equivalents of English but; adjectives that indicate acceptance; adverb(ial)s, particularly concessive expressions; interjections and exclamations, etc. There is further a substantial amount of zero correspondence or omission. The results are broadly in line with the findings of Aijmer and Simon-Vandenbergen (2003). Apart from vel, which has been dealt with in some detail above, there is good agreement between Norwegian and Swedish. The affirmative response particle ja is a common translation of well in both of the Scandinavian languages (unlike German and Dutch). German is like Dutch in that most of the frequent translations have some relationship to words for ‘now’. The main difference is the common use of also, which as far as I can see has no counterpart in the Dutch material. It is significant that similar means have been appropriated in different languages, but I will not go into this, as it has been dealt with in some detail by Aijmer and Simon-Vandenbergen. All in all, the interpretations given by the translators seem to provide a good picture of the type of meaning carried by the discourse particle well. We find both expressions of acceptance and concession, of agreement and disagreement, emotional expressions, etc. Some 14

Note that both of these particles are hearer-oriented. In his discussion of the Danish modal particles (referred to in the paper as ‘discourse particles’), Davidsen-Nielsen (1996: 296) groups jo and da together. Both are described as polyphonic and presupposing, but while jo indicates attachment, da signals separation.

132 Pragmatic Markers in Contrast instances where the meaning is spelled out more explicitly are worth noting, in particular: the attention-getting forms hör mal, nun sag mal, hør her (lit. ‘listen here’); the hedging expressions mag sein, nicht direct, and jeg vet ikke (‘I don’t know’). Space does not permit a detailed discussion of examples. I will just focus on two points in relation to the English-Norwegian-German material: zero correspondences and individual variation. 4.1

Zero correspondences

Although there is so much to draw on in both languages, well is often omitted in the translation. As shown in Table 1, the amount of omission is somewhat larger in German than in Norwegian. In a number of cases there is omission in both translations. Some examples are: 47. "How far?" Matilda said. "Well, I don't really know, Miss Honey. …" (RD1) "Wie weit?" fragte Matilda. "Das weiß ich nicht genau, Fräulein Honig. ..." Hvor langt? sa Matilda. — Det vet jeg faktisk ikke, frøken Honey. 48. "Well, come in or go out, you great big gorm face. …" (ST1) "Rein oder raus, du großer Klumpatsch. …" "Du får kommerei inn eller pellerei ut, din digre grobian. …" 49. "I always thought," he commented, "that drug company detail men were always, well, men." (AH1) "Ich dachte immer, bei den Arzneimittelfirmen gäbe es nur männliche Vertreter." "Jeg har alltid trodd at medisinselgere var menn," sa Andrew. 50. "… He was so good and sweet and friendly!" – "Well, fine," Macon said.... (AT1)"… Der war so brav und süß und lieb!" – "Wie schön", sagte Macon. "… Han har vært så søt og snill og vennlig!" – "Flott," sa Macon. The first example is similar to (37) above in that it has a fronted object pronoun. The omission in (48) may have something to do with the rather rude command, in (49) with the medial position, while in (50) a rendering of well was probably felt to be redundant. In most of the cases, there is omission in one of the translations only, and it is difficult to find a good reason for the difference. It is notable, however, that omission in the German translation is found repeatedly where Norwegian has a response particle, as in: 51. "Well, I can't see the sense of leaving here, even if there isn't any work. …" (PDJ3) "Ich sehe nicht ein, warum du wegziehen mußt, selbst wenn es hier keine Arbeit gibt. …" "Ja, jeg skjønner ikke vitsen ved å dra herfra, selv om det ikke er arbeid å få. …" 52. "… Do you remember what you answered?" – "Well, not offhand," Macon said. (AT1) "… Weißt du noch, was du geantwortet hast?" – "Im Moment nicht." "… Husker du hva du svarte?" – "Nei, ikke på stående fot," sa Macon.


Compare also: 53. "Well, good morning. I'm substituting for Mrs.... and this period we're supposed to have Norwegian. My name is Per Kristian, Payk for short, Tomter... " Guten Tag. Ich vertrete also Frau....und in dieser Stunde haben wir also Norwegisch. Mein Name ist Per Kristian Tomter, abgekürzt Peik... "Ja, goddag. — Jeg vikarierer altså for fru... og i denne timen skal vi altså ha norsk. Mitt navn er Per Kristian, Peik for korthets skyld, Tomter... " (KF1) In this example, taken from a Norwegian text and its English and German translations, a substitute teacher imagines meeting his class for the first time. The original text has ja, conveyed in English by well but omitted in the German translation.15 As shown in Table 1, the Norwegian response particles are more common than their German counterparts as translations of well. In view of these results, and the similar relationship reported for Swedish vs. Dutch by Aijmer and Simon-Vandenbergen (2003), there seems to be a need for a contrastive study of the response particles in Norwegian/Swedish vs. German/Dutch. Although there are differences as regards omission, there is complete agreement between the Norwegian and German translations as regards non-omission in cases where well is unattached and is marked off as a separate unit; cf. the comments on examples (38) and (39) in Section 3.4. 4.2

Individual variation

There is little doubt that the comparison above gives an adequate overall picture of the correspondence patterns for the discourse particle well in Norwegian and German translations. It is just as clear, however, that there are difficulties in comparing specific examples, as they may reflect individual translator preferences. The degree of individual variation which we might expect is indicated by a comparison of two OMC text extracts written by the same author but translated by different individuals; see Table 2. Note the differences in the choice of also, nun ja, and omission in the two German translations and in the use of ja, vel, and omission in the two Norwegian translations. A comparison of this kind should ideally be based on alternative renderings of the same texts, but we do not often find translations of the same texts made under similar conditions by more than one professional translator. For illustration I provide an example of how well was rendered recently by three professional Norwegian translators:16

15 In this context, ja and well presumably serve as some kind of ice-breaker facilitating the opening of the interaction. 16 The translations were made for the project Norvengelsk, arranged in 2003 by Foreningen !Les in cooperation with Norsk Språkråd, Norsk Oversetterforening and Læringssenteret. Pupils in Norwegian schools were invited to tackle a variety of language tasks, including voting on the best of three translations. The English text was a brief extract from “Something about Harry” by Annie Leibovitz and Leslie Bennet, published in Vanity Fair in 2001. The article deals with the filming of the bestseller Harry Potter and the Philosopher’s Stone by J. K. Rowling. In the text extract there was just a single instance of the discourse particle well.

134 Pragmatic Markers in Contrast Table 2 Well in two text extracts: 17 German and Norwegian correspondences German correspondences DL1

DL2

aber also eben gut hm ja jedenfalls na na ja na schön na und? nun nun ja tja und und nun?

1 2 1

omission

2

8

21

25

Total

5 1 1 2

1 1 3 2 1 2 1 2 1 1

2

6

Norwegian correspondences DL1 DL2 da i hvert fall ja (ja) ja … jo ja … jo … da men nei nå nåja og sannelig så tja tro? vel vel … da omission Total

6 1 1 1

1 1 1

1 2 1 1 1 1 2 6

1 11 1

1

5

21

25

54. Rowling was also forthcoming in offering guidance to the actors, according to Robbie Coltrane, who plays the half-giant Hagrid. “When I was trying to get a fix on Hagrid’s character, she said, Well, think of him as one of those really big Hell’s Angels that gets off a motorbike and then starts talking about how his garden is coming,” recalls Coltrane. Translator 1: ‘Høyr her, tenk på han … (lit. ‘listen here, think of him …’) Translator 2: ‘Tenk på ham som … (lit. ‘think of him as …’) Translator 3: ‘Du kan jo sjå for deg at han …’ (lit. ‘you can of course see for yourself that he …’) In other words, three different choices were made: an attention-getter, omission, and the modal particle jo.

5.

HOW WELL CAN WELL BE TRANSLATED?

The correspondences which I have shown vary in meaning and use, but they clearly pick up aspects of the meaning of the discourse particle well. Is it possible to find a unity behind the diversity? There is certainly a great deal to be said for Aijmer and Simon-Vandenbergen’s claim that well is a heteroglossic option signalling the speaker’s awareness of other, possibly 17

Doris Lessing, The Fifth Child (DL1), 21 occurrences of well in the text extract. German translation by Manfred Ohl and Hans Sartorius, Norwegian translation by Bodil Roald. Doris Lessing, The Good Terrorist (DL2), 25 occurrences of well in the text extract. German translation by Eva Schönfeld, Norwegian translation by Kia Halling.

Well and its correspondences in Norwegian and German 135 discordant voices. This interpretation seems to work best where well is attached to a statement, as long as the notion of heteroglossia is extended to include an internal dialogue within the mind of the speaker; cf. my comments on example (10). But is it equally applicable in cases where well is unattached or initiates a question, a command, a greeting, or an exclamation? In Biber et al. (1999: 1086) well is described as a deliberation marker. It seems to me that any account of well must include the speaker’s perspective as well as his/her accommodation to the addressee. Another possibility is to follow Blakemore’s relevance-theoretic account suggesting that well “could be regarded as a signal in the sense that it provides a green light for the hearer, a sign to go ahead with the inferential processes involved in the derivation of cognitive effects” (Blakemore, 2002: 147). This seems to be sufficiently vague to accommodate most uses. But again I miss the speaker’s own perspective. Well seems to mark both negotiation in relation to the addressee and deliberation in the speaker’s mind. Sometimes the use of well seems to be merely phatic, serving to open or maintain the channel of communication, as in this example taken from Roman Jakobson’s discussion of language functions: Dorothy Parker caught eloquent examples: “‘Well!’ the young man said. ‘Well!’ she said. ‘Well, here we are,’ he said. ‘Here we are,’ she said, ‘Aren’t we?’ ‘I should say we were,’ he said, ‘Eeyop! Here we are.’ ‘Well!’ she said. ‘Well!’ he said, ‘well.’” (Jakobson, 1960: 355f.) Cf. also example (53) above. Whatever the best interpretation may be, it is clearly the case that neither Norwegian nor German, nor the languages considered by Aijmer and Simon-Vandenbergen (Swedish and Dutch) and Bazzanella and Morra (Italian), possess fully-fledged counterparts of the English discourse particle well. Many different means are used to pick up facets of its meaning, but they are far more restricted as discourse markers, even the Norwegian discourse particle vel, which looks very similar both in form and meaning. The many correspondences and the high degree of zero correspondence show that the translation of well is far from straightforward. But translation is hardly ever without complications. My first linguistics teacher, the late Professor Bertil Malmberg from Lund University – he was a staunch structuralist – used to put it in this way: “translation is theoretically impossible, but feasible in practice”. It is feasible in spite of cross-linguistic and cross-cultural non-equivalence, because the translator serves as a negotiator across languages and cultures (cf. the recent book by Umberto Eco, 2003). Note also Harald Weydt’s comment on translation in his discussion of particle functions and Gestalt recognition: Hier kann nur kurz angedeutet werden, dass gerade dies das Problem der Übersetzung ist. Das Verständnis eines Textes besteht nur in einem geringen Teil in dem Verständnis dessen, was in ihm wörtlich gesagt wird, es spielt sich nicht einmal auf der Ebene der wörtlichen Bedeutungen ab. Die wörtlichen Bedeutungen sind vielmehr als kleine Ersatzstücke zu verstehen, die im Verein mit anderen Elementen ein Gesamtverständnis ermöglichen, das dem Aufnehmen einer Gestalt verglichen werden kann. Die Aufgabe einer Übersetzung ist es, einen derartigen

136 Pragmatic Markers in Contrast Gestalterkennungsprozess auf die Art und mit den Mitteln einzuleiten, die in der Zielsprache dazu benutzt würden und die in aller Regel nicht den Mitteln der Ausgangssprache entsprechen. (Weydt, 1989: 338) ‘Gestalt recognition’ is probably especially important in connection with discourse particles, which do not carry an easily definable meaning. Nevertheless, it is my conviction that we can learn a great deal from examining the pieces that make up texts. This is what we do in using our multilingual corpora, and in the process we learn both about languages and about the problems of translation.

REFERENCES Abraham, W. (1984). De betekenis en de functie van het Nederlandse wel – een vergelijking met het Duits. In: Studies over Nederlandse partikels (J. Van der Auwera and W. Vandeweghe, eds), Antwerp Papers in Linguistics, 35, 17-46. Aijmer, K. and A.-M. Simon-Vandenbergen. (2003). The discourse particle well and its equivalents in Swedish and Dutch. Linguistics, 41,1123-1161. Bazzanella, C. and L. Morra. (2000). Discourse markers and the indeterminacy of translation. In: Argomenti per una linguistica della traduzione. On Linguistic Aspects of Translation. Notes pour une linguistique de la traduction (I. Kortzen and C. Marello, eds), pp.149-157. Edizioni dell’Orso, Alessandria. Biber, D., S. Johansson, G. Leech, S. Conrad, and E. Finegan. (1999). Longman Grammar of Spoken and Written English. Longman, Harlow. Blakemore, D. (2002). Relevance and Linguistic Meaning. The Semantics and Pragmatics of Discourse Markers. Cambridge University Press, Cambridge. Davidsen-Nielsen, N. (1996). Discourse particles in Danish. In: Content, Expression and Structure. Studies in Danish Functional Grammar (E. Engberg-Pedersen, M. Fortescue, P. Harder. L. Heltoft, and L. Falster Jacobsen, eds), pp. 283-314. Benjamins, Amsterdam & Philadelphia. Eco, U. (2003). Mouse or Rat? Translation as Negotiation. Weidenfeld & Nicholson, London. Heinemann, I. (1985). Die Niederländische Partikel wel und ihre Deutschen Entsprechungen. Linguistica Antverpiensia, XVIII-XIX, 38-58. Jakobson, R. (1960). Closing statement: Linguistics and poetics. In: Style in Language (Th. A. Sebeok, ed.), pp. 350-377. The M.I.T. Press, Cambridge, Mass. Johansson, S. (1998). On the role of corpora in cross-linguistic research. In: Corpora and Cross-linguistic Research: Theory, Method, and Case Studies (S. Johansson and S. Oksefjell, eds), pp. 3-24. Rodopi, Amsterdam & Atlanta, GA. Johansson, S. and B. Løken. (1997). Some Norwegian discourse particles and their English correspondences. In: Sounds, Structures and Senses. Essays Presented to Niels Davidsen-Nielsen on the Occasion of His Sixtieth Birthday (C. Bache and A. Klinge, eds), pp. 149-170. Odense University Press, Odense.

Well and its correspondences in Norwegian and German 137 Östman, J.-O. (1979). On the pragmatic meaning of textual particles. In: Kontrastiv lingvistik och sekundärspråksforskning (B. Hammarberg, ed.), pp. 171-191. Department of Linguistics, Stockholm University, Stockholm. Westheide, H. (1985). Eine kontrastive Analyse der Partikeln Dt. wohl und Nl. wel. Zeitschrift für Germanische Linguistik, 13, 186-202. Weydt, H. (1989). Partikelfunktionen und Gestalterkennen. In: Sprechen mit Partikeln (H. Weydt, ed.), pp. 331-345. Walter de Gruyter, Berlin & New York.



CONTRASTIVE ANALYSIS OF ADVERSATIVE RELATIONAL MARKERS, USING COMPARABLE CORPORA Diana M. Lewis, University of Lyon2

1.

INTRODUCTION

This paper deals with relational pragmatic markers; that is, those lexical or semi-lexical expressions which can indicate coherence relations between adjacent ideas or sequences of ideas. For convenience, the two ideas or sets of ideas related by such a pragmatic marker will be termed the ‘related ideas’1. A given coherence relation between two ideas may viewed as belonging to one of two broad types: consonant relations, where the related ideas are presented as compatible and in harmony, or dissonant relations, where the related ideas seem to clash because they are incompatible or their co-occurrence is otherwise unexpected. The following markers of dissonant relations are discussed and contrasted: (i) the possible translation equivalents French au contraire and English on the contrary; (ii) the adversative expressions en revanche and on the other hand, whose usage is shown to overlap with that of au contraire; (iii) adversative in fact, which occurs in English data in contexts very similar to those where au contraire occurs in French data. For the contrastive study of relational pragmatic markers (or indeed of any other linguistic expressions), a tertium comparationis is needed; this may be the meanings, the referents, the usages, and/or the etymologies, common to the expressions being contrasted. Some recent studies within typology and cognitive linguistics use comparable expressions in different languages to build ‘semantic maps’ which map linguistic expression to conceptual areas.2 The semantic map approach uses the metaphor of ‘conceptual space’ to designate the conceptual content that can potentially be expressed through language.3 Conceptual space is conceived of as a multidimensional, universal space. Notional areas that might form part of this space, such as the one posited here, of ‘adversative coherence relations’, are initially identified on intuitive or traditional grounds only. However, these intuitions can later be tested against the evidence from contrastive analyses and refined. For example, where the usages (or parts of them) of two expressions of language A correspond to the usages of a single expression in language B, we can hypothesize that the semantics of the two expressions 1

The related ideas are sometimes termed the ‘conjuncts’. See, for instance, Anderson (1986), van der Auwera and Plungian (1998), Croft (2001), Haspelmath (2003). Most of this work has focused on the expression of grammatical meanings, but the approach is in principle equally applicable to lexical semantics. 3 Following Croft (2001:93), the term ‘conceptual space’ is preferred here to the term ‘semantic space’, the better to distinguish what is universal from what is language-specific. 2


map to contiguous or overlapping areas of conceptual space, and that the single expression in language B is polysemous. The idea is to try to identify, by cross-linguistic analysis, what parameters seem to act as boundary markers between one expression and another. This will help to gradually build a semantic map of the meanings of particular expressions by plotting their uses in relation to those parameters. Semantic maps, then, are built using the results of contrastive analyses. Their purpose is to gain insights into the structure of conceptual space. But once sketched, they may in turn assist contrastive analysis by providing a reference for comparison. Corpus data can suggest the kinds of parameters that will be useful in exploring a coherence-relational space. They allow us to identify and compare contexts of use for particular markers. The contextual similarities and differences observed can reveal patterns of use and help us to discover what parameters seem to constrain speakers’ choice of expression.4 Corpus data also show which are the relatively frequent uses of an expression and which are the more peripheral ones in a given genre. This paper suggests four candidate parameters on which au contraire and on the contrary and other adversative pragmatic markers may be contrasted. The first is the nature of the adversativity that obtains between the related ideas, independently of any marker; it concerns the likelihood that the related ideas coexist. The second is the status - realized or unrealized - of the related ideas, according to the speaker. The third and fourth candidate parameters have to do with information structure. ‘Thematic structure’ involves the way speakers organize the flow of information or the discourse thread. A discourse segment typically contains two elements: a theme (T) and something that is said about it, or rheme (R) (cf. Halliday, 1994:37ff). An element introduced in one unit may occur again in the next. Often, themes recur, as in The tortoise was slow but he won the race (T1-R1 T1-R2). Both theme and rheme may change, as in The tortoise plodded while the hare slept (T1-R1 T2-R2). Occasionally a rheme recurs with a new theme. The fourth possible parameter concerns the relative salience of successive segments, which is known in Rhetorical Structure Theory as ‘nuclearity’ (Mann and Thompson, 1987). Where two adjacent discourse segments or groups of segments in coherent text are equally salient they are both said to be Nuclear (N-N); where one is more salient than the other, there is an asymmetrical relation: one is the Nucleus, the other the Satellite (N-S) or vice versa (S-N).

2.

DATA

For contrastive analysis of coherence relations, comparable corpora (that is, two corpora consisting of native text and matched for genre) are preferable to parallel corpora (where one of the corpora is a translation of the other one) for several reasons. First, translation can vary considerably, ranging from quasi-glosses and overt translations to covert translations or free adaptations, depending on the purposes to which the target text is to be put. Second, the reliability of data from parallel corpora depends on the quality of the translation, which 4 It has proven difficult to taxonimize coherence relations from a purely semantic point of view (Hovy and Maier, 1995, Knott and Mellish, 1996). This has led to attempts to ‘parameterize’ coherence relations (e.g. Sanders et al, 1993, Louwerse, 2001), which seems to be a profitable approach.

Contrastive Analysis of Adversative Relational Markers 141

cannot be taken for granted. Third, there is evidence that translated text generally differs in kind from original text (Baker, 1993; see also Borin, 2002). Finally, and importantly, the frequency and distribution profiles of lexemes and constructions in translated text are likely to be constrained by the source text and therefore may not match those of the target language for the comparable genre. This point will be reiterated in section 4. The main data are taken from a comparable corpus of political speeches, written to be spoken. This corpus is described in table 1.

Table 1: Comparable corpus of political speeches

No. of speeches No. of words approx. No. of speakers Period

French

English

316 0.8m 52 1995-2002

508 1.4m 75 1995-2001

The French speeches were all given by French politicians; the English ones by politicians of the United Kingdom or Ireland. Political speeches were chosen, first because they are comparable cross-linguistically; the form, the distribution, the content and the purpose of speeches made by government ministers of European Union states are very similar. Second, it was assumed that political speeches would typically be persuasive in style; that they would aim to defend the speaker’s case by presenting arguments for or against particular options and opinions and would therefore be a rich source of inter-ideational relations. This assumption was only partially borne out by the data. Third, quantities of the genre were available in electronic form and were easy to collect. Additional data come from the following collections of newspaper text: Le Monde Diplomatique (1997), The Guardian (2002) and The Times/Sunday Times (2002) (electronic editions).

3. 3.1

ADVERSATIVE PRAGMATIC MARKERS Overlaps among markers

The many-to-many relationship between pragmatic markers and the functions they perform is well known. When comparing markers, therefore, whether within or across languages, it is interesting to look both at the markers themselves to see how they are used, and at their contexts of use, to see what alternative expressions, if any, can serve the same or a similar function in the same context. Examples (1) and (2) show that French au contraire can overlap with en revanche in some contexts, but not in others. In example (1), both au contraire and en revanche can serve to express a speaker-determined contrast between the two ideas. In example (2), however, au


contraire cannot be replaced by en revanche. This suggests that in (2) we may be dealing with a different type of relation of contrast from that of (1). 1. La surveillance prudentielle vise à traiter les difficultés avant qu’elles ne soient rendues publiques. La régulation des marchés vise {au contraire / en revanche} à ce que tout soit sur la place publique. (Laurent Fabius, 14.06.2001, au contraire in original) ‘Careful monitoring aims to deal with any difficulties before they become public. The regulation of the markets aims au contraire at making everything public.’ 2. Cet appel à la cohésion nationale ne nie pas l’exigence du débat. {Au contraire / *En revanche} j’y attache une grande importance. (Jean-Pierre Raffarin, 03.07.2002, au contraire in original) ‘This call to national cohesion in no way denies the need for debate. {Au contraire / *En revanche}, I consider it most important.’ In the same way, French au contraire overlaps with English on the contrary in some contexts, such as example (3), but not in others like example (4). 3. C’est une réalité. Nous ne devons pas la nier. Nous devons, au contraire {on the contrary}, l’appréhender dans toutes ses dimensions … (Pierre Moscovici, 26.10.1999) ‘It’s a reality. We must not deny it. We must, au contraire {on the contrary}, grasp it in all its complexity’ 4. On voit trop souvent des personnes en difficultés perdre tout lien avec l’administration: c’est bien au contraire {*on the contrary} vers ces personnes que l’administration doit concentrer ses efforts. (Michel Sapin, 06.07.2000) ‘Too often we see people in difficulties losing touch with the administration: it’s au contraire {*on the contrary} on these people that the administration must focus its efforts’ In (3), on the contrary is a plausible translation equivalent of au contraire, whereas in (4) on the contrary seems most unlikely as a translation. There would seem, then, to be notional areas that can be expressed by either au contraire or en revanche and others that cannot; contexts where au contraire can be translated by on the contrary, and contexts where it cannot. These data reveal partial overlap between pragmatic markers, both within and between languages. For cross-linguistic comparison, we need to identify what the similarities and differences are between the two expressions and what it is that forms the boundary between the acceptable and unacceptable translations of au contraire by on the contrary. 3.2.

‘Au contraire’ and ‘on the contrary’

Au contraire is considerably more frequent in French political speeches than on the contrary is in English ones: approximately 153 per million words (pmw) and 29 pmw respectively,


based on 123 occurrences of au contraire and 41 occurrences of on the contrary. 5 (In the newspaper text, by comparison, estimated frequencies were 123 pmw for au contraire and 8 pmw for on the contrary.)6 Perhaps au contraire is compatible with different contexts that in English require different expressions; or perhaps the particular rhetorical strategy that seems to call for au contraire is less frequently employed in English speeches. There is some evidence that both factors are relevant. We shall assume that a reading of an utterance containing a relational pragmatic marker is obtained by computing the readings of the discourse segments linked, the sense of the marker, the arrangement of the information units of the utterance and the context of utterance. Both au contraire and on the contrary inherit from contraire / contrary the sense of oppositeness. But although au contraire cannot always be translated by on the contrary, it is hard to find an instance of on the contrary that cannot be translated by au contraire. Au contraire would appear, therefore, to have a wider use than on the contrary. But we need to identify in what way it is ‘wider’. 3.3.

Adversative coherence relations: adversativity of the related ideas

We shall consider first the readings of the discourse segments on either side of the adversative relational marker. These related ideas – i.e. the situations described in the segments linked by au contraire or on the contrary – contrast to varying degrees. These range from contradiction, through objective incompatibility (where the co-occurrence in the real world of the situations referred to is impossible or improbable), to situations where the speaker evokes unexpectedness of co-occurrence between situations that are not obviously related. For convenience, we will describe this continuum of contrast in terms of relation types: (i) contradiction, where there are two possibilites that are mutually exclusive and exhaust the domain; (ii) contrary, where the situations (states or events) referred to are opposites; (iii) (perceived) incompatibility or contrast, where the co-occurrence of the situations referred to is possible but the difference between them is deemed striking or unexpected; (iv) (implicated) incongruity, where the speaker/writer views the situations as incongruous, despite there being no objective incompatibility between them. These labels are merely a convenience to allow us to describe the apparent relations between situations in terms of degree of objective compatibility.7 The scale has objective contrast at one end and subjective (speaker-oriented) contrast at the other. At the objective end of the scale, the situations referred to differ along one dimension only (cf. Cruse, 1986:197 on antonymous relations among lexemes), whereas at the subjective end of the scale, the situations may be quite unlike one another, except for one dimension perceived by the speaker and used as the basis for the contrast. Contradiction is illustrated in (5):

5

Related expressions contrary to (n=1), au contraire de (n=4) and contrairement à (n=5) were excluded. These figures are based on 108 occurrences of au contraire in Le Monde Diplomatique (1997), 211 occurrences of on the contrary in The Guardian (2002) and 253 in The Times/Sunday Times (2002). 7 Clearly there is a sense in which all interideational relations are projected by the speaker/writer, but there is a case for positing a scale of subjectivity. 6

144 Pragmatic Markers in Contrast 5. a.

Ce n’est donc pas un désengagement. Au contraire. Et l’intégralité des moyens, humains et financiers, qui étaient ceux mobilisés chacun de son côté par l’ENA et l’IIAP, est maintenue. (Michel Sapin, 29.08.2001) ‘So this is not a withdrawal. Au contraire. And all the human and financial resources that the ENA and the IIAP each had at their disposal are retained.’

b. What conclusion should we draw? That it is time to abandon our support for reform in Russia? On the contrary. Successful reform in Russia would be a huge prize. (Douglas Hurd, 28.02.1995) In (5a) au contraire refers back to un désengagement and reinforces the denial that it is a désengagement. Likewise, on the contrary in (5b) refers back to that it is time to abandon our support for reform in Russia and affirms that our conclusion should not be that it is time to abandon our support for Russia. It is equivalent to ‘no’. In neither case is an opposite (the ‘contrary’) spelled out, so that the markers are not in fact relational here. Rather, they constitute the second idea, which differs from the first along the sole dimension of polarity. The idea they refer back to is false, of course, and therefore marked as unrealized (by a negative in (5a), by an interrogative in (5b)). Contrary relations are illustrated in (6): 6. a.

nous voulons réformer l’Etat non pas pour l’affaiblir ou l’affadir, mais au contraire pour le rendre plus fort (Michel Sapin, 16.05.2000) ‘we want to reform the State not in order to weaken it or diminish it, but au contraire to make it stronger’

b. I cannot ignore this fact. On the contrary, I am obliged, in the public interest, to address this issue. (Michael Smith, 27.11.1997) c.

entre 1998 et 1999, l’amélioration de l’emploi n’a pas été très sensible,… En revanche, l’amélioration du marché de l’emploi commence aujourd’hui à bénéficier aux quartiers (Claude Bartolone, 09.10.2000) ‘between 1998 and 1999, the improvement in the employment situation was not very noticeable, … En revanche, the improvement in the employment market is now beginning to benefit the districts’

d. most continental countries have extensive permeable land borders. Britain on the other hand has no land border except with Ireland. (Malcolm Rifkind, 19.03.1997) Two related ideas are not mutually exhaustive. In example (6a), not weakening the state does not entail strengthening it; in (6b), being unable to ignore the fact does not entail being obliged to address the issue; in (6c) the minimal improvement in the past does not entail the greater improvement now; in (6d) not having extensive land borders does not entail having no land borders. In each case, however, there is mutual exclusion: an antonymous relationship between elements of the related ideas such that both ideas cannot simultaneously hold. In (6a) affaiblir and affadir contrast with rendre plus fort; in (6b) ignore contrasts with address; in (6c) entre 1998 et 1999 contrasts with aujourd’hui and amélioration pas très sensible


(possibly a euphemism for ‘no improvement’) contrasts with amélioration. In (6d) extensive contrasts with almost none. Perceived incompatibility, where the contrast is less obvious and less objective, is illustrated in (7): 7. a.

Ce n’est pas entretenir en permanence les clivages et les divisions. C’est au contraire proposer un sursaut national … (Jean-Pierre Raffarin. 31.05.2002) ‘This is not to perpetuate splits and divisions. It’s au contraire to propose a nation-wide leap forward’

b. We have no intention of importing the excessive social security and labour market restrictions that apply in some other parts of Europe. On the contrary, we are confident that signing up to the Social Chapter … will have positive effects for the UK economy … (Margaret Beckett, 23.09.1997) The related ideas in cases such as those in (7) have much less in common than those in cases of contradiction and contrary. (7a) involves opposition between the static (entretenir en permanance) and the dynamic (sursaut) but at a general, abstract level. These expressions are not antonymous in the sense that faible and fort in (6a) are antonyms. The rhetorical effect of (7b) is to suggest incompatibility by emphasizing the contrast between the damage that one course of action would cause and the benefits another will bring. Again, the pejorative excessive restrictions is not directly antonymous with positive benefits, but there is a clear opposition. Finally, there are cases where the related ideas seem to have little in common other than that they both occur on some scale of evaluation or of emotional reponse in the speaker (example 8). 8. Comment ferait-on partager aux adolescents d’aujourd’hui nos souvenirs de potaches? La remise des prix dans la salle des actes ou au contraire la douloureuse convocation dans le bureau du censeur ou du surveillant général, … (Xavier Darcos, 10.07.2002) ‘How could we share with today’s adolescents our memories of when we were schoolkids? Prize-giving day in the hall or au contraire the painful summons to the proctor’s or the deputy head’s office’ The two ideas in (8) are related on a scale of emotion: the contrast is between the presumably happy emotions associated with prize-giving and the unhappy ones associated with a summons. Examples (5) to (8), then, exemplify pairs of situations that are more or less naturally incompatible. Table 2 shows which types of contrast several relational pragmatic markers occur with in the corpus.


Table 2: Types of contrast occurring in the corpus with selected adversative pragmatic markers au contraire

CONTRADICTION CONTRARY INCOMPATIBILITY INCONGRUENCE

| | | | | | | |

on the contrary | | | | | |

en revanche

| | | | | |

on the other hand

| | | |

in fact [CONTR.]

| |

On this dimension (objectivity of contrast), some markers are seen to be compatible with a wider range of contrasting situations than others. We might want to say, then, that it is in this sense that au contraire has a ‘wider’ meaning than on the contrary. Table 2 can be read as a one-dimensional semantic map where the adversative expressions are plotted on a scale of adversativity.

Fig. 1 shows the distribution of the corpus occurrences of au contraire and on the contrary across a range of contrast.

Although the figures for on the contrary (n=41) are unreliably small, the chief difference with au contraire appears to be the absence of on the contrary from the more subjective adversative contexts. In none of the ‘incongruence’ situations linked by au contraire would on the contrary be an acceptable translation. And in the newspaper text, only the occasional, ambiguous case was a candidate for an ‘incongruence’ contrast. This suggests that the


absence is not just an accident of the small sample. Fig. 1 shows how the overall distribution of on the contrary mirrors that of au contraire, but at much lower frequencies. The nature of the contrast between the two related ideas is an aspect of the contexts of the markers. The next sub-section looks at the semantics of the markers themselves. 3.4.

The status of the related ideas

Both au contraire and on the contrary encode the notion ‘the opposite is true’ and both relate ideas that oppose each other along at least one dimension. In the comparable corpus of speeches, the idea preceding on the contrary is always marked as unrealized (irrealis) by, for instance, a negative, an interrogative or a modal expression. In the newspaper data, almost 90% of the contexts of on the contrary have a negation of the first idea, and most of the rest have some modal marker of unrealized situation. On the contrary refers back to the idea itself; it does not include the irrealis marker in its scope. The second idea is nearly always a realized situation (or other assertion of truth). In (9), for instance, on the contrary refers back to, and denies, the idea emphasizing the role of the nation state is somehow contrary to the aspirations of the founding Fathers of the Community, which is marked as irrealis by the evidential it is sometimes said that. 9. It is sometimes said that emphasizing the role of the nation state is somehow contrary to the aspirations of the founding Fathers of the Community. On the contrary. The Treaty of Rome was a Treaty between nation states. (Malcolm Rifkind, 10.03.1997) The reason on the contrary does not fit in examples (4) and (8) seems to be that it necessarily denies the previous idea, as well as or rather than introducing the opposite. This would account for its being restricted to contexts where it follows an idea that is marked as dubious or false. By contrast, the only requirement for au contraire to occur felicitously is an opposition: the presence of two antonymous ideas (a condition met by examples (4) and (8)). Although denial of the idea preceding au contraire may be inferred in many contexts, denial does not seem inherent to au contraire. The claim, then, is that on the contrary always conveys denial while au contraire does not. However, in the newspaper text, a couple of ambiguous cases that could be exceptions were found, such as example (10). 10. …poor countries are now told rapid liberalisation is the key to faster growth. On the contrary, the west should open up its own protected markets before it starts giving lectures to the developing world. (The Guardian, 30.11.2002) At first sight, there is no natural incompatibility in (10). Rapid liberalisation and open up protected markets are clearly very similar ideas. On the contrary can hardly, therefore, refer back to rapid liberalisation is the key to faster growth. It is possible that in this example stance is the basis for a contrast between two views of what the west should or should not do. The intention behind on the contrary here may be to contrast poor countries are now told [by the west]… with the west should not give lectures yet. If so, then the ideas are reversed in that here it is the second idea, introduced by on the contrary, that is an unrealized situation (should

148 Pragmatic Markers in Contrast open … before it starts), while the first idea is realized (are told). The contrast, then, may be on a scale of stance, in which case this pattern is parallel to that of au contraire in example (4). An alternative reading is that on the contrary somehow points to an opposition between poor countries and the west, such that the implicated message would be “it’s not poor countries who should liberalize; on the contrary, it’s the west”, with rhematic poor countries and the west, in which case we have the typical on the contrary pattern of ‘false idea + on the contrary + correct idea’. In either case, this seems to be a highly unusual context for on the contrary. 3.5.

Thematic structure and nuclearity of ‘au contraire’ and ‘on the contrary’ constructions

The most frequent thematic structure (32/41 or almost 80%) of the contexts in which on the contrary occurs in the data has a first idea consisting of a theme (T1) bearing a rheme (R1), followed by a second idea consisting of the same theme (T1) followed by a contrasting rheme (R2).8 What contrast are the two rhemes: typically, speaker/writer claims that R1, which might be entertained by hearer/reader, is false, while R2 is the actual state of affairs. The theme stays constant, providing a strong common element, so that the situations are similar and the contrast will therefore tend to be objective. In 23 of the 32 cases of this type of construction (or around 56% of the total),9 there is a contrast of the kind described in section 2 as ‘contrary’. A parallel pattern was found for au contraire (45% of the total). The tendency, therefore is for both on the contrary and au contraire to occur in contexts of objective contrast between two rhemes of a single theme where the first is denied and the second asserted. The pattern is illustrated in table 3.

Table 3: On the contrary and au contraire: most frequent pattern (Pattern 1)

Type of contrast

Contrary

Thematic structure

T1 - R1

on the contrary T1 - R2

Status

Unrealized

Realized

Nuclearity

Satellite

Nucleus

Example Eng.

an enlarged NATO will threaten no one

on the contrary it will enhance stability (Malcolm Rifkind, 04.09.1995)

Example Fr.

le destin de la France, je le répète, n'a jamais été de se replier sur l'hexagone

il est au contraire de se projeter vers l'extérieur (Jacques Chirac, 26.08.1999)

8 ‘Theme’ and ‘rheme’ as used here are pragmatic categories similar to those known as ‘topic’ and ‘comment’ in some frameworks; ‘theme’ refers to what the message is about; ‘rheme’ to what is said of the theme. 9 For comparison, 60% of occurrences in the English newspaper corpus conformed to this pattern.


Both au contraire and on the contrary usually signal an asymmetrical relation (unlike, for instance, by contrast, on the other hand, or en revanche which tend to link two equally salient ideas). That is, one idea (the ‘Nucleus’ of the relation) is favoured and is given informational salience, while the other (the ‘Satellite’) is disfavoured and is backgrounded. As pointed out by Oberlander and Moore (1999), a relational pragmatic marker only attaches to a Nucleus when the Satellite precedes the Nucleus, and this is the case for au contraire and on the contrary, which attach to the second idea, the one favoured by the speaker. This asymmetry naturally results from Pattern 1, ‘false idea + on the contrary/au contraire + correct idea’, where the speaker will clearly want to give more salience to the correct idea. A further 20% of the contexts of au contraire are accounted for by a second pattern (Pattern 2), shown in table 4 with example (1).

Table 4: Au contraire: second most frequent pattern (Pattern 2)

Type of contrast

Perceived incompatibility

Thematic structure

T1 - R1

au contraire T2 - R2

Status

(Un)realized

(Un)nrealized

Nuclearity

Nucleus

Nucleus

Example Fr.

La surveillance prudentielle vise à traiter les difficultés avant qu'elles ne soient rendues publiques.

La régulation des marchés vise au contraire, à ce que tout soit sur la place publique. (Laurent Fabius, 14.06.2001)

A further example is given in (11): 11. [… beaucoup pensent à tort que la guerre d’Algérie n’est pas étudiée dans les classes,] … certains s’indignent, également à tort, que la pratique de la torture ou la répression du 17 octobre 1961 n’y soient pas évoquées, alors que pour d’autres au contraire, y seraient systématiquement occultés les massacres de Pieds-noirs , d’Algériens et de harkis, perpétrés par le FLN. (Jack Lang, 29.08.2001) ‘[… many people mistakenly think that the Algerian war is not studied in class,] … some object, also mistakenly, that torture and the repression of 17 October 1961 are not mentioned, while others au contraire think that the massacres by the NLF of Pieds-noirs, Algerians and ‘harkis’ are systematically concealed’ The contrast is between two different beliefs (R1 and R2) held by two different groups of people (T1 and T2). It is unlikely, though not inconceivable, that a person could hold both beliefs or that both could be true, and the speaker presents them as incompatible. The speaker holds that both beliefs are mistaken, so both ideas are marked as unrealized, they have equal salience, and are merely contrasted. In this pattern, then, both themes and rhemes are different. Contradiction is therefore not possible, and it is unlikely that the ideas will be mutually exclusive. The ideas may be realized or not, but they have the same status. En revanche also fits this pattern (example (1)),

150 Pragmatic Markers in Contrast and it is the commonest pattern to occur with on the other hand in the English data (example (6d)). If on the contrary fits the pattern at all, it will be a marginal use. No examples were found in the comparable corpus or in the newspaper data. The non-equivalence of au contraire and on the contrary is not primarily due, we suggest, to the degree of contrast between the related ideas or to the types of construction they can enter into, but to the differing senses of the markers themselves. On the contrary is inherently negative. Historical evidence, however, suggests that on the contrary was until recently more like au contraire (example 12): 12. a.

for as the Bear is conceiv’d and born suddenly; and being born is but short-liv’d: So on the Contrary, the Elephant is said to be two years in his dams belly (Walton, 1653-1676)

b. any Tax laid on Foreign Commodities in England, raises its price, … But on the contrary a Tax laid on your Native Product, … lessens their price, (Locke, 1691) (12a) and (12b) conform to Pattern 2 described in table 4. They are strange to present-day English, which suggests a semantic shift has taken place in the direction of on the contrary becoming more negative. With regular use in Pattern 1, the implicature ‘not x’ from the reading ‘the opposite of x’ might semanticize and induce a reanalysis from ‘[not x] [on the contrary, y]’ to ‘[not x] [on the contrary] [y]’ to ‘[not x, on the contrary] [y]’. We have already seen in the examples of contradiction (5) that both au contraire and on the contrary can simply negate an idea rather than state that its opposite is the case, a use that may result from reanalysis on the above lines. Prosodic information is needed as a guide to the structure in each case, though punctuation provides some clue to the speaker’s analysis, as in example (13): 13. Mais les causes d’instabilité mondiale n’ont pas disparu avec le développement économique, au contraire. La crise qui s’est déclenchée en Asie en 1997 et 1998 a montré la fragilité de notre système international. (Jacques Chirac, 14.06.2000) ‘But the causes of global instability have not disappeared with economic development, au contraire. The crisis triggered in Asia in 1997 and 1998 showed that our international system is fragile’. Here, as in (5), the sense also suggests the pragmatic marker goes with the first idea, not the second; that it is intended to reaffirm denial of the previous idea rather than to assert its opposite. Au contraire and on the contrary, then, can both signal denial. But unlike au contraire, on the contrary has lost the usages where no denial is implied. Possibly the frequencies of these usages were too low to sustain them.

4.

FREQUENCY DIFFERENCES

It has been seen that a possible explanation for the considerably higher frequency in the data of au contraire compared with on the contrary is that the former is appropriate in a wider range of contexts. In particular, it is often found in contexts of even-handed contrast between two ideas, while on the contrary seems restricted to the more argumentational contexts in


which the speaker/writer rejects the first idea as false and replaces it with the preferred, second, true idea. A second possible explanation is that other pragmatic markers (such as instead, rather, quite the opposite, etc.) are common in English in the same argumentational context. Third, it may be that contrast is less used as a rhetorical strategy in one language than the other in the comparable genre. Looking at the rhetorical pattern in which one idea is declared false in order to highlight the importance of another, right idea, we often find adversative in fact (example 14). It occurs about as frequently in the comparable corpus (33pmw) as on the contrary (29pmw). 14. the age profile of our troops, frequently seen as a problem, was in fact a benefit in SFOR (Michael Smith, 08.01.1998) In this context, in fact could adequately be replaced by on the contrary and translated by au contraire, despite the different scope. In fact refers back to the epistemic status of the first idea, that is, to seen as, whereas au contraire would refer back to the idea itself, that is, a problem. In fact claims that those people who thought it was a problem were wrong; au contraire would claim that what those people thought was wrong. Where the first related idea contains an evidential, as in (14), in fact (with rise-fall intonation) carries an adversative sense. But where the first idea contains a negation, as in (15), in fact (with level intonation) is used in an elaborative sense, and refers back to the negation of the first idea.10 15. a.

Building a free society from over 70 years of centralised rubble was never going to be easy, …In fact it requires sustained and thoughtful reform. (Michael Portillo, 30.06.1994)

b. The recent leaks have not allayed my concerns. In fact, they have significantly increased my concerns. (Joe Walsh, 10.03.1998) Elaborative in fact in examples (15) is a marker of a consonant relation. However, these are precisely the types of context (containing contrasts such as easy vs sustained and thoughtful; allay my concerns vs increase my concerns) where au contraire is typically found in the French data (cf. example (3)). This suggests that some situations that are typically described adversatively in French may be more frequently described elaboratively in English. A danger of using translation corpora is that this type of rhetorical preference may be obscured in translated text.

5.

CONCLUSION

This paper has suggested that in order to establish a basis for cross-linguistic comparison of relational pragmatic markers, we need to envisage a language-independent coherencerelational space on which the semantic range of expressions can be mapped. For adversative markers, four language-independent parameters have been suggested as potentially useful for 10

It is possible that elaborative in fact, which postdates adversative in fact, arose via a scope reanalysis. On in fact, see Aijmer (1996), Lewis (1998), Schwenter and Traugott (2000), Traugott (1999, 2003).


describing what constrains the choice of marker, and for describing partial translation equivalence across languages. It may well turn out that these parameters are not independent of one another. It has also been argued that, in order to distinguish central or prototypical uses from peripheral ones, we need comparable corpora, where we can look for interrelations and patterns among the pragmatic contexts that the markers occur in, the senses of the marker and the constructions they can enter into.

REFERENCES Aijmer, K. (1996). Conversational Routines in English. London: Longman. Anderson, L. B. (1986). Evidentials, paths of change, and mental maps: typologically regular asymmetries. In: Evidentiality: The Linguistic Coding of Epistemology (W. Chafe and J. Nichols, eds.), pp. 273-312. Norwood, N.J.: Ablex. Baker, M. (1993). Corpus linguistics and translation studies: implications and applications. In: Text and Technology (M. Baker et al, eds.), 233-250. Amsterdam: John Benjamins. Borin, Lars 2002 “… and never the ‘twain shall meet?” In: Parallel Corpora, Parallel Worlds, (L. Borin, ed.), pp.1-43. Amsterdam: Rodopi. Croft, W. (2001). Radical Construction Grammar: Syntactic Theory in Typological Perspective. Oxford: Oxford University Press. Cruse, D. A. (1986). Lexical Semantics. Cambridge: Cambridge University Press. Halliday, M. A.K. (1994). An Introduction to Functional Grammar 2nd edn. London: Arnold. Haspelmath, M. (2003). The geometry of grammatical meaning: semantic maps and crosslinguistic comparison. In: The New Psychology of Language: Cognitive and Functional Approaches to Language Structure, vol.2, (M. Tomasello, ed.), pp. 211242. Mahwah: Erlbaum. Hovy, E. H. and E. Maier (1995). Parsimonious or profligate? How many and which discourse structure relations? Unpublished ms. University of Southern California Information Sciences Institute. Knott, A. and C. Mellish (1996). A feature-based account of the relations signalled by sentence and clause connectives. Language and Speech 39(2-3), 143-183. Lewis, D. M. (1998), From modal adverbial to discourse connective: some rhetorical effects in present-day English. In: Pragmatics in 1998: Selected Papers from the 6th International Pragmatics Conference, vol.2, (J. Verschueren ed.), pp. 363-375. Antwerp: International Pragmatics Association. Locke, J. (1691). Some considerations of the consequences of the lowering of interest and the raising of the value of money. In a letter sent to a Member of Parliament, 1691. London. Printed for Awnsham and John Churchill, at the Black Swan in Pater-NosterRow. Louwerse, M. (2001). An analytic and cognitive parameterization of coherence relations. Cognitive Linguistics 12, 291-315. Mann, W. C. and S.A.Thompson (1987). Rhetorical Structure Theory: A Theory of Text Organization. ISI Report RS-87-190. Marina del Rey, CA: Information Sciences Institute, University of Southern California.


Oberlander, J. and J.D. Moore (1999). Cue phrases in discourse: further evidence for the core:contributor distinction. Paper presented at the workshop Levels of Representation in Discourse, University of Edinburgh, July 7-9, 1999. Sanders, T. J. M., W. P.M. Spooren and L.G.M. Noordman (1993). Coherence relations in a cognitive theory of discourse representation. Cognitive Linguistics 4, 93-133. Schwenter, S. and E.C. Traugott (2000). Invoking scalarity: the development of in fact. Journal of Historical Pragmatics 1(1), 7-25. Traugott, E. C. (1999). The rhetoric of counter-expectation in semantic change: a study in subjectification. In: Historical Semantics and Cognition (P. Koch and A. Blank, eds.), pp. 177-196. Berlin: Mouton de Gruyter. Traugott, E. C. (2003). Constructions in grammaticalization. In: The Handbook of Historical Linguistics (B.D. Joseph and R.D. Janda, eds), pp. 624-647. Oxford: Blackwell. van der Auwera, J. and V. A. Plungian (1998). Modality’s semantic map. Linguistic Typology 2, 79-124. Walton, I. (1653-1676). The Compleat Angler. Edition by J. Bevan. Oxford: Clarendon Press, 1983. Reproduced in The Helsinki Corpus of English Texts.



THE SPANISH DISCOURSE MARKERS O SEA AND PUES AND THEIR ENGLISH CORRESPONDENCES Anna-Brita Stenström, Bergen University

1.

INTRODUCTION

The number of studies devoted to discourse markers is steadily growing due to the awareness of their crucial importance in spoken interaction. But although discourse markers are bound to occur in all languages in one form or another, the contrastive aspect has so far attracted little linguistic attention. This paper, which compares the use of two common Spanish discourse markers with what I assume to belong to their closest correspondences in English, is a modest attempt to bridge the gap. The paper begins with a brief overview of the ten most frequent discourse markers used in Spanish and English natural conversation as manifested in two corpora representing adolescent conversation, El Corpus de Lenguaje Adolescente de Madrid (COLAm) and The Bergen Corpus of London Teenage Language (COLT), and two representing adult conversation, El Corpus Oral de Referencia del Español Contemporáneo (COR92) and the spoken part of The British National Corpus (BNC).1 In the following sections, I concentrate on the use of the Spanish markers o sea and pues and suggest what I take to be their nearest English counterparts. As regards the functional typology, I start out from the different functions of o sea and pues already established in earlier studies, in particular, Schwenter (1996) and Briz (2001) for the use of o sea, and Briz & Hidalgo (1988), Martín Zorraquino (1991), Porroche Ballesteros (1996) and Briz (2001), for the use of pues. I then suggest a correspondence in English for each of these functions. Examples from the above-mentioned corpora, as well as from the studies referred to, accompany the account. The translations of the Spanish examples are my own, with the exception of the examples taken from Schwenter (1996).

2.

THE TOP TEN

As illustrated in Table 1, both o sea and pues belong to the ten most common discourse markers in Spanish conversation, and so do well and cos, which seem to be among their closest counterparts in English conversation.

1

For a description of the corpora, see www.cola.tk, www.hit.uib.no/colt, ftp://ftp.lllf.uam.es/pub/corpus/oral, http://escorp.unizh.ch.


Table 1: The top ten discourse markers per 1,000 words

Spanish COLAm 5,7002 words no sí/si4 o sea pues bueno eh tía ya mira claro

42 17 6 5 5 4 4 4 3 2

COR92 500,000 words no sí pero ya bien o sea pues hombre bueno mira

English

33 12 9 8 3 3 3 3 2 1

COLT 435,000 words yeah no know oh like just well right cos really

17 11 10 9 7 6 5 4 4 4

BNC/spoken 4,200,0003 words yeah 14 oh 10 no 8 well 6 but 6 just 5 know 5 mm 5 yes 4 like/ 3 cos

(The number of know instances includes you know.) Due to the very different sizes of the corpora (presented in the table heading), the raw frequency counts have been normalized in order to give a more reliable picture. A word of caution is in place, however. Since the number of words is based on untagged versions of the corpora, the figures in Table 1 are calculated on the total number of instances of each word. This means that some of the figures should not be taken at face value, in that a certain word may serve other functions in addition to its functions as a discourse marker. Incidentally, the altogether most frequent word in Spanish, que, has not been included among the top ten, since separating the discourse functions from its functions as a conjunction, a relative pronoun, an interrogative pronoun (written qué) and qué used in exclamations (e.g.¡Qué pena! ‘What a pity!’) would simply have been too laborious for the purpose of this paper. Had it been included, it would have made up almost half of the total occurrence of the top ten items. Example (1) gives an idea of its use: 1. Y les dijo que … que nada, o sea que … que ni hablar, que … que para nada, vamos; que era una operación totalmente desaconsejada, que no se le ocurriera y tal, que tampoco ‘And he told them that … that nothing, that is that … that no way, that … that there was no question of, come on, that it was a totally inadvisable action, that it would never occur to him and things like that, nor that (incomprehensible)’ (COR92:022B) Table 1 indicates that the adult and the adolescent speakers use roughly the same top ten markers in Spanish as well as in English, but that the distribution differs somewhat between 2

This figure represents the approximate number of words transcribed when the paper was written. This figure represents the portion of the spoken material that consists of conversation only. 4 No distinction is made here between the conjunction si, spelt without the accent, and the response marker sí. 3

The Spanish Discourse Markers O Sea and Pues 157 the age groups. The positive and negative markers belong to the most common ones in all four corpora, but unlike in English, the Spanish negative marker (no) is strikingly more frequent than the positive marker (sí). This discrepancy can be explained by the fact that Spanish no serves not only as a response and a response marker but also as a frequent appealer for feedback, in addition to its use as a negation marker (no sé ‘I don’t know’), as illustrated in (2): 2. H1: H4:

Tienes por aquí cerca el coche ¿no? ‘you have the car near here, right?’ Sí … ‘Yes …’ (COR92:006B)

The positive marker sí is only occasionally found in turn-final position. One example is (3), where ¿o sí? turns a declarative utterance into a request for confirmation: 3. Ana: lo que pasa es que no tiene dos pisos ¿o sí? ‘the thing is that it doesn’t have two floors, or does it?’ María: sí sí que tiene dos ‘yes yes it has two’ (COLA01-1) Notice that o sí only appears after negative utterances, while no, which typically appears after a positive utterance, can also occur after a negative utterance (e.g. no participan me imagino ¿no? ‘they don’t participate I imagine, do they?’). Eh corresponds to a tag-question regardless of the polarity of the preceding clause, whether it is negative (e.g. no sabes escribir ¿eh? ‘you can’t write can you?’) or positive as in (4), where the speakers talk about conditions in schools run by nuns: 4. H1: H3: H1:

… pero hay otras que son unas brujas ‘but there are others who are real bitches’ Igual pasa con los curas, ¿eh? ‘it’s the same with the priests, isn’t it’ Pues sí … ‘Well yes …’ (COR92006C)

Eh – which also occurs in English conversation though far less frequently – is only represented in one of the age groups in the top ten ranking lists, and so are some of the other markers, notably the vocative tía (‘girl’) and the response claro (‘of course’), which predominate in the Spanish adolescents’ speech, while the connective pero (‘but’), the response marker bien (‘good’, etc.) and the vocative/expletive hombre (‘man’/’goodness’) are more common in the adults’ speech. As regards English, really and right are more often used by the adolescents, while the connective but is more often used by the adults – like Spanish pero – and so are the response signals mm and yes. It is also worth noticing that attention signals, here represented by mira (‘look’), are far more common in Spanish than in English conversation. This agrees with the findings reported in Romero Trillo (1997), which showed that adult Madrid speakers’ use of mira was five times as frequent as adult London speakers’ use of look.

158 Pragmatic Markers in Contrast More importantly, considering the object of this study, both o sea and pues, and cos and well, are found among the top ten markers, with o sea and pues in the same order of frequency in the two Spanish corpora, while cos is used somewhat more often by the adolescents and well considerably more often by the adults in the English corpora. The extent to which cos and well might be suitable renderings of o sea and pues will be dealt with in Section 3.

3.

O SEA AND PUES AND THEIR ENGLISH EQUIVALENTS

Due to their frequent use in spoken interaction, o sea and pues, like discourse markers in general, have undergone grammaticalization (cf. e.g. Brinton 1996: 50 ff.), which goes hand in hand with their multifunctional use. There is consequently more than one way of translating them into English. As will be seen, the functional typology differs in many respects between the two markers, but they have at least one discourse function in common, that of serving as a cohesive device and maintaining discourse coherence. Clearly, the inherent multifunctionality of the markers, which has the effect that they are doing more than one thing at once, makes them very difficult to classify. Therefore, my attempt to present a classificatory overview of the functions of o sea and pues and what I suggest as their closest English correspondences (displayed in Figure 1, Section 3.2.5) should be taken with a pinch of salt. To facilitate the comparison, I will discuss the various functions in terms of three main levels: the discursive level, which has to do with the organization of the discourse, the pragmatic level, which has to do with the interpretation of what is being said, and the interactional level, which is related to the interplay between speaker and addressee. 3.1 3.1.1

O sea General

O sea is a lexicalized phrase consisting of the conjunction o (‘or’) and the 3rd person singular present subjunctive form of the verb ser (‘to be’) (cf. Schwenter 1996). In his article, Schwenter discusses two main uses of o sea, the connective use, which initiates a clarification of a preceding utterance and links the clarification to what preceded while marking a semantic link between the utterances, and the epistemic use, which reflects the speaker’s degree of commitment to what s/he says and increases the modal content (cf. 1996: 855). Cortés Rodríguez (1991) interprets some very frequent uses of o sea as ‘muletillas’ (‘crutches’), which are void of semantic content and generally used unconsciously, i.e. more or less like pre-fabricated planning ‘fillers’ in the terminology of Brown and Yule (1985: 15, 17), i.e. a function that belongs to the discursive level. In a similar vein, Christl (1996) discusses discourse markers in general in terms of muletillas. He makes an important distinction between two types, ‘muletillas de función expletiva’, which help the speaker overcome moments of hesitation with a stalling effect, and ‘muletillas de función comunicativa’, with important functions for the organization of the discourse. O sea belongs to the latter type and is referred to, more specifically, as a ‘muletilla de reformulación’, which

The Spanish Discourse Markers O Sea and Pues 159 signals a different way of saying what has just been said, or a ‘muletilla de corrección’, which helps the speaker correct what he just said (1996: 127–128). Serrano (2002) regards o sea only as a ‘conector reformulativo’ (‘reformulating connective’) and sees it on a par with items such as digamos (‘let’s say’), es decir (‘that is’) and por ejemplo (‘for example’), which link utterance units on the pragmatic-discursive level, and whose functions, like those of discourse markers in general, can only be determined through a detailed analysis of the contexts in which they occur (2002: 151). As regards the position of o sea in the discourse, Schwenter compares it to you know, which ‘may occur at virtually any point in the syntactic structure’ (cf. Erman1987: 68). With the turn as a starting-point, this means that o sea can occur at the beginning and the end of the turn as well as in turn-medial position. From a formal and structural point of view, Schwenter finds that o sea is prosodically independent and morphologically fused The difficulty of translating o sea is reflected in the fact that, in his examples, he has translated all but the marker itself. His suggestions more generally regarding possible renderings are that is and so and sometimes maybe or I guess – but neither cos nor well, which I venture to propose for some of the occurrences. 3.1.2

The discursive level

This level involves o sea as a connective device, the core function of which, following Schwenter (1996) and Casado Velarde (1988), is to introduce an explanation or a reformulation. This, according to Schwenter, can be done by clarifying a preceding utterance: 5. A:

Mi tía Pepita viene el día 20, o sea el viernes que viene. ‘My aunt Pepita is coming on the 20th. O SEA next Friday.’ (Schwenter 1996: 856)

by marking background in a narrative: 6. Y entonces llegamos .. nos bajamos del autobús y vemos a la Mari Carmen o sea ella ya estaba allí, no? ‘and so we arrive .. we get off the bus and we see Mari Carmen O SEA she was already there, you know?’ (Schwenter 1996:860) or by adding information to a preceding answer to a question: 7. A: B:

[…] sí, conoces a ése? ‘[…] yes, do you know that guy?’ no, o sea, le conozco de oídas. ‘no, O SEA, I’ve heard of him.’ (Schwenter 1996: 860)

English that is would be the obvious rendering in (5), cos would not be unlikely in (6) besides well, while well would be the nearest corresponding marker in (7), reflecting the non-specific response. In example (8), where o sea introduces a conclusion (cf. Schwenter 1996: 863), the best, or maybe only, English rendering is so:

160 Pragmatic Markers in Contrast 8. M: R:

le han dicho, que ya no le dan {prórroga}. ‘they’ve told him, that they won’t give him another extension.’ {o sea:} ya tiene que: irse a Valencia , no? ‘O SEA now he has to go off to Valencia, right?’ (Schwenter 1996: 863)

As mentioned in Section 3.1.1, both Cortés Rodríguez (1991) and Christl (1996) regard o sea as a filler (‘muletilla’), with little semantic content but with an important function in the discourse organization, helping the current speaker to hold the turn while planning what to say next. This is illustrated in extract (9), where two girls are worried about who will listen to the recordings and the speakers’ anonymity: 9. A:

B:

las profesoras no lo escuchan ¿eh? y ellos – en plan saben o sea a mi por ejemplo me ponen yo soy la número uno tu la número dos (nombre) la tres en plan ‘the teachers don’t listen, do they? and they – like know well me for example they put me I am number one you number two (name) number three like …’ ¿te estás grabando a tí misma? ‘are you recording yourself?’(COLA06-1)

This is a good example of o sea as an organizing device in a situation where the speaker is in need of planning. It is accompanied by hesitation features such as a dash indicating a pause, en plan indicating uncertainty, and reformulation. As regards the variant o sea que, Schwenter regards it as a marker of conclusion or consequence to be translated by so or therefore, while Terrádez Guerra (2001), who studied a concordance of o sea based on a 100,000-word subcorpus from the VALESCO5 material, suggests several interpretations. He discovered that o sea + que appeared among the most characteristic collocations and found that this form is used, for instance, to reinforce what is being proposed or to introduce a new piece of argumentation. But he also found that it can be used by the current speaker, in turn-final position, to show that s/he expects the hearer to complete the argumentation by supplementing missing information. 3.1.3

The pragmatic level

This level involves what Schwenter refers to as the epistemic use of o sea, which reflects the speaker’s degree of commitment to what s/he says and typically occurs with 1st person singular subjects and a belief verb (cf. Schwenter 1996: 864–865). This usage is discussed in terms of hedging, as in (10), where o sea occurs in the middle of a turn: 10. Pues que pienso que es positivo, porque, o sea, a mi modo de ver … ‘well I think that it is positive, because, O SEA, in my view …’ (Schwenter 1996: 864)

and in (11), where it is found at the end of a turn:

5

VALESCO is an acronym for VALencia ESpañol COloquial.

The Spanish Discourse Markers O Sea and Pues 161

11. A:

Creo que el menu es un poco caro, o sea ‘I think that the menu is a little bit expensive, O SEA.’(Schwenter 1996: 865)

Well seems to be a suitable rendering of o sea in (10) and sort of in (11), or maybe isn’t it, although a little bit already has a hedging effect, which suggests a zero rendering. The view of o sea as an epistemic device, hedge, differs then from that of Cortés Rodríguez (1991), who regards o sea reflecting hesitation as a planning device, which simply helps the speaker to go on speaking (cf. turnholder, section 3.1.5). As regards o sea occurring at the end of an argumentation, Briz (2001: 218) suggests that it sometimes reinforces what was being said: 12. Es que no tiene explicación/ no es/ simplemente no/ o sea Ya// Pero bueno . pero bueno/ pero es que algo te debe pasar ¿no?o sea ‘There is no explanation, it’s certainly not simply no. But OK . but OK but something has to happen to you, don’t you think surely.’ (Briz 2001:218) or minimizes what was being said: 13. Yo no diría eso o sea ‘I wouldn’t say this, I think’ (Briz 2001:218) This function resembles what Schwenter calls hedging in (10) and (11), the difference being that in (13) o sea refers to the entire preceding utterance, while o sea in (10) and (11) refers only to a preceding adjective. In all three cases, the use of o sea reflects the speaker’s degree of commitment to what he says. O sea prefacing an imperative is considered to have a diminishing effect on the force of the command, as in (14) (abbreviated) (Schwenter 1996: 862) where I suggest well as a plausible English rendering: 14. A:

B:

… tienes que acordarte de quiénes estaban ‘you have to remember who were in them’ …………… O sea, que no memorices tanto. ‘O SEA, don’t study too much.’ (Schwenter 1996: 862)

Another interesting use is that of introducing a direct quote, which Schwenter (1996: 866– 867) says is related to the epistemic use in that the speaker does not take full responsibility for the ‘correctness’ of the upcoming quotation, as in example (15) (abbreviated): 15. R: que ha: sonado el teléfono y: cuando he contestado, = la mujer dice, o sea::.’está, está Paco? ‘the telephone rang and when I answered,= the woman says, O SEA. is, is Paco there? (Schwenter 1996: 866–867) The corresponding quotation marker in English is like. There is a difference in usage, however: in Spanish, o sea is preceded by a reporting verb (dice), while in English, especially in teenage talk (cf. Andersen 1998: 147–170; Stenström et al. 2002: 108 ff.), the reporting

162 Pragmatic Markers in Contrast verb say is replaced by BE+ like, sometimes by GO + like or the verb GO on its own, as in (16) – (18): 16. And I’m like, and I’m like, scum! (COLT 141707: 343) 17. And then he goes like, sorry man, close the door and get out (COLT 139003: 21) 18. He goes, nah mate! Ain’t you lot ever heard of tea bags? (COLT 132617: 101) SAY + like is occasionally met with, however, which would then correspond fairly closely to the Spanish usage: 19. She slides down the banister and says like blurgh, la blah la blah loo! (COLT 135805: 87) 3.1.5

The interactional level

O sea is also used as a structural device on the interactional level, where it serves as a turntaker (20), a turnholder (21), and (22) a turnyielder, as illustrated in the following examples from COLA and COR92: 20. J2y: en las olimpiadas puedes llevar pantalón corto pero aquí no puedes llevar pantalones cortos sino que tienes que llevar pantalón largo ‘in the Olympics you can wear short trousers but here you cannot wear short trousers but you have to wear long trousers’ J2E: o sea salir del colegio tienes que salir con pantalones cortos ‘that is outside the college you have to wear short trousers’ (COLA04-1) 21. H1: Además es que, yo y otra chica, o sea, pero, hija, bueno, son cosas … mira ‘Besides, me and another girl, that is, but girl, well, there are things … look … ‘ (COR92 018B) In (20), A yields the turn, and B can just take over by continuing the previous argumentation, as it were. In (21), it is obvious that the speaker does not really know how to proceed, but o sea helps her to hold the turn by showing that she wants to continue. Extract (22) is an example of o sea serving as a turnyielder: 22. A: B: A:

… tía tía mañana es las olimpiadas y luego el viernes ya está o sea … ‘tomorrow is the Olympics and Friday is already here in other words …’ ¿cuanto dura el disco? ‘how long does the disc last’ ochenta minutos tía ‘eighty minutes Ø’ (COLA38)

Speaker B is apparently not interested in the Olympics, and this topic comes to an abrupt end when she asks about the duration of the disc. The meaning of o sea is uncertain. Tía, which is

The Spanish Discourse Markers O Sea and Pues 163 a colloquial word for ‘girl’, is an extremely frequent term of address in some girls’ conversations, but has no direct correspondence in an English conversation. 3.2

Pues

3.2.1

General

Pues derives from the Latin adverb and preposition post, which developed into a causal and consecutive conjunction in Spanish and later into a discourse marker (cf. e.g. Martín Zorraquino 1991: 273). The core function of pues, according to Martín Zorraquino and others, is to serve as a cohesive element in spoken discourse, where it signals communicative continuity while reinforcing the rhematic elements. But as Porroche Ballesteros (1996: 72) points out, there is still no agreement in traditional Spanish grammar books as to its status. Some of the functions of pues correspond to the functions of the English short form cos from (because). Both are, for instance, used as a causal connector, introducing a subclause, as in (23): 23. Mañana no habrá clase, pues es fiesta. ‘There will be no class tomorrow, cos it’s a holiday’ (Porroche Ballesteros 1996:73) and as a thematic link at the level beyond the clause, as in (24): 24. H5: Si no tiene chimenea . No tiene chimenea . Pues entonces no comes, lo sentimos mucho ‘If she doesn’t have a fireplace . She doesn’t have a fireplace. Cos then you won’t eat, we are very sorry’ (COR 92) But pues has a much wider range of functions than cos, which are realized by other discourse markers in English, as in (25), where pues is used as a response marker, signalling a dispreferred answer: 25. 3:

Bueno. ¿Qué te vas a poner tú? ‘What are you going to wear?’ J3c: Pues no lo sé. Creo que <nombre> me deja su falda que se puso en la fiesta de<nombre> ‘Well I don’t know. I think will let me have the skirt she was wearing at ‘s party’ (COLA05-1)

or when signalling an answer that reflects some insufficiency by providing information indirectly, for instance when responding to an invitation such as ‘Can I offer you a drink’ by ‘I’ve got my car outside’ (cf. e.g. R. Lakoff, 1973; Svartvik, 1980; Levinson, 1983; Pomerantz, 1984; Stenström, 1984; Schourup, 1985; Briz, 2001). In some cases, pues would probably not be translated into English. One example is (26): 26. Andrea: Sí sé quién es. ‘Yes I know who she is’

164 Pragmatic Markers in Contrast Ana:

Pues sí, la prima de Marta ‘Ø yes, Marta’s cousin’ (COLA01-3)

where pues has a reinforcing effect. Both pues and cos (like o sea and well) can serve as an interactional device, which enables the speaker to hold the turn and signals that s/he intends to continue or further elaborate a prior statement (cf. Schleppegrell, 1991); it can act as a ‘post-completion extension’ (cf. Ford, 1993), and it can serve as a ‘take-off for further talk’ (cf. Stenström, 1998: 128 ff.). In the following, I will discuss the various functions of pues, except the clausal functions referred to above, on the basis of the descriptions by Briz & Hidalgo (1988), Martín Zorraquino (1988), Porroche Ballesteros (1996), Briz (2001), and others. 3.2.2

The discursive level

On this level, the connector pues serves to structure what the speaker says and maintains the discourse coherence. It is used, for instance, to mark a conclusion, as a restarter, as a topic transition and thematic link, a filler and a chunking device. In (27) pues corresponds to English then, signalling conclusion: 27. Si estás de acuerdo/ pues llamo a casa ‘If you agree/ then I will phone home’ (Briz 2001: 175) As a conversational restarter, pues introduces an utterance/turn after a brief silence, as in (28), where ahora (‘now’) is followed by a fairly long silence, indicated by the three dots: 28. H1: A ver si viene ahora … ‘Let’see if she comes now …’ H3: Pues haz una cosa, pásate por casa y quedamos … luego por allí ‘Well do one thing, go home and we stay here’ (COR92:006C) In English conversation, especially adult conversation, well is generally the most common choice in similar situations. Cos would be possible, but only if the following utterance were related to what the preceding conversation was about, or to something present or recoverable in both speakers’ minds. Well, on the other hand, can introduce a completely unrelated utterance. When used as a discourse organizer, pues has a demarcating effect by chunking units of information, as well as a unifying effect by linking what is currently being said to what preceded. Both well and cos can be used as demarcators, as is shown in (29) and (30), although the ability of cos to serve this purpose is relatively limited. Pues in (28) links B’s utterance to what A said. The English equivalent is well: 29. Ana:

… está en tu colegio ‘she is at your school’ Andrea: sí ya ya sé quién es ‘yes I know who she is’


Ana:

pues esa morena con rizos ‘well this dark (one) with curls’ (COLA01-3)

In (30), pues is uttered within the same speaker’s turn, linking two units of information: 30. H2: ¡Qué bueno! ¡Ay qué bueno! Pues yo es que me quiero ir a casa ya. Que quiero ir a ver la nota. ‘Good! Very good! Cos I – I would like to go home. I would like to see the message’ (COR92:034A) As a topic transition marker, pues helps the speaker to go smoothly from one topic to the next, as in (31) from Briz (2001), where pues is related to the discourse organizing function and probably best rendered by well: 31. … creo que en un concesionario de coches de lujo de esos que llevan los jeques pues después se casó con una sueca … ‘I think it was in connection with a licence for luxory cars of the type owned by sheiks well later he married a Swedish girl … ‘(Briz 2001: 208) It can also introduce additional information, as in (32), where cos might be an English alternative, besides well, linking up with and providing a background for what preceded: 32. H3: Se han dado de frente contra un camión. Pues sería en alguna curva sin buena visualidad. ‘They broke into a lorry. Well/Cos it apparently happened in a curve … without good visibility.’ (COR 92:018D) Pues as a thematic link is illustrated in (33), where the temporarily abandoned theme, voz bonita (‘beautiful voice’) initiated by H3 is resumed again after quite a few exchanges: 33. H3: … Tiene una voz muy bonita, ella. ‘She has a very beautiful voice.’ H1: Esa tía … puede tranquilamente competir con con sa voz. ‘This girl … can easily compete with with this voice.’ H5: Es profesora de <siglas EGB> y tiene dos niños. ‘She teaches acronyms EGB and she has two children.’ H3: Sí? ‘Really?’ H5: Está casada … con el de las gafas? ‘Is she married to the one with spectacles?’ H1: Sí. ‘Yes.’ H5: Y el otro es el hermano. ‘And the other is the brother.’ H2: El hermano de quién? ‘Whose brother?’ H5: De … ella.

166 Pragmatic Markers in Contrast ‘Hers.’ <silencio> H5: Fuimos una vez … y es que de verdad … ‘We went once … and really … H3: Pues tiene la voz muy bonita …’ ‘Well she has a very beautiful voice …’ (COR92:024A) Here pues is best rendered by well. Cos is a frequent thematic link in English adolescent conversation, as illustrated in extract (34) from COLT, where cos links the continuation of Celia’s narrative with what she said before the parenthetical remark, Sampson had just gone past, etc. (the incident is from a mixed boarding school): 34. Celia: … so I tap on the window. Sampson had just gone past and I was just about to go in his room, you know, cos he’s standing there and just tapped on the window (COLT: 142304) The continuative function is unmistakable when pues initiates a question that is closely related to what preceded, as in (35), where it links up with the ongoing discussion about weight: 35. H1: … en comparación con las demás está gordita. Pero fíjate, 38 kilos ‘compared to the others she is fat. But listen, 38 kilos.’ H2: Pero ¿ como es de alta? ‘But how tall is she?’ H1: Y tiene dieciseis años. ‘And she’s sixteen years old.’ H2: Pues es poco treinta y ocho kilos ¿no? Tú ¿cuánto pesas? ‘Well thirty-eight kilos is little, isn’t it? How much do you weigh?’ H1: Yo cincuenta. ‘Me fifty.’ (COR92:018B)

The nearest English equivalent in this case would be well, which also reflects B’s personal opinion about weight – or maybe but. Like Christl (1996), Porroche Ballesteros (1996) uses the term ‘muletilla’, i.e. ‘filler’, for ‘superfluous empty words’, which help the speaker to go on speaking (cf. Cortés Rodríguez, Section 3.1.1, for certain uses of o sea). This is illustrated in example (36), where pues, preceded by hesitation (que … que), has a stalling effect, signalling that the speaker does not know how to go on and needs time for more planning: 36. pues Zaragoza es una cuidad que … que, pues hace quinientos años debía ser una una ciudad muy bonita ‘well Zaragoza is a town that … that, well five hundred years ago must have been a very beautiful town’ (Porroche Ballesteros 1996: 78) However, most scholars agree that fillers are not just empty words, and Porroche Ballesteros (1996: 78), too, points out that, even in the filler capacity, discourse markers are at the same


time marking the rhematic quality of what follows. In English, well is frequently used in the same way, especially in combination with filled and/or unfilled pauses or other discourse markers, such as I mean or you know. Cos, as well, could sometimes be used in a similar context, by giving the speaker time and showing that s/he intends to go on speaking. According to Briz (2001: 208), one role of pues is as a kind of pre-posed punctuation marker. This is illustrated in (37), where the slash indicates that all five instances of pues initiate a clause:66 37. el otro día me pasó a mí un caso que se lo dije a Roberto no se lo quería decir pero había también me di cuenta /pues se lo voy a comentar el otro día pudimos oír la película porque se oía una conversación que mi hijo dice mamá /pues eso es de- el aficionao paraba al momento otra vez y así toda la película entonces ya no ha vuelto a hacer y yo le digo /pues yo le digo nada porque ya no ha vuelto a pasar pero ayer por la mañana le digo /pues me pasó eso y dice /pues eso a lo mejor de tarde en tarde y efectivamente no- la primera vez y ya no me ha vuelto a pasar os ha pasao algune vez esto a vosotros? (Briz 2001: 208) In English, both cos and well occur in a similar pre-posed position, initiating new units, but a far better candidate for the label punctuation marker is the post-posed yeah, as in extract (38) from COLT: 38. My dad used to work erm know this bloke in the pub yeah/ and this bloke used to like a pint of erm oh I can’t remember, export bitter, yeah/ I can’t remember what it’s like, Right? And there’s a little drop of lemonade yeah/ right that’s what I like to drink, right? Oh yeah, and his dog yeah/ liked it as well yeah/ they used to leave a tray out for him every night yeah/ when it was closing time the dog knew every time it closed. (COLT. 132405) In addition, yeah intensifies the relationship between speaker and listener in very much the same way as right, okay, innit and you know would have done, by acting as an appealer (cf. Section 2). 3.2.3

The pragmatic level

This level involves the use of pues as a response initiator and a reinforcing marker. As illustrated in example (25) in Section 3.2.1, both pues and well are typically used as response markers, known to initiate ‘dispreferred’ responses, i.e. responses that are insufficient in some way, and which do not meet the questioner’s expectations. Typical examples are (39), where speaker B declines an invitation (the lengthened vowel indicates hesitation): 39. A:

oye vente mañana a casa que hemos montado una fiestecita … ‘listen come to our place tomorrow cos we have organized a little party …’

6 Translation into English is irrelevant, since the only purpose of the example is to illustrate the point made by Briz.

168 Pragmatic Markers in Contrast B:

puees es que tengo un montón de trabajo ‘weell I’ve got loads of work’ (Briz 2001: 175)

and (40), where pues initiates a response that goes beyond the questioner’s expectations by giving more information than was asked for (tengo varios): 40. A:

¿Tienes alguna afición o algún hobby? ‘Do you have any preferences or a hobby?’ Pues sí, tengo varios. ‘Well yes, I’ve got several.’ (Poroche Ballesteros 1996: 75)

B:

According to Porroche Ballesteros (1996: 76), pues in (40) has a reinforcing effect: it reinforces the rhematic elements that follow the marker and puts special emphasis on the part of the theme already introduced. She further shows that, while responses to Spanish polar questions (‘preguntas absolutas’) tend to be realized simply by pues sí or pues no, open questions (‘preguntas relativas’) can be answered in three different ways: by pues followed by repetition of the topic of the question, which is followed by the new information (41): 41. A:

¿Dónde has pasado la infancia? ‘Where did you spend your childhood?’ Pues la infancia la pasé en un pueblo de Logroño. ‘Well my childhood I spent it in a Logroño village’ (Porroche Ballesteros 1996:

B: 75)

by a repetition of the topic preceding pues, followed by the new information (42): 42. A: B:

Y sobre el mercao ¿qué te parece el mercao central? ‘And as regards the market, what’s your opinion about the central market? ¿El mercao central?, pues lo han dejao muy majo y bastante higiénico y tal. ‘The central market, well it’s become very pleasant and hygienic and so on.’ (Porroche Ballesteros 1996: 76).

or by the new information without repetition of the topic (43): 43. A: B:

3.2.4

¿Y ello por qué? ‘And why this?’ Pues porque el nivel cultural es muy bajo … ‘Well because the cultural level is very low’ (Porroche Ballesteros 1996: 75) The interactional level

Like o sea, pues is used as a turntaker, a turnholder and a turnyielder. In (44), the girls talk about what to wear at a party. Here, pues serves as an ‘uptake’ (cf. Stenström, 1994: 72), typically introducing a response: 44. A:

tú ¿qué te vas a poner? ‘you what are you going to put on?’


B:

pues yo me voy a poner una falda de vaquera y luego por arriba es que no lo sé ‘well I am going to put on a jeans skirt and then at the top I don’t know’ (COLA05-1) (44) is an example of ‘smooth’ turntaking. A yields the turn, and B simply takes over. Pues can also be used to interrupt the current speaker. Cos, which, unlike well, never carries syllabic stress, is a less likely English marker used for this purpose than well. In (45), the two instances of pues are used differently. A is not certain about which school two friends go to: 45. A:

B:

a pues yo pensaba que estaba en el Santa María pero no sé quién de estas va al al Santa María ‘ah because I thought that she was at Santa María but I don’t know which of them goes to to Santa María’ bueno pueees esta niña es anoréxica yyy entonces sabes – que solamente looo lo sabe ella antes su mejor amiga ‘okey cooos this girl is anorectic and then you know – that she only knows it before her best friend’ (COLA31)

The first instance of pues is used in the same way as pues in (44), i.e. as an uptake, while the lengthened vowel in the second instance is a clear indication that B intends to go on speaking, which means that pueees acts as a turnholder. The turnyielding function, finally, is demonstrated in (46), which is an extract from a discussion about the art of selling: 46. H1: … es que hay que saber venderse . es decir hay que decir ‘Yo soy el mejor’ y ‘Yo soy el que más pido’ . entonces pues … ‘one has to know how to sell oneself . that is one has to say ‘I am the best’ and ‘I am the one who asks for most’ . then well … ‘ H2: no … es verdad hay que decir … en todos los temas ‘no … it’s true one has to say … in all cases’ (COR92-006) A use of pues that does not seem to be documented in the literature is illustrated in (47) to (50), where it collocates with nada (‘nothing’), resulting in a phatic device: 47. María: pues nada muy bien hala hija adiós ‘well anyway very good that’s it girl bye (COLA01-3) 48. María: no te he oído tronca no oigo no oigo ‘I didn’t hear you pal I can’t hear I can’t hear’ Ana: pues nada you me subo a mia casa … ‘never mind I’ll go back home’ (COLA01-3) 49. Ana: y bueno entonces pues nada llamé a Belén que tampoco iba a ir a la fiesta ‘and okay then well anyway I called Belén who wasn’t going to the party either’ (COLA01-3)

170 Pragmatic Markers in Contrast 50. A: B: A:

… pues nada ‘well anyway’ hasta luego ‘see you later’ hasta luego (fabricated) ‘see you later’

In (47) María has nothing more to say and uses pues nada as a way of ending the conversation before saying good bye, in (48) Ana gathers that the conversation cannot go on any longer and decides to leave, and in (49) Ana uses pues nada to get to the point. Thus, pues nada is used as a kind of transition marker between stages in a conversation. And like well anyway in English small talk, pues nada can be used to go from talk to action, as in (50). 3.2.5

Summing up

I suggest the following tentative and very simplistic overview of how the main uses of the items discussed are realized in the corpora: MARKER DISCURSIVE PRAGMATIC INTERACTIONAL link chunk stall hedge modify quote take hold yield phatic __________________________________________________________ o sea + + + + + + + + + pues + + + + + + + + --------------------------------------------------------------------------------------well + + + + + + + + cos + + + + other + + __________________________________________________________ Figure 1: The main functions of o sea and pues compared to cos and well

4.

CONCLUSION

The main purpose of this paper has been to suggest the most likely correspondences in English to the Spanish discourse markers o sea and pues, with previous accounts of their multiple functions in spoken interaction as a starting-point for the functional typology and basis for discussion. A preliminary look at word frequencies in the two corpora representing adult speech and the two representing adolescent speech revealed that o sea and pues belong to the ten most frequent discourse markers in Spanish, while cos and well, which I assumed would be two of the closest renderings, belong to the ten most frequent English markers. Not unexpectedly, the use of discourse markers was found to be particularly rich in adolescent speech, where the inventory of markers does not correspond entirely to that in adult speech, and where some markers have even developed new functions.

The Spanish Discourse Markers O Sea and Pues 171 The core function of both o sea and pues is said to be to contribute to the cohesiveness and coherence of spoken discourse (cf. e.g. Martín Zorraquino, 1991; Porroche Ballesteros, 1996; Briz, 2001), and there is general agreement that, though both tend to be overused by many speakers and are seemingly meaningless, they are never entirely devoid of pragmatic content and, consequently, do not deserve to be referred to simply as fillers ( ‘muletillas’). O sea has been described as a connective and epistemic marker by Schwenter (1996), who emphasizes that translation into English is difficult due to overlaps between functions and meanings. Possible renderings may vary, he says, between that is, so, maybe and I guess. Using his examples, I have suggested the English markers that is or well for o sea used as a connective marker and well, sort of or like when it is used as an epistemic marker. The studies by Briz & Hidalgo (1988), Martín Zorraquino (1988, 1991), Porroche Ballesteros (1996) and Briz (2001) show that pues serves a wider range of functions than o sea. In my opinion, reflected in Figure 1, well is the nearest English correspondence in the majority of cases, while cos corresponds to pues in a few cases, a different English marker in some cases, or none at all in other cases.

REFERENCES Andersen, G. (1998). The pragmatic marker like from a relevance-theoretic perspective. In: Discourse markers: Description and Theory (A. Jucker and Y. Ziv, eds.), pp. 147– 170. Benjamins, Amsterdam. The Bergen Corpus of London Teenage Language (COLT) Brinton, L. (1996). Pragmatic Markers in English. Grammaticalization and Discourse Functions. Mouton de Gruyter, Berlin and New York. The British National Corpus (BNC) Briz, A. (2001). El español coloquial en la conversación. Ariel, Barcelona. Briz, A. and A. Hidalgo (1988). Conectores pragmáticos y estructura de la con-versación. In: Los marcadores del discurso (M. A. Martín Zorraquino and E. Montolío Durán eds.), pp.121–142. Arco Libros, Madrid. Brown, G. and G. Yule (1985). Discourse Analysis. Cambridge University Press, Cambridge. Casado Velarde, M. (1988). Lingüistica del texto y marcadores del discurso. In: Los marcadores del discurso (M. A. Martín Zorraquino and E. Montolío Durán eds.), pp. 71–92. Arco Libros, Madrid. Christl, J. (1996). Muletillas en español hablado. In: El español hablado y la cultura oral en España e Hispanoamérica (T. Kotschi, W. Oester-reicher and K. Zimmermann eds.), pp. 117–146. Vervuert Verlag, Frankfurt am Main. Cortés Rodríguez, L. (1991). Sobre conectores, expletivos y muletillas en el español hablado. Editorial Librería Ágora, Málaga. Corpus de Lenguaje Adolescente de Madrid (COLA) El Corpus Oral de Referencia del Español Contemporáneo (COR92). Erman, B. (1987). Pragmatic expressions in English: A study of you know, you see and I mean in face-to-face conversation. Almqvist & Wiksell, Stockholm. Ford, C. (1993). Grammar in Interaction. Adverbial Clauses in American English Conversations. Cambridge University Press, Cambridge.

172 Pragmatic Markers in Contrast Lakoff, R. (1973). Questionable answers and answerable questions. Issues in Linguistics. Papers in Honor of Henry and Renée Kahane (B. Kachru, ed.), pp. 453–467. University of Illinois Press, Urbana. Levinson, S. (1983). Pragmatics.: Cambridge University Press, Cambridge. Martín Zorraquino, M. A. (1988). Los marcadores del discurso desde el punto de vista gramatical. In: Los marcadores del discurso (M. A. Martín Zorraquino and E. Montolío Durán eds.), pp. 19–53. Arco Libros, Madrid. Martín Zorraquino, M. A. (1991). Elementos de cohesión en el habla de Zaragoza. I Curso de Geografía Lingüistica de Aragón (J. Enguita ed.), pp. 253–286. Institución Fernando el Católico, Zaragoza. Pomerantz, A. (1984). Agreeing and disagreeing with assessments: some features of preferred/dispreferred turn shapes. In: Structures of Social Action: Studies in Conversation Analysis (K. Atkinson and J. Heritage eds.), pp. 57–101. Cambridge University Press, Cambridge. Porroche Ballesteros, M. (1996). Las llamadas conjunciones como elementos de connexión en el español conversacional. In: El español hablado y la cultura oral en España e Hispanoamérica (T. Kotschi, W. Oester-reicher and K. Zimmermann eds.), pp. 71–93. Vervuert Verlag, Frankfurt am Main. Romero Trillo, J. (1997). Your attention, please: Pragmatic mechanisms to obtain the addressee’s attention in English and Spanish conversations. Journal of Pragmatics 28, 205–221. Schleppegrell, M. (1991). Paratactic because. Journal of Pragmatics 16, 323–337. Schourup, L. (1985). Common discourse particles in English conver-sation: like, well, y’know. Academic Press, New York. Schwenter, S. (1996). Some reflections on o sea: A discourse marker in Spanish. Journal of Pragmatics 25, 855–874. Serrano, M. J. (2002). Aproximación a la grámatica del discurso del español. Lincom Europa. München. Stenström, A-B. (1984). Questions and Responses in English Conversation. Lund University Press. Lund. Stenström, A-B. (1994). An Introduction to Spoken Interaction. Longman, London. Stenström, A-B. (1998). From sentence to discourse: cos (because) in teenage talk. In: Discourse markers: Description and Theory (A. Jucker and Y. Ziv, eds.), pp. 127– 146. Benjamins, Amsterdam. Stenström, A-B., G. Andersen and I. K. Hasund. (2002). Trends in Teenage Talk. Corpus compilation, analysis and findings. Longman, London. Svartvik, J. (1980). Well in conversation. In: Studies in English Linguistics for Randolph Quirk (G. Leech and J. Svartvik eds.), 167–177. Longman, London. Terrádez Guerra, M. (2001). Frecuencias léxicas del español coloquial: análisis cuantitativo y cualitativo. Facultat de filología, Valencia.


ASPECTUAL PARTICLES IN SOME EUROPEAN LANGUAGES Willy Vandeweghe, Department of Translation Studies Hogeschool Gent

1.

INTRODUCTION

This paper deals with aspectual particles in some of the languages in Europe. The ‘aspectual particles’ discussed here are the equivalents of English already / not yet / still / not anymore. The central representatives of the quartet are already and still, which in some approaches are believed to be dual opposites (e.g. still = NOT already NOT). Evidence from the languages of Europe is adduced to confirm this duality hypothesis. Put more generally, the questions addressed in this paper are the following: 1. How do we conceive of ‘aspectuality’ or phasal aspect? And how does it interact with phenomena like polarity, time, scales, etc.? 2. How can this be related to the structure of the lexical paradigm? 3. What patterns are there to be found in the lexical representatives for ‘aspectuality’, or ‘phasal aspect’, in the languages of Europe? 4. And does the patterning lend support to some of the hypotheses that have been put forward concerning the structuring of the aspectual lexical field?1

2.

THE SEMANTICS OF ASPECTUALITY OR PHASAL ASPECT

When talking about ‘aspectual particles’, we have in mind the equivalents of English already / not yet / still / not anymore. Adverbials are called aspectual “when they carry information about the start or end of an event, which are themselves essentially events of polarity transition.” (Smessaert & Ter Meulen, 2004: 210). Various labels have been used in the literature to refer to these four expressions: perspectivity operators or particles (Vandeweghe, 1979, 1992; van Baar, 1990), particles of change and continuation (van der Auwera, 1991, 1993), phasal adverbials (Löbner, 1989, 1990; van der Auwera, 1998), phasal polarity items (van Baar, 1997), aspectual particles (Smessaert, 1999a), adverbs of aspectual focus (Smessaert, 1999b). Aspectuality is expressed (among other things) by adverbials or particles taking scope over the sentence or part of the sentence, and relating the state of affairs (SoA) as described by that sentence to its possible alternative. In that sense, aspectuality has to do with 1

I thank the anonymous referee whose remarks led to several improvements in the text.


comparison and evaluation, as is common with focussing operators in general. An important component of aspectual meaning is that a perspective is imposed on a state of affairs (SoA). One aspectual use, the imperfective, implies that this SoA and its contrasting alternative are to be seen as embedded in a succession of SoA’s. These SoA’s are evaluated with respect to a chosen point of view furnishing a parameter for comparison and subsequent evaluation. When this point is situated on a temporal scale, which is the ‘most common’ perspective, I will mark this evaluation point as te, in the same way as Löbner does in his work on phasal aspect (Löbner, 1989, 1990). Let us look at a first example: 1. Peter is AL in Brussel Peter is already in Brussels Grasping the meaning of this sentence requires an interpretation of the sentence without its particle. Let’s represent it with ‘p’. The interpretation of AL ‘already’, relies on a contrast between p and NOT-p, whereby the situation described by NOT-p precedes the one described by p. This succession is looked at and evaluated from a moment in time te (= evaluation moment), and the observer takes both alternatives into consideration at the same time. This can be represented as in Figure 1. The arrows of the two alternatives depart from the evaluation time, the solid line represents the actual value ActV (the one which really obtains for the SoA), and the dotted line represents its alternative value AltV.

te

+

ActV

-

AltV

°

p

Figure 1 The semantics of AL (p) ‘already’

When this proposition is denied, both Dutch and English use another lexeme, Dutch nog preceding the negation, or English yet following it: 2. Peter is NOG NIET in Brussel Peter isn’t in Brussels yet The semantics of NOG NIET ‘not yet’ is represented in Figure 2.

Aspectual Particles in some European Languages 175

te

+

AltV

-

ActV

°

p

Figure 2 The semantics of NOG NIET (p) ‘not yet’

The proposition in both cases, as well as in the cases to be discussed, can be rendered as 3. a b

AL (te, P) NOG NIET (te, P)

The predicate P” is the focus of the construction in the syntactic sense as well as in the semantic sense (since it contains the polarity focus of the sentence)” (Löbner 1999: 87). When the sequence is reversed, the initial situation remains ‘Peter is in Brussels’ (P) and is now followed by its negation ‘~P’ (see Figure 3 The semantics of nog and Figure 4 The semantics of niet meer). 4. a b

Peter is NOG in Brussel Peter is still in Brussels Peter is NIET MEER in Brussel Peter isn’t in Brussels anymore

te

-

AltV

+

ActV

°

p

Figure 3 The semantics of NOG (p) ‘still’


te

-

ActV

+

AltV

°

p

Figure 4 The semantics of NIET MEER (p) ‘not anymore’

An aspect of the semantics of NOG in (4a) is the presupposition of an earlier moment for which p was true, a presupposition shared by its negative counterpart NIET MEER in (4b). The evaluation point itself is situated on the time-axis (we then speak of te) or some other scale with properties comparable to those of the t-scale. In fact, aspectuality has to do with two scales which are interrelated. One of them comprises an element which is taken as a vantage point te to evaluate the progression on the other scale which displays a succession or sequence of events (see Figure 5).

+

te

°

time scale (t)

sequence of events

Figure 5 The semantics of AL ‘already’ with binary sequence


Note, by the way, that the idea of an alternative (ALT) may be induced by several discourse factors. Differences in discourse constraints may lead to differences in the eventual interpretation. To give a few examples: (i) an independent source for ALT can be the expectation that some state of affairs is due to come about anyhow, such as for instance the arrival of the members at a congress venue. AL characterizes the SoA as relatively advanced with respect to the time point taken as a vantage point, which in its turn is characterized as earlier than might be expected. Example (1) is repeated below: 1. Peter is AL in Brussel Peter is already in Brussels (ii) Of course, (1) may also be interpreted in the context of a journey. When Peter is undertaking a pilgrimage to Compostella for instance, (1) might be the description of a stage in that process, which defines a scale for the SoA to be part of. Sample (1) then tells us that ‘Peter’ has made quite some progression on that scale. (iii)a more context-bound interpretation might arise when someone has just noted that something is due to be the case as soon as Peter is in Brussels. The reaction might be: 5. Maar Peter ís al in Brussel But Peter is in Brussels alreády Summing up: the perspectives defined for AL ‘already’ and NOG ‘still’ are each other’s opposites (cf. Löbner 1999: 76): •

AL: “ the perspective (is) determined by the negative presupposition (i.e. the first-not-P-then-P perspective)” ; this same perspective is shared by NOG NIET ‘not yet’, having the opposite polarity

•

NOG: “ the perspective is determined by the positive presupposition (i.e. the first-P-then-not-P perspective)”

This difference in presuppositional background leads to the basic lexical opposition in the aspectual paradigm: that of DUALITY.

3. 3.1

THE LEXICAL PARADIGM The Duality Hypothesis

Lexical paradigms realizing the four situations depicted above in Dutch require three different words: AL, NOG, MEER. Of these AL and NOG are basic, as they function in the affirmative contexts: they explicate the confirmation of P against their respective backgrounds. In an attempt to generalize over other languages, we started out by representing them by their Latin counterparts, viz. IAM and ETIAM, thus following a tradition initiated by Välikangas (1982).


Yet occasionally, we will also draw on the terminology initiated by van der Auwera (1998: 35). The latter also distinguishes between INCHOATIVE for the representatives of already, while associating the other three elements with the notion of CONTINUITY: still is CONTINUATIVE, not anymore, no longer is DISCONTINUATIVE and not yet is CONTINUATIVE NEGATIVE. It was Löbner who in the eighties drew attention to the basic duality in the opposition at hand. Duality is a notion originating in philosophy and model theory (cf. Barwise & Cooper, 1981) and used in the contrasting of quantifiers. The dual opposite of a quantifier Q is taken to be the outer negation of its inner negation, so Q is dual to ~Q~. A case in point is the pair ALL (x) / SOME (x), exhibiting the universal and the existential quantifier (see Figure 6): Subnegation ~x x~

x ~x~ (external)

(external) Negation Negation ~x x~

x ~x~ Subnegation

Figure 6 Duality in the lexical domain of quantification

6. a b c d

x x ~x ~x

For all x: p There is an x: p For no x: p Not for all x: p

For no x: not p Not for all x: not p For all x: not p There is an x: not p

It turns out that in the pair IAM/ETIAM both lexemes constitute each other’s dual, as is shown in the equivalences:2

2

7. a b

IAM ETIAM

8. a b

Peter is already in Brussels Peter is still in Brussels

For some criticism: see Smessaert (1999a: 71).

~ETIAM~ ~IAM~ = =

Peter is NOT [still [NOT in B]] Peter is NOT [already [NOT in B]]


3.2

Criticism of the Duality Hypothesis

The analysis of the lexical paradigm in terms of duality, as it was first proposed by Löbner, has been challenged by several authors (Mittwoch, 1993, Michaelis, 1996, van der Auwera, 1993). One of van der Auwera’s arguments is the existence of a ‘finally’-lexeme which in the inchoative sphere is very akin to already, in the sense that •

both have to do with a transition not-P / P

•

while AL (already) signals relative earliness, EINDELIJK (‘finally’) is the signal of relative lateness

Now EINDELIJK (‘finally’) is not given a proper place in the duality square, and that is why van der Auwera has developed a system of alternative scenarios. In his view already cannot be the dual of still: If already were a dual [of still], it would only express the discontinuation of a negative state (not … still … not). But the semantics is more complicated. Already’s newly obtaining positive state is different from two negative ones, one that is anterior and one that is counterfactually simultaneous. In the first case, the change is neither early nor late, in the second it is early. Already furthermore excludes a third scenario, viz. the one in which the change is late, this being the province of finally. (van der Auwera 1993: 622f) 3.3

Refutation of (some of) the Criticism

In my view the ‘finally’ argument is a mistaken one. The ‘finally’ lexeme is not linked up with polarity in the same way as the ‘already’ lexeme. I would prefer to classify it as extrapropositional, superseding polarity, linking it up with aspectuality and having it play a role in the modal sphere. See the example adapted from Vandeweghe (1992: 90): 9. The professor is proud to announce that a. the problem has already been solved b. the problem has finally been solved In (9a) the professor’s happiness may have to do with the relative earliness of the solution, but in (9b) it cannot plausibly have to with its relative lateness: his happiness simply concerns the solution itself. For the same reason, (10) would sound odd (at least in the normal ‘sigh-ofrelief’ reading of finally): 10. The professor resents announcing that a. the problem has already been solved b. ?the problem has finally been solved


The professor’s resentment may have to do with the early solution of the problem, but in b the professor’s mood is in contradiction with what is suggested by finally, viz. the fact that it was a most welcome solution, be it a long-awaited one. Löbner (1999: 87) suggests that the difference between the two adverbials has to be seen in terms of foregrounding and emphasis:3

4.

4.1

-

endlich (‘finally’) P: foregrounds the phase not-P (the ending of which is welcomed)

-

already P: foregrounds the phase P

THE PARADIGM AS REPRESENTED IN THE EUROPEAN LANGUAGES The Basic Aspectual Lexemes

In 1986 I undertook a study of the lexical representatives for the aspectual duality square in some European languages. When selecting them I did not rely on any rigid criterion, but rather on the informants I happened to know and on the literature that dealt with aspectual particles at that time. The late eighties then saw the start of the EUROTYP project, sponsored by the European Science Foundation, which set up a cross-linguistic study of adverbials in as many European languages as possible. Phasal adverbials were among them and Johan van der Auwera headed this part of the project (van der Auwera 1991, 1998). The project had history on its side, because these selfsame years the Iron Curtain came down, a fact which created more opportunities for researchers to present their questionnaires to informants in the far east of the former Soviet Union. My own information from these parts went as far as Russian, but the Eurotyp project enabled the field workers to gather far more data for the corpus from the Caucasian, Altaic and Uralic languages. The final report was included in the 1998 volume Adverbial Constructions in the Languages of Europe, edited by van der Auwera and Ó Baoill. The conclusions I drew from my earlier sample were not invalidated by the much bigger sample of the Eurotyp researchers and by van der Auwera’s conclusions. There are very few languages with no aspectual particles at all, like Kalmyk, an Altaic language, and Laz, a Caucasian language. Chechen, also Caucasian, seems to have only one lexeme, viz. for the discontinuative IAM NON. Most of the European languages do have lexical representatives for the four corners of the duality square. If only one representative is lacking, it is to be found in the ‘already’ area, as is the case for Icelandic (one of the Germanic languages), Albanian, Mansi (an Ob-Ugric language) and maybe Turkish. Van der Auwera takes this circumstance as evidence for the plausibility of a “ Euroversal” accessibility hierarchy, implying that the concept which has the highest chance of being adverbialized is the most accessible one: 3

For Löbner 1999 (81-3) Gm. endlich, Du. eindelijk actually form a duality group of their own with Gm. noch immer, Du. nog altijd.


11. Accessibility hierarchy (van der Auwera 1998: 37) discontinuative > continuative > inchoative > continuative negative Referring to elements that support this hypothesis, van der Auwera points to the fact that “most of the languages that use foreign stock material for an inchoative adverbial exclusively rely on this” (1998: 70), and INCH often appears in the system at a later stage (1998: 63). He claims that higher accessibility attracts more lexemes (1998: 82), and more variation is a signal of greater unmarkedness (1998: 73-4): INCH is only retrospective, CONT is retrospective as well as prospective. It seems reasonable to add that INCH is more redundant than most other features, and as such remains unexpressed in some languages. The fact that IAM is lacking in some languages may have to do with this redundancy, which IAM shares with ‘now’, or ‘affirmativity’: IAM in a sense gives aspectual (= time- and perspectiverelated) ‘colour’ to positive or affirmative polarity. Here is a table with the lexemes from the languages in my own sample (Vandeweghe, 1986) and the much larger sample in van der Auwera (1998).

Table 1. Aspectuality: lexical paradigm in European languages INCHOATIVE

CONTINUATIVE

NEGATIVE CONTINUATIVE

DISCONTINUATIVE

IAM

ETIAM

NON IAM ETIAM NON

NON ETIAM NON PLUS Neg meer, langer A Neg mehr Neg anymore, any longer Neg längere, mere

GERMANIC Dutch German English Danish Norse Swedish Gothic

3 3 4 3 3? 3? 2 (3)

al, reeds schon, bereits already allerede allerede re(da)n ju

nog noch still endnu, stadig(væk) ennå än(nu) naúh (-an[uh], anamais/-seis

nog Neg noch Neg Neg yet endnu Neg ennå Neg än(nu) Neg naúh (-an[uh] Neg

CELTIC Irish

3

cheana (féin)

fó, go fóill, ar fad, I geónai, le eto

mwy(ach)

ju Neg B anamais/-seis A!

Welsh ROMANCE French Italian Portuguese Spanish Rumanian

4

yn barod, eisoes

fó, go fóill, ar fad, I geónai, le o hyd

3 3 3 3 2

déjà già já ya deja

encore ancora ainda todavía, aún inca, mai

Neg encore Neg ancora Neg ainda todavía, aún Neg inca, mai Neg

Neg …plus A Neg più A já Neg B, [Neg mais A] ya Neg B Neg mai A(!)

SLAVIC Russian

2 (3)

ue

je ë

je ë Neg

2 (3) 3 2 (3) 3

ve e ve Ju u / ji

oste jo jeszcze jete

oste jo jeszcze jete

ue B, bol’e A Neg ue bol’e B + A pove e A, ve e B e e ve e ju B, ju wi cej B + A hio B, wjace A

2 (3)

jau

dar

dar

Bulgarian Serbian Polish Czech BALTIC Lithuanian

mó, faide, sia

jau B, daugiau A, jau daugiau B + A


akómi akoma, edhe, ende

pja B, pléon A më A

arden; aylyeves

akómi akoma, edhe, ende, hala der, deryeves, takavin

der, deryeves, takavin

aylyeves B, arden B

3 3 2 (3) 2+Ø

jo (i)ni már(is) -

vielä na, e o még I

vielä na, e o még i

enää A öböl A már B, tovább A Wassi

(artık)

daha, hâlâ

daha, hâlâ

(artık) B

Chuvash Nogai Karaim

2 (3) Ø? 2 (3) 2 (3) 2 (3)

ente endi ue

hale t’e, tata, -xa ali daa, hanuz

-xa ali daa, hanuz

ente B, tek A? endi B, artyk A? ue B, artych A?

URALIC Nenets

2

mal’e

tamna

tamna -

mal’e B

CAUCASIAN Georgian

2 (3)

uk’ve

er

er

Svan Abkhaz Lezgian Dargwa

2 3 2 2 + Ø?

ser sta hele(lig) -?

ike mak”ana hele, mad xalla, hannara

ike mak”ana hele xalla, hannara

uk’ve B, a ar A, met’i A, a ar met’i A ser B wa a A mad A! Ur hat’i- ?

Basque

2 (3)

ya, déjà, ja(dan), jadanik

oraindik, anartean

oraindik

ya,, jadanik B, gehiago A

INDO-IRAN. Kirmanji

2 (3)

Hew

hên

êdi, hew B, (ne)ma A

Ossetic

2 + Ø?

Ø?

hên/hîn/hîna, hêj, hela hê nrma, ma

nrma, ma

(n)al / (ma)ual

Yiddish

3

shoyn

nokh

nokh

nit mer, mer nit, A shoyn nit B

Greek Albanian

2 (3) 2+Ø

idhi, pja, kjolas -

Armenian

2

FINNICUGRIC Finnish Udmurt Hungarian Mansi ALTAIC Turkish

4.2

Division of Labour in the Lexical Field

The “glue” for the aspectual paradigm is to be found in the second and third column: the vast majority of languages using the ETIAM-lexeme for positive continuativeness when negating the inchoative IAM. This fits in well with the hypothesized dual character of IAM and ETIAM. We could imagine a perfectly “economical” system with only two aspectual lexemes, the dual opponents IAM and ETIAM, only making use of the NEG particle in internal position:


Sp ya Ru ue Hu már

IAM

IAM NON / ya no = NON ETIAM ue ne már nem

Sp todavía no ETIAM NON Ru jeë ne = NON IAM Hu még nem

ETIAM todavía jeë még

Figure 7 An “economical” system exploiting NEG and duality relations

This “economical” system is realized in European languages like Russian, Polish, Bulgarian; Spanish, Portuguese; Hungarian. Furthermore it exists in Nenets, a Uralic language, in Caucasian Svan, in the Altaic languages from the sample with occasional influence from Russian or Turkish. It is also to be found in languages outside Europe, like Japanese: m (IAM), mada Neg, mada (ETIAM), m Neg. The “least economical” of systems is the one which has a different lexical item in every corner of the square. Of the languages under discussion, only English is a case in point, probably together with Welsh:

already

NEG yet

Figure 8 The duality square in English

NEG anymore

still


The relative lexical complexity corresponds to grammatical simplicity: negation is always external, and the relevant lexemes have developed into NPIs.4 The lexical split-up between still and yet with respect to polarity has produced a straightforward and very simple lexical solution to express aspectuality. In conclusion, there are two important ways of negating: (i) One way is to replace the aspectual lexeme by an item with opposite duality, and replacing external negation by internal or subnegation. *~X Y~ [Y= dual of X]

12. a

(ii) The other way is leaving external negation unchanged, but replacing the aspectual lexeme by an item with opposite polarity sensitivity, a so called ‘negative polarity item’ (NPI) *~PPI ~NPI

12. b 4.3

Negation of IAM

We noted that in the vast majority of languages inchoative IAM is negated by the ETIAM lexeme + Neg, which fits in nicely with the hypothesized dual relation between IAM and ETIAM. A proviso for English is that the former continuative PPI yet has been reinterpreted as an NPI5, thus allowing the scope relationship between NEG and ASP to be inverted. In all the languages represented, IAM – when present – is tied up with affirmative polarity, and as such it is a positive polarity item (PPI), which means that in unmarked cases the combination with external negation results in ungrammaticality. One of the ways to come to terms with this is to replace the aspectual lexeme by its dual counterpart; this naturally implies inversion of scope relations, in such a way that we end up with a cluster consisting of the dual opposite aspectual particle accompanied by internal negation: 13. *~IAM

ETIAM~

Examples: 14. Dutch: *niet AL German: *nicht SCHON Spanish: *no YA6

niet nicht TODAVÍA no NOG

NOCH

It should be pointed out, in passing, that left-to-right ordering is not the unequivocal reflection of operator-scope relationship: others factors play a role apart from scope, such as the word order type (SVO/SOV) and the position of NEG with respect to the finite verb. There is a

4

Rumanian seems to be a special case as well, in the sense that mai is possible in both environments, positive as well as negative. See also van der Auwera (1998) on this. 5 On the diachrony of yet as an ETIAM lexeme, see Traugott/Waterhouse (1969), Ladusaw (1977), König/ Traugott (1982), van der Auwera (1998). 6 With the caveat that Spanish ya no can alternate with no … ya, having the same meaning.


universal tendency to place NEG immediately next to it7, before or after. Most probably, the factors referred to underlie, e.g., the mirroring relationship between French on the one hand, and Dutch or German on the other: 15. a b

French (SVO): Dutch / German (SOV):

ne V] pas] ENCORE NOG [niet [V, NOCH [nicht [V

= =

Neg ETIAM Neg

ETIAM

The ordering of the operators is each other’s reverse: 16. a b

French (SVO): Dutch / German (SOV):

V] NEG] ASP ASP [NEG [V

The matter of placing the negation is dealt with at great length in van der Auwera (1998: 102– 116). The factors he mentions, include beside scope, what he calls categorial harmony: i.e. “elements of the same category follow the same positional rules” (1998: 107). Obviously, areal influence may also play a part: Yiddish is a case in point, as the only ‘Germanic language’ with mer nit (PLUS Neg) order, most probably in imitation of Slavic (as in Ru bol’e ne). 4.4

Negation of ETIAM

Just like IAM, ETIAM normally does not support external negation, at least in the majority of cases. Most languages resort to other means of expressing the negative variant. (i) IAM NON type languages A minority of European languages opt for IAM NON to negate ETIAM. These are the Slavonic languages (albeit not the only alternative, cf. bol’e ne), alongside Hungarian (a Sprachbund phenomenon?) and Iberian Romance languages. 17. *~ ETIAM

IAM ~

Examples: 18. Spanish: *no TODAVÍA Russian: *ne EË

YA no UE ne

It is exactly the same strategy as the one used to negate inchoative IAM, and both strategies exploit the duality relationship between the two lexemes. IAM NON type expressions seem to have removed the expectancy presupposition associated with IAM from their meaning, and what is kept prominent is the (perspective) idea of crossing a given transition point from a given phase (in this case p) to its complement ~p. In a discourse context where p belongs to the shared knowledge, the directionality of the

7 This is shown in Dryer (1988), who investigated the position of NEG in a sample of 297 languages. In a few exceptional cases, however, like the one of nu mai in Rumanian, the aspectual particle is among the few elements to be inserted in between NEG and V.


change is likely to be from p to ~p. The IAM lexeme basically means ‘now’, with an added implicature of ‘now’ as the beginning of something (‘from now on’). (ii) NON PLUS type languages (comparative lexeme) By far the most frequent type of lexicalization used for the negation of ETIAM is a third item based on a comparative for adjectives denoting quantity (‘plus’: ne … plus in French, no … più in Italian, niet… meer in Dutch, nicht …mehr in German). It is often materialized on dimensions such as length (‘longer’ is found in Scandinavian, and also as a variant for PLUS in Dutch and English), height (Serbian vie = ‘higher’), tallness (Rum. mai < Lat. magis ‘bigger’; Russ. bol’e). The comparative operates semantically much like an existential quantifier,8 in that it quantifies over moments for which p holds, but its comparative morphology indicates that the moments are to be seen in relation to comparable moments which have preceded with respect to te.

5.

CONCLUSION

The survey of the aspectual adverbials or particles in the languages of Europe lends support to the linguistic relevance of dual negation in the composition of this lexical field, next to ‘normal’ negation and subnegation. The basic building blocks used for the expression of aspectuality consist of inchoativity (AL) and continuity (NOG), the abstract elements that every language may well materialize from different lexical sources.

REFERENCES Barwise, J. and R. Cooper (1981). Generalized quantifiers in natural language. Linguistics and Philosophy, 4, 159–219. Dryer, M. S. (1988). Universals of negative position. In: Studies in Syntactic Typology (M. Hammond, E. Moravcsik and J. Wirth, eds.), pp. 93–124. John Benjamins, Amsterdam. König, E. and E. C. Traugott (1982). Divergence and apparent convergence in the development of yet and still. Proceedings of the Tenth Annual Meeting of the Berkeley Linguistics Society (BLS 10), pp. 170–179. Berkeley Linguistics Society, Berkeley. Ladusaw, W. A. (1977). Tracking a semantic change: still and yet. Unpublished paper delivered to the annual meeting of the L.S.A. (ms.) Löbner, S. (1989). German ‘schon’, ‘erst’, ‘noch’: an integrated analysis. Linguistics and Philosophy, 12, 167–212. Löbner, S. (1990). Wahr neben Falsch. Duale Operatoren als die Quantoren natürlicher Sprache. Max Niemeyer Verlag, Tübingen. Löbner, S. (1999). Why German schon and noch are still duals: a reply to van der Auwera. Linguistics and Philosophy, 22, 45–107.

8

In English this is signalled even more explicitly by the addition of any.


Michaelis, L. A. (1996). On the use and meaning of already. Linguistics and Philosophy, 19, 477–502. Mittwoch, A. (1993). The relationship between schon/already and noch/still : a reply to Löbner. Natural Language Semantics, 2, 71–82. Smessaert, H. (1999a). Perspectief en vergelijking. Aspectuele partikels in het Nederlands, (Studies op het gebied van de Nederlandse Taalkunde 4). Belgisch Interuniversitair Centrum voor Neerlandistiek/Uitgeverij Peeters, Leuven. Smessaert, H. (1999b). Adverbials of aspectual focus and negation in Dutch. In: Tense and Aspect (S. Vogeleer, W. De Mulder and I. Depraetere, eds). Belgian Journal of Linguistics, 12, 61–75. Smessaert, H. and A. ter Meulen (2004). Temporal reasoning with aspectual adverbials. Linguistics and Philosophy, 27, 209–262 Traugott, E. C. and J. Waterhouse (1969). “Already” and “yet”: a suppletive set of aspect markers? Journal of Linguistics, 5, 287–304. Välikangas, O. (1982). La notion de ‘déjà’ et les mots qui servent à la rendre dans quelques langues européennes. Neuphilologische Mitteilungen, 83, 371–404. van Baar, T. (1990). The Dutch Perspectivity Particles in FG. Working Papers in Functional Grammar, 36. UvA (Amsterdam University), Amsterdam. van Baar, T. (1997). Phasal Polarity. Studies in Language and Language Use, 30. IFOTT, Amsterdam. van der Auwera, J. (1991). Beyond duality. In: Adverbs and Particles of Change and Continuation (J. van der Auwera, ed.), Eurotyp Working Papers, Vol. 2, 131–159. European Science Foundation, Strasbourg. van der Auwera, J. (1993). ‘Already’ and ‘still’ : beyond duality. Linguistics and Philosophy, 16, 613–653. van der Auwera, J. (1998). Phasal adverbials in the languages of Europe. In: J. van der Auwera and D. P. Ó Baoill, (eds), pp. 25–145. van der Auwera, J. and D. P. Ó Baoill (eds) (1998). Adverbial Constructions in the Languages of Europe. Mouton de Gruyter, Berlin. Vandeweghe, W. (1979). Perspectivity operators in Dutch: al, nog, nogn,m. In: Bedeutung, Sprechakte und Texte. Akten des 13. Linguistischen Kolloquiums, Gent 1978 (W. Vandeweghe and M. Van de Velde, eds), Vol. 2, pp. 111–122. Niemeyer, Tübingen. Vandeweghe, W. (1986). Complex aspectivity particles in some European languages. Groninger Arbeiten zur Germanistischen Linguistik, 27, 220–231. Groningen: Groningen University. Vandeweghe, W. (1992). Perspectivische evaluatie in het Nederlands : de partikels van de AL/NOG/PAS-groep. Koninklijke Academie voor Nederlandse Taal- en Letterkunde, Gent.



DIALOGISTIC POSITIONS AND ANTICIPATED AUDIENCES – A FRAMEWORK FOR STYLISTIC COMPARISONS Peter R.R White, University of Adelaide, South Australia Motoki Sano, University of Wollongong, New South Wales

1.

INTRODUCTION

This paper is directed towards one important aspect of stylistic variation – variability in how speakers/writers position themselves intersubjectively with respect to other speakers and with respect to potential respondents to the current communication. More specifically it is directed towards this positioning in the context of public, mass communicative (rather than private, person-to-person) discourse. Here such positioning acts to construct for the text a putative addressee or audience as the speaker/writer1 is presented as assuming that this addressee operates with particular beliefs and values, and as anticipating certain responses and reactions on this addressee’s part. The meanings by which such positioning is conducted linguistically are many and varied, including locutions which would perhaps typically be regarded as ‘pragmatic markers’ as well as those which would not. Within the appraisal framework of Martin, White, Iedema, Feez, Rothery, Coffin, Macken-Horarik and Fuller under which this paper operates (see for example, Iedema et al., 1994, Christie & Martin, 1997, Coffin, 1997 and 2003, Rothery & Stenglin, 1997, White, 1998, Martin, 2000, White, 2002, MackenHorarik & Martin, 2003 and Martin & White, in press, Fuller, 1995, 1998), this set of meanings is dealt with under the heading of ‘engagement’. Our purpose in this paper is to offer some further insights into locutions which operate with what can be termed this ‘dialogistic’ functionality. More narrowly the paper is concerned with a subset of these engagement meanings – those formulations which signal a recognition that the current proposition is a contentious or a dialogically problematic one which is likely to be in tension with alternative propositions, and those by which the speaker implicitly lays claim to agreement and solidarity with the putative addressee. While these meanings are considered primarily in the context of English, our subsidiary purpose is to attend to issues of crosslinguistic comparison, exploring how we might conduct analyses which would compare and contrast dialogistic positioning as it operates in comparable texts in different languages.

1

In this paper we employ the term ‘speaker’ to cover both speakers and writers, unless otherwise indicated.


Towards this end we consider dialogistic positioning in a comparison of press conferences delivered by the British and Japanese prime ministers. Our primary focus is upon markers which, in the literature, are typically treated under such headings as ‘epistemic modality’ (see for example, Coates, 1983, Palmer, 1986), ‘evidentiality’ (see, for example Chafe & Nichols, 1986), hedging (see for example Markkanen & Schröder, 1997), and less specifically under the heading of ‘meta-discourse’ (see, for example, Crismore, 1989). Specifically we attend to, •

locutions which construe different degrees of probability – e.g. in English, adjuncts such as possibly, perhaps, probably and definitely; related impersonalizing formulations such as it’s possible that, it’s probable that; modal auxiliaries such as may, might, must; and personalizing locutions such as I suspect that…, I believe, I think, I’m sure, I know, etc,

•

evidentials such as it seems, apparently

•

formulations by which the speaker overtly announces propositions as self-evident or very generally known or accepted – e.g. in English, of course, naturally, obviously, needless to say, it goes without saying, as you know, etc

Our approach to these locutions is informed by Bakhtin/Voloshinov’s now widely influential notions of dialogism and heteroglossia under which all verbal communication, whether written or spoken, is ‘dialogic’ in that to speak or write is always to reveal the influence of, refer to, or to take up in some way, what has been said/written before, and simultaneously to anticipate the responses of actual, potential or imagined readers/listeners. As Voloshinov states, The actual reality of language-speech is not the abstract system of linguistic forms, not the isolated monologic utterance, and not the psychological act of its implementation, but the social event of verbal interaction implemented in an utterance or utterances. Thus, verbal interaction is the basic reality of language. Dialogue…can also be understood in a broader sense, meaning not only direct, face-to-face, vocalised verbal communication between persons, but also verbal communication of any type whatsoever. A book, i.e. a verbal performance in print, is also an element of verbal communication. …[it] inevitably orients itself with respect to previous performances in the same sphere… Thus the printed verbal performance engages, as it were, in ideological colloquy of a large scale: it responds to something, affirms something, anticipates possible responses and objections, seeks support, and so on. [Voloshinov 1995: 139] And as Bakhtin similarly observes, all utterances exist

Dialogistic Positions and Anticipated Audiences 191

…against a backdrop of other concrete utterances on the same theme, a background made up of contradictory opinions, points of view and value judgements … pregnant with responses and objections” [Bakhtin 1981: 281]. This perspective leads us to attend to aspects of the communicative functionality of these locutions which have frequently not been considered by those whose theoretical orientation has been shaped by British/American analytical philosophy and a concern with truth functionality (see for example, Lyons, 1977 and Palmer, 1986). Thus while earlier treatments have tended to interpret modals and evidentials as signs of lack of commitment by the speaker to the truth value of the proposition, we are directed, rather, to attend to the intersubjective, dialogistic effects associated with such meanings. In the sections which follow, we set out an account of these dialogistic effects and offer some preliminary observations about how such accounts can be applied to comparative analyses of the evaluative styles of different texts and different speakers, both within texts of the same language and across different languages.

2.

DIALOGISTIC POSITIONING

2.1

The monoglossic and the heteroglossic

Before we attend to these locutions individually, it is necessary to outline the broader parameters by which dialogistic positioning may vary both within a language and between languages. In English we observe a contrast between utterances in which propositions are barely asserted without any explicit acknowledgement of the dialogistic and heteroglossic setting in which the text is operating, and those which do reference and engage with this heteroglossic backdrop. For example the following sequence of barely asserted propositions: 1. Saddam is a threat. That is why for 12 years the United Nations has been trying to get him to disarm Iraq peacefully of chemical, biological and nuclear weapons… We live today in a world beset by international terrorism, whose groups are desperate to acquire ever more dangerous weapons, and who are already using chemical and biological poisons. [from a press conference by the British Prime Minister in the lead up to the US and UK governments’ invasion of Iraq in 2003] can be contrasted with, 2. The resolution stated clearly that Saddam was already in material breach of UN disarmament demands, it gave him what it was said was a final opportunity to comply or face serious consequences…. It is plain, in our judgement, that Saddam continues to be in breach.

or with:


3. The basis of our action is disarmament. That is the UN mandate. Of course I understand the concerns of the thousands that marched on Saturday, and of course I should and do listen to those concerns.

In extract 1 the speaker makes bald assertions negatively evaluating the Iraqi leader Saddam Hussein and the state of security in the world. The propositions are formulated in such a way that that there is no recognition of the possibility of alternatives to the speaker’s viewpoint, nor anticipation by the speaker that any of these propositions will be in any way problematic for those being addressed. In extract 2, in contrast, the speaker executes a somewhat complex dialogistic manoeuvre with respect to the proposition that the Iraqi leader Saddam Hussein was in breach of the United National resolution. By the locution it is plain, he indicates an anticipation that the addressee will find this an unproblematic viewpoint which he/she will share with the speaker. However, he then complicates the positioning by means of in our judgement, acknowledging that this is a matter of personal opinion and hence potentially just one viewpoint among a range of potential alternative positions. In extract 3, the speaker is presented as anticipating that the information about to be conveyed will be so self-evident for the addressee that the statements barely needs to be uttered at all. There is a tradition within English linguistics within which the barely asserted proposition, as exemplified by extract 1, is characterised as intersubjectively neutral, objective or ‘factual’. Lyons, for example, sets up a contrast between the supposed ‘objectivity’ of the bare assertion, which he terms ‘factive’, and the ‘subjectivity’ of the modalised utterance, which he terms ‘non-factive’ (Lyons 1977: 794). But of course once we understand all verbal communication as occurring against a heteroglossic backdrop of other voices and alternative viewpoints, then we are likely to view the ‘bare’ assertion in a rather different light. Following Bakhtin we see the bare assertion as constituting a particular, potentially highly charged intersubjective position – one by which the current proposition is construed as either having no alternatives or challenges at all, or as having no alternatives or challenges which need to be acknowledged or engaged with in the current communicative context. We classify such formulations as ‘monoglossic’, once again following Bakhtin, in recognition of the way they present the speaker as a solitary voice unengaged with any dialogic partners or alternatives. These contrast broadly with the formulations exemplified by extracts 2 and 3, where the speaker presents himself variously as (a) acknowledging that there is a heteroglossic backdrop of alternative views and multiple voice with which he is likely to be in tension (in our judgement) and, (b) as sensitive to where others are likely to stand on the proposition being advanced (it’s plain and of course). We employ the term ‘heteroglossic’ to label all formulations which, in these and other ways, acknowledge that the utterance operates against a heteroglossic backdrop and present the speaker as recognising or engaged with other voices and other viewpoints within this backdrop. English, then, operates with this broad, binary distinction between the monoglossic and the heteroglossic. As a starting-point for comparative analyses of speakers’ interpersonal styles we might begin by exploring where, how often, and with what type of utterances they employ monoglossic versus heteroglossic modes of expression. We should note, however,


with an eye to the application of the framework to cross-linguistic analysis, that while this two-way division operates in English, this may not always be the case in other languages. Here we have in mind those now well-documented languages in which there is no obvious equivalent to the English ‘bare assertion’. These are languages in which all utterances are in some way marked for their evidential status or for the grounds on which the speaker offers the information or opinion. We can mention here, by way of just one example selected largely at random from the literature, the Wanka dialect of Quechua. According to Floyd (1996), in Wanka propositions are associated with one of three evidential suffixes indicating either that the speaker bases the statement on conjecture (-chr(a)), on reported evidence (-sh(i)) or on direct, first-hand personal experience (-m(i)) (Floyd, 1996: 70). The crucial point here is that what, from the perspective of English, we might think of as the most ‘direct’, ‘objective’ or ‘factual’ mode of expression, the proposition with the suffix –mi(i)), nevertheless still involves speakers explicitly declaring their personal evidential ‘interest’ in the proposition. It would appear that even with this ‘direct’ form, speakers are still explicitly presenting the proposition as grounded in their own individual and contingent subjectivity. To the extent that this characterisation is accurate, we can say, then, that in Wanka (and similar languages) all utterances are heteroglossic in that even the most directly assertive formulations still present the proposition as subjectively contingent and hence as but one of a range of possible alternative propositions. If we were to seek to compare the interpersonal styles found in such languages with the interpersonal styles of English, we would need to allow for the possibility that the underlying dialogistic dispositions of the two languages may be fundamentally different, with Wanka apparently not operating with the binary distinction between the monoglossic and the heteroglossic which we observe in English (and similar languages). 2.2

Dialogistic expansion and contraction

Within those formulations which operate heteroglossically we observe a further broad distinction according to whether they are ‘dialogically expansive’ or ‘dialogically contractive’ in their intersubjective functionality. The distinction turns on the degree to which an utterance, by dint of the use of one or more of these ‘engagement’ options, actively makes allowances for dialogically alternative positions and voices (dialogic expansion), or alternatively, acts to challenge, fend off or restrict the scope of such (dialogic contraction). The distinction is illustrated in the following extracts which, like the previous extracts, have been taken from press conference statements by the British Prime Minister Mr Tony Blair. (Formulations which ‘contract’ or ‘expand’ in this way have been underlined.) 4. (dialogic expansion) And I remember when we didn’t intervene in Bosnia back in the early 1990s and tens of thousands of innocent people died as a result, and I think probably it would have been better had we intervened.


5. (dialogic expansion) But some of it is connected with thuggery and violence of people who are drunk. Now in my view we need to take specific measures against those types of people and the premises that they are using. 6. (dialogic contraction) On the rest, and it is interesting in a sense if I can say this respectfully to everyone here, that someone from regional television actually asks the question about antisocial behaviour, because I do say to you, in communities in this country this is absolutely at the core of their concerns, and anti-social behaviour orders have been very successful where used, but the truth of the matter is they have been used less readily than we want them to be and that is because the court procedures have been too cumbersome, the bureaucracy too great, so we are going back and trying to strip all that down and make it easier. In extract 4 and 5, the speaker employs the formulations I think, probably, and in my view and thereby explicitly grounds the proposition in his own contingent subjectivity. This is not to indicate lack of commitment to the truth value of these propositions, nor to indicate their epistemic unreliability, but rather to signal a recognition that, in the current communicative context, these are contentious matters and thereby to signal recognition that those being addressed may query, reject or at least find such propositions novel or otherwise problematic. By this mechanism, the interpersonal cost to any who might advance alternative views is lowered as their position is recognised as a valid one in the current ongoing colloquy. In this way, such formulations expand the space for dialogistic alternatives. In extract 6, the Prime Minister is responding to a series of questions which have been raising doubts with respect to the government’s proposal to introduce new measures to counter so-called ‘anti-social behaviour’. Previous questioners have suggested that the proposals amount to an overreaction to a relatively minor issue and that the new measures are not needed because laws already in place are well able to deal with such alleged breaches of the peace. The Prime Minister employs the formulations I do say, absolutely and the truth of the matter is. Such locutions constitute overt interventions or interpolations in the text by the speaker by which they emphasise or insist upon the warrantability of the proposition. By intervening in this way, the speaker makes more salient his subjective role through this effort at ‘vouchsafing’ the proposition. And such insisting or emphasising implies the presence of some resistance, some contrary pressure of doubt or challenge against which the speaker asserts him/herself. (In this case, the contrary view was that being referenced by earlier questions.) It is only necessary to insist when there is some counter viewpoint against which the insistence is directed. Accordingly, such formulations do acknowledge the heteroglossic diversity of the current communicative context. But unlike the previous ‘expansive’ options, they set the speaker against that diversity, presenting that voice as challenging or heading off a particular dialogistic alternative. Thus the interpersonal cost to any who would advance such a contrary position is increased and the dialogic space for this alternative in any

Dialogistic Positions and Anticipated Audiences 195 upcoming dialogic interaction is reduced. Accordingly, such formulations are classified as ‘dialogically contractive’. 2.3

Dialogic expansiveness – ‘entertain’

The range of dialogistic contractive and expansive resources is too large for us to consider in any detail in this paper. Beyond so-called epistemic modality, evidentials, and adjuncts such as in my view, dialogic expansion also includes hearsay, various types of attribution and some types of rhetorical question, while the resources of dialogic contraction include other types of attribution, adversatives, concession, negation and other types of rhetorical question. For a full account, see Martin & White, in press. For the purpose of this paper we direct our attention to one sub-type of dialogic expansion, that typically realised through so-called epistemic modals and evidentials and given the label ‘entertain’ in the appraisal literature. (See White, 1998, White, 2000/2000a, White, 2003, Martin & White, in press.) The dialogistic functionality of these values has already been addressed in the earlier discussion of in our judgement, I think, probably and in my view. As indicated there, these are locutions which explicitly ground the proposition in the speaker’s subjectivity and thereby construe the view being advanced as contingent and potentially in tension or alternation with dialogistically divergent positions. These formulations thus ‘entertain’ dialogic alternation. By considering where and how often a speaker employs values of ‘entertain’ it is possible to model one key aspect of stylistic variation – the degree to which, and the contexts in which, the speaker is presented as acknowledging, and opening up the dialogic space for, alternative voices and positions. Such analyses can be applied to develop comparisons of different stages in a given text, different texts, different speakers and different collections of texts. We demonstrate such analyses by means of the following short treatment of an extract from a BBC radio topical discussion programme, Any Questions. In the analysis, all instances of ‘entertain’ are identified and all clause-level propositions (whether or not they are associated with a value of ‘entertain’) are classified according to whether they involve: • assessments of obligation • positive or negative evaluation • counter-expectation • prediction • generalisation • statements about causes or effects • surmise with respect to the thoughts or feelings of others The text is firstly presented with instances of ‘entertain’ identified with underlining and then an analysis of propositions is provided.

196 Pragmatic Markers in Contrast 7. (from BBC radio 4, Any Questions) Well the question was whether Tony Blair would wish Mr Hoon to be gone now or later and one does have the feeling that Mr Hoon is being saved up for the end of the Hutton inquiry when he will be presented as the animal on the sacrificial pyre. Personally I think that he is tarnished, I think his behaviour was – at the intelligence committee – was inexcusable. He didn’t tell a lie but he quite deliberately I think gave a misleading impression. And his behaviour at the Hutton inquiry was even more extraordinary in the way that he ducked and dived and you can’t really have a minister exposing his evasiveness to that extent, let alone a defence minister. I think however that the hue and cry over Mr Hoon is however a monumental distraction. Everybody wants a scalp and it’s good fun, it’s good sport but I think that the underlying issue is much more serious, the underlying question which has been concerning all of us for several months is whether we were lied to and whether the government exaggerated the case for war in taking us to war in Iraq. And my own personal view is so far, so far, from the evidence that’s been produced, I think that it is completely plain that we were not lied to, the government made a perfectly responsible case based on the available intelligence, that we went to war quite properly.

The interpersonal profile revealed by the analysis is one in which the speaker passes numerous quite highly charged value judgements but typically does so in a dialogistically expansive way. All but four of her 14 evaluative propositions are framed with a value of ‘entertain’. As well, the speaker makes one prediction (1), one assessment of obligation (7) and three times claims knowledge of other people’s thoughts, intentions or feelings (5,9,13). Both the prediction and the assessment of obligation are construed in dialogistically expansive terms. Two of the three claims about other people’s thoughts and feelings are monoglossed, making this the only type of proposition for which the speaker does not favour a dialogically expansive mode of expression. One point requiring further discussion does emerge from this analysis – the dialogistic status of I think. This particular pragmatic marker has received a good deal of attention in the literature (see, for example, Urmson, 1952, Hooper, 1975, Lysvåg, 1975, Palmer, 1986, Aijmer, 1997, Simon-Vandenbergen, 1998, 2000). Under the systemic functional linguistics of Michael Halliday and his colleagues (see, for example, Halliday, 1994, Martin, 1992, Matthiessen, 1995, Halliday & Matthiessen, 1999) it is analysed (along with related locutions such as I believe that, I suspect, I know that) as realising the modal value of ‘probability’ (see Halliday, 1994: 254). Palmer, following a similar line of reasoning, holds that the formulation conveys an ‘epistemic judgement’ (Palmer 1986:168) There has been some debate about the effect on propositions framed by I think, with some of the earlier work holding that the effect is a weakening or softening one (see Hooper, 1975). However, later work by Aijmer (1997) and Simon-Vandenbergen (2000) has demonstrated compellingly that I think occurs just as readily with forthright, confident and strongly assertive statements as with utterances which can be seen as in some way tentative or uncertain. Thus while I think may suggest uncertainty when used in the context of uncontentious, ‘factual’ propositions (for example, I think Mary


teaches French) it is typically entirely compatible with authoritativeness and a high degree of conviction when used with evaluative propositions. The speaker’s use of I think in extract 7 above is illustrative of what Aijmer and Simon-Vandenbergen have termed ‘deliberative’ function.

Table 1: Analysis; dialogic expansiveness (Propositions which do not fall within the scope of

you can’t really

I think

I think

X

X

(3) his behaviour was – at the intelligence committee was inexcusable.

X

(4) He didn’t tell a lie (5) he quite deliberately gave a misleading impression

X X

(6) And his behaviour at the Hutton inquiry was even more extraordinary in the way that he ducked and dived (7) have a minister exposing his evasiveness to that extent, let alone a defence minister

(8) however that the hue and cry over Mr Hoon is however a monumental distraction.

X

X X

X

X

X

(9) Everybody wants a scalp (10) it’s good fun

X

(11) it’s good sport (12) the underlying issue is much more serious

X X

(13) [the question] has been concerning all of us for several months

prediction

proposition

X

X

X

X

counter-expect

(16) it is completely plain that we were not lied to, (17) the government made a perfectly responsible case based on the available intelligence, (18) that we went to war quite properly

X

X

obligation

(14) we were lied to; (15) the government exaggerated the case for war in taking us to war in Iraq.

evaluation

the underlying question is whether my own personal view is; from the evidence that’s been produced; I think ‘entertain value’

X

cause or effect

I think

X

other’s cognitive

personally I think

I think

(1) Mr Hoon is being saved up for the end of the Hutton inquiry when he will be presented as the animal on the sacrificial pyre (2) he is tarnished

generalisation

one does have the feeling that

obligation counterexpect prediction

proposition

evaluation

‘entertain value’

generalisation cause or effect other’s cog nit ive

an ‘entertain’ value have been underlined.)


All these accounts are broadly compatible with our view of I think (and related formulations) as explicitly grounding the proposition in the speaker’s subjectivity and thereby construing the proposition as contingent and but one of a number of possible propositions in the current communicative context. Our research, however, points us towards another possibility with regard to the communicative effect of I think which has not, to our knowledge, so far been identified or discussed in the literature. In our study of a corpus of transcripts of some 40 hours of spontaneous topical discussion and debate from the BBC radio programme Any Questions (transcripts of 54 45-minute programmes), we found that I think occurs with a dramatically higher frequency than any of the other values of ‘entertain’. Thus I think occurred with an average frequency of 6.74 instances per 1000 words, while the frequencies for other ‘entertain’ values were as follows: I believe: 0.22 per 1000 words (around 30 times less frequent than I think) I suspect: 0.14 per 1000 words (around 48 times less frequent than I think) I know: 0.31 per 1000 words (around 22 times less frequent than I think) perhaps: 0.36 per 1000 words (around 19 times less frequent than I think) possibly: 0.12 per 1000 words (around 56 times less frequent than I think) probably: 0.45 per 1000 words (around 15 times less frequent than I think)

Table 2: Frequency of I think across individual BBC programmes text with the highest frequency (instances per 1000 words) I think

15.5

text with the lowest frequency (instances per 1000 words) 3.1

I believe

0.55

0.12

I suspect

0.79

0.12

I’m sure

0.66

0.12

perhaps

1.07

0.12

possibly

0.69

0.12

probably

1.29

0.13

may (all uses)

2.55

0.14

might (all uses)

1.58

0.12

must (mostly obligation rather than probability)

1.72

0.12

It was not possible to ascertain the frequency for epistemic uses of the modal auxiliaries, since, obviously, automatic corpus lookup is unable to distinguish, for example, between may when used to realise probability and may when used to realise permission, and we were not in a position to manually tag occurrences. But even when all uses of the modal auxiliaries are included in the count, they are significantly less frequent than instances of I think. For example,

Dialogistic Positions and Anticipated Audiences 199 all uses of may: 0.89 per 1000 words (around 8 times less frequent than I think) all uses of might: 0.6 per 1000 words (around 11 times less frequent than I think) all uses of must: 0.52 per 1000 words (around 13 times less frequent than I think) Even larger discrepancies are revealed when we consider frequencies across individual texts (transcripts of individual programmes). See Table 2. The following extract demonstrates a typical use of I think in texts of this spontaneous, opinionated and argumentative type. 8. [BBC radio 4 – Any Questions, 14/11/2003] The February anti-war march caused real friction in our family because my daughter went off to the march with my husband and I was very, very pro-war and stayed at home with my son. But I could understand why they did it because I think then it was legitimate to say what you felt about the war because it hadn’t yet started and there was a reason to go on a march and to say what you felt and to try and make your opinion heard. Now I think we’ve had the war, I think it’s ludicrous to try and block it anymore, I think what we should be doing is rallying together to try and make sure that it works in Iraq, that we can work together, that we can look to the future and that we have got servicemen out there and that we should be supporting them and that they are doing a fantastically good job and they are risking their lives every single day to do this. And I think that’s where we should be supporting them. I think also also if you look at the interview that President Bush gave today in the Telegraph and the Financial Times he was actually falling over himself to be as really as relaxed as he could and especially about the march and he was saying that he thought there was a great deal of freedom in Britain and he admired the fact that we could go on marches. He was also saying he wasn’t going to try and take this war any further, that he wasn’t going for the axis of evil. And I think that that’s the great benefit that the British can have is that this relationship with America means that we can actually have a restraining influence if we want to and I think that’s very important for us. I think the French went the other way and it means they have no influence at all. There is plenty of evidence to suggest that I think is used in this way, and with this frequency, in spontaneous, opinionated or argumentative spoken language, but not in comparable types of written language. For example, we find that I think occurs at an average frequency of 5.2 instances per 1000 words in a corpus of ten largely off-the-cuff press conferences by the British Prime Minister Tony Blair, but at a frequency of 0.23 per 1000 words (just the one instance) in a corpus of ten of Mr Blair’s prepared speeches (taken at random from the Downing Street website). Similarly we find I think occurring at a rate of 8.92 per 1000 words in the spontaneous contributions of Ziauddin Sardar, one of the participants in one of the Any Questions discussions, but no instances in a corpus of ten of his commentary pieces published in the New Statesman magazine. This clear-cut difference between spontaneous and nonspontaneous discourse does not appear to apply, or at least is not of the same magnitude, for


other values of ‘entertain’. By way of demonstration consider the comparative ratios for the Blair spontaneous corpus (ten press conferences) and the Blair non-spontaneous corpus (ten prepared speeches.) See Table 3:

Table 3: Comparative ratios for the Blair spontaneous and non-spontaneous corpus Blair spontaneous (rate per 1000 words)

Blair non-spontaneous (rate per 1000 words)

I think

5.2

0.23

I believe

0.43

0.59

I suppose

0.00

0.10

I’m sure

0.00

0.11

perhaps

0.38

0.25

possibly

0.00

0.15

probably

0.23

0.33

definitely

0.00

0.00

may/might

0.82

0.97

must (mostly obligation)

1.70

0.20

For values other than I think, frequencies are roughly equivalent across the two modes of language, with some locutions occurring more frequently in the spontaneous discourse (perhaps, must) and others more frequently in the non-spontaneous discourse (I believe, I’m sure, possibly, probably, may/might). It is also noteworthy that in the non-spontaneous corpus, I think occurs with a lower frequency than I believe, perhaps, probably and may/might. This evidence is strongly suggestive that in spontaneous, opinionated, spoken language of a type represented by these corpora, I think is in some significant way different in its functionality, or at least in some aspect of its functionality, from the other ‘entertain’ values. The fact that its behaviour is only anomalous with respect to the other ‘entertain’ value in the context of spontaneous, spoken language is suggestive that it may be a reflex or marker of opinionated spontaneous speech. It suggests that there is a base line which is set for this type of language by which a relatively high frequency is the norm. One possible consequence of this is some de-lexicalising or de-semanticising effect. Sinclair has discussed this effect in the context of collocational co-selections: The meaning of words chosen together is different from their independent meanings. They are at least partly delexicalized. This is the necessary correlate of co-selection. If you know that selections are not independent, and that one selection depends on another, then there must be a result and effect on the meaning which in each individual choice is a delexicalization


of one kind or another. It will not have its independent meaning in full if it is only part of a choice involving one or more words. [Sinclair 1994: 23] It seems plausible that the principle is more generally applicable and that the more ‘automatic’ the use of a particular linguistic item in a given context (i.e. opinionated spontaneous speech), then the less ‘meaning’ will be carried by its presence in the text. It is plausible, therefore, that, in comparison with the other ‘entertain’ values, I think is to some degree de-semanticised and accordingly does not so strongly convey dialogistic expansiveness as these other options. It must be stressed that this effect would only apply with ‘default’ uses of the locution – that is to say when the I think is unstressed and in clause initial position. When, in contrast, the I think is stressed, has its own tone group and/or occurs clause medially or finally, then full optionality is restored and accordingly so is the full dialogistically expansive effect. We observed one such non-default use of I think in an earlier example, where the I think occurred clause medially with an obviously dialogically expansive effect. He didn’t tell a lie but he quite deliberately, I think, gave a misleading impression However, this is not to suggest that I think is so automatised in this type of text that it has lost all its interpersonal functionality. In this regard we note that, in our corpus of Any Questions broadcasts, the frequency of use of I think varies very substantially between different speakers. The following variation was observed across the participants in five programmes (see Table 4). This indicates that a significant optionality still applies to the use of I think and that a significant variation in style may result as a consequence of speakers more frequently or less frequently using this locution. As a consequence, we believe any exploration of dialogistic expansiveness as an axis of stylistic variability needs to attend to instances of I think, but that it is wise to distinguish uses of I think from uses of the other values of ‘entertain’ on the grounds that I think may be less strongly dialogistically expansive than these other resources. Thus we would note, in the analysis conducted above with respect to extract 7, that it was I think, rather than other options, which was employed by the speaker in the majority of cases. Similarly, we would expect significantly different dialogistic effects in texts where, for example, locutions such as I believe, in my view, definitely, perhaps, probably, as I see it, and in my judgement were used in place of I think.


Table 4: The frequency of I think for different speakers in the Any Questions broadcasts

2.4

speaker

total of instances

rate per 1000 words

thomson

49

31.37

alli

28

23.37

bercow

26

17.99

fortier

12

14.25

willetts

25

13.56

phillips

25

12.99

duncan

17

12.51

steel

17

10.73

oaten

11

10.39

jowell

19

10.06

foulkes

18

9.87

fabiani

13

9.19

sardar

8

8.92

hughes

18

8.31

mercer

7

6.83

moisi

7

6.07

keetch

10

5.76

graffy

6

5.26

cashman

6

4.88

roberts

4

3.38

guerot

4

2.53

Dialogic contraction – ‘concur’

The resources of dialogistic expansion – of which ‘entertain’ is one sub-type – stand in contrast to resources of dialogistic contraction, as outlined above. Once again we do not have the space here to consider the full repertoire available in English and for illustrative purposes focus on just one sub-type. (For a full account of all resources of dialogic contraction, see Martin & White, in press: chapter 3.) These are what, within the appraisal framework are termed values of ‘concur’ – meanings by which the speaker is presented as anticipating that the information about to be conveyed will already be known by the addressee or that the view about to be expressed is one which the addressee will inevitably agree with and share with the speaker. Thus these are meanings by which the speaker construes a relationship of alignment and solidarity – of ‘concurrence’ – with the purported addressee he/she is construing for the text. In English this dialogistic position is conveyed via locutions such as of course, naturally, obviously, as you know, needless to say, and it goes without saying. For example,


9. (anticipating that the information about to be conveyed will be known) Dimbleby: (Any Questions presenter, introducing the listeners to the church hall from which the programme is being broadcast.) Welcome to Petersfield in Hampshire which is decked out for Christmas and where we’re in St. Peter’s Church, which is renowned architecturally for its fine Norman tower and socially for its concerts, plays, exhibitions and civic events, as well as being of course a place of Christian worship. 10. (anticipating that the view about to be expressed will be shared by the addressee) Dimbleby: (programme presenter and interviewer) What do you make then of the point that Peter Hitchens was mating – making, to the effect that the marriage is fundamental to the belief of the church and fundamental to its identity, as he believes it also to be in a coherent civilised society? Bryant: (in response) Of course marriage is absolutely essential to a coherent and a good society and for the vast majority of people it’s the way they’re going to live their lives but there are some people, like myself, who are gay or are lesbian who are never going to have the opportunity of marriage, who might want to live in long trusting loving relationships and I think the church should be helping people to do that rather than making it more difficult.

11. (anticipating that the addressee will find the forecast being made to be self evident.) Prime Minister: I don’t think I should comment on any intelligence that we have. Our objective remains as it is, to remove Saddam’s regime, and the people that are closest to him will naturally want to cling on to power. This relationship of ‘concurrence’ may also be realised via certain types of rhetorical or ‘leading’ questions – those by which the writer/speaker is presented as assuming that no answer needs to be supplied for a particular question on account of that answer being so ‘obvious’. Addresser and addressee are thus presented as so thoroughly in alignment, and the proposition at issue so ‘commonsensical’, that agreement can be taken for granted. Consider by way of example the following leading question from the front cover of the March 25, 2002 edition of the New Statesman magazine. A full-page coloured picture depicts a group of young, friendly and happy Iraqi children in close-up. Looking up towards the camera, they wave bunches of flowers and with welcoming smiles directly engage with the viewer. The text of a headline superimposed over the picture reads: Iraq: Should we go to war against these children? Here the question leads the reader to an ‘unavoidable’ answer. The text operates under the assumption that the reader will inevitably supply, ‘No, of course we shouldn’t go to war against these children.’


These various concurring formulations, then, are dialogistic in that they present the speaker as ‘in dialogue’ with their audience generally. They are contractive in that they represent the shared knowledge, value or belief as universally, or at least as very widely, held in the current communicative context. Thus they have the effect of excluding any dialogistic alternatives from any ongoing communicative exchange in that they position anyone who would advance such an alternative as at odds with what is purportedly generally agreed upon or known. Accordingly, they construe for the text a backdrop which is heteroglossic in that it contains multiple voices (the authorial voice and those it is purportedly in concord with), but from which dissident voices and positions are excluded. By attending to how often and with respect to which propositions the speaker employs such locutions, we are able to map another highly significant line of stylistic variation. Such analyses provide for comparisons of speakers with respect to the degree they lay claim to a knowledge of what the addressee knows and takes for granted. In order to demonstrate such an analysis we supply below all the instances of ‘concurrence’ from the press conference by British Prime Minister Mr Tony Blair which we referenced above. We follow the same procedure as adopted in our earlier analysis, although in this case we have not analysed continuous spans of text but rather isolated just those propositions which are associated with a ‘concurrence’ value. We have added the category of ‘experiential’ for propositions which can be seen as describing real world events without also providing evaluation, assessments of obligation, surmise or prediction. The category of ‘inclination’ has been added alongside that of ‘obligation’ to capture those propositions where the speaker is outlining his own willingness, inclinations or desires (see Table 5):

obviously

(1) Obviously I will take questions on any issues you

(2) Of course I understand the concerns of the thousands (3) and of course I should [listen]

of course

(4) and do listen to those concerns

own/ other’s cognitive X

X

that marched on Saturday, of course

cause or effect

X

want to raise,

of course

generalisation

evaluation

proposition

future / prediction

value’

experiential

‘concurrence

obligation or inclination

Table 5: Types of concurrence value associated with propositions

X X


obviously

(5) Most of the people who are concerned, who went on the march, are not actually in the position of being

X

against conflict in all circumstances, though obviously some are.

leading question

(6) Is it reasonable, after 12 years of flouting previous UN resolutions, that Saddam is given a final chance to

X

fulfil them? leading question

(7) Is it reasonable to ask Saddam to account for the weapons that independent UN inspectors have said

X

remain unaccounted for?

leading question

(8) Is it reasonable to ask Saddam to let weapons inspectors speak to the scientists and experts that worked on his programmes, without those people

X

fearing for their lives or being intimidated? as you know

(9) And as you know there are situations in which I suppose in politics you have to manoeuvre your way

X

round certain issues obviously

(10) This is something obviously that has to be

X

discussed not just with allies, but with the UN and with people inside Iraq.

obviously

(11) Obviously my preference is that the more that we can at least set ourselves on a path towards greater

X

democracy the better of course

(12) Well my reaction to that is that of course there are

X

Iraqi children suffering from malnutrition today. of course

(13) So yes of course there are consequences we have to

X

guard against, and we will guard against those of course

X

(14) If I thought we were going to unleash something in which hundreds of thousands of people were going to die, we were going to have more Bin Ladens, the Middle East was going to go up in flames, no of course I don’t believe that that is the case

obviously

X

(15) It is important at least, and that is why obviously we can only communicate in the end through the people

own or other’s cognitive

cause or effect

generalisation

future event / prediction

evaluation

experiential

proposition

value’

obligation or inclination

here,

‘concurrence

206 Pragmatic Markers in Contrast The analysis provides a picture of a dialogistic orientation in which the speaker quite regularly signals (15 instances in one press conference) an assumption that he and his interlocutors share knowledge, beliefs, expectations and values. We note with interest that many of these values of ‘concurrence’ construe an addressee who has an intimate knowledge of the Prime Minister’s own inclinations, desires, inner thoughts, level of understanding and the obligations he is under (i.e. 1, 2, 9, 10, 11, 13, 14).

3.

CROSS-LINGUISTIC APPLICATIONS

The framework we have been setting out, then, is one in which the functionality of these various locutions is accounted for by reference to broad axes of variation in dialogistic positioning. As outlined above, these include, at the broadest level of analysis, the distinction between the monoglossic and the heteroglossic, and then at the next level down in delicacy, between the dialogically expansive and the dialogistically contractive. Within these two further categories, subtypes are identified according to finer grained differences in the dialogistic relationship being negotiated (e.g. the subtype of ‘entertain within dialogic expansion, and the subtype of ‘concur’ within dialogic expansion, and then further subtypes). This approach enables analyses in which we interrelate a diverse array of locutions according to similarities and differences in their dialogistic functionality. The framework has the potential to provide a methodology for the comparison of texts, registers and individual interpersonal styles across languages. It provides a methodology for the discovery of similarities and differences according to the type of intersubjective, dialogistic relationships negotiated by speakers and the ways in which they go about construing for the text a putative addressee. We are not assuming, of course, that that the same repertoire of dialogistic options which operates in English will apply in other languages. The case of Wanka (and similar languages) discussed above points to the need to allow that languages may operate with significantly different dialogistic arrangements, even at this basic level of the distinction between monoglossic and heteroglossic modes of expression. We thus treat as an open question the issue of whether or not the taxonomy of subtypes of ‘engagement’ proposed for English can usefully be applied when working with other languages. We should, however, foreshadow that our own preliminary work on Japanese and English strongly suggests that the broader axes of variation in dialogistic functionality outlined above do apply in Japanese as well as in English. Thus, it has been relatively unproblematic to establish that Japanese operates with the broad-based distinctions between the monoglossic and the heteroglossic, and between the dialogistically expansive and contractive, even while our preliminary study suggests that the ‘valeur’ of these options (i.e. the rhetorical effects associated with the choice of the bare assertion) will almost certainly differ between the two languages. Our initial work also suggests that substantial differences will emerge between the two languages at more delicate levels of analysis – when, for example, we explore the range of options which are available, for example, for ‘entertaining’ or for ‘concurring’.


To illustrate something of how cross-linguistic applications of the framework may proceed, we offer in the following section a brief account of an analysis in which we compare an extract from a press conference by the British Prime Minister Tony Blair (introduced earlier in the discussion) with a press conference by the Japanese Prime Minister, Mr Junichiroo Koizumi. 3.1

Comparing the dialogistic rhetoric of two prime ministers

By choosing to work with transcripts of press conferences by the two prime ministers we arrange, of course, to compare texts which are substantially similar with respect to their social setting and their communicative purposes, although we acknowledge that there may still be substantial differences between the conduct of Japanese and British politics and hence in the social ‘meaning’ of such events. The conference by the Japanese Prime Minister (delivered on January 4, 2005 and available at www.kantei.go.jp/jp/koizumispeech/2005/01/04press.html) was concerned with the Japanese cabinet’s plans for the coming year, the government’s proposals for the privatisation of the Japanese Ministry of Posts and Telecommunications, the revision of the Japanese constitution, the situation in North Korea, and other foreign affairs matters. The press conference by the British Prime Minister (delivered on February 18, 2003 and available at www.number-10.gov.uk/output/Page3006.asp) was mostly concerned with the issue of Iraq and the terms under which Britain might join with the United States in an invasion of that country. Given the objectives of this paper, our purpose is not to reach any clear conclusions about general similarities and differences between Japanese and British prime ministerial rhetoric, but simply to illustrate the application of the ‘engagement’ framework in this type of cross-linguistic work. We selected for analysis the contributions of the two Prime Ministers, excluding the journalists’ questions. All of the Japanese Prime Minister’s words were analysed (some 185 top level clauses) and the first half of the British press conference (some 142 clauses). All propositions were tagged according to the taxonomy of proposition types set out above and reported in table 1 (i.e. identifying whether the proposition involves value judgements, assessments of obligation, predictions, surmise about other’s mental processes and so on). We identified and classified all instances of ‘engagement’, and recorded the type of proposition over which they had scope. Here we confine ourselves to discussing our findings with respect to those propositions which were (a) monoglossically formulated (i.e. by means of bare assertion), which (b) fell within the scope of a value of ‘entertain’, and which (c) fell within the scope of a value of ‘concur’. 3.1.1

The monoglossic – construing propositions as dialogistically inert

There is one potential complication with respect to the monogloss versus heterogloss distinction. In Japanese, it is usual for utterances to be grammatically marked for level of what is typically termed ‘politeness’. Thus the form of the verb (as well the forms of some

208 Pragmatic Markers in Contrast other parts of speech) will vary according to the social identity and status of the speaker, and degrees of deference, distance or familiarity between interlocutors. This act of choosing a level of politeness can be seen as dialogistic to the extent that speakers thereby construe for themselves an addressee with whom they enter into a particular relationship of equality/inequality and distance/familiarity. Nevertheless, the monogloss versus heterogloss distinction holds in Japanese in that there are ‘bare’ forms which, even when marked for politeness, do not present the speaker as referencing or responding to what others might have said previously, as recognising the possibility of other viewpoints, or as anticipating particular responses or reactions on the part of the actual or putative addressee. There was a clear difference between the texts in terms of the use the two speakers made of the monogloss option. Seventy-four propositions in the British text were barely asserted (52 percent of the total propositions) whereas only 31 propositions (17 percent) in the Japanese text were construed in this way. There were also differences in the profiles of the proposition which were formulated in this way. In the English text this monoglossic mode of expression was regularly used with evaluative propositions (27 instances, 36 percent of total monoglossic utterances). For example, Saddam is a threat. They have a rightful hatred of the consequences of war.

Iraqi co-operation is the reasonable and easily delivered requirement of the international community. it is the very nature of Saddam, how he operates is history, how he treats his people, that mean that in his hands these weapons of mass destruction are all the more dangerous.

and also with propositions by which the Prime Minister asserted what should and should not happen (12 instances, 16 percent of monoglossic assertions). For example, the basis upon which we act has to be the disarmament of Iraq of weapons of mass destruction These are threats best dealt with by a unified international community.

it is worth people understanding that there are also consequences of not taking action in terms of bloodshed Thirteen (18 percent) of the monoglossic utterances involved content which could be classified as essentially ‘informational’ or ‘experiential’. For example: he is the only leader who has used them [weapons of mass destruction] he is the only leader still in power that has twice declared war on his neighbours

Dialogistic Positions and Anticipated Audiences 209 [groups] who are using chemical and biological poisons However we note that even here many of these ‘experiential’ meanings had a clear potential for triggering attitudinal assessments on the part of the listener. In contrast, in the Japanese text, monoglossed ‘informational/experiential’ propositions substantially outweighed the monoglossed evaluative propositions. Specifically, 69 percent of the monoglossed propositions in the Japanese Prime Minister’s contributions were essentially ‘informational’ in this way. For example, kotoshi wa, 3 gatsu kara, Aichi ken de banpaku ga kaisai-saremasu. (In this year, from March, in Aichi prefecture, an international exposition will be held) sudeni yuubin ni shitemo, aruiwa kodutsumi ni shitemo, cyokin ni shitemo, hoken ni shitemo, minkan de yatteiru zigyoo de arimasu. (Already, postal service, delivery service, banking, insurance are the jobs that are managed privately.) Some 31 percent of monoglossed propositions were evaluative. For example, kore wa nihon ni tottemo kooki de aru, chansu de aru to. (This comes at a good time and is a [good] opportunity for Japan.) tuduite honyosan, koremo ima, daizi na keizai no kyokumen ni kiteorimasu. (and about the main budget, this issue is also occurring at an important point in the economic cycle). None of the seven instances in the Japanese text where the Prime Minister made assessments of what should or should not happen were monoglossed (compared with the 12 instances of monoglossic assessments of obligation in the British text). The contrast which emerges from this analysis is one in which it is predominantly only ‘factual’ uncontentious propositions which the Japanese Prime Minister treats as dialogically inert (no recognition required of alternative view points and voices) while the British Prime Minister, with great regularity, construes potentially contentious evaluations and assessments in these dialogistically inert terms. 3.1.2 Dialogistic expansiveness – ‘entertain’ Japanese has available a range of resources by which the proposition is explicitly grounded in the speaker’s contingent subjectivity and by which, thereby, the possibility that the speaker is in tension with alternative voices and viewpoints is recognised. These include, by way of some examples,

210 Pragmatic Markers in Contrast soo da (look, appear to be), rasii (it seems), yoo da (it appears that), mitai da (it appears that); -kamo sirenai (might), - kasira [kanaa] (I wonder), (ni) cigainai (must, certainly) and omoo (I think/believe) (See Teruya 1998, from which we have taken some of these examples, for further discussion.). The British Prime Minister employed values of ‘entertain’ on 15 occasions (around 11 percent of total propositions) while the Japanese Prime Minister employed them on 41 occasions (around 22 percent of total propositions.) For the British Prime Minister, they are typically used in the context of evaluative propositions (7 out of 15 uses). For example, It is plain, in our judgement, that Saddam continues to be in breach. I think they are very, very powerful testimonies as to the nature of the regime Tony Blair

He also used them with some regularity (5 instances) in combination with assertions of what should or should not happen. I think there are certain situations in which you have simply got to say to people look this is what I believe The Japanese Prime Minister also typically used these dialogistically expansive formulations in the context of evaluative propositions (22 out of 41 instances) and with assertions of what should and shouldn’t happen (7 instances). For example, genzai no koosha no keitai yorimo minkanzin ni keiee wo makaseta hoo ga, yori kokumin no saabisu koozyoo ni tunagaru no dewa naika to. (Rather than the current system of the public corporation, if (we) leave the management to private ownership, that would provide a better service for the public.) minasamagata mo taihen gokuroo no ooi toshi data to omoimasu. ((I) think that the last year was a tough/hard year for lots of you too.)

The analysis demonstrates a high degree of similarity in the uses to which these two speakers put these meanings (signaling openness to dialogistic alternatives when advancing evaluative propositions), though clearly they differ in the frequency with which they adopt this dialogically expansive stance, with the Japanese Prime Minister almost twice as likely to employ values of ‘entertain’ as the British Prime Minister. A further point of similarity is revealed when we focus on the I think form (as discussed at length above) and its nearest Japanese equivalent omoo (and its various conjugational variants). The I think/omoo form is by far the most frequent value of ‘entertain’ in both texts (9 out of 15 instances of ‘entertain’ in the Blair text; 36 instances out of 41 in the Koizumi text), though the preference for this form is substantially greater in the Japanese than in the

Dialogistic Positions and Anticipated Audiences 211 British text. The Japanese Prime Minister’s use of this form is reminiscent of the use made of I think in the Any Questions discussions cited above. For example, compare extracts 7 and 8 above with the following extract from the opening of the Japanese Prime Minister’s address. Shinnen, omedetoo gozaimasu. sakunen wa, nihon ni okimashitemo, taihuu, zishin, shuuchuu-goou too, ookina higai wo uke, minasamagata mo taihen gokuroo no ooi toshi data to omoimasu. higaisha no minasan-gata mo, ima, hukkyuu, hukkoo katsudoo ni tutusinde orareru to omoimasu ga, zehi tomo kono konnan kara tatiagatte, aratana shinnen wo mukaemasite, kiboo wo motte tiiki no hukkoo ni torikunde itadakitai to omoimasu. (Happy New Year. Last year, Japan (as in other countries) suffered damage from typhoon, earthquake, heavy rain, etc. (I) think that the last year was a tough/hard year for lots of you too. Although (I) think/imagine that those who suffered from the damage are working hard for restoration, (I) think that you will overcome this difficult situation, enter a new year and work on with the restoration with hope.) This leads us to postulate that the same de-semanticisation effect which we hypothesized as operating in English with I think, may be operational in Japanese with omoo. It is interesting to note, that if we exclude instances of I think/omoo from our analysis, then the Japanese Prime Minister makes less use of values of ‘entertain’ than the British Prime Minister. Obviously further research is required to establish the degree of variation in the use of I think/omoo across different social settings and different types of texts. 3.1.3

Dialogic contraction – ‘concur’

Our earlier discussion of ‘concur’ established that the British Prime Minister makes quite regular use of this meaning, thereby constructing a discursive persona which assumes substantial solidarity and common knowledge with those being addressed. The Japanese text did not contain any clear-cut instances of this value. There was just one marginal instance: “minkan ni dekirukoto wa minkanni”, “gyoo-zaisei-kaikaku wo dankoo seyo”, “koomuin wo hearse” to iu koto ni tuite wa, hotondo subeteno too ga sansei siteorimasu Most parties agree with “privatising institutions under public management”, “carrying out the reformation of administrative and financial system” and “the reduction of the number of public servants.

It would seem, then, that Japanese Prime Minister differs from his British counterpart in construing a more socially distant relationship between himself and his journalistic questioners – one in which there is little overt signalling of an assumption of commonality and mutual knowledge. We do note that the majority of Mr Blair’s assumptions of solidarity


are self directed. He assumes that his questioners will be entirely familiar with his thoughts and inclinations and that his good intentions will be self-evident to all. In this regard, Mr Blair might be seen as more self-involved than Mr Koizumi.

4.

CONCLUDING REMARKS

By this account, then, we hope to have demonstrated the new insights provided by a dialogistic understanding of pragmatic markers such as I think, indeed, of course, and needless to say. Under this perspective, the division which is often drawn between these socalled ‘markers’ and the ‘grammar’, on one side, and the ‘lexis’ on the other, dissolves as we observe these apparently different types of meaning all contributing to the negotiation of dialogistic positioning. We have also demonstrated how this framework can provide for crosslinguistic comparisons in which the focus is not upon isolated, individual meanings but upon the systemic opportunities languages provide for the speaker to construe for their texts particular heteroglossic backdrops and putative addressees.

REFERENCES Aijmer, K. (1997). I think – an English modal particle. In: Modality in Germanic Languages. Historical and Comparative Perspectives (Swan,T. and O. Westvik, eds),pp. 1–47. Mouton de Gruyter , Berlin/New York. Bakhtin, M. M. (1981). The Dialogic Imagination (translated by C. Emerson & M. Holquist). University of Texas Press, Austin. Chafe, W. & J. Nichols (eds). (1986). Evidentiality: the Linguistic Coding of Epistemology. Ablex, Norwood, N.J. Christie, F. and J. R. Martin (eds). 1997. Genres and Institutions: Social Processes in the Workplace and School. Cassell (Open Linguistics Series), London. Coates, J. (1983). The Semantics of Modal Auxiliaries. Croom Helm, London & Canberra. Coffin, C. (1997). Constructing and giving value to the past: an investigation into second school history. In: Genre and Institutions – Social Processes in the Workplace and School (Christie, F. and , J.R. Martin ,eds), pp. 196–230. Cassell, London. Coffin, C. (2003). Reconstruals of the past – Settlement or invasion? The role of JUDGEMENT Analysis. In: Re/reading the Past: Critical and Functional Perspectives on Discourses of History (J. R. Martin and R. Wodak, eds), pp. 219–246. John Benjamins, Amsterdam. Crismore, A. (1989). Talking with Readers: Metadiscourse as Rhetorical Act (American University Studies Series XIV : Education), Peter Lang Publishing, New York. Floyd, R. (1996). Experience, certainty and control, and the direct evidential in Wanka Quechua questions. Functions of Language, 3.1, 69–93. Fuller, G. (1995). Engaging Cultures: Negotiating Discourse in Popular Science. University of Sydney Ph.D. Thesis.

Dialogistic Positions and Anticipated Audiences 213 Fuller, G. (1998). Cultivating science: Negotiating discourse in the popular texts of Stephen Jay Gould. In: Martin & Veel 1998, 35–62. Halliday, M.A.K. (2004/1994). An Introduction to Functional Grammar. (2004 third edition revised by C.M.I.M. Matthiessen). Edward Arnold, London. Halliday, M.A.K. & C.M.I.M. Matthiessen. (1999). Construing Experience through Meaning: a Language-based Approach to Cognition. Continuum, London. Hooper, J.B., (1975), On assertive predicates. In: Syntax and Semantics vol 4 (J.P. Kimball, ed.), pp. 91–124. Academic Press, New York. Iedema, R., S. Feez & P.R.R. White. (1994). Media Literacy (Write it Right Literacy in Industry Research Project – Stage 2). Metropolitan East Disadvantaged Schools Program: Sydney. Lyons, J. (1977). Semantics. Cambridge University Press, Cambridge. Lysvåg, P. (1975), Verbs of hedging. In: Syntax and Semantics vol 4 (J.P. Kimball, ed.), pp. 125–154. Academic Press, New York. Macken-Horarik, M. & J.R. Martin (eds) (2003). Negotiating Heteroglossia: Social Perspectives on Evaluation. (Special Issue of Text 23.2). Markkanen, R. & H. Schröder (1997). Hedging and Discourse: Approaches to the Analysis of a Pragmatic Phenomenon in Academic Texts. Walter De Gruyter, The Hague. Martin, J.R. & White, P.R.R., in press, The Language of Evaluation – Appraisal in English. London & New York, Palgrave. Martin, J. R. (1992). English Text: System and Structure. Benjamins, Amsterdam. Martin, J. R. (2000). Beyond exchange: Appraisal systems in English. In: Evaluation in Text – Authorial Stance and Construction of Discourse (S. Hunston & G. Thompson,eds), pp. 142–175. Oxford University Press, Oxford. Matthiessen, C.M.I.M. (1995), Lexicogrammatical Cartography : English Systems. International Language Sciences, Tokyo. Palmer, F.R. (1986). Mood and Modality, Cambridge University Press, Cambridge, UK. Rothery, J. & M. Stenglin (1997). Entertaining and instructing: Exploring experience through story. In: Christie & Martin 1997, 231–263. Rothery, J. & M. Stenglin (2000). Interpreting literature: the role of appraisal. In: Researching Language in Schools and Communities: Functional Linguistic Perspectives (L. Unsworth ,ed.), pp. 222–244, Cassell, London. Simon-Vandenbergen, A.-M. (1998). I think and its Dutch equivalents in parliamentary debates. In: Corpora and Cross-Linguistic Research. Theory, Method and Case Studies (S. JohanssonS. Oksefjell, eds), pp. 297–331. Rodopi: Amsterdam/Atlanta. Simon-Vandenbergen, A.-M. (2000). The function of I think in political discourse, International Journal of Applied Linguistics 10 (1), 41–63. Sinclair, J.M. (1994). Trust the Text. In: Advances in Written Text Analysis (M. Coulthard, ed.), pp. 12–25. Routledge, London Teruya, K. (1998), An Exploration into the World of Experience: a Systemic-functional Interpretation of the Grammar of Japanese, unpublished Ph.D dissertation, Macquarie University

214 Pragmatic Markers in Contrast Urmson, J.O. (1952). Parenthetical verbs. Mind LXI, 480–496. Voloshinov, V.N. (1995). Marxism and the Philosophy of Language, Bakhtinian Thought – an Introductory Reader. S. Dentith, L. Matejka & I.R. Titunik, (trans), Routledge, London. White, P.R.R. (1998). Telling Media Tales: the News Story as Rhetoric. Ph.D dissertation, Sydney, University of Sydney. White, P.R R. (2000). Dialogue and Inter-Subjectivity: Reinterpreting the Semantics of Modality and Hedging. In: Working With Dialogue (M. Coulthard, Cotterill, J. & F. Rock, eds), pp. 67–80. Niemeyer, Tübingen. White, P.R.R. (2002). Appraisal – the language of evaluation and stance. In: The Handbook of Pragmatics (J. Verschueren, Östman, J.-O., Blommaert, J. & C. Bulcaen, eds), pp. 1– 27. John Benjamins, Amsterdam. White, P.R.R. (2003). Beyond modality and hedging: a dialogic view of the language of intersubjective stance. Text – Special Edition on Appraisal, 259–284. White, P.R.R. (2004). Appraisal web site: www.grammatics.com/appraisal White, P.R.R., to appear, Modality as dialogue – a Bakhtinian reanalysis of “epistemic” stance. Word – Journal of the International Linguistics Association.


VRAIMENT AND REALLY IN CONTRAST. WHEN TRUTH AND REALITY MEET Dominique Willems, Ghent University Annemie Demol, Ghent University Our contribution starts with a presentation of the multiple uses of the particle vraiment in spoken and written French, followed by a brief comparison of vraiment with some other particles from the same semantic field (réellement, en réalité, à vrai dire). The descriptive monolingual approach, based on the analysis of some 1000 examples in written and spoken French, not only throws new light on the function of these particles in modern French, but also throws up several specific methodological problems one is confronted with while working with real language material, especially with spoken language data. In the second part of our contribution we present a synthesis of the uses of English really as proposed by the available linguistic literature and make a first comparison with the results of our analysis of vraiment. Finally, we present the results of a contrastive analysis, based on a multilingual translation corpus, not only of vraiment and really, but also of their Dutch counterparts echt and werkelijk.

1. 1.1.

VRAIMENT1 Frequency

Vraiment is particularly frequent both in written and spoken French: we found no less than 2155 occurrences in Le Monde (1994) (21 million words) and 1049 in the Corpaix corpus of spoken French (1 million words). According to the statistics of M. Bilger (2003), vraiment occupies the first place of all modalising adverbs in the reference corpus of contemporary spoken French.2 In her written corpus (extracts from Le Monde), vraiment occupies only the 5th rank.3 It is also interesting to note that modern French vraiment seems to be three times more frequent than it was in 1600 (cf. corpus TLF4). It has consequently known a growing success over the centuries. 1 Our analysis is based on the material compiled by Brecht De Cooman in his MA thesis “Une description des différents emplois de vraiment, dans le cadre des tentatives de classification des adverbiaux”, University of Ghent, 2001. 2 With 25% of all occurrences, vraiment is 3 times as frequent as the second modalising adverb justement in the “Corpus de Référence de Français Parlé”. For a description of this corpus cf. Bilger (2000). 3 After seulement, notamment, également and actuellement. 4 The corpus of the Trésor de la langue française (TLF) is constituted of more than 2000 texts of various types from the XVIIth to the XXth centuries.


1.2.

Multifunctionality

As is usually the case for frequent words, it is not really surprising to note the multifunctional character of contemporary French vraiment. Table 1 offers a synthesis of its different uses, taking into account (1) the degree of syntactic and prosodic autonomy of the particle; (2) its functional level, syntactic distribution and scope; (3) its position in the sentence and (4) its semantic content. The various uses are presented on a continuum, going from left to right, from the more integrated and grammaticalized functions to the less integrated, pragmaticalized and contextualized uses.

No separating features Phrase level Attributive

Predicative

Validation5 of the modified element (Adj, V): – in relation to the cognitive prototype; – in relation to reality intensifier (with properties or gradable processes) metalinguistic function (lexical search)

Separated

Autonomous

Sentence level Dialogue level Initial Medial Final Responses

Clause level Other verbal constructions

Validation of the proposition in relation to reality: truth identifier (emphaziser) privileged link to – various modalities – subjectivity (opinion, evaluation, feeling verbs)

Validation of utterance –speaker’s involvement and sincerity –? invitation for the hearer to validate his/her sincerity initial: emphasizer final: softener ? doubt

Validation of previous context (+ oui, non) ?! need for validation hope, surprise, irony, interest (…)

?6: doubt of the validity of the proposition rhetorical questions

– more integrated – syntactic-semantic function – more grammaticalized/lexicalized

– less integrated – semantic-pragmatic function – discursive/contextual

Figure 1: functions of vraiment

The multiple uses of vraiment can be classified in three main categories: (i) a typical adverbial use, where vraiment functions at clause level, modalizing the predicate as an epistemic truth identifier (“what is being said is true”). In this case, vraiment then emphasizes the correspondence between what is said and

5

The notion of “validation”, used by Gezundhajt (2000:279) to characterize one of the uses of vraiment, can in fact be extended to all its uses. The core meaning of vraiment always involves a notion of conformity, correspondence, or adequacy, which we propose to subsume under “validation”. 6 ? = in an interrogative sentence;! = exclamative sentence.

Vraiment and Really in Contrast 217 extralinguistic reality (especially when this reality is unexpected, or could seem exaggerated to the hearer)7: 1. tu vis avec une personne tu la découvres vraiment et euh et puis tu tu découvres vraiment la réalité quoi à travers cette personne [Corpaix, CITU 1,16] 2. et c'est on dirait vraiment des dessins animés quand tu vois les dessins [Corpaix, RENCONT 4,12] 3. (…) leur donner un nom qui s'adapte vraiment à ce qu'elles représentent [Corpaix, 30SYL 8,13] In this use, vraiment often co-occurs with verbs of opinion, evaluation or feeling, and with certain modalities (possibility, necessity). The link with subjective modality is made even more evident by the frequent use of personal particles: moi, pour moi (…). In these cases, vraiment often has a mere emphasizing function. 4. j'avais d'ailleurs horreur de l'école je detes- je vraiment je detestais détestais ça et j'hésitais mais je crois qu' il faut le dire vraiment” [Corpaix PORTUGA3 2,13] 5. moi, j'ai vraiment trouvé que ils ils maîtrisaient la langue relativement bien [Corpaix, ACCENT 9,7] 6. il y a vraiment enfin je pense qu'il y a vraiment une une grande différence entre (…) [Corpaix, 8CUBALEX 7,1] 7. bon ben euh c' – pour moi c'est c'est vraiment le le point noir aujourd'hui [Corpaix, MAURIEPC 2,15] 8. ils s'attendaient vraiment pas à ce que ce soit euh un client quoi… [Corpaix, VENGEANC 6,9] In interrogative clauses, vraiment expresses doubt concerning the validity of the propositional content (9) and frequently signposts a rhetorical question (pragmatic derivation (10)): 9. Mais est-ce que, vraiment, la crédulité va aussi loin? [Le Monde, 1994] 10. Est-il vraiment nécessaire d'ajouter une petite pierre au monument élevé par le Figaro (…). Peut-être pas. [Le Monde, 1994] 11. Est-ce que vous aviez vraiment euh le le désir d'emporter cette montagne [Corpaix, EXPEDITIO 5,2] (ii)

7

a use at the phrase level as modifier of an adjective, a noun or a noun phrase: from the original meaning of adequacy to the cognitive prototype (12), or to the reality (13), vraiment, combined with gradable adjectives

Most examples will be taken from spoken French (Corpaix).


or properties, has evolved to a more grammaticalized use as a mere intensifier (semantically close to très) (14,15). In this use vraiment can be followed by any grammatical category, including the negation particle pas (contrary to très, fort and other intensifiers): 12. Véronique Prigent, une journaliste avec une carte de presse, donc vraiment journaliste [Le Monde, 1994] 13. et là on a réalisé là c'est c'est vraiment là qu'on a réalisé en fait (…) [Corpaix, ATTENTAT 8,13] 14. un gros bouquet de fleurs d'Interflora mais vraiment gros hein + [Corpaix, PARO2 11,12] 15. en a fait un un magnifique bassin à poissons mais vraiment magnifique avec des rocailles avec des fontaines [Corpaix, GEORGES 5,4] 16. (…) euh non non je parle vraiment pas bien [Corpaix, CYRIL 14,3] It is interesting to note that a similar grammaticalized use of the French adjective vrai is at the origin of the English adverb very.8 In a metalinguistic use, vraiment functions as a reformulation device, correcting a previous lexical choice (17, 18) 17. / et, eh / là c' était pas vraiment une société mais une association qui regroupait (…) [Corpaix, JEANNE 5,11] 18. euh c'est pas un lotissement quoi c' est + vraiment euh + les voisins sont + sont très espacés euh [Corpaix, CAMPAGN 3,1] By far the most frequent structure in which we have found vraiment in the spoken language corpus is the predicative construction “c'est vraiment x”9: in this structure vraiment emphazises the truthfulness of a definition, a qualification or an evaluation. It is sometimes difficult to distinguish between the truth emphasizing function relating to the whole proposition (i) and the intensifying function with the predicative adjective as scope (ii):

8

Cf. Guimier (1988: 58) “La grammaticalisation est totale lorsque l’adverbe est compatible avec tous les adjectifs gradables. Tel est le cas de very, intensif par excellence en anglais moderne. Issu de l’adjectif verrai emprunté au vieux français (=vrai), il emporte, lui aussi, dès l’origine, l’image d’un mouvement positivant (…) Ce n’est pas un hasard si, en anglais américain tout particulièrement, l’adjectif real (qui évoque au départ, tout comme verrai, l’idée d’authenticité) est engagé dans le même processus et est de plus en plus employé comme adverbe intensif. Or ce very ou real seront souvent employés pour traduire vraiment, qui en français contemporain aussi semble se grammaticaliser et ‘dématérialiser’, l’emploi intensif devenant apparemment de plus en plus dominant”. 9 This has also been noted by Bilger (2003: 15): “En fait, la seule distribution qui semble remarquable (…) est celle qui concerne cet adverbe (= vraiment) avec le verbe c’est; en effet, on note que dans le corpus de parole “privée”, les exemples du type c’est vraiment particulier (…) représentent 25% des emplois de vraiment.”

Vraiment and Really in Contrast 219 19. c'est vraiment à la dernière minute c'est vraiment le bout de la Maurienne c'est vraiment une utopie complète c'est vraiment intéressant c'était c'était vraiment le départ quoi là c'est vraiment la sortie des feuilles c'était vraiment euh c'était vraiment euh très incompréhensible This ambiguity between the two functions (i) and (ii) is quite frequent, especially when the predicate already has an intensifier: 20. les français ont une idée des Portugais qui est vraiment totalement fausse [Corpaix, PORTUGA6 2,3] 21. ma palette (…) elle est vraiment très très pauvre en couleurs [Corpaix, PEINTURE 6,9) 22. il est vraiment très désagréable [Corpaix, CORPUS3, 3,12] (iii)

a use as discourse particle: in this case, vraiment functions at the sentence level and is separated prosodically from the rest of the sentence. It is no longer an epistemic modalizer modifying the predicate, but holds the whole utterance in its scope and the interpersonal relation involved: the speaker emphasizes his/her personal commitment towards the message, the sincerity of his/her beliefs and s/he invites the hearer to admit the truthfulness of his/her message. In final position, vraiment functions as a softener, giving a subjective quality to a previous more general statement:

23. parce que bon moi vraiment ça m'intéresse je veux dire j'ai envie de(…) [Corpaix, PORTUGA4 10,12] 24. enfin il y a des choses vraiment c'est aberrant et quand tu vois bon (…) [Corpaix, RENCONT 3] 25. donc euh + ils ont beaucoup de devoirs + vraiment L1 les très grands ils ont un + tas de devoirs [Corpaix, CORPUS3] 26. j'hésitais mais je crois qu' il faut le dire vraiment [Corpaix, FGARCIN 10,9] 27. mais sinon des problèmes quand même pas + pas vraiment [Corpaix, PORTUGA8 10,3] In dialogues, vraiment can also appear in an even more autonomous way in responses, giving or seeking validation of the previous message. Many other “nuances” can be conveyed such as hope, surprise, irony etc:

220 Pragmatic Markers in Contrast 28. L2 oui les enfants sont gentils L1 ah oui vraiment hein c'est L2 c'est rare [Corpaix, METIER2 14,11] 29. Q. Lisez-vous beaucoup de littérature de voyage? R. Non, pas vraiment, pas systématiquement ou alors si vous appelez “littérature de voyage” les sophistes de la Grèce antique ou Montaigne(…) [Le Monde, 1994] 30. “Une fois”, reprit-elle, “dans le Palais-Royal, je suis passé à côté de vous!” – “Vraiment?” Et il lui dit le bonheur qu'il avait eu en la retrouvant chez les Dambreuse. [Flaubert, L'éducation sentimentale] 1.3.

Core meaning and double movement of meaning extension (grammaticalization and pragmaticalization)

The whole spectrum of uses of vraiment can semantically best be described as a case of multiple polysemy starting from a core meaning (vraiment as “validation” operator), which has given rise to a double movement of meaning extension in relation to the syntactic level. From the core meaning of truth identifier at clause level, which is still the most frequent use in modern French, through a grammaticalization process involving a frequent use in highly constrained local contexts, vraiment becomes a mere intensifier. Through another movement in the opposite direction of disintegration and contextualisation, vraiment occupies a clause external position, undergoes semantic bleaching and functions as pragmatic discourse marker emhasizing the interpersonal commitment to the message. Each of those meanings can give rise to a certain number of side effects, as listed in figure 2 below:

Grammaticalization phrase level, local scope intensifier

Core meaning Clause level, predicative scope truth identifier

Pragmaticalization Utterance level, wide scope Interpersonal function

emphasizer subjectivity

planner, corrector

Figure 2: core meaning and meaning extensions

emphasizer, softener irony etc

Vraiment and Really in Contrast 221

1.4.

Lexicalization and clustering

An interesting characteristic of vraiment in spontaneous spoken language use10 is its tendency to cluster with other elements (adverbs or particles): this tendency concerns essentially its use as intensifier (31, 32) and its pragmatic use as discourse marker (33, 34, 35): 31. et celles-ci se vendent vraiment très très chères [Corpaix, Parfum 14,13] 32. une noire du Texas et qui chantait qui chantait vraiment super bien [Corpaix, CHRIS 3,14] 33. n'importe quelle époque mais pourvu parce que bon moi vraiment ça m' intéresse je veux dire j ' ai envie de [Corpaix, PORTUGA4 10,12] 34. L1 non L2 ah non pas du tout alors vraiment alors là pas du tout [Corpaix, PORTUGA5 4,11] 35. L1 ah oui oui mais vraiment ah oui oui bon alors ils ne sont pas bons mais [Corpaix, RENCONT 5,3] As a subjective modalizer (function (i)), vraiment is often followed by hesitation markers as in: 36. ben on peut dire que c'est vraiment euh c'est un monde à part [Corpaix, INFIRM 5,8] 37. -- vraiment je sais pas + c'est c'est pas ju- ju- juste [Corpaix, PORTUGA2 7,14] 38. (…) il y a certaines façons de parler + c'est vraiment euh subjectif + qui te plaisent ou te déplaisent” [Corpaix, PHILO 4,14] 39. et c'est vraiment euh c'est c'est ça ça vous obnubile quoi [Corpaix, ATTENTAT 9,3] The apparent paradox of a truth emphasizer followed immediately by a hesitation marker has already been pointed out by Halliday (1994: 340): “We only say we are certain when we are not. Only controversial propositions are in need of modalisation”. 1.5.

Methodological questions

From a methodological viewpoint, we are aware of the somewhat idealised nature of the scheme proposed in figure 1. (i) When compared to our written data, the analyses of the examples from spoken French have shown that the relation between function and position is much more arbitrary in this latter mode. There seem to be very few constraints on the position of vraiment in the sentence and no real link with the semantic 10

In the Triptic-Namur corpus we have investigated (see below), only two occurrences of clustering have been found (vraiment très).


classification proposed11. This has already been noted by Gezundhajt (2000: 277): “Du point de vue de la structure de l'énoncé, on note que pour que l'adverbe vraiment porte sur l'énonciation, il n'a pas besoin de précéder ou de suivre l'énoncé sur lequel il porte. Il peut parfaitement être intégré.” Vraiment as utterance modifier does not have to figure in initial or final position but can be perfectly integrated in the sentence, as is shown in the following examples: 40. Et bien je vais dire quelque chose vraiment de la rose [Corpaix, ANTHROP9,2] 41. Bon parce que c'est vraiment on voudrait pouvoir s'en passer [Corpaix, ANTHROP 7;5] 42. J'ai envie de retrouver vraiment mes racines [Corpaix, PORTUGA4 10,8] (ii) The scope of the adverb is often quite difficult to determine and without a prosodic analysis12, many examples resist disambiguation. The fact that discourse particles are oriented in two directions (backwards to the prior discourse and forwards to the upcoming message (cf. Aijmer 2002: 37), does not make the interpretation easier: 43. Je donc l'année du Bac j'ai dû vraiment travailler [Corpaix, PORTUGA2 4,9] (= j'ai dû vraiment /v/ vraiment travailler) 44. J'ai envie vraiment de connaître le Portugal [Corpaix, 5PORTUGA] (= j'ai envie vraiment /v/ de vraiment connaître) 45. C'est-à-dire qu'ils écrivaient vraiment ce qu'ils entendaient [Corpaix, STEPH 1,10] (= ils écrivaient vraiment/v/ ce qu'ils entendaient vraiment) Especially when combined with an opinion verb, vraiment seems to be subject to contradictory tendencies. On the one hand, it seems to have a predilection for taking the opinion verb as scope, even when it figures at some distance from the verb: 46. Et je pense pas qu'ils vraiment qu'ils partagent un amour dingue pour le Portugal [Corpaix, PORTUGA6 1,7] (je pense pas vraiment /v/ qu'ils partagent vraiment?) On the other hand we could sometimes speak of a “raising” of vraiment in the structural dependency of the verb, even when the real scope is another element of the sentence: 47. Je pense vraiment que nos intérêts, peut-être pas vitaux mais sans doute essentiels, sont en cause dans cette affaire [Le Monde 1994] (= je pense que nos intérêts sont vraiment en cause dans cette affaire)

11 12

In fact the privileged position is the medial position (cf. also Bilger 2003, Gezundhajt 2000 among others). To our knowledge, there is no corpus of spoken French with prosodic annotation available at this time.


48. Moi, j'ai vraiment trouvé que ils ils maîtrisaient la langue [Corpaix,1ACCENT, 9,7] (= j'ai trouvé qu'ils maîtrisaient vraiment la langue?)

2.

VRAIMENT /V/ REELLEMENT, VERITABLEMENT, EN REALITE, A VRAI DIRE

Although a thorough comparison of vraiment with other particles of the same semantic field cannot be drawn within the limits of this article, it is nevertheless worthwhile, in the light of the contrastive analysis that follows, to briefly situate vraiment in relation to its closest competitors: réellement, véritablement and à vrai dire. 2.1

Global frequency

A quick search taking into account three months of the press corpus (Le Monde, 1994) has revealed that the truth identifying modalizers are distributed as follows: Vraiment: Réellement: Véritablement: A vrai dire:

539 157 89 32

According to the statistics of the research centre for the TLF13, vraiment is 5 times as frequent as réellement in written French. In our sample of Le Monde it occurs 3.4 times more often. In the entire Triptic-Namur corpus14, 4.3 times, in its Debates subcorpus, 2.7. 2.2

Vraiment /v/ réellement

According to Danjou-Flaux, in her outstanding analysis of réellement (1982), vraiment can be considered as the nearest synonym of réellement15. She draws, however, the reader's attention to some restrictions on the discourse use of réellement, which seems less natural than vraiment when appearing in isolation as a response (49) or in discourse initial position (50): 49. A. Tu tiens tant à ta liberté? B. – Vraiment ?Réellement 50. Vraiment, tu n'as pas l'air de bonne humeur ce matin! ?Réellement, tu n'as pas l'air de bonne humeur ce matin! (Danjou-Flaux 1982: 133)

13

As mentioned by Sabourin & Chandioux (1977). Cf. note 15. 15 From the 35 adverbial tests elaborated by Sabourin & Chaudioux (1977), vraiment and réellement give identical results in 31 cases ( Danjou-Flaux, 1982: 136). 14

224 Pragmatic Markers in Contrast Réellement (protected by its syllabic structure?) shows a lesser tendency to be delexicalized and pragmaticalized: “C'est que vraiment fonctionne un peu comme une interjection (cf. Ben vraiment alors etc.), tandis que réellement reste un morphème très lexicalisé (…). C'est je crois la seule différence sémantique entre les deux adverbes: différence de degré et non de nature, qui manifeste la pertinence du concept de continuum en linguistique” (Danjou-Flaux 1982: 141). Gezundhajt (2000: 272) also argues the more lexical nature of réellement by contrasting the two adverbs in relation to the possibility of gradability: très réellement /v/ *très vraiment. She emphazises (2000: 273) the discursive and interpersonal function of vraiment as opposed to the more extralinguistic emphasis of réellement. 2.3

Vraiment /v/ véritablement

In contrast to vraiment, véritablement seems to be limited to the function of predicate modifier and cannot be used in an autonomous way: 51. Vraiment! Vous pourriez faire attention *Véritablement! Vous pourriez faire attention. (cf. Gezundhajt, 2000: 271) Like réellement, véritablement is essentially concerned with the propositional content rather than the discourse situation or the speaker's appreciation of the event, which is the domain of vraiment. 2.4

Réellement /v/ en réalité

En réalité, morphosemantically close to réellement, is functionally quite different: as an opposition marker, it is closely related to the previous discourse (cf. English actually) and is used to refute the proposed argument by virtue of a better knowledge of the world. In this anaphoric and ‘refuting’ function, en réalité is very close to en vérité and à vrai dire. In this last expression we note, however, the explicit reference to the locutionary aspects of the speech act. Vraiment Réellement En réalité A vrai dire

3. 3.1

refuting + +

anaphoric + +

autonomous uses + (+) -

VRAIMENT AND REALLY IN CONTRAST Really

In figure 3, we present a synthesis of the uses of really as we found them in the available linguistic literature (Simon-Vandenbergen 1988, Stenström 1986):


NO SEPARATING FEATURES Phrase level 'intensifier' 'planner' (speech specific function)

Clause level 'truth identifier' / 'emphasizer' [makes explicit the speaker's view that the statement being made is true]

SEPARATED Sentence level 'evaluator' [speaker's attitude to what he is saying] (rarely initial) 'softener' (clause final)

AUTONOMOUS Dialogue level Response to questions Comment to a previous utterance 're-opener' 'continuer' (speech specific functions)

Figure 3: functions of really

Really appears to be an almost perfect equivalent of vraiment, similar not only in its core meaning but also in its various extensions. In this sense the pair vraiment/really constitutes an exception to the widespread observation that “particles are very often highly idiosyncratic, 'untranslatable', in the sense that no exact equivalents can be found in other languages” (Wierzbicka 1991: 341). Following the terminology of Altenberg & Granger (2002: 22) vraiment and really can be described as a case of “overlapping polysemy”, the items in the two languages presenting the same meaning extensions. Bolinger (1972: 94–95) describes how really, among a number of other sentence adverbs has undergone a gradual change from truth identifier to intensifier. Quirk et al. (1972: 440) point to the ambivalence of really as emphasizer as to whether the emphasis is on the part or on the whole. The two particles are also very similar in terms of frequency: Stenström (1986: 152) stresses the frequency of really in spoken language. In Biber et al. (1999: 562), really figures with particularly high frequency in conversation (the most frequent epistemic adverb) and in fiction. Really is also the only adverbial marked in this grammar as semantically ambiguous between an epistemic reading and an interpretation as degree adverb. This similarity with vraiment holds for positional preferences: as for vraiment, the dominating position of really is “immediately after the operator or BE” (Stenström 1986: 154). Really, like its French counterpart, is also rare in initial position (Stenström 1986: 151). 3.2

Vraiment and really in a translation corpus

In order to contrast our two particles, we have analysed their use in a bidirectional translation corpus: the Triptic-Namur corpus, compiled by Hans Paulussen.16 The following table gives 16 The Triptic-Namur Corpus is a trilingual parallel computer corpus which was initially compiled for the analysis of prepositions and particles in English, Dutch and French, as part of a doctoral thesis (Paulussen 1999). This translation corpus is aligned at paragraph level and contains almost 2,000,000 words. It is divided into two subcorpora i.e. fiction and non-fiction, each containing approximately 1,000,000 words. The non-fiction subcorpus consists of extracts from the UNESCO Courier and from the verbatim transcription of the Debates of the European Parliament. The fiction subcorpus consists of extracts of approximately 25,000 words each from 12 original fiction texts. These extracts are taken from four original English texts and the same number of

226 Pragmatic Markers in Contrast us the absolute frequency of each particle and of the closest synonym in both languages in the debate part and in the fiction part of the corpus. Triptic-Namur corpus Vraiment Réellement Really Truly

Debates 89 39 123 9

Fiction 122 10 166 24

Really has the highest frequency in both corpora: in the debates its frequency is equivalent to the combined frequency of vraiment and réellement. Truly is the least frequent equivalent, at least in the debates. In the fiction part, really and vraiment have the same frequency relative to each other as they do in the debates; réellement and truly reveal opposite tendencies. The translations are instructive in many ways: the following tables show the target language alternatives corresponding to vraiment (table 1) and really (table 2) as sources, as well as the different sources that are translated by vraiment and really respectively:

Table 1: vraiment as source and as translation Vraiment as a source vraiment (19)

translations into English really: 7 : 5 other (stronger) verb: 2 properly:1 fully: 1 seriously: 1 extremely: 1 sincerely: 1

Vraiment as a translation vraiment (51)

English source really : 23 : 6 actually: 5 “do”: 3 properly: 2 indeed: 2 very: 2 hardly (+neg): 1 fully: 1 surely: 1 little: 1 clearly: 1 truly: 1 seriously: 1 only: 1

original French and Dutch texts. Twelve original extracts (selected from books published between 1982 and 1992) translated into two languages give a total of 36 extracts. Our thanks are due to Hans Paulussen for making his corpus available and for his help in retrieving data.

Vraiment and Really in Contrast 227 Table 2: really as source and as translation really as a source really (68)

translations into French vraiment: 22 (+ vrai: 1) réellement: 12 (+ réel: 2) : 9 véritablement: 4 (+ véritable: 2) en fait: 5 en réalité: 2 franchement: 1 totalement: 1 parfaitement: 1 à dire vrai: 1 effectivement: 1 à y regarder de plus près: 1 absolument: 1 bien: 1 “condit.”: 1

Really as a translation really (23)

French source vraiment: 8 (+ vrai:1) : 4 véritablement: 3 (+véritable:1) réellement: 2 plutôt: 1 tout à fait: 1 effectivement: 1 en effet: 1

Really is the most frequent translation of vraiment, although there are a few alternatives, most notably the zero translation (frequent in more formal translations cf. Simon-Vandenbergen & Aijmer 2002/2003). There are also two occurrences of a stronger lexical verb and properly, fully, seriously, extremely and sincerely occur once each, highlighting some specific meaning aspect. Vraiment as translation covers a wide range of sources: really is the privileged source, but many other candidates appear from a broader semantic field (emphasizing particles and intensifiers, but also less obvious equivalents). Vraiment is the prototypical equivalent of really (in 30% of the cases), followed by its closest synonyms réellement, véritablement, en fait and en réalité. Nine other alternatives occur only once each within the limits of our corpus. Really occurs as the translation of vraiment in 30% of the cases. Its other translation sources have roughly the same meaning (réellement, véritablement, en fait, en réalité… ). From these tables, vraiment and really appear as prototypical translation equivalents and the correspondence holds in both directions. Vraiment however seems to cover a broader semantic field than its English equivalent and appears as the translation of other discourse particles (actually, indeed etc.).

4.

VRAIMENT, REALLY AND THEIR DUTCH COUNTERPARTS

We have also examined which Dutch counterparts appear in the Debates subcorpus of the Triptic-Namur corpus. One important remark has to be made here: we have only considered the Dutch material in relation to vraiment and really. In other words, we have analysed the occurrences of the Dutch sources translated by vraiment and really as well as the Dutch translations of vraiment and really17. On the other hand, a comparison of the global frequency 17

This means that the subcorpus still contains other occurrences of the Dutch words listed in the tables 3 and 4 below.

228 Pragmatic Markers in Contrast of echt and werkelijk, which, on a morpho-semantic basis, could be seen as the closest equivalents of vraiment/réellement and really/truly, is made difficult due to the morphological ambiguity between the adverbial and adjectival functions of these forms. 4.1

Vraiment and its Dutch equivalents

Table 3 gives an overview of the Dutch equivalents of vraiment chosen by the translators of the debates and of the Dutch sources translated by vraiment.

Table 3: vraiment as source and as translation vraiment as source

translations into Dutch

vraiment as translation

Dutch sources

19

echt (8) werkelijk (5) (4) eigenlijk (1) juist (1)

19

echt (7) wel (4) daadwerkelijk (2) werkelijk (1) heel erg (1) nog eens (1) bepaald (1) maar niet (1) in wezen (1)

It shows that echt is the most frequent translation of vraiment and is also its privileged source. Other frequent translations are werkelijk and the zero translation, with eigenlijk and juist occurring only once each. When comparing the frequency of the various Dutch translations of vraiment (19)18 to that of the English translations (table 1), we observe that echt (8) and really (7) are about equally frequent. Whereas really was followed by the zero translation (5), echt is followed by its close lexical equivalent werkelijk (5). A more detailed analysis shows that in 5 cases really and echt also appear as translations of the same source. Secondly, we observe that vraiment appears 19 times as the translation of Dutch sources, of which echt (7) and wel (4)19 are the most frequent. Werkelijk occurs only once as source, daadwerkelijk twice. Five other Dutch lexical phrases occur only once as sources for vraiment. 4.2.

Really and its Dutch equivalents

In table 4 we find the Dutch sources and translations corresponding to really.

18

The numbers between brackets indicate the number of occurrences. Abraham (1984) offers a description of the different meanings and uses of wel, in relation to its German counterparts. 19

Vraiment and Really in Contrast 229 Table 4: really as source and as translation really as source 68

translations into Dutch (23) echt (18) werkelijk (18) eigenlijk (4, including 2 adj) in feite (2) waarlijk (1) maar (1) in wezen (1)

Dutch sources really as translation (6) 32 echt (4 + 1adj) eigenlijk (4) wel (4) (1x “nou wel”, 1x “toch wel”) werkelijk (3) daadwerkelijk (2) nog (1) enkele (geen enkele) (1) inderdaad (1) in feite (1) uiteindelijk + als zodanig (1) heel erg (1) toch (1) in gemoede (1)

The closest Dutch lexical equivalents of really are echt en werkelijk, occurring both 18 times as translations of really and five and three times respectively as sources. However, as sources, both are preceded by the “zero source”, with eigenlijk and wel also featuring equally prominently (see also Aijmer and Simon-Vandenbergen 2004 on Dutch translations and sources of really). More generally we see that really – as source and as translation – covers a rather broad range of Dutch equivalents some having a quite specific meaning, others functioning as mere emphasizers and intensifiers. The most striking finding from table 4 is the fact that in 33.8% of all cases the translators have chosen not to translate really. In some cases the translation deviates from the source and the global meaning of the sentence is paraphrased , as in the following example: 52. E. You cannot really say that a report concerned only with a small aspect of one part of Community programme somewhere should be treated in the same way as a report that encompasses a very wide-ranging aspect of policy. [Source] D. Er bestaat een duidelijk verschil tussen een verslag dat slechts een klein aspect van een bepaald onderdeel van een communautair programma betreft en een verslag dat een verstrekkend beleidsaspect omvat. [English gloss: There is a clear difference between a report concerned only with a small aspect of one part of Community programme and a report that encompasses a very wide-ranging aspect of policy.] 20 F. Vous ne pouvez pas vraiment dire qu'un rapport portant seulement sur un aspect limité d'une partie d'un programme communautaire doit être traité de la même manière qu'un rapport qui porte sur un aspect de la politique ayant une large portée.

20

The examples preceded by “E.” are the original ones we have found in the translation corpus; those preceded by the mention “English gloss” are more literal translations of the Dutch examples, respecting the scope, place and sometimes slightly deviating meaning of the Dutch example in the translation corpus. If the former type of examples are not different from the Dutch ones with respect to scope, place and meaning, then no “English gloss” is given.


On the other hand, we observe that in 6 examples really has been inserted in the translation of Dutch sources containing no lexical phrase equivalent to really: 53. D. U lacht, maar het is cynisch! [Source] [English gloss: You are laughing, but it is cynical!] F. Vous riez, mais c’est cynique! E. You may laugh, but it really is ironic! When we compare the frequency of the Dutch and French translations of really (68), we notice that the zero translation is a lot more prevalent in Dutch (23) than in French (9). Furthermore we observe that vraiment (22 + 1) is slightly more frequent than echt (18) and that réellement (12 + 1) is somewhat less frequent than werkelijk (18). In summary, vraiment is translated in 42 % of the examples by echt and in 26 % of the cases by werkelijk. The most frequent translations of really are also echt (26 %) and werkelijk (26 %), but these are preceded by the zero translation (34 %). Really seems to cover a broader semantic field in Dutch than vraiment, whereas vraiment covers a broader semantic field in English than really in French (cf. above). 4.3.

Translation of multifunctional particles

A closer analysis of the translation data is also very revealing for some interpretation problems caused by the polyvalent and ambiguous nature of our particles. In the following examples, the translator modifies the scope and/or position of the particle (54,56) or uses a particle with a particular semantic nuance (55): 54. E. Surely, this is to ignore Parliament (…)” [Source] F. L'on fait vraiment peu de cas du Parlement (…) D. Dat is beslist miskenning van het Parlement (…) [English gloss: This is definitely an underestimation of Parliament.] 55. E. (…) without knowing where we are actually going to end up” [Source] F. (…) sans savoir vraiment où cela nous mènera.” D. (…) zonder dat we weten waar wij uiteindelijk zullen belanden. [English gloss: (…) without knowing where we finally are going to end up.] 56. E. There is really no difference in principle between Parliament and Mrs Scrivener, (…).” [Source] F. Il n'y a en principe pas de véritable divergence de vues entre le Parlement et Mme Scrivener et (…). D. Er bestaat werkelijk beginselmatig geen verschil tussen het Parlement en mevrouw Scrivener en ik (…). [English gloss: There is really in principle no difference between Parliament and Mrs Scrivener (…).” The following example is quite complex, involving two focusses and different lexical means for the three languages involved:


57. D.. (…) maar ik vraag me heel erg af of dat nu op de beste manier gebeurt. [Source] [English gloss: (…) but I wonder very much if that is being done in the best possible way.] E. (…) but I really wonder wether it is going about it in the most effective way. F. (…) mais je me demande avec beaucoup de perplexité si la méthode qu'elle applique est vraiment la meilleure. In (58) a negative sentence is translated in French by a rhetorical question containing the emphasizer vraiment (and appealing for a negative confirmation). The Dutch translation is closer to the source: 58. E. But he will not need to be reminded of the justified criticism from the transport industry (…) [Source] F. Mais est-il vraiment besoin de lui rappeler les critiques justifiées qui émanent de l'industrie du transport (…)?” D. Maar hij en de overige commissarissen, als zij luisteren, behoeven niet te worden herinnerd aan (…) [English gloss: But he and the other commissioners, if they are listening, will not need to be reminded of (…)] In (59) there is an interesting combination of two particles from the same semantic field: the translation of really by vraiment and truly by sincèrement confirms the strong correspondence and similarity of our particles. The Dutch variant of the utterance, however, contains no real equivalents: 59. E. But really and truly, we must have answers to questions (…) [Source] F. Mais vraiment, sincèrement, il nous faut des réponses à nos questions (…) D. Wij verwachten echter antwoord op onze vragen (…) [English gloss : But we expect an answer to our questions (…)] 4.4.

Core meaning, grammaticalization and pragmaticalization of the Dutch counterparts

To our knowledge, there are no thorough analyses of echt21 and/or werkelijk that would permit us to compare the core meaning and the meaning extensions of these two adverbs to those of vraiment and really. In addition to this, our corpus is too restricted22 for us to make any solid claims in this domain. However, we can present the results of our analyses as valid within our limited subcorpus.

21 For a description of the behaviour of echt in contrast with gewoon, see Vanden Eynde (2003), according to whom “In using the hedge echt, the speaker focuses (…) on an intercategorial contrast. The speaker, in effect, narrows down the category under consideration.” (Vanden Eynde 2003:441) This use corresponds to the use of vraiment described in 1.2 (ii) above (see example 12). 22 Containing only 35 examples of echt and 27 examples of werkelijk, all from the same rather specific Debate subcorpus.


We observe that echt functions in about 21 cases at the clause level, modalizing the predicate as a truth identifier or as an emphasizer. 60. Alleen op die manier kan blijken welke landen en regio's de hormonenbelangen echt bestrijden en welke niet. [E. Only in this way it will become apparent which countries and regions are really resisting the hormone lobby and which are not.] 61. Het maakt duidelijk dat het rijke Europa echt begaan is met het lot van de arme landen. [E. It shows that affluent Europe really cares about those who have very little indeed.] The distinction between these two functions is especially hard to make when opinion verbs are involved (6 examples): 62. Ik meen echt dat het Parlement beter verdient, te meer daar het debat reeds twee keer was verdaagd. [E. I really do believe that this House deserves better, particularly as the debate has been put off twice already.]

In seven examples echt is used at the phrase level, modifying a gradable adjective or a negation particle and functioning as intensifier: 63. We staan dus voor twee echt belangrijke vraagstukken (…) [E. We have to consider two really important questions.] 64. (…) lijkt echt technisch en (…) [E. (…) sounds very technical and (…)] 65. Ik vind het echt niet juist dat wij het verslag bespreken zoals het gisteren in de commissie is goedgekeurd, omdat (…).” [English gloss: I feel that it is really not appropriate that we should look at the report as it was adopted in the committee yesterday because (…)] [E. I really feel that it is inappropriate that we should look at the report as it was adopted in the committee yesterday because (…)] We may tentatively conclude that the core meaning of vraiment and really as well as their more grammaticalized use as local scope intensifier can also be verified in the case of echt. On the other hand, in our corpus, we found no clear examples of a pragmaticalized use23. As for werkelijk, in the 29 examples, this particle functions at the clause level, either as a reality identifier or as an emphasizer: 66. (…) om deze dodenmars werkelijk te beëindigen. [E. To actually stop that death march and (…).]

23

In autonomous uses, as question or response, a combination of echt and waar (echt waar) seems to be most commonly used.

Vraiment and Really in Contrast 233 67. Het is niet zo dat we de zaken werkelijk verbeteren. [English gloss: It is not so that we are really improving things.] [E. We are not really improving things.] But even in its more emphasizing uses, werkelijk keeps a certain nuance of reality marker. It is also striking that in 8 of the examples, werkelijk appears in the identifying or qualifying structure “X is werkelijk Y”. 68. (…) en dat is werkelijk een schandaal, een ramp. [English gloss: (…) and that is really a scandal, a disaster.] [E. (…) and that is a scandal and a disaster.] 69. De leden die zich hierover zorgen maken, verzeker ik dat het werkelijk vrij gevaarlijk is af te wijken van de NUTS II-grondslag. [E. I assure Members who are concerned about this that it is really rather dangerous to get away from the NUTS II basis.]

It seems that werkelijk has a rather strong core meaning comparable to that of echt. Our limited set of examples offers no transparent indications as to the existence of meaning extensions (grammaticalization and pragmaticalization) similar to those existing for vraiment, really and, to a lesser degree, echt.

5.

VRAIMENT, REALLY, ECHT AND WERKELIJK.

Taking into account the totality of our data, we would like to present the following conclusions: (1) vraiment and really appear to be very similar and can be considered as cases of overlapping polysemy: they share the core meaning as well as the various sideeffects. They appear also as prototypical translation equivalents and the correspondence holds in both directions. (2) the closest competitor of vraiment, namely réellement, although a morphosemantic cognate of really, appears much less frequently in the French data, as well as in the translation corpus. It is especially less used in the more pragmaticalized and grammaticalized uses. On the other hand, the closest competitor of really, and the semantic cognate of vraiment, namely truly, is even less frequent than the French réellement. (3) the Dutch situation is an interesting one: two elements can claim the status of prototypical translation equivalents: echt, from the semantic field of truth, and werkelijk from the semantic field of reality. Both seem to occupy, on an almost equal basis, the translation domain of vraiment and really. They do not, however, seem to share with vraiment and really the more pragmaticalized meaning extensions. The situation can be roughly represented by figure 4:


French

vraiment

English

really

Dutch

echt

werkelijk

réellement truly

Figure 4: vraiment, really, echt and werkelijk

REFERENCES Abraham, W. (1984). De betekenis en de functie van het Nederlandse wel – een vergelijking met het Duits, pp. 19–46. Antwerp Papers in Linguistics 35, Universiteit Antwerpen, Aijmer, K. (2002). English Discourse Particles. Evidence from a Corpus. John Benjamins, Amsterdam / Philadelphia. Aijmer, K. and A.-M. Simon-Vandenbergen (2004). A model and a methodology for the study of pragmatic markers: the semantic field of expectation. Journal of Pragmatics 36, 1781–1805. Altenberg, B. and S. Granger (2002). Lexis in Contrast. Corpus-based Approaches, Studies in Corpus Linguistics 7, John Benjamins, Amsterdam / Philadelphia. Biber, D., S. Johansson, G. Leech, S. Conrad and E. Finegan (1999). Longman Grammar of Spoken and Written English. Longman, London. Bilger, M. (2000), Autour du projet “Corpus de référence du français parlé”. In: Le français de référence (M. Francard et alii, eds), Cahiers de l’institut linguistique de Louvain 26 (1–4). Bilger, M. (2003), Quelques données sur les adverbes en –ment dans le Corpus de Référence du français parlé, in Recherches sur le français parlé 18, pp. 1–19. Université de Provence. Bolinger, D. (1972). Degree Words. Mouton: The Hague, Danjou-Flaux, N. (1982). Réellement et en réalité: données lexicographiques et description sémantique. In: Adverbes en –ment, manière, discours (N. Danjou-Flaux and M.N. Gary-Prieur, eds), pp. 105–150. Lexique 1, Lille, Presses Universitaires de Lille. De Cooman, B. (2001). Une description des différents emplois de vraiment, dans le cadre des tentatives de classification des adverbiaux, Master thesis, University of Ghent.

Vraiment and Really in Contrast 235 Gezundhajt, H. (2000). Adverbes en –ment et opérations énonciatives. Peter Lang, Bern / Berlin / Bruxelles / Frankfurt. Guimier, C. l. (1988). Syntaxe de l'adverbe anglais, Presses de l'Université de Lille, Lille. Halliday, M.A.K. (1994). An Introduction to Functional Grammar, second edition. Edward Arnold, London, Paulussen, H. (1999). A Corpus-based Contrastive Analysis of English on/up, Dutch op and French sur within a Cognitive Framework. PhD dissertation. University of Ghent. Quirk, R., S. Greenbaum, G. Leech and J. Svartvik (1972). A Grammar of Contemporary English. Longman, London, Sabourin, C. and J. Chandioux (1977). L'adverbe français. Essai de catégorisation (classification statistique des adverbes en -ment), Editions Jean-Favard, Paris. Simon-Vandenbergen, A.-M. (1988). What really really means in casual conversation and in political interviews, Linguistica Antverpiensia 22, 206–225. Simon-Vandenbergen, A.-M. and K. Aijmer (2002/2003). The expectation marker of course in a cross-linguistic perspective. Languages in Contrast 4/1, 13–43. Stenström, A.-B, (1986). What does really really do? Strategies in speech and writing. In: English in Speech and Writing. A Symposium (G. Tottie and I. Bäcklund, eds), pp. 149–163. Uppsala, Almqvist & Wiksell, Vanden Eynde, C. (2003). The Dutch hedges echt and gewoon: Markers of prototypicality?. In: Cognitive Approaches to Lexical Semantics, (H. Cuyckens, R. Dirven and J.R.Taylor, eds), pp. 427–446. Mouton De Gruyter, Berlin/New York. Wierzbicka A. (1991), Cross-cultural Pragmatics: The Semantics of Human Interaction, Mouton De Gruyter, Berlin/NewYork.



CONSTRUCTIONS IN CROSS-LANGUAGE RESEARCH: VERBS AS PRAGMATIC PARTICLES IN SOLV Jan-Ola Östman, University of Helsinki

1.

PREAMBLE

The search for an adequate standard of comparison, a tertium comparationis, in the field of cross-cultural research on pragmatic particles has always had top priority. In a very real sense, the closer we get to finding the appropriate metalanguage, the closer we get to understanding the function of pragmatic particles. As in all linguistic studies, the issue has been approached in two complementary manners, either starting from the form (in a wide sense) of the particles, or starting from the pragmatic function they have in discourse. In both approaches, similarities with other elements and functions have been brought to bear on the argumentation: prosody and modals on the one hand, and sequential organization and polyphony on the other – to name but a few. In this study I argue for a view of pragmatic particles as constructions (in the sense of Fried & Östman, 2004), where the notion of construction is not secondary to other means of categorization (part of speech, distribution, information structuring, grammatical or semantic role; genre, sequential organization, sense-negotiation), but is the primary means of categorization and thus for understanding their pragmatic function in language in general. This view implies that cross-language comparison of any aspect of language is only possible at the level of constructions (cf. Croft, 2005). This is the backdrop of the views brought forth in this study. The issues mentioned above are not argued for in the abstract, but are illustrated through a detailed study of some pragmatic particle phenomena in one particular language.

2.

THE ISSUE

The study is done in the framework of a system for describing and formalizing pragmatic aspects of language that I have been developing during the last thirty years (cf. Östman, 1977, 1979, 1981a, 1981b, 1982a, 1982b, 1986, 1987a, 1995a; Östman & Wårvik, 1994; Fried & Östman, 2005). In this view, pragmatics deals with implicit aspects of how speakers and interactants anchor their messages and attitudes and relations to others and to their cultural context. There are three parameters for explicating the implicitness of such anchoring, relating to culture, interaction, and the personal-affective. (For details, see Östman, 1981b, 1986, 1987c, 1995a.) This view is tied to a constructional approach to pragmatic phenomena: linguistic manifestations are central to a systematic study of pragmatics, and the analysis of


linguistic manifestations in general has proven to be especially fruitful within models like Construction Grammar (e.g., Fried & Östman, 2004). The empirical data for the study comes from Solv, a dialect of Finland Swedish. I will first give a brief overview of how the functions of particles in Solv can be elucidated in terms of the notion of implicit anchoring. I will then briefly discuss the use and function of Question particles in Solv, as linguistic elements being typically ambivalent with respect to cognitively codified meaning and socio-culturally codified function. The specific focus of this study is a group of verbs which have taken on particle-like characteristics. The verbs under discussion are vaa, jäär, and mono. A description in traditional terms might describe these as modal auxiliaries. However, they also occur together with other (modal) auxiliaries, as in Va jäära dö iss ga hejm?, which would translate as ‘Why did you bother to go home?’, except that jäära is then left untranslated. In this case, a more appropriate translation would be something like ‘Why on earth did you bother to go home?’. That is, the finite verb jäära could here best be translated as if it was a pragmatic particle. Theoretically, the study also addresses the question of what elements of language can take on the functions that we typically consider part of the domain of pragmatic particles.

3.

SETTING THE SCENE

Solv – the language – is genetically a Swedish dialect spoken in the village Solf on the west coast of Finland, some 400 kilometers north-west of Helsinki. Solf is a village with some 1,000 inhabitants, and now part of the Korsholm county. Solv is one of over 80 traditionally recognized dialects of Swedish in Finland.1 In this study I will continue to use the term ‘pragmatic particle’, despite the fact that other terms like ‘discourse marker’ and ‘contextualization cue’ have become more frequent in the literature. What were called pragmatic particles in the 1970s are to my mind still pragmatic particles; ‘particles’ because they are not declined or conjugated, and ‘pragmatic’ because their implicit functions in discourse are primarily at issue. They are ‘particles’ rather than ‘markers’ or ‘cues’ as they have been crystallized into verbal manifestations – as opposed to gestures, pausing, and intonation contours.2 (See Östman, 1995a.) The term that is not in my title, but could have been there is ‘modal’. The kinds of ‘particles’ I want to discuss in this study are in a sense ‘modal particles’, except that I do not consider it informative to suggest a priori that their meanings/functions need to be tied to epistemic, deontic or dynamic modality. However, the phenomena I will be discussing are in a sense modal particles proper: verbs, and in particular modal verbs or auxiliaries that seem to 1 Finland has two national languages, Finnish and Swedish, of which Finnish is the majority language and Swedish is the mother tongue of close to 6% (300,000) of the population; the population of Finland is just over five million. In addition, the Finnish constitution makes special provisions for speakers of Sami, Romany, and Signed languages; and there are furthermore some 100,000 inhabitants in Finland with mother tongues other than these. I am a native speaker of Solv. 2 Pragmatic particles can nevertheless have the ‘same’ functions as gestures and prosodic contours, but I consider it significant to make a distinction between what a language community has crystallized (and thus potentially codified) verbally, and what it has not.

Pragmatic Particles as Constructions 239

function as pragmatic particles. They are dynamic and pragmatic, but they retain their finiteness as predicates. In Östman (1995a), I have objected to the prevalent view that particles are multifunctional, since this view suggests that any particle can, in principle, have any number of functions. As such, the idea that anything can mean anything is not to be frowned upon; I can call a chair a tree and talk about that tree throughout a whole discourse. To me, this possibility of words having a flexible semantics indicates and underlines the point that, in the last resort, misunderstanding3 is what we start out with in any communicative interaction; to reach mutual understanding is a matter of negotiation, and successful interaction can thus be regarded as a special case of communication. Context is always important – language and context/culture, however defined, always go together. However, it does not follow that anything can mean virtually anything. Thus, pragmatic particles are indeed multifunctional in the very real sense that interaction – and thus language – is built on negotiation and sensecreation, but in this sense, all elements of language are multifunctional; particles are not more multifunctional than other elements in language. I believe there is still some sense in trying to find the function (the ‘core’ function, if you want) of any particular particle. Despite the undisputed occasional instances of homonymy and a wealth of polysemy in language, the starting point for understanding codification in language is the particular relationship that holds between a form and a meaning, where – following Bolinger (1977) – we can work on the assumption that one form corresponds to one meaning. In this view, a pragmatic particle is a crystallization of a particular meaning or function. Having said that – but only after having said that – a particle can be seen as multifunctional in the sense that it has, or can have, a vast number of meanings or functions, as long as you work out a system for how to represent – even formalize – something that looks multifunctional. If such a system is developed, it is possible to specify, even define, multifunctionality and provide constraints on the extent to which the meaning and function of a pragmatic particle can vary. Thus, for me, ‘core’ function is not one explicit meaning, but an array of functions. In previous studies, I have attempted precisely to develop a system for how to talk about pragmatic aspects of language and about the function of language and pragmatics in general. Section 4 presents this system.

4.

PRAGMATICS AS IMPLICIT ANCHORING – PIA

In approaching linguistic function as a dynamic object of analysis, it is crucial to make a heuristic, but basic and systematic distinction between meaning and function. The manner I have found to be most operational when doing this is to distinguish between meaning as the explicit in language (what has been codified: the lexical, propositional, semantic, and discourse-level ‘meaning’), and function as the implicit in language (what takes place ‘between the lines’ of what one says: the implicated, and – primarily – aspects that the speaker is not accountable for propositionally). In order to systematize or at least 3

Or rather, ‘understanding differently’.


systematically approach the implicit in language, one has to accept that such an endeavor is akin to searching for ‘true understanding’ even though one knows one will never find it; the important thing is to go in the ‘right’ direction in some way.4 Based on empirical and ethnographic research, I have suggested that a division into three types of restraints is operational and useful in the study of language function. The hypothesis is that speakers in general make interpretations in relation to (i) their cultural coherence, their tradition and history, the society they live in, and its institutions; (ii) the interactive restraints, the conversations and norms of politeness and tact that they have to take into account when they are in interaction with other speakers; and (iii) the constraints on emotions, feelings and opinions, on the expression of affect and attitudes, and the prejudices that surround them as interactants and speakers. For these three patterns of constraints – parameters – I use the abbreviations C (for ‘coherence’); P (for ‘politeness’); and, I (for ‘involvement’), respectively. Together, these parameters constitute the backbone of the model of language function known as PIA – Pragmatics as Implicit Anchoring. That is, rather than using the same types of categories that are used to describe structural and semantic aspects of language, I take these three parameters as the basis on which to build a systematic account of language function. Pragmatic particles, as well as any other contextual markers or cues, are then seen as pragmatic crystallizations, where the word ‘pragmatic’ is to be understood precisely in relation to these three parameters. Expressions are ‘anchored’ to C, P, and I in the sense that these three parameters constrain the use of linguistic expression; expressions are not, however, anchored in a static fashion, but contain dynamic cues that indicate how they are to be interpreted and understood. During the last thirty years, there has been a wide-ranging debate within Conversation Analysis about the extent to which one can take any categories for granted when attempting to describe a linguistic phenomenon in a particular language. Within dialogical and ethnomethodological research, solid and well-founded arguments have been advanced in support of the view that linguistic categories have to emerge from the material that one is studying and analyzing. That is, one should not take categories like ‘gerundive’ and ‘conjunctive’ as givens and try to find them in all languages, but the language itself, its structure and function, suggest which categories are important for that particular language, on its own premises. As a principle, I accept this. At the same time, I submit that we have to be able to take something for granted. What we can take for granted may be only proposed general human properties, but if we do not investigate what these possible aspects are, we can easily be blinded and tricked by our own material and our own superficial objectivity. I thus work on the assumption that there are three domains in the lives of human beings, and even that we ourselves have some notion that these are three different aspects on life. These three domains are (i) society-culture, (ii) interaction-politeness, and (iii) emotions-attitudes. These are the three fundamental corner stones in the PIA way of viewing the function of language: C, P, 4 ‘Going in the right direction’ here means that one needs to collaborate with scholars and researchers from other fields and adapt – discursively and in a negotiating manner – to their results and ponderings (or, minimally, take into account what they are doing); i.e., the direction indicated by the collected knowledge from one’s own field and from other fields of research.


and I, respectively. And thus, these are also the three parameters with whose help we get at the implicit in language, i.e., the primary object of study for pragmatics. This way of thinking implies that codification of meaning on the syntactico-semantic level – i.e. that chair prototypically means ‘chair’ and not ‘tree’ – de facto constitutes a fourth parameter on language and behavior, the parameter that linguists traditionally have paid most attention to in detailed analyses of language. The abbreviation ‘S/S’ for ‘syntax-semantics’ is used here to signify ‘codification’. What all of this amounts to is the contention that when we study the function of language, we can presume that we have four main parameters/domains to use as tools5 when we look at concrete texts and discourses.

5.

PRAGMATIC PARTICLES IN SOLV

An initial attempt at classifying the pragmatic particles in Solv with the help of the four parameters established, suggests the following. Particles whose usage seems primarily to be constrained by cultural coherence I will call ‘Discourse particles’. C: Discourse particles Va sko ja sej, ......, annos Hödö (+ vocative) Höni; Sidö/Sydö; Sini Kaar! Va sko ja sej is a complex turn-taking particle that typically introduces a brand new topic; annos is a turn-final, turn-giving particle; hödö (and its variants höni, sidö, sydö and sini – depending on the addressee) potentially followed by a vocative is an attention getter; and kaar is akin to English man or blood, primarily indicating male bonding. Particles that are primarily constrained by interaction, rules of politeness, and which are typically sequentially based, I will call ‘Dialogical particles’. Some examples from Solv are given below. P: Dialogical particles jaa, hmm, <jyy … å / men… ..., så he så [pronouns] The most typical dialogical particles are the backchanneling supports, jaa, hmm, <jyy6. Another frequent type of dialogical particles are the turn-final å and men plus pause, which

5

I.e., as major ‘attributes’ in a constructional approach to the formalization of linguistic findings; see below.


function similarly to English turn-final and and but, on the one hand signaling (an illusion) that there is more to add, but at the same time creating a transition-relevance-place with the pause that follows. The turn-final så he så is more definitely turn-ending, sometimes coming close to a coda in a narrative, but also corresponding to Finland Swedish att sånt and Finnish ettäs sillai – in English something like and stuff. A further type of dialogical marker is the frequent use of turn-final pronouns; these are close to grammaticized when they are repetitions of the subject of a clause, but they can also function in a turn-giving manner. Particles that are primarily constrained by the expression of attitudes and emotions I will call ‘Affective particles’. Some examples follow. I: Affective particles Hyvvadå Nämen / Nää (va segä dö) ... Hyvvadå is an expression of disgust; nämen or nää (va segä dö) typically express an attitude of incredulity. In most of the characterizations of the pragmatic function of the particles above, I have qualified the description with words like ‘primarily’ or ‘typically’. This is not merely in order to safeguard the results of my observations and analyses, but to point to the prototypical nature of such ascribed functions. Another – but similar – problem with these kinds of classifications is that many of the particles mentioned are not particularly ‘implicit’. If you say Hyvvadå, it is clear that you are openly expressing your negative opinion or attitude towards something. Hyvvadå is thus a partly codified – a semi-codified – expression. The general observation here is that the clearer a particle could be classified under only one of the parameters, the more codified it has become, and the closer it is to the analyst’s semantic sphere of investigation. Many particles, seemingly the most frequently used particles, cannot easily be specified or categorized under only one of the three pragmatic domains; their use is constrained by a complexity of factors. This is particularly the case with particles that have vaguely modal uses (both epistemic and deontic), which are typically found turn-medially or even utteranceinternally. However, their function is (at least in some loose sense) responsive and thus sequentially constrained. I here call them ‘Utterance particles’7 and indicate their level of pragmatic codification as ambivalent.8 6 The symbol ‘ role finite frame JÄÄR

Figure 4. Characteristics of the Solv verbal/particle jäär – revised version

19

A fourth function, that of being a pro-verb, might be added to this characterization – cf. A dö jöft hondin? – Jaa, ja joor he (‘Have you helped the dog? – Yes, I did that.’). And we can even find a potential fifth function of jäär if we distinguish its meaning of ‘do’ from its potential to mean ‘make’. In standard Swedish, göra means both ‘do’ and ‘make’, but in the dialects, the ‘make’-meaning is not typically associated with this verb; thus, I have not dealt with it here. In Solv, laag is the typical verb for ‘make’.


8.

CONCLUSIONS

In this study I have wanted to stress the importance of being systematic in research on pragmatic particles. I have suggested that the insight offered by the kind of formalism used in Construction Grammar is a good alternative also for the purpose of capturing pragmatic restraints in language. I have also wanted to suggest how a detailed and systematic account of the pragmatic factors of verbs and modal auxiliaries can reveal that some ‘verbals’ are closer to being pragmatic particles than to being verbs or modal auxiliaries. The data I have used come from Solv, a dialect of Swedish, and the study has adopted the view that dialects are languages, i.e., have their own systematicity, and (sub-systems within dialects also) need to be analyzed as such, in order to get a deeper understanding of the typological possibilities in language. This way of approaching dialects will, in turn, greatly enhance the possibility for adequate crosslinguistic comparison. I have suggested that constructions function well as tools in cross-language and contrastive linguistic research, and that this also applies to studies on particles in contrast. What I have not done in this study is to take up for discussion why Finland-Swedish dialects like the Solv dialect have developed a complex system of modals and particle-verbals which is lacking in standard (Finland) Swedish. One answer that suggests itself is the close contact of the Swedish dialects to Finnish, which has a rich inflectional verbal morphology that can express many more modalities than what we typically find in Germanic languages. The similarities between Solv and Finnish (dialects) are not primarily to be found on the surface, but in the pragmatics of the two languages (cf., e.g., the comparison of the pragmatics of demonstrative adverbs in Solv and in Finnish as discussed in Östman, 1995b). One point of reference here is no doubt the study by Kiuru (1977) on verbs in Finnish that require a negative complement – cf. Finnish viitsiä and Solv iss. But in order to be able to compare different manifestations in typologically very different languages, we need to start by describing the language-internal systems as such. This does not mean that we have to wait with doing comparative work, since it is precisely comparisons with other languages that help us see similarities, differences and relations in the internal structure of specific languages.

REFERENCES Bolinger, D. L. (1977). Meaning and Form. Longman, London & New York. Bolinger, D. L. (1989). Extrinsic possibility and intrinsic potentiality: 7 on may and can + 1. J. Pragmatics, 13, 1–23. Croft, W. (2005). Logical and typological arguments for Radical Construction Grammar. In: Construction Grammars. Cognitive grounding and theoretical extensions [Constructional Approaches to Language 3] (J.-O. Östman and M. Fried, eds), pp. 273–314. John Benjamins, Amsterdam & Philadelphia, PA. Eide, K. M. (2002). Norwegian modals. PhD dissertation. Department of Linguistics, NTNU Trondheim.


Fillmore, C. (1985). Frames and the semantics of understanding. Quaderni di semantica VI(2): 222- 254. Fillmore, C. and P. Kay (1995). Construction Grammar, Lecture Notes, Copy Central, Bancroft Avenue, Berkeley, CA. Fried, M. and J.-O. Östman (2003). The explicit and the implicit in the Suasion Frame. In: Proceedings of the International Congress of Linguists 17 (E. Hajiová, A. Kotovcová & J. Mírovsk, eds), CD-ROM. Matfyzpress, Prague. Fried, M. and J.-O. Östman (2004). Construction Grammar: A thumbnail sketch. In: Construction Grammar in a Cross-language Perspective [Constructional Approaches to Language 2] (M. Fried and J.-O. Östman, eds), pp. 11–86. John Benjamins, Amsterdam & Philadelphia, PA. Fried, M. and J.-O. Östman (2005). Construction Grammar and spoken language: The case of pragmatic particles. Submitted to a Special issue on different approaches to conversation of J. Pragmatics ed. by K. Aijmer and A.-B. Stenström. Harling-Kranck, G. (1998). Från Pyttis till Nedervetil. Tjugonio prov på dialekter i Nyland, Åboland, Åland och Österbotten [SLS 610]. Svenska Litteratursällskapet i Finland, Helsinki. Kay, P. and C. Fillmore (1999). Grammatical constructions and linguistic generalizations: The What's X doing Y? construction, Language 75.1, 1-33. Kiuru, S. (1977). Suomen kielen kieltohakuiset verbit. SKS, Helsinki. Leino, J. and J.-O. Östman (2005). Variability in Construction Grammar. In: Construction Grammar: Back to the Roots [Constructional Approaches to Language 4] (M. Fried and H. Boas, eds). John Benjamins, Amsterdam & Philadelphia, PA. Östman, J.-O. (1977). On English also, Finnish -kin, and related matters: A contrastive approach. In: Contrastive Linguistics and Error Analysis (R. Palmberg and H. Ringbom, eds), pp. 165–196. Publications of the Research Institute of the Åbo Akademi Foundation (no. 19), Turku. Östman, J.-O. (1979). On the pragmatic meaning of textual particles. In: Kontrastiv lingvistik och sekundärspråksforskning (B. Hammarberg, ed.), pp. 171–191. Department of Linguistics, University of Stockholm. Östman, J.-O. (1981a). A functional approach to English tags. Studia Anglica Posnaniensia XIII, 3–16. Östman, J.-O. (1981b). You know: A Discourse-Functional approach [Pragmatics and Beyond II:7]. John Benjamins, Amsterdam. Östman, J.-O. (1982a). The symbiotic relationship between pragmatic particles and impromptu speech. In: Impromptu Speech: A Symposium (N. E. Enkvist, ed.), pp. 147–177. Publications of the Research Institute of the Åbo Akademi Foundation (no. 78), Turku. Östman, J.-O. (1982b). Pragmatic particles in an applied perspective. Neuphilologische Mitteilungen LXXXIII, 135–153. Östman, J.-O. (1986). Pragmatics as Implicitness. An Analysis of Question Particles in Solf Swedish, with Implications for the Study of Passive Clauses and the Language of Persuasion. Ph.D. dissertation. Department of Linguistics, University of California, Berkeley.


Östman, J.-O. (1987a). Pragmatic markers of persuasion. In: Propaganda, Persuasion and Polemic (J. Hawthorn, ed.), pp. 90–105. Edward Arnold, London. Östman, J.-O. (1987b). Implicit involvement in interactive writing. In: The Pragmatic Perspective [Pragmatics and Beyond, Companion Series 5] (J. Verschueren and M. Bertuccelli-Papi, eds), pp. 155–178. John Benjamins, Amsterdam. Östman, J.-O. (1987c). Pragmatic aspects of persuasion: Coherence. In: CDEF 86. Papers from the Conference of Departments of English in Finland (H. Nyyssönen, R. Kataja and V. Komulainen, eds), pp. 93–109. Publications of the Department of English, University of Oulu (no. 7), Oulu. Östman, J.-O. (1989). Testing iconicity: Sentence structure and politeness. Belgian J. Linguistics 4, 145–163. Östman, J.-O. (1991). On the language-internal interaction of prosody and pragmatic particles. In: Levels of Linguistic Adaptation [Pragmatics and Beyond New Series 6:2] (J. Verschueren, ed.), pp. 203–221. John Benjamins, Amsterdam. Östman, J.-O. (1995a). Pragmatic particles twenty years after. In: Organization in Discourse (B. Wårvik, S.-K. Tanskanen and R. Hiltunen, eds), pp. 95–108. Department of English, University of Turku, Turku. Östman, J.-O. (1995b). Recasting the deictic foundation, using physics and Finnish. In: Essays in Semantics and Pragmatics [Pragmatics and Beyond New Series 32] (M. Shibatani and S. Thompson, eds), pp. 247–278. John Benjamins, Amsterdam & Philadelphia, PA. Östman, J.-O. (2005). Construction Discourse: A prolegomenon. In: Construction Grammars. Cognitive Grounding and Theoretical Extensions [Constructional Approaches to Language 3] (J.-O. Östman and M. Fried, eds), pp. 121–144. John Benjamins, Amsterdam & Philadelphia, PA. Östman, J.-O. and B. Wårvik (1994). The Fight at Finnsburh: Pragmatic aspects of a narrative fragment. Neuphilologische Mitteilungen XCV, 207–227. Palmén, H. (2003). Uppbackningar i Liljendal. Paper presented at Dialektseminarium, Nordica, University of Helsinki, March 7, 2003. Weydt, H. (1969). Abtönungspartikel. Gehlen, Bad Homburg.


Pragmatic Markers in Contrast (SiP 2), Volume 2 (Studies in Pragmatics) (Studies in Pragmatics)

Lexical Markers of Common Grounds (SiP 3), Volume 3 (Studies in Pragmatics) (Studies in Pragmatics)

Pragmatic Competence (Mouton Series in Pragmatics)

Oxford Studies in Metaphysics: Volume 2 (Oxford Studies in Metaphysics)

Pragmatics

Pragmatics

Relevance and Linguistic Meaning: The Semantics and Pragmatics of Discourse Markers (Cambridge Studies in Linguistics)

Assertion (Palgrave Studies in Pragmatics, Language and Cognition)

Pragmatics in Neurogenic Communication Disorders

Assertion (Palgrave Studies in Pragmatics, Language and Cognition)

Pragmatic Markers in Oral Narrative: The Case of English and Catalan (Pragmatics and Beyond New Series)

Explicit Communication: Robyn Carston's Pragmatics (Palgrave Studies in Pragmatics, Language and Cognition)

Constraints in Discourse 2 (Pragmatics and Beyond. New Series)

Lexis in Contrast: Corpus-based Approaches (Studies in Corpus Linguistics)

ADVANCES IN CATALYSIS VOLUME 2, Volume 2

Lexis in Contrast: Corpus-based Approaches (Studies in Corpus Linguistics)

Meaning Through Language Contrast, Volume 2 (Pragmatics & Beyond New Series)

Illinois Classical Studies - Volume 2

Studies in History and Jurisprudence vol 2

Harvard Studies in Classical Philology. Vol. 2

Music and Patronage in Sixteenth-Century Mantua: Volume 2 (Cambridge Studies in Music) (v. 2)

Papers in Metaphysics and Epistemology: Volume 2 (Cambridge Studies in Philosophy) (v. 2)

Current Trends in the Pragmatics of Spanish

Modality (Oxford Surveys in Semantics and Pragmatics)

Doing Pragmatics

Death and Renewal: Sociological Studies in Roman History, Volume 2

HISTORICAL JESUS, VOLUME 2 (Critical Concepts in Religious Studies)

Death and Renewal: Sociological Studies in Roman History, Volume 2

New Horizons in Bon Studies (Bon Studies-2)

Experimental Pragmatics

ADVANCES IN GEOPHYSICS VOLUME 2, Volume 2 (v. 2)

Pragmatic Markers in Contrast (SiP 2), Volume 2 (Studies in Pragmatics) (Studies in Pragmatics)

Lexical Markers of Common Grounds (SiP 3), Volume 3 (Studies in Pragmatics) (Studies in Pragmatics)

Pragmatic Competence (Mouton Series in Pragmatics)

Oxford Studies in Metaphysics: Volume 2 (Oxford Studies in Metaphysics)

Pragmatics

Pragmatics

Relevance and Linguistic Meaning: The Semantics and Pragmatics of Discourse Markers (Cambridge Studies in Linguistics)

Assertion (Palgrave Studies in Pragmatics, Language and Cognition)

Pragmatics in Neurogenic Communication Disorders

Assertion (Palgrave Studies in Pragmatics, Language and Cognition)

Pragmatic Markers in Oral Narrative: The Case of English and Catalan (Pragmatics and Beyond New Series)

Explicit Communication: Robyn Carston's Pragmatics (Palgrave Studies in Pragmatics, Language and Cognition)

Constraints in Discourse 2 (Pragmatics and Beyond. New Series)

Lexis in Contrast: Corpus-based Approaches (Studies in Corpus Linguistics)

ADVANCES IN CATALYSIS VOLUME 2, Volume 2

Lexis in Contrast: Corpus-based Approaches (Studies in Corpus Linguistics)

Meaning Through Language Contrast, Volume 2 (Pragmatics & Beyond New Series)

Illinois Classical Studies - Volume 2

Studies in History and Jurisprudence vol 2

Harvard Studies in Classical Philology. Vol. 2

Music and Patronage in Sixteenth-Century Mantua: Volume 2 (Cambridge Studies in Music) (v. 2)

Papers in Metaphysics and Epistemology: Volume 2 (Cambridge Studies in Philosophy) (v. 2)

Current Trends in the Pragmatics of Spanish

Modality (Oxford Surveys in Semantics and Pragmatics)

Doing Pragmatics

Death and Renewal: Sociological Studies in Roman History, Volume 2

HISTORICAL JESUS, VOLUME 2 (Critical Concepts in Religious Studies)

Death and Renewal: Sociological Studies in Roman History, Volume 2

New Horizons in Bon Studies (Bon Studies-2)

Experimental Pragmatics

ADVANCES IN GEOPHYSICS VOLUME 2, Volume 2 (v. 2)

Recommend Documents