The Routledge Handbook of Pragmatics. 9781317362579, 1317362578

2,128 194 8MB

English Pages [599] Year 2017

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

The Routledge Handbook of Pragmatics.
 9781317362579, 1317362578

Table of contents :
Cover
Title
Copyright
Contents
List of illustrations
Notes on contributors
Acknowledgements
1 Pragmatics broadly viewed: introduction
2 Interdisciplinarity in pragmatics and linguistics
Part I Methods and modalities
Data collection
3 Naturally occurring data
4 Elicited data
5 Corpora
Nonverbal communication
6 British Sign Language (BSL)
7 Gesture and pragmatics: from paralinguistic to variably linguistic
8 Paralanguage
Part II Established fields
Pragmatics and variation
9 Variation and change: historical pragmatics
10 Variational pragmatics
11 Postcolonial pragmatics
12 Gender and sociopragmatics
13 Bilingualism and multilingualism
Pragmatics and culture
14 Interlanguage pragmatics: a historical sketch and future directions
15 Intercultural pragmatics
16 Identity and membership
17 Folk pragmatics
Linguistic pragmatics
18 Intention (including speech acts)
19 Temporal reference
20 Formal and natural languages: what logic tells us about natural language
21 Presupposition and accommodation
22 Grammaticalisation
Cognition and pragmatics
23 Metarepresentation
24 Relevance
25 Metaphor in pragmatics
26 Enrichment
Interactional pragmatics
27 Conversation
28 Discourse
29 Politeness
30 Reported speech
Part III Pragmatics across disciplines
31 Clinical pragmatics
32 Pragmatics and neurolinguistics
33 Doing ethnography
34 Language use in a social semiotic perspective
35 Linguistic pragmatics from an evolutionary perspective
Part IV Applications
36 Pragmatics and ontology
37 Pragmatics and translation/interpreting
38 Pragmatics in legal interpretation
39 Social media
40 Teaching pragmatics
Index

Citation preview

THE ROUTLEDGE HANDBOOK OF PRAGMATICS

The Routledge Handbook of Pragmatics provides a state-of-the-art overview of the wide breadth of research in pragmatics. An introductory section outlines a brief history, the main issues and key approaches and perspectives in the field, followed by a thought-provoking introductory chapter on interdisciplinarity by Jacob L. Mey. A further thirty-eight chapters cover both traditional and newer areas of pragmatic research, divided into four sections: • • • •

Methods and modalities Established fields Pragmatics across disciplines Applications of pragmatic research in today’s world.

With accessible, refreshing descriptions and discussions, and with a look towards future directions, this Handbook is an essential resource for advanced undergraduates, postgraduates and researchers in pragmatics within English language and linguistics and communication studies. Anne Barron is Professor of English Linguistics at the Leuphana University of Lüneburg, Germany. Recent publications include Public Information Messages (2012) and Acquisition in Interlanguage Pragmatics (2003). She has also co-edited several volumes, including Pragmatics of Discourse (2014), Variational Pragmatics (2008) and Pragmatics of Irish English (2005), all three co-edited with Klaus P. Schneider. She is a member of the editorial boards of the Journal of Pragmatics and Intercultural Pragmatics. Yueguo Gu is a Special Title Professor of Linguistics at Beijing Foreign Studies University, China. Recent publications include Using the Computer in ELT (2006) and Pragmatics and Discourse Studies (2010). He has also edited several series of textbooks and collections of academic papers such as Initial Exploration of Online Education (2004) and Second Exploration of Online Education (2005). Gerard Steen is Professor of Speech Communication, Argumentation and Rhetoric at the University of Amsterdam, the Netherlands. He is also the founding director of the Metaphor Lab Amsterdam.

Routledge Handbooks in Applied Linguistics

Routledge Handbooks in Applied Linguistics provide comprehensive overviews of the key topics in applied linguistics. All entries for the handbooks are specially commissioned and written by leading scholars in the field. Clear, accessible and carefully edited, Routledge Handbooks in Applied Linguistics are the ideal resource for both advanced undergraduates and postgraduate students. For a full list of titles in this series, please visit www.routledge.com/Routledge-Handbooksin-Applied-Linguistics/book-series/RHAL The Routledge Handbook of Instructed Second Language Acquisition Edited by Shawn Loewen and Masatoshi Sato The Routledge Handbook of Migration and Language Edited by Suresh Canagarajah The Routledge Handbook of Pragmatics Edited by Anne Barron,Yueguo Gu and Gerard Steen The Routledge Handbook of English Language Teaching Edited by Graham Hall The Routledge Handbook of Language Learning and Technology Edited by Fiona Farr and Liam Murray The Routledge Handbook of Language and Identity Edited by Siân Preece The Routledge Handbook of English for Academic Purposes Edited by Ken Hyland and Philip Shaw The Routledge Handbook of Language and Digital Communication Edited by Alexandra Georgakopoulou and Tereza Spilioti The Routledge Handbook of Literacy Studies Edited by Jennifer Rowsell and Kate Pahl The Routledge Handbook of Interpreting Edited by Holly Mikkelson and Renée Jourdenais The Routledge Handbook of Hispanic Applied Linguistics Edited by Manel Lacorte The Routledge Handbook of Educational Linguistics Edited by Martha Bigelow and Johanna Ennser-Kananen The Routledge Handbook of Forensic Linguistics Edited by Malcolm Coulthard and Alison Johnson The Routledge Handbook of Corpus Linguistics Edited by Anne O’Keeffe and Mike McCarthy

THE ROUTLEDGE HANDBOOK OF PRAGMATICS

Edited by Anne Barron, Yueguo Gu and Gerard Steen

First published 2017 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 711 Third Avenue, New York, NY 10017 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2017 selection and editorial matter, Anne Barron,Yueguo Gu and Gerard Steen; individual chapters, the contributors The right of Anne Barron,Yueguo Gu and Gerard Steen to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. 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 Names: Barron, Anne, editor. | Gu,Yueguo, editor. | Steen, Gerard, editor. Title: The Routledge handbook of pragmatics / edited by Anne Barron, Yueguo Gu and Gerard Steen. Description: Milton Park, Abingdon, Oxon ; New York, NY : Routledge, [2017] | Series: Routledge Handbooks in applied linguistics | Includes bibliographical references and index. Identifiers: LCCN 2016028580 | ISBN 9780415531412 (hardback) | ISBN 9781315668925 (ebook) Subjects: LCSH: Pragmatics—Handbooks, manuals, etc. Classification: LCC P99.4.P72 R68 2017 | DDC 401/.45—dc23 LC record available at https://lccn.loc.gov/2016028580 ISBN: 978-0-415-53141-2 (hbk) ISBN: 978-1-315-66892-5 (ebk) Typeset in Bembo by Apex CoVantage, LLC

CONTENTS

List of illustrations ix Notes on contributors xi Acknowledgementsxviii   1 Pragmatics broadly viewed: introduction Anne Barron, Yueguo Gu and Gerard Steen

1

  2 Interdisciplinarity in pragmatics and linguistics Jacob L. Mey

4

PART I

Methods and modalities

19

Data collection

19

  3 Naturally occurring data Andrea Golato

21

  4 Elicited data J. César Félix-Brasdefer and Maria Hasler-Barker

27

 5 Corpora Martin Weisser

41

v

Contents

Nonverbal communication

53

  6 British Sign Language (BSL) Gary Quinn

55

  7 Gesture and pragmatics: from paralinguistic to variably linguistic Alan Cienki

61

 8 Paralanguage Tim Wharton

69

PART II

Established fields

77

Pragmatics and variation

77

  9 Variation and change: historical pragmatics Andreas H. Jucker and Daniela Landert

79

10 Variational pragmatics Anne Barron

91

11 Postcolonial pragmatics Eric A. Anchimbe and Richard W. Janney

105

12 Gender and sociopragmatics Janet Holmes and Brian W. King

121

13 Bilingualism and multilingualism Jasone Cenoz

139

Pragmatics and culture

151

14 Interlanguage pragmatics: a historical sketch and future directions Naoko Taguchi

153

15 Intercultural pragmatics Alessia Cogo and Juliane House

168

16 Identity and membership Dorien Van De Mieroop

184

vi

Contents

17 Folk pragmatics Dennis R. Preston and Nancy Niedzielski

197

Linguistic pragmatics

213

18 Intention (including speech acts) Jesús Navarro

215

19 Temporal reference Kasia M. Jaszczolt

227

20 Formal and natural languages: what logic tells us about natural language Jacques Moeschler

241

21 Presupposition and accommodation Jacopo Romoli and Uli Sauerland

257

22 Grammaticalisation María José López-Couso and Elena Seoane

277

Cognition and pragmatics

293

23 Metarepresentation Nicholas Allott

295

24 Relevance Stavros Assimakopoulos

310

25 Metaphor in pragmatics Miriam Taverniers

323

26 Enrichment Alison Hall

341

Interactional pragmatics

355

27 Conversation Hansun Zhang Waring

357

28 Discourse Rodney H. Jones

371

vii

Contents

29 Politeness Dawn Archer

384

30 Reported speech Isabelle Buchstaller

399

PART III

Pragmatics across disciplines

417

31 Clinical pragmatics Louise Cummings

419

32 Pragmatics and neurolinguistics Elisabeth Ahlsén

433

33 Doing ethnography Dorothy Pawluch, Arthur McLuhan and William Shaffir

447

34 Language use in a social semiotic perspective Christian M. I. M. Matthiessen

459

35 Linguistic pragmatics from an evolutionary perspective Nikolaus Ritt

490

PART IV

Applications503 36 Pragmatics and ontology Laurent Prévot

505

37 Pragmatics and translation/interpreting Nicole Baumgarten

521

38 Pragmatics in legal interpretation Alan Durant and Janny H. C. Leung

535

39 Social media Francisco Yus

550

40 Teaching pragmatics Helen Basturkmen and Thi Thuy Minh Nguyen

563

Index575 viii

ILLUSTRATIONS

Figures 5.1

5.2 5.3 5.4 5.5 19.1 20.1 20.2 20.3 20.4 20.5 25.1 25.2 25.3 25.4 25.5 25.6 25.7 29.1 30.1 30.2 34.1 34.2

The first fourteen results (hits) of a KWIC concordance of the word well, nicely illustrating the distinction between its uses as a discourse marker (henceforth DM) and an ordinary adverb N-Gram list SPICE-Ireland sample DART XML format SPAADIA (ver. 1) coding sample Representation of the pragmatically inferred reading of (15) as (15a) in Default Semantics Wide scope reading (‘not all students passed’) Narrow scope reading (‘no students passed’) From sentence to meaning The logical square ⟦Some . . . not⟧ as a complement set Early pragmatic views on the relation between literal and non-literal meaning Sperber and Wilson’s (1986a) explicature and its relation to implicature Grice’s circle Two sides of an implicature: detection and resolution Implicature as a recursive process Three ways in which a first intermediate level of meaning between ‘what is said’ and ‘what is meant’ is ‘special’ Ad hoc concept formation as based on a mutual adjustment of two inputs Facework scale, including accidental FTA zone Multivoicedness in reported speech/thought sequences The evaluative stances that hold in a quotative situation Systemic orders of phenomenal realms and alternative interpretations of systems of the fourth order – semiotic or cognitive Frequency of the terms social semiotic and social semiotics and of social semiotician from 1970 to 2008 according to Google’s Ngram Viewer ix

42 42 46 48 49 237 242 242 243 249 251 326 328 329 329 330 333 335 393 404 407 460 465

Illustrations

34.3

The development of social semiotics as a theme within SFL and as a separate “discipline” in relation to other movements in social and semiotic (including linguistic) theories 34.4 Social semiotics and related developments – key scholars 34.5 The location of “text” along the dimensions of the cline of instantiation and the hierarchy of stratification in SFL 34.6 Fourth-order systems – cognitive or semiotic 34.7 Semiotic systems differentiated according to the sensory system involved in the processing of the expression plane 34.8 Diagrammatic representation of Halliday’s distinction between primary and higher-order semiotic systems 34.9 Example of a context-specific microfunctional meaning potential in protolanguage with articulatory realizations 34.10 Types of semiotic system – seen “from above” in terms of content functions, and “from below” in terms of expression modalities (used by the addressee) 36.1 Metamodel for dialogue act annotation 36.2 A sample of natural language-motivated ontological distinctions 36.3 A taxonomy of abstract objects 36.4 An attempt of integration using formal ontology

468 471 475 477 479 480 481 482 507 508 509 515

Tables 20.1 20.2 32.1

Truth conditions for logical connectives Lexical realisation of the logical square vertices Overview of typical objects of study and typical research methods in neurolinguistics and pragmatics

x

247 250 436

CONTRIBUTORS

Elisabeth Ahlsén is Professor of Neurolinguistics at the SSCCIIL Interdisciplinary Research Center and the Division of Communication and Cognition at the Department of Applied Information Technology, University of Gothenburg, Sweden. Her main areas of research are neurolinguistics, pragmatics and communication disorders. She has been a Research Fellow at The ZiF Center for Interdisciplinary Research in Bielefeld (2005–2006) and at Stellenbosch Institute for Advanced Study (STIAS) (2011). Nicholas Allott is Senior Lecturer in English Language at the University of Oslo, Norway. He has worked on the roles of inference and metarepresentation in utterance interpretation, on word meaning and lexical modulation, on the pragmatics of legal language, and on the philosophy of linguistics. Publications include Key Terms in Pragmatics (2010) and, with Neil Smith, the third edition of Chomsky: Ideas and Ideals (2016). Eric A. Anchimbe is a Lecturer in English Linguistics at the University of Bayreuth, Germany. His current research is on offers and offer refusals in postcolonial communities from a postcolonial pragmatics perspective, political discourse from below in Africa, sociolinguistic identity, and World Englishes. His recent publications include the monograph Language Policy and Identity Construction (2013). Dawn Archer is Professor of Pragmatics and Corpus Linguistics, based at Manchester Metropolitan University, UK. She has written extensively on activity types such as the (historical) courtroom. She is also involved in the development of (semi-)automatic methods of pragmatic analyses. Stavros Assimakopoulos is Senior Lecturer in Linguistics at the University of Malta. His research lies at the interface of linguistics, philosophy and cognitive psychology and focuses on the implications that cognitive approaches to inferential pragmatics carry for the study of linguistic meaning and thought. He recently guest-edited a special issue of The Linguistic Review (vol. 29(4), 2012) on ‘cognitive pragmatics and its interfaces in linguistics’.

xi

Contributors

Anne Barron is Professor of English Linguistics at the Leuphana University of Lüneburg, Germany. Her research has focused on variational pragmatics, the pragmatics of Irish English, interlanguage pragmatics and contrastive genre analysis. Recent publications include Public Information Messages (2012) and Pragmatics of Discourse (2014, co-editor Klaus P. Schneider). Helen Basturkmen is Associate Professor in Applied Language Studies and Linguistics at the University of Auckland, New Zealand. She has published articles in a number of international journals and has served as assistant editor of the journal Language Teaching Research and as an editorial review board member of the Journal of English for Academic Purposes. Nicole Baumgarten is a Lecturer in Applied Linguistics at the University of Sheffield, UK. Her work is on interpersonal communication and identity construction-related language use in a variety of settings, including business, science and technology communication, teaching, classroom discourse, conversation, interviewing, and social media discourse. Her special focus lies on communicative settings that are multilingual, intercultural, and transnational. Isabelle Buchstaller is Professor of Varieties of English at Leipzig University, Germany. She is author of Quotatives: New Trends and Sociolinguistic Implications (2013), has edited Quotatives: Crosslinguistic and Cross-disciplinary Perspectives (with Ingrid van Alphen 2012) and has published many articles on reported speech and thought. Jasone Cenoz is Professor of Research Methods in Education at the University of the Basque Country, UPV/EHU. Her research has focused on multilingual education, bilingualism and multilingualism combining psycholinguistic, social psychological, sociolinguistic and educational perspectives. She has published extensively on multilingualism. Alan Cienki is Professor of Language Use and Cognition at the Vrije Universiteit Amsterdam, the Netherlands, and Director of the Multimodal Communication and Cognition Lab at Moscow State Linguistic University, Russia. He co-edited the book Metaphor and Gesture (2008) and the two-volume handbook Body–Language–Communication (2013, 2014). Alessia Cogo is a Lecturer in Applied Linguistics at Goldsmiths, University of London, UK, where she teaches modules in sociolinguistics and global Englishes. She is Reviews Editor of the English Language Teaching Journal as well as co-founder and co-convenor of the AILA Research Network on ELF. Her current research concerns ELF multilingual practices in professional, academic and immigration contexts. Louise Cummings is Professor of Linguistics at Nottingham Trent University, UK. She is the author of several books in pragmatics and clinical linguistics, the latest including Pragmatic Disorders (2014), Pragmatic and Discourse Disorders: A Workbook (2015) and Case Studies in Communication Disorders (2016). Alan Durant is Professor of Communication in the School of Law at Middlesex University, UK. Recent books include Meaning in the Media: Discourse, Controversy and Debate (2010), the RELI textbook Language and Media, co-written with Marina Lambrou (2009) and Language and Law co-authored with Janny Leung (2016). The edited collection of essays Language, Law and Power, also co-authored with Janny Leung, is forthcoming.

xii

Contributors

J. César Félix-Brasdefer is Associate Professor at Indiana University, USA. His research interests include pragmatics, discourse analysis, cross-cultural and interlanguage pragmatics, pragmatic variation, and (im)politeness theory. He has published numerous research articles in a variety of scholarly journals. His most recent book is titled The Language of Service Encounters: A Pragmatic-Discursive Approach (2015). Andrea Golato is Professor of Modern Languages and Dean of The Graduate College at Texas State University, USA. Her research interests lie primarily in conversation analysis, particularly in the connection between grammar and interaction, on one hand, and language use in different cultures, on the other. She has published on reported discourse, compliment, response tokens, and various functions of repair in German. Yueguo Gu is a Special Title Professor of Linguistics at Beijing Foreign Studies University, China, and Director of the China Multilingual & Multimodal Corpora of Big Data Research Centre. He is also Head of the Corpus Linguistics Department at the Chinese Academy of Social Sciences. His research interests include pragmatics, discourse analysis, corpus linguistics, and online education. Recent publications include Studies in Pragmatics and Discourse Analysis and Chinese Painting by Foreign Language Teaching and Research Press (FLTRP). Alison Hall is a Lecturer in English Language at De Montfort University, UK. Her research has focused on contextualism, lexical pragmatics and ellipsis. Recent publications include ‘ “Free” enrichment and the nature of pragmatic constraints’ (International Review of Pragmatics 2014). Maria Hasler-Barker is Assistant Professor at Sam Houston State University, Texas, USA, where she teaches courses in Spanish language and linguistics. She publishes on interlanguage pragmatics and pedagogy, and also has current research interests in second language acquisition and bilingual service encounters. Janet Holmes is Emeritus Professor of Linguistics and Associate Director of the Wellington Language in the Workplace Project at Victoria University of Wellington, New Zealand (www. victoria.ac.nz/lwp/). She has published on many aspects of workplace discourse, as well as language and gender. Juliane House is Professor Emeritus, University of Hamburg, Germany, and currently Distinguished University Professor, Hellenic American University, Greece, where she set up and directs a PhD programme in applied linguistics. Her current research focuses on English as a lingua franca in business contexts, translation theory and (im)politeness in computer-mediated communication. Richard W. Janney is Professor Emeritus of Modern English Linguistics in the Department of English and American Studies at the University of Munich, Germany. His research areas include postcolonial pragmatics, film pragmatics, Internet pragmatics and bodily communication. He is former co-editor of the Journal of Pragmatics. Kasia M. Jaszczolt is Professor of Linguistics and Philosophy of Language at the University of Cambridge, UK. She has published extensively on various topics in semantics, pragmatics

xiii

Contributors

and philosophy of language. Her authored books include Meaning in Linguistic Interaction (2016), Representing Time (2009), Default Semantics (2005), Semantics and Pragmatics (2002) and Discourse, Beliefs and Intentions (1999). Rodney H. Jones is Professor of Sociolinguistics at the University of Reading, UK. Among his recent books are Health and Risk Communication: An Applied Linguistic Perspective (2103) and The Routledge Handbook of Language and Creativity (2016). Andreas H. Jucker is Dean of the Faculty of Arts and Social Sciences and Professor of English Linguistics at the University of Zurich, Switzerland. His current research interests include historical pragmatics, politeness theory, speech act theory, and the grammar and history of English. Brian W. King is Assistant Professor of English Language Studies at City University of Hong Kong. He has published in various areas under the umbrella of language, gender and sexuality. Daniela Landert is a postdoctoral researcher at the University of Zurich, Switzerland. Her research interests include historical pragmatics, corpus pragmatics, modality and mass media communication. In her current research project, she investigates epistemicity and evidentiality in early modern English from the perspective of historical corpus pragmatics. Janny H. C. Leung obtained her MPhil and PhD in English and Applied Linguistics from the University of Cambridge, and holds an LLB from the University of London, UK. She is Associate Professor in the School of English, the University of Hong Kong. Her published work ranges across linguistic and legal topics, and a monograph entitled Multilingual Legal Order is forthcoming. María José López-Couso is Associate Professor in English Linguistics at the University of Santiago de Compostela, Spain. Her research focuses on historical syntax, historical pragmatics, grammaticalisation, and (inter-)subjectification. She has published extensively on these topics in leading linguistics journals and contributed to the Mouton handbooks Historical Pragmatics (2010) and English Historical Linguistics (2011), as well as to the Cambridge Handbook of English Historical Linguistics (2016). Christian M.I.M. Matthiessen is Professor in the Department of English at the Hong Kong Polytechnic University, Hong Kong. Recent publications include Systemic Functional Grammar: A First Step into the Theory (2009), co-authored with M.A.K. Halliday, and Key Terms in Systemic Functional Linguistics (2010), co-authored with Marvin Lam and Kazuhiro Teruya. Arthur McLuhan is Assistant Professor of Sociology at York University, Canada. Broadly, his research focuses on the sociologies of (1) self and identity; (2) deviance and crime; and (3) subcultures and everyday life, with an emphasis on developing more general theories and concepts of social processes that are empirically grounded in the ethnographic examination of particular social worlds. Jacob L. Mey is Professor Emeritus at the University of Southern Denmark. He has received honorary doctorates from the University of Zaragoza (1993) and from the University of Bucharest (2006). He served as the editor in chief of the Journal of Pragmatics (1977–2010) and is currently editor in chief of Pragmatics and Society (2010–present). xiv

Contributors

Thi Thuy Minh Nguyen is Assistant Professor in Applied Linguistics at the English Language and Literature Academic Group, National Institute of Education, Singapore. Her research interests are in the area of pragmatics and language pedagogy. She has published articles in international journals and book chapters in edited volumes. Jacques Moeschler is Professor of French Linguistics at the Department of Linguistics, University of Geneva, Switzerland. His research has focused on tenses, causality, discourse and logical connectives, as well as negation. Recent publications include Initiation à l’étude du sens (2012) and Initiation à la linguistique française (2010), both with S. Zufferey. He is currently preparing a textbook on implicatures (also with S. Zufferey). Jesús Navarro is Associate Professor of Philosophy at the University of Seville, Spain. He has authored three books and various articles on scepticism, epistemology, and the philosophy of mind and language. A translation of his book How to Do Philosophy with Words: Reflexions on the Searle–­Derrida Debate is forthcoming. Nancy Niedzielski is Associate Professor of Linguistics at Rice University, USA. Her main areas of interest are speech perception, sociophonetics, and urban sociolinguistics. Her current body of research is encompassed by the Houston Urban English Survey (HUES) project, which examines Houston language varieties from both perception and production viewpoints. Dorothy Pawluch is Associate Professor of Sociology at McMaster University, Canada. Her interest in meaning-making processes has led over the years to studies of people living with HIV/AIDS and changing understandings of health, illness and deviance. Her book The New Pediatrics: A Profession in Transition (2009) examines the expansion of paediatrics into behavioural problems and the role this expansion played in medicalising children’s deviance. Dennis R. Preston is Regents Professor of Linguistics at Oklahoma State University and Distinguished Professor of Linguistics Emeritus at Michigan State University, USA. His main research interests are sociolinguistics and language attitudes, and he has helped found and contributed to the specialisations of folk linguistics, perceptual dialectology, and sociolinguistic aspects of second language acquisition. Laurent Prévot is Professor in Linguistics at Aix Marseille University, France, and a junior member of Institut Universitaire de France. He is carrying out his research at the Laboratoire Parole et Langage, where he focuses on natural language processing techniques to discourse and dialogue. Recently, he has been working on the development of annotation schema for characterising discourse and interaction phenomena in spontaneous conversations. Gary Quinn is Assistant Professor and Head of the British Sign Language (BSL) section in the Department of Languages and Intercultural Studies at Heriot-Watt University, UK. His research interests include sign language curricula, pedagogy of BSL teaching and teacher education, applied sign linguistics, language variation, vocabulary development and specialised lexicons in the fields of science, mathematics, history and geography. Nikolaus Ritt is Professor of Historical English Linguistics at the University of Vienna, Austria. His research deals with phonological and morphological changes and their interaction, as well as with the theory of language variation and change. He attempts to account for languages and their xv

Contributors

properties in Darwinian terms by viewing them as populations of constituents. The principles of that approach are described in Selfish Sounds and Linguistic Evolution (2004). Jacopo Romoli is a Lecturer in Linguistics and the course director of the master’s programme in English Language Linguistics at Ulster University, UK, and associate investigator of the Australian Research Council’s Centre of Excellence for the Study of Cognition and Its Disorders (CCD) at Macquarie University, Australia. His research interests are in formal semantics and pragmatics for natural languages, language processing and language acquisition. Uli Sauerland leads the semantics/pragmatics research team at the Centre for General Linguistics in Berlin, Germany. He is interested in all areas of semantics and pragmatics and has made a particular contribution to the theories of binding, implicature and presuppositions.To substantiate his theoretical research, he has done language acquisition studies and fieldwork in Indonesia and Brazil. His published work includes 110 reviewed publications. Elena Seoane is Associate Professor in English Linguistics at the University of Vigo, Spain. Her research interests lie in English historical syntax, historical pragmatics, grammaticalisation theory and, more recently, in morphosyntactic variation in World Englishes. Recently, she has contributed to the English Historical Linguistics: An International Handbook (eds. Bergs and Brinton 2011) and the Oxford Handbook of the History of English (eds. Nevalainen and Traugott 2012). William Shaffir is Professor of Sociology at McMaster University, Canada. He is the author of books and articles on professional socialisation, Hassidic Jews, ethnic violence and qualitative research methods. His books include The Riot at Christie Pits (co-authored with C. Levitt), Leaving Religion and Religious Life (co-authored with M. Bar-Lev) and Ethnographies Revisited: Constructing Theory in the Field (co-edited with A. Pudepphat and S. Kleinknecht). Gerard Steen is Professor of Speech Communication, Argumentation and Rhetoric at the University of Amsterdam, the Netherlands. He is also the founding director of the Metaphor Lab Amsterdam. Naoko Taguchi is Associate Professor in the Modern Languages Department at Carnegie Mellon University, USA. Her research interests include pragmatics, classroom-based research, and learning contexts (study abroad, English-medium universities). She is the author of Context, Individual Differences, and Pragmatic Competence (2012) and Developing Interactional Competence in a Japanese Study Abroad Context (2015). She also co-authored a forthcoming monograph entitled Second Language Pragmatics. Miriam Taverniers is Associate Professor of Functional Approaches to English Linguistics at Ghent University, Belgium. Her research interests include structural-functional models of language, metaphor and grammatical metaphor, argument structure and diathesis relations, voice in relation to aspect (secondary) predication and syntax-semantics, semantics-pragmatics and other interfaces in the architectures of linguistic theories. Dorien Van De Mieroop is Assistant Professor in Dutch Linguistics at the Faculty of Arts of the University of Leuven, Belgium. Her research focuses on the discursive analysis of identity construction, which she studies in narratives and institutional contexts. She has published more

xvi

Contributors

than twenty articles in a number of peer-reviewed journals, including Discourse & Society, Narrative Inquiry and Journal of Pragmatics. Hansun Zhang Waring is Associate Professor of Applied Linguistics and TESOL at Teachers College, Columbia University, USA, and the author (with Jean Wong) of Conversation Analysis and Second Language Pedagogy (2010), as well as Theorizing Pedagogical Interaction: Insights from Conversation Analysis (2016). Martin Weisser is Professor of Linguistics and Applied Linguistics in the Center for Linguistics and Applied Linguistics, Guangdong University of Foreign Studies, China. His main research areas are general corpus linguistics and corpus pragmatics. He is also author of the Dialogue Annotation and Research Tool (DART). Tim Wharton is Principal Lecturer in Linguistics at the University of Brighton, UK. His research explores how ‘natural’, non-linguistic behaviours – tone of voice, facial expressions, gesture – interact with the linguistic properties of utterances (broadly speaking, the words we say). His main theses are outlined in his book Pragmatics and Non-verbal Communication (2009). Francisco Yus currently heads both the inter-university Institute of Applied Modern Languages of the Valencian Community (IULMA) and the Professional and Academic English research group at the University of Alicante, Spain. His latest research concerns the production and interpretation of humorous discourses (Humour and Relevance, 2016).

xvii

ACKNOWLEDGEMENTS

Several people have contributed to the production of this volume. First, we, the editors, would like to thank the editorial staff at Routledge and in particular Louisa Semlyen, Rosemary Baron, Sophie Jaques and Laura Sandford for their support, expert assistance, efficiency and interest in the project throughout this venture, but especially during some particularly difficult moments. We are also grateful to Peter Grundy for his early initiatives with this title. A special thanks also to Beatrice Brodesser, Irina Pandarova, Kerstin Single, Romy van den Heerik and Insa Wippermann for their invaluable and competent support in various stages of the editing, proof-reading and preparation of the final manuscript. Finally, we would also like to thank the following organisations for granting permission to reproduce. Excerpt (3) in Holmes and King: Blackwell-Wiley for permission to use the original excerpt in Holmes (2006); Figure 30.2: Wiley Blackwell for the adapted version used in this Handbook (Buchstaller 2014: 49) and Benjamins for the original figure (Spronck 2012: 72); Figure 36.1: ELRA (European Language Resources Association) for permission to use the original figure (Bunt et al. 2010) published within the proceedings of the LREC 2010 Conference and also Figure 36.2: Elsevier for permission to use the original figure in Dölling (1995). Every effort has been made to contact copyright holders. Please advise the publisher of any errors or omissions, and these will be corrected in subsequent editions. Anne Barron (Leuphana University of Lüneburg) Yueguo Gu (Chinese Academy of Social Sciences) Gerard Steen (University of Amsterdam) January 2016

xviii

1 PRAGMATICS BROADLY VIEWED Introduction Anne Barron, Yueguo Gu and Gerard Steen

Pragmatics deals very broadly with the study of language in the social contexts in which it is used. This area of study has roots stretching into earlier centuries (albeit not under the term pragmatics), but pragmatics as a field of research in linguistics was not established until the 1970s. It developed outside of linguistics in ordinary language philosophy, in particular under the influence of philosophers such as Ludwig Wittgenstein, John Austin, John Searle and Herbert P. Grice. Rather early in its development, two main strands of pragmatics began to be identified: an Anglo-American tradition and a Continental tradition (cf. Nerlich 2010; Jucker 2012; Mey 2013; Huang 2015: 4–6 on the history of pragmatics). The Anglo-American tradition, also termed ‘the component view’ or ‘the pigeon-hole view’ (Huang 2015: 4), was particularly influenced by Geoffery Leech’s textbook, Principles of Pragmatics (1983), a renowned publication which determined the basis of the early discipline. This approach is rather narrowly focused and notably favours a pragmalinguistic over a sociopragmatic approach. It sees pragmatics as one core ‘component’ within linguistics, on the same level as other core components such as semantics, syntax, morphology and phonology, and values clear task delineation from these other levels of language analysis. It deals, in particular, with implicature, inferencing, presupposition, deixis, reference and speech acts. The other tradition which emerged, the Continental European tradition (macropragmatics, in Mey’s [2001] terms), is broader in scope. It does not view pragmatics as one component of linguistic analysis, and is not concerned with marking boundaries with other components of language. Rather, it sees pragmatics as dealing with social meaning and representing a superordinate perspective on linguistic communication in general (hence the approach is also termed ‘the perspective view’ [Huang 2015: 5]). The perspective view sees pragmatics as not only limited to linguistic research, but also as practised across neighbouring fields such as sociology or psychology. The range of topic areas dealt with include those addressed in the component perspective, but also such areas of interest as intercultural pragmatics and literary pragmatics, among many others. In addition, this view extends the focus from language alone to behaviour and thus also focuses on nonverbal communication. This division of focus between the Anglo-American and Continental European traditions is upheld today. Indeed, many of the major textbooks on pragmatics, of which scholars have published a growing number in recent years, reflect this tradition. Two very recent textbooks by Birner (2013) and Huang (2015), for instance, concentrate on the Anglo-Saxon tradition alone. 1

Anne Barron, Yueguo Gu and Gerard Steen

Others, such as Culpeper and Haugh (2014), however, seek to unite both. This is also the aim of the present handbook. Specifically, this collection is an attempt to unite these different pragmatic traditions in one volume and to bring together experts from all strands of the discipline to present a state-of-the-art overview on the wide breadth of research in pragmatics.Thus, while some chapters take a more microlinguistic approach and others a more macro sociocultural approach, others deal narrowly with linguistic structure and also at the same time with the broader macro context. Underlining the unifying aims of this handbook, the present collection begins with a chapter on interdisciplinarity in pragmatics and linguistics by Jacob Mey in which he, looking back on the development of the field of pragmatics, highlights its status as an interdisciplinary undertaking. Following this introductory chapter, the remaining chapters are grouped into four major sections, focusing on methods and modalities, established fields, pragmatics across disciplines and, finally, on applications of pragmatic research in today’s world, respectively. Each of these content areas is briefly described later. Section I, Methods and modalities, contains two subsections. The first addresses data collection methods employed in pragmatics today, as well as recent innovations and trends in the area. The second deals with nonverbal communication, underlining that pragmatics may be also concerned with behaviour in general, not just with verbal communication alone. The Established fields section follows. It provides a state-of-the-art approach to pragmatics in all of the major areas of pragmatics research. As the section given most weight in the volume, it is subdivided into five further sections, namely: • • • • •

Pragmatics and variation Pragmatics and culture Linguistic pragmatics Cognition and pragmatics Interactional pragmatics

The third section, Pragmatics across disciplines, enables readers familiar with one approach to pragmatics to gain an understanding of other approaches. Particularly readers with a linguistic background will see here the opportunity to experience a range of ways in which pragmatics has been approached across related disciplines, from ethnography, neurolinguistics and clinical linguistics through to evolutionary approaches. This in turn will provide readers with an understanding of the applicability of their field of study to related disciplines. Finally, the fourth section, Applications, concerns applications of pragmatic research in today’s world. It comprises a series of chapters dealing with areas such as pragmatics in ontology, translation, interpreting, legal interpretation, social media and language teaching, and its inclusion is an attempt to broaden the base of pragmatics and to encourage cross-dissemination across theoretical and applied disciplines. This handbook is designed to be relevant and accessible to a wide readership, from upperlevel undergraduate and taught postgraduate students through to researchers and academics. The chapters explain the significance of the topic or approach at hand, provide a critical survey of the principal approaches to it and sketch future directions. Contributors to this volume are from around the world and include established figures in the field, on occasion teamed up with younger, up-and-coming voices. Similarly, the topics addressed aim at representing traditional areas of pragmatic research, but also newer ones, such as the newly emerged fields of variational pragmatics and postcolonial pragmatics. We hope, in conclusion, to have compiled a volume which will not only further blur the lines between the component and perspective traditions of 2

Pragmatics broadly viewed

pragmatic research, but which will also provide readers with accessible, refreshing descriptions and discussions and with future perspectives and challenges which will shape and inform the pragmatics of tomorrow.

References Birner, B. J. (2013) Introduction to Pragmatics. West Sussex: Wiley-Blackwell. Culpeper, J. and Haugh, M. (2014) Pragmatics and the English Language. Basingstoke: Palgrave Macmillan. Huang,Y. (2015) Pragmatics, 2nd ed. Oxford: Oxford University Press. Jucker, A. H. (2012) ‘Pragmatics in the history of linguistic thought’, in K. Allan and K. M. Jaszczolt (eds.) The Cambridge Handbook of Pragmatics. Cambridge: Cambridge University Press. 495–512. Leech, G. (1983) Principles of Pragmatics. London: Longman. Mey, J. L. (2001) Pragmatics: An Introduction. Oxford: Wiley-Blackwell. Mey, J. L. (2013) ‘A brief sketch of the historic development of pragmatics’, in K. Allan (ed.) The Oxford Handbook of the History of Linguistics. Oxford: Oxford University Press. 587–612. Nerlich, B. (2010) ‘History of pragmatics’, in L. Cummings (ed.) The Pragmatics Encyclopedia. London/New York: Routledge. 192–193.

3

2 INTERDISCIPLINARITY IN PRAGMATICS AND LINGUISTICS1 Jacob L. Mey

Introduction At the Second International Conference ‘Zeichen und System der Sprache’ (Magdeburg, September 1964), a certain East German professor took the floor during a discussion of one of the linguistic presentations. He started his comments by saying: ‘Als Mathematiker weiß ich zwar von der Sache nichts, kann aber immer etwas Prinzipielles dazu sagen’ (‘Being a mathematician, I don’t know anything about the subject matter, but I can always make some principled remarks.’). The case is interesting as an instance of what I call ‘cross-disciplinarity’: the ‘crossing over’ from one discipline to another, applying the ‘principles’ one is familiar with from one’s own discipline to problems raised in another disciplinary context. By contrast, consider the situation of a linguist investigating a case of ‘language death’ (a particular language or dialect falling into oblivion and being pronounced ‘dead’ when the last native speaker passes away – sometimes attested historically, as in the case of Cornish, when its last speaker, a woman called Dolly Penreadh, age 102, died in 1777; Pedersen 1924: 48). Such instances are hard to explain by appealing to internal linguistic factors; usually, the social context and the life circumstances of the speakers are brought to bear on the phenomenon, or individual or communal psychological factors are invoked. In our days, the phenomenon of ‘domain loss’ comes to mind, where whole segments of a language are surrendered to a foreign idiom (mostly English, as with ‘computer speak’). In such cases, the ‘principles’ of one’s discipline, considered as an imminent, coherent whole, will not be sufficient to furnish an explanation, and an interdisciplinary approach is needed. When we talk about ‘interdisciplinarity’ in linguistics and pragmatics, we often confuse these two concepts: the cross-disciplinary and the (true) interdisciplinary. Whereas the former notion, when we look at pragmatics, involves a ‘crossing over’ into the disciplines of linguistics (e.g. as regards the pragmatic aporias that one encounters within semantics, or the pragmatic inconsistencies within conversation analysis and the ways its data are collected, selected and evaluated), interdisciplinarity, on the other hand, is concerned with the two disciplines’ mutual interpenetration and dialectic interdependence. In particular, linguistics needs pragmatics to obtain a proper perspective for its findings (a perspective sometimes captured under the somewhat misleading label of ‘pragmatic intrusion’; see further later); but also, pragmatics needs linguistics to obtain the data required to prove its pragmatic points. 4

Interdisciplinarity

To take an example from speech act theory: here, the interdisciplinary emerges as the need to localize the speech act in space and time, to ‘contextualize’ it in the widest possible sense of the term: local and temporal (sequential). Likewise, in conversation analysis, the conversational structures need to be pragmatically grounded in the societal relations, rather than having these relations emerge, automatically as it were, in the neo-constructivist conception of society adopted by a number of conversation analysts. In addition, a need recently has arisen to re-orient the debate on the semantics–pragmatics interface, which up to now mostly has been informed by a linguistically based, exclusively semantic approach. In a wider perspective, not to be pursued in the present context, scientific practice in general illustrates the need of interdisciplinarity (e.g. by its use of metaphoric thinking in biology, physics or even mathematics). What follows here is an attempt to assign the distinction its proper place within pragmatics, and (among other things) to point out its relevance to the ongoing semantics–pragmatics debates.

Of string players and bricklayers: immanence and identity A question that has pervaded linguistic debates from their very start is the problem of ‘immanence’, understood as the requirement to consider language as a scientific object, presenting itself ‘as a self-sufficient totality, a structure sui generis’ (Hjelmslev [1943] 1993: 7; 1953: 2). On the surface, this statement seems to exclude any influences from, or relationships to, other disciplines. Louis Hjelmslev in fact explicitly excluded the study of what he called ‘the disiecta membra of language’; by contrast, for Hjelmslev, ‘linguistics [is] the study of language and its texts as an end in itself ’ (Hjelmslev [1943] 1993: 7; 1953: 2). What is not clear from this quotation was the degree to which the author’s ideas themselves were influenced by other disciplines (such as philosophy, the natural sciences, among these what used to be called ‘natural philosophy’, aka physics) and how his theories show a clear relationship to those other fields of scientific endeavor. And in this sense, the very ‘immanence’ that was postulated as a decisive criterion for any linguistic theory may turn out not to be all that sui generis. When I started my linguistic studies at the University of Copenhagen in 1953, the famous ‘New Testament’ of Danish structural linguistics, the original version of Hjelmslev’s Prolegomena (from which the foregoing quotations are taken), was just ten years old (Hjelmslev [1943] 1993). To me, coming from a more traditional form of linguistics, commonly known as ‘philology’, Hjelmslev’s approach was as startling as it was inspirational. For the first time, I met a scientist who endeavored to combine results from other disciplines with his own efforts to create a modern philology, a linguistics not bound up with earlier generations’ presentations and representations of linguistic facts. Hjelmslev (1899–1965) was clearly influenced by the thoughts and concepts promoted by the Neopositivist Vienna School, in Denmark represented by the philosopher Jørgen Jørgensen (1894–1969), with whom Hjelmslev kept close contact; thus, the latter’s approach to the science of language was clearly a case of ‘cross-disciplinary’ pollination. As noted earlier, Hjelmslev’s ambition to constitute linguistics as an ‘immanent’ science, a realm of inquiry that had no other responsibility than to make sense in itself and by itself, showed its cross-disciplinary origins by appealing to the three famous principles of non-­contradiction, completeness, and simplicity (Hjelmslev [1943] 1993: 17–19; 1953: 10–11), which adherents of the Vienna School, such as Moritz Schlick (1882–1936; see e.g. Schlick [1936] 1979; for a critical evaluation, cf. Mey 1951), had recognized for a decade. While the need to secure the integrity and independence of the discipline was a cornerstone of Hjelmslev’s theorizing, his dependence on the Neopositivist doctrine gainsaid his pronouncements declaring the need for an ‘immanent’ linguistic research. Ironically, and using a somewhat anachronistic term, his 5

Jacob L. Mey

method could be dubbed a mere ‘heuristic technique’, a notion that would become popular especially after the appearance of Chomsky’s early works (1957). Methodically, Hjelmslev’s was an analytic approach: starting from the total text, he divided and subdivided the textual whole into lesser and lesser units, bound together by relationships of presupposition and other dependencies, to arrive at a complete inventory of what he called ‘glossemes’ (the theory was called ‘glossematics’, and as such it became quite influential, though mostly locally, for a couple of decades – roughly, until Hjelmslev’s untimely death in 1965). By analyzing the language as a complex array of syntagmatic concatenations and paradigmatic categories, glossematics purported to give a semantically neutral, language-independent characterization of any possible linguistic phenomenon, even though the initial operation (dividing the linguistic ‘chain’) was based on the actual occurrence of linguistic elements in a text, consisting of larger and smaller concatenations, bound together by selection and other textual (inter) dependencies. Linguistic categories were not a priori ‘givens’, or (for any particular language) borrowed from other languages; they would arise automatically and spontaneously, when the linguists operating on the strings of words performed their analytic task correctly. When doubts arose as to the legitimacy of a particular linguistic category of individual descriptive items, one had recourse to the textual chain again, by devising a maximally exclusive, yet minimally contrastive operation of ‘commutation’: the item in question was confronted with a number of possible contexts, in which it could or could not function; this is the celebrated ‘commutation test’, the keystone of any glossematic analysis. The commutation test, in general, is defined as a procedure by which a difference in the one ‘plane’ of the language (the expression plane) is related to a difference on the other plane, that of the content: ‘There is a difference between invariants in the expression plane when there is a correlation (e.g., the correlation between /e/ and /a/ in pet-pat), to which there is a corresponding correlation in the content plane; and vice versa for content invariants’ (Hjelmslev [1943] 1993: 59; 1953: 41). Towards the end of my doctoral studies, I became interested in another version of formalized linguistics, the nascent generative-transformational theory of Noam Chomsky (1957). Here, another kind of ‘string playing’ was going on: in Chomsky’s system, the initial partition and subsequent sub-partitions of the linguistic chain, or ‘string’, followed the rise of abstract grammars, developed in a computational environment. Chomsky applied one of these, a so-called context-free grammar, to the well-known linguistic model of ‘immediate constituent grammar’. Chomsky’s formalization, called ‘generative grammar’, was essentially a context-free grammar, supposed to ‘generate’ the basic linguistic strings traditionally associated with constituent grammar (now called ‘phrase-structure grammar’). Chomsky’s work thus provides another typical example of cross-disciplinary influence (here: from mathematics and computer science) on a humanistic area of investigation (here: linguistics). And even if Chomsky himself repeatedly denied any computer-related influence on his works, the outcomes have clearly confirmed the effects of this cross-pollination, as seen in the universal adoption of the ‘Chomsky hierarchy’ (Ginsburg 1975) of formal languages by the computer science community, and its subsequent role in the development of computer languages proper and its theoretical counterpart, automata theory. Whereas both Hjelmslev and Chomsky still kept an eye on the realities of the language they wanted to analyze, others were having recourse to more independent formalizations, where the ‘linguistic algebra’ (for Hjelmslev, the endpoint of his endeavors) was instead taken as a point of departure. In ‘categorical grammars’ (such as originally devised by the Polish logician Kazimierz Ajdukiewcz [1890–1963], and further developed by the Canadian mathematician Joachim Lambek [1922–2014] and the Israeli philosopher Yehoshua Bar-Hillel [1915–1975]), the first step is an arbitrary assignment of symbolic values to linguistic items, which then are combined, 6

Interdisciplinarity

following consecutive, explicit rules of ‘unification’, to result in a string that can be inserted as representing a full sentence. The movement of the investigation here is ‘bottom up’, rather than ‘top down’: one starts with the individual linguistic items, assigning them a categorical value, and threading them together to finally arrive at ‘Sentence’ (rather than expanding a Chomskyan top-symbol such as ‘Sentence’ into smaller and smaller constituents). This is why I call these latter cross-­disciplinarians ‘bricklayers’ rather than ‘string players’ – but in the end their common problem remains the same. The strings or structures they generate can only be identified with ‘real brick’ data by applying the results of the analysis to concrete phenomena. Similar considerations apply to the most famous of the ‘model-based’ linguistic theories, the one known under the name of ‘Montague grammar’, a self-contained linguistic algebra named for its inventor, the US logician and philosopher Richard Montague (1930–1971). Montague grammar is intended to be ‘universal’, that is, to be valid across the board for all linguistic domains, syntax, semantics, and pragmatics. From a pragmatics point of view, the main problem is that such a ‘universal grammar’, that is, a grammar that in principle combines all rules of grammar valid for any particular language, is solely based on the way linguistic elements correlate in the abstract (‘grammar without content’). In reality, however, people use grammars and grammatical rules in many different ways, and the criteria for ‘correct’ use of language vary from community to community, from language to language. When a grammar is descriptively adequate (as in the systems projected by linguists and philosophers such as Hjelmslev, Chomsky, and Montague), the question of how people handle these ‘adequacies’ is still open. Noam Chomsky tried to capture this cleavage by introducing the notions of ‘competence’ vs. ‘performance’, but that distinction too has come under fire as too general and, in fact, itself based on abstraction notions. And as to ‘universal grammar’, it is still among the most hotly debated and least successfully pursued projects in present-day linguistics.

Saussure’s express trains and the identity problem Ferdinand de Saussure, by many considered the father of modern linguistics, used the now famous image of an express train to clarify his thoughts on language and how it works. Talking about how trains function within a fixed schedule, he remarks that we establish identity with regard to, for example, ‘the Geneva–Paris express at 8:45 p.m.’, whose trains leave Geneva regularly every evening at a specified time and arrive at Paris in the morning, likewise at a preset time of day (Saussure 1949: 151). In this connection, it does not concern us, says Saussure, that the trains probably are very different indeed in make-up and population: the passengers, the engine, the wagons, the engineer and the rest of the train crew, everything and everybody may vary from day to day, yet we say that it is the ‘same’ express. This notion of ‘identity’, which is crucial to Saussure’s thinking on language, contains one grave error (its Aristotelian prôton pseûdos): it omits the human factor. The people using and working the trains during various days are mostly different; as to the passengers, they use the trains for different purposes and do very different things while on the train (some take a nap, some relax or sleep on a couchette or even in a real bed, some eat and drink, or do their work, and so on). All this is of no importance for Saussure: his trains abstract from every facet of human activity (even the people operating the trains, although different, are considered ‘the same’, that is to say, one could replace them with other people without the express in the least being ‘another’). For the purposes of running a railway (which includes scheduling, accounting, ticketing, maintenance, passenger services, and so on), the ‘differential identity’ of the express may be a necessity of operation. But to transpose this practical thinking to the domain of language 7

Jacob L. Mey

use conceals an important fact: humans, unlike trains, are not repeatable incidents. And their language shares this fate. Another famous image, dear to Saussure, is that of the game of chess, whose individual pieces possess their identity exclusively by force of their difference and of their placement on the board (1949: 125–126). Even so, Saussure has to acknowledge that, in order for the parallel to work perfectly, ‘for the chess game to be similar in every respect to the language game, we would have to presuppose a player without awareness or intelligence’ (1949: 127; my translation) – a perfect illustration of the way Saussure’s analogy neglects the human activity involved in the comparison. The Russian philosopher and semiotician Mikhail Bakhtin (1895–1975) has in many of his works taken a critical stance towards the Saussurean view of language, in particular Saussure’s notion of identity as it was (cross-disciplinarily) borrowed from the discourse of physics. An experiment in physics must be repeatable to be valid; one must be able to replace all its constituent components with their corresponding replicas and obtain the same results. By contrast, language is essentially non-repeatable; Bakhtin insists that one has to separate out the ‘replaceable’ elements that make up a sentence, for instance, from the non-repeatable use we make of these elements. In addition, a sentence is not an utterance; thus, a speech act such as ‘asserting’ is not a true act until it is ‘acted’, that is to say, uttered. ‘Sentences are repeatable. Sentences are repeatable,’ as Gary Morson and Caryl Emerson iconically remark in their important study of Bakhtin’s work; yet utterances are not: ‘Each utterance is by its very nature unrepeatable. Its context and reason for being differ from those of every other utterance, including those that are verbally identical to it’ (1990:126). And the reason is that utterances are produced in real time, by real speakers, in real contexts, rather than in the comfortable armchair environment of most linguistic thinking, including Saussure’s, Chomsky’s, and Hjelmslev’s. The pragmatic contexts are never the same. But why did I call this the prôton pseûdos of structural linguistics? Because its forced identities do not function as a true model of our thinking. Having ‘crossed over’ from another area (hence being ‘cross-disciplinary’), the model is alien to a humanistic discipline such as the study of human language use, and therefore does not belong there. The notion that elements in a system have to be identified and differentiated according to their idealized systemic functions did not originate in the human sciences; neither did the postulate that human behavior should be modeled on the pattern of physical occurrences, to be quantified, measured, contrasted, and identified in accordance with abstract characterizations. To take an example: the notion of a linguistic item’s ‘context’, as it is usually defined and used, takes aim at what is present in the linguistic chain adjacent to the particular item (representing a certain phonological, morphological or syntactic category). Certain items are found only in particular contexts (a preposition needs to co-occur with a substantive), and conversely: the occurrence of a particular item may signal the need for a corresponding context (a preposition triggers the need for, ‘presupposes’, an inflected substantive or other item that can be ‘selected’, as Hjelmslev used to call it [1943] 1993: 23–25).This kind of ‘immanent’ thinking about identity in language (as it is also reflected in Hjelmslev’s commutation test, discussed earlier) is a direct offspring of the physicalist demand to establish identifiable and measurable quantities of description; as such, it has been instrumental in helping linguists to identify and classify the sounds of languages and their grammatical units. But it has also forced linguists to adopt a model in accordance with which not only the grammatically acceptable sentences, but human utterances in general are considered as similar, or even identical, to those physical occurrences which it is imperative to aggregate into systems uniquely defined by ‘identities and differences’, as was Saussure’s ideal (1949: 151) and remains the holy grail of structural linguistics until today. 8

Interdisciplinarity

Hybrid monsters? Bakhtin’s great merit (and before him, that of his colleague Valentin N. Voloŝinov [1895–1936 (?)]) was to introduce the idea of a ‘dialogue’ into what until then essentially had been a monologue, shaped in the head of the linguist. One cannot respond to a sentence, says Bakhtin, one can only respond to an utterance; that is to say that dialogue is the founding principle of all use of language. ‘Dialogue’ is to be taken here in the pragmatic sense of ‘words uttered by humans in a real-world context’, and not in the sense of calling up words or sentences and imagining a context in which those words or sentences could be uttered. As Morson and Emerson remark, ‘Linguists sometimes choose sentences as examples and then without taking account of their assumption imagine particular simple situations in which those sentences are utterances. Thus, they “smuggle into” the sentence characteristics of an utterance’ (1990: 126; the quoted expression is from Bakhtin’s essay ‘The problem of the text’; 1994: 123). To quote Bakhtin himself: A great many linguists . . . are held captive by this confusion [between sentence and utterance] and what they study as a sentence is in essence a kind of hybrid of the sentence (unit of language) and the utterance (unit of speech communication). (1994: 75) It is this ‘hybrid monster’ that has terrorized the study of language ever since the introduction of the ‘original lie’ and its physicalist ancestry. It has prohibited students of language to see human speech as an essentially communicative activity, to be studied on the premises of the ‘human condition’ in all its varieties and shapes. It was not until the philosophers got engaged in the business of language that one gradually notices a shift in interest and momentum: away from the purely descriptive, ‘pasteurized’ accounts of systems and structures to the more interesting question of ‘what one can do’ with language.

‘Waiting for the philosophers’: enter pragmatics In a mode different from the cross-disciplinary approaches outlined earlier, a new wave of interest in language originated among practitioners of other disciplines, first of all the philosophers. What the members of the ‘Ordinary Language Philosophy’ group at Oxford and Cambridge did in the 1950s was to approach philosophical problems using non-technical terms, eschewing the jargon and symbolisms that for many people had obstructed the understanding of these problems. Here, they joined up with earlier British philosophical views, such as the empiricist tradition originating with John Locke (1632–1704), who had maintained (in the ‘Prologue’ to his Essay concerning human understanding, 1689) that our intellect was enough of a ‘candle’ to by itself illuminate even the darkest recesses of our minds, including the cobweb-infested areas of language norms and linguistic prescriptions, where method often took a front seat to content and aim. This recent, truly interdisciplinary approach to language studies is first and foremost connected to the name of John L. Austin (1900–1958), whose posthumous but epochal work appeared in 1962 under the intriguing title of How to do things with words – as much a declaration of principle as a notification of content.The emphasis has shifted from immanent description of languages for the sake of scientific exactitude (compare Hjelmslev’s ideal of a ‘linguistic algebra’) to looking at language as a tool for people to realize their everyday objectives. However, it seems fair to say that the success of speech act theory (as it came to be called after Austin’s emphasis on ‘doing’ things, ‘acting’, with words) would not have been as spectacular 9

Jacob L. Mey

as it eventually turned out but for the work of further generations of interdisciplinary thinkers (including not only philosophers, but also anthropologists, psychologists, and sociologists), who rethought, refined, and developed Austin’s ideas into a more or less consistent body of thought. Philosophers such as John Searle, Herbert Paul Grice (1913–1988), or Robert Stalnaker, working within the burgeoning framework of speech acts come to mind (see Searle 1969; Stalnaker 1977). From psychology, seminal contributions are found in the works of James J. Gibson (1904–1979), ‘the seer of Ithaca’, as one reviewer of his 1979 work called him (Restle 1980). Basing himself on his work on visual perception, Gibson studied what he called ‘affordances’, that is, the ways the mind pre-acts, and in fact pre-establishes, what the eye can see. In anthropology, a figure such as Stephen Levinson looms large, not least through his ground-breaking Pragmatics (1983), his subsequent work (together with Penelope Brown) on phenomena of politeness in language use (1987) and, prominently, his magisterial development of the complex cluster of problems connected with Grice’s concept of ‘implicature’ (that is, the way we understand what is implied, but not mentioned or expressed directly, in the words of an utterance; Grice 1989; Levinson 2000). A further interdisciplinary approach has its roots in the sociology of speech; Harvey Sacks (1936–1975) and his (often posthumously aligned) collaborators worked out a new way of analyzing the phenomenon of human conversation, by Levinson dubbed ‘the matrix of language acquisition’, that is to say, human talk-in-interaction seen as the fertile origin of our existence as communicating, socially bonded individuals (1983, chapter 6, esp. p. 284). By paying attention to what people really say when talking (and not what they are supposed to be saying, according to the prescriptive linguists), these interdisciplinarians introduced a new paradigm called ‘Conversation Analysis’, which also incorporated ideas originating in contemporary sociology, such as that the facts of sociality, including language and its use, were not ‘givens’, in the sense of Platonic ideas, but had to be constructed ‘on the spot’, so to say, by the participants in the conversational action. From the linguistic camp, the reactions to all this were initially rather mixed, but eventually a pattern emerged also here: attention came to be paid to the ways the semantics of a language related to what was being said in conversation and other daily practices. Tying in with Austin’s notion of the ‘total speech act situation’ (1962: 8) of language use as the proper context for delimiting and describing speech acts, the earlier notion of ‘pragmatics’ (due to philosophers such as Charles S. Peirce [1839–1914], William James [1842–1910], and Charles Morris [1901– 1979]) was instated as a full member of the triadic ‘canon’ of syntax-semantics-pragmatics. The situation of speech acting was conceived of as this ‘total context’ in which language was being used; pragmatics now had its first investiture as a humanistic discipline. From the very beginning, pragmatics has emphasized the importance of the language user in all thinking about and dealing with language. In an interdisciplinary approach, pragmatics relies on and calls on the various sciences of human behavior to be incorporated in the quest for this total context; semi-autonomous disciplines dealing with human linguistic behavior, such as psycholinguistics and sociolinguistics, are now partners in the pragmatic project inasmuch as they deal with the language user from their respective stances in psychology or sociology. In recent linguistic studies, the renewed emphasis on semantics as opposed to syntax as the major focus of linguistic inquiry that had started already in the 1960s with the ‘defection’ of a fair number of young generative linguists from the Chomskyan orthodoxy – names such as John Robert (‘Haj’) Ross, George Lakoff, and James McCawley (1938–1999) come to mind – turned out to be more than an ephemeral phenomenon. Even before his move to Berkeley, Lakoff had started a new trend in language studies, in which he incorporated cognitive phenomena such as metaphor, earlier neglected or mistreated by most linguists and philosophers alike (see Lakoff and Johnson 1980). His ventures into what was to become ‘cognitive linguistics’ constitute a 10

Interdisciplinarity

prime example of interdisciplinarity, also against the backdrop of the reception of his ideas across the field of human behavioral studies; philosophers and linguists, sociologists and psychologists all started to work within the new paradigm, adopting the most recent advances in the individual disciplines and joining them under the umbrella of ‘Cognitive Science’. Also in the field of pragmatics itself, many related approaches were seen to emerge (politeness research is just one of them; literary pragmatics another). A further development occurred when the theory of ‘pragmatic acts’ was born: it is based on the idea that no ‘act’ in itself carries meaning, but that it has to be situated in Austin’s total context (1962) to obtain its significance and validity. Pragmatic acts subsume speech acts, but they also incorporate interdisciplinary aspects from text theory, sociolinguistics, psycholinguistics, and even phonetics and phonology. In addition to the initial insistence on ‘location’ as a defining aspect of the situation, more recent developments have stressed the ‘sequentiality’ of pragmatic acting, by which a particular act gains full acceptance only through its development over time (or conversely, may lose its ‘power’ due to temporal restrictions or chronological developments; more on this at the end of the next section). From what we have seen in this section, it ought to be clear that pragmatics, being essentially an interdisciplinary undertaking, has certain problems demarcating itself against the surrounding disciplines (like semiotics, anthropology, sociolinguistics, psycholinguistics, conversation analysis, and so on). For many linguists, pragmatics is a part of a semiotically oriented linguistics (compare the well-known tripartition, due to Charles Morris, of semiotics into semantics, syntax, and pragmatics). For others, pragmatics is what we invoke when linguistics, especially linguistic semantics, fails to deliver the descriptive goods; since semantics ought to be able to stand on its own legs, some consider pragmatics an unwelcome but unavoidable ‘intrusion’: a disruption of the semantic peace, yet indispensable in the whole picture of language. And finally, and this is my position, pragmatics seeks to incorporate the entire situation as it reflects itself in language and in other forms of human social behavior; linguistics, inasmuch as it tries to describe this linguistic behavior, is thus a part of pragmatics, and not conversely, as it is often maintained. (See Mey 2013a for more details on how pragmatics developed from its [linguistic and other] beginnings; see also Mey 2012, and the short chapter 2 in Mey 2001.)

Cerberus and the ‘King of France’ The American philosopher Willard van Orman Quine (1908–2000) has (in his book Methods of Logic, 1958) an interesting discussion of a mythological figure, Cerberus. The point he makes is that, although we know Cerberus does not exist or has ever existed, being a figure from Greek mythology (the hell-hound guarding the gates of Hades), we still may speak meaningfully about Cerberus and make consistent (non-contradictory, exhaustive, and simple, in Hjelmslevian parlance) remarks about the beast. Quine observes that ‘there happens to be nothing such as the word “Cerberus” purports to name, near or remote, past, present, or future’ (1958: 198). The catch here is in the word ‘name’: if, for Quine, ‘naming’ means pointing to an object that exists in the ‘real’ world of things, then clearly ‘Cerberus ’ does not name anything. Another favorite example philosophers use in this connection is the ‘King of France’, a stalwart ever since the days of Bertrand Russell and Peter Strawson.We cannot say, for example, that ‘The present King of France is bald’: there is no current king of France; hence the expression ‘King of France’ by itself does not name any existing object. Even so, it makes sense to speak about and refer to such objects as ‘Cerberus’ or the ‘King of France’. In philosophical terms, they ‘denote’ something, even if what they denote does not exist, except in the usage of the speakers. And, as Recanati remarks, commenting on the problems involved in the acquisition of vocabulary, ‘How do language learners find out what the words they hear denote? They have to 11

Jacob L. Mey

rely on context and pragmatic factors’ (2010: 7) – factors which include, but by no means only or even mainly, ‘ostensive’ procedures such as pointing – which in our case, would clearly make no sense, unless we were talking about a picture of a Cerberus, found on a Pompeian fresco, or if we (with regard to a certain French king) were referring to Hyacinth Rigaud’s famous painted representation of Louis XIV in the Louvre. But even in the absence of proper ‘ostension’, it behooves us to examine more closely what could be meant by those ‘pragmatic factors’. Pragmatics studies language in use, and in particular the users of language and how they go about using it. As the earlier cited W.V.O. Quine remarks (in the revised edition of his perhaps most influential work, Word & Object), ‘language is a social art’ (2013:7). That implies that we, in our search for meaning, are bound to rely on what the users, as social beings, are doing in the course of their everyday linguistic activities. In other words, the meaning of what we say cannot be deduced from some abstract principles, or explained in philosophical terms alone; our semantics needs pragmatics in order to function. In some of the recent literature, this has been discussed fact under the label of ‘pragmatic intrusion’ (see e.g. Levinson 2000; Recanati 2010): pragmatic considerations are felt to be ‘intruding’ into the domain of semantics; and, as with most intruders, they do not seem to be welcomed with open arms. On the contrary, the fact of ‘intrusion’ has been explained away by reducing it to some variety of semantics (or ‘pragmantax’; see Levinson 2000: 192–193), or by having it formally represented – using formal models of language use based on set theoretical principles and methods from predicate logic – in the text itself (the ‘discourse’, as in Kamp and Reyle’s ‘Discourse Representation Theory’; 1993). Turning away from these linguistic-philosophical characterizations of the meaning of abstract sentences or propositions, let us take the case of pragmatic acts vs. speech acts, mentioned in ‘Waiting for the philosophers’: Enter pragmatics, and consider how pragmatic acts depend on their contexts of use. Elsewhere (Mey 2001: 219ff), I have argued that there are, strictly speaking, no speech acts as such; the only ‘acts’ performed, using language, happen to be performed in a concrete situation of user activity (as actually already pointed out by John Austin in the 1950s, and later by Levinson in an early article; 1979).This actual situation of speaking is often thought of as anchored locally, in space: it has to do with the speakers’ and hearers’ being located in a particular societal environment, with all of its affordances for action: openings as well as limitations.This context of use (taken in its widest possible interpretation) covers the personal and social conditions of language users, and the ways that users have been conditioned through socialization into one particular, or sometimes several, societal complexity/ies.The way we speak reflects the ways in which we are part of those complex social wholes; it presupposes knowledge not only of ‘ways of speaking’, but also of ways of acting; in a pragmatic perspective, all these features combine interdisciplinarily. The pragmatic act is thus seen as a complex whole linking together elements from various, varying (and sometimes conflicting) societal, economic, geographical, religious and other spheres, in ever-changing but even so remarkably stable patterns of speaking and behaving, often loosely referred to as ‘the culture’. An aspect of this cultural context that is frequently overlooked is that of time – a fact which in itself is amazing, given that both time and place are among the most universal categories of the human experience, as reflected in linguistics and language philosophy. That time has an impact on how we speak and perceive speech is not only a matter of synchronizing our interchanges in orderly conversation (following the Gricean maxims) or respecting conversation-analytic rules for turn-taking; it touches on the very nature of our speech activity as informed not only by its local contexts, but also, and important, by its temporal anchoring – that which I have called the sequentiality of our acting. A presumed sexual innuendo of an invitation such as ‘Do you want to go to bed?’ (Mey 2013b) is not construable from conversational principles or speech act maxims; 12

Interdisciplinarity

it only realizes its pragmatic potential through the ensuing activity. In other words, it has to be judged as occurring, or having occurred, on the basis of what comes after: it is interpretable only in sequentiality. The speech act in question changes its character depending on the outcome of the interchange (Mey 2013b: 266; see also, where I mention another illustration, taken from the domain of flight control). So, while it is true, as Recanati observes, ‘what an expression means at time t in a given community depends on the history of its uses before t in the community’ (2004: 445), the statement should be expanded by extending the temporal dimension to comprise also the future: history is not only what has happened; it includes what is going to happen. Given that the future once was past, the future also is the only historically valid instrument that we have at our disposal for evaluating our acts. In flight control (Nevile 2006), a pilot has to know what is being said by the control tower and the co-pilot in order to ratify it and execute a particular maneuver; but also, the ratification itself is part of the process: a maneuver which cannot be completed unless ratified. In other words, Recanati’s statement should include a reference to what comes after ‘time t’: what an expression means at time t in a given community depends on the history of its uses not only before, but also after time t. Such a ‘revisionist’ conception of speech acting naturally relies heavily on interdisciplinary input: it is always hard to make predictions, especially when it comes to future happenings (as the Danish humorist Robert Storm Petersen once expressed it) – but it is not just hard, but well-nigh impossible, without having recourse to the other disciplines dealing with human behavior, both locally and globally, both pre- and post-historically. Conversation analysts were among the first to realize the potential of sequential acting for describing and characterizing the present and the past of language as spoken, of language in use. A conversational turn, by definition, never stands alone: it derives its meaning from the context and (as was the case in the flight control example) from the subsequent ratification of the interchange. By ratifying their turns, the pilots-conversationalists ‘co-construct’ the interchange: together, they actively and collaboratively build up the conversational activity in real time. Pragmatics, by appealing to the sequentiality that plays a prominent role not only in Conversation Analysis but also elsewhere (it was shown, inter alia, for the pragmatics of reading in Mey 2000 and 2013b), sets out to capture the interdisciplinary and dialectic momentum that is characteristic not only of conversation, but of speech acting in general, by appealing to the notion of the ‘pragmatic act’ as the basic element in our societally based and -exercised linguistic activities. As I have stated elsewhere, [T]he theory of pragmatic acts does not try to explain language use from the inside out, that is, from words having their origin in a sovereign speaker and going out to an equally sovereign hearer (who then becomes another sovereign speaker, and so on and so forth). Rather, its explanatory movement is from the outside in: the focus is on the environment in which both speaker and hearer have their affordances, such that the entire situation is brought to bear on what can be said. . . ., as well as on what is actually being said. (Mey 2001: 221)

The bottom line: inter- vs. cross-disciplinarity In all talk about inter- vs. cross-disciplinarity, the main feature that stands out is that of the context of the situation. This notion, originally developed in anthropology and sociology (an early reference being Malinowski 1923), is the guiding lamp by whose light we may identify the beneficial 13

Jacob L. Mey

effects of being inspired by other disciplines’ varying contexts, as against being dominated (not to say overrun) by the alien, and often alienating, traditions of thought originating in a mostly physicalist-oriented mindset. Unfortunately, the latter seems to have been the case in much of the humanities during the twentieth century, following the spectacular successes the ‘exact’ sciences had been able to show for themselves during the preceding century and a half.The popularity of the methods behind those successes also gave rise to an unhealthy tendency towards a universal, one-size-fits-all, ‘unified’ scientific methodology, by which quantitative measuring and abstract calculation were supposed to be the only valid and valuable ways to assay human progress. The gradual encroachment of the physicalist mindset on our thinking started as early as the period when the first ‘great discoveries’ of science came to the scientific and popular fore in the late 1700s to the early 1900s. People rushed to identify with this newfangled pattern of thinking, originally still conceived of as a ‘natural philosophy’ (the earlier name for the science of physics). At the same time, the notion that only that which we can express in quantities and can measure, obligatorily converting all human knowledge to a numerical representation, went viral, as we would say today. An early voice that could be heard in favor of an alternative way of looking at what came to be caught under the universal label of ‘progress’, a voice that wanted to combine scientific insights with an appeal to ’wonder’, conceived of as the ability to see beyond numbers and statistics, was that of the Scottish philosopher, satirical writer, essayist, historian and teacher Thomas Carlyle (1795–1881). His serious satire Sartor Resartus (1838) creates the persona of a German professor with the curious name of ‘Diogenes Teufelsdröckh, “Professor der AllerleyWissenschaft”, or, as we should say in English, “Professor of Things in General” ’ (Carlyle 1930: 14; one notes that nomen est omen here: the name ‘Teufelsdröckh’ is a slightly masked variant of the German Teufelsdreck, meaning ‘satanic shit’, so the professor/author aligns himself here with the traditional role of the diabolus, the ‘confuser’, messing up things with disrespectful verbal eliminations). Accordingly, Carlyle’s professor exhibits a truly ‘devilish’ delight in unmasking his contemporaries’ flight into abstruse formulations and high-falutin terminology (a certain Jena professor comes to mind), and he does this through the intermediary of developing a ‘philosophy of clothes’ (Carlyle 1930: 43; hence the title of the work: Sartor Resartus, in English, ‘The clothier re-clothed’). ‘Are we Opossums; have we natural Pouches, like the Kangaroo?’, asks the professor rhetorically; Carlyle 1930: 55); our specific ‘unspeakable differences’ (Carlyle 1930: 54) come not from our bodily composition (which is the same for all humans), but from the ways we can create and put on clothing; for how else, lacking a natural pouch, would we cope with the absence of an organ of such importance as ‘the master-organ, soul’s seat, and true pineal gland of the Body Social: I mean a purse?’, asks the ironic author (Carlyle 1930: 55). It is the dialectic interdependence between the naked body and its clothing that makes this line of thinking relevant to our present purpose: that of assigning the interdisciplinary elements of the context their proper roles and places. Not only do the clothes constitute our bodies’ immediate context, it is also the case that they in a sense ‘make’ the body, and inform the human person: Kleider machen Leute (‘Clothes make people’), as the nineteenth-century Swiss author Gottfried Keller (1819–1890) expressed it in one of his best-known novellas (Keller 1874). But there is a subtle dialectic at work also here: clothes do make the person, but at the same time may reduce him or her to an Olympia, a mere displaying device à la the moving and speaking doll in Jacques Offenbach’s opera Les contes d’Hoffmann (1880). In Carlyle’s words: ‘Clothes gave us individuality, distinction, sociability; Clothes have made Men of us; they are threatening to make clothes-screens of us’ (Carlyle 1930: 33). The (active as well as passive) ‘wonder’ of clothes, as also the wonder of the context, consists for us in keeping in mind not only the conditions 14

Interdisciplinarity

under which a particular item of clothing is used, a particular utterance-in-context is delivered. While keeping an eye on the wearers or utterers, on their garments or words, we should in equal measure attend to the ways the clothing is acquired and worn in society, the utterances are produced and received in the community of speech, and to the possibilities and limitations of these processes of production and reception. What Carlyle and Keller unconsciously prelude upon is the very materiality of our thought and speech – a materiality which in its essence is dialectic, such that the clothes and the speech indeed ‘make Men of us’. Men (read ‘humans’) are then not thought of as immutable and impermeable (in the guise of ‘clothes-screens’), but rather as participants in an interactive process of communication – an idea which in the end is indebted, for its conception and propagation, to the very German Herr Professor against whom Carlyle/Teufelsdröckh aim their satirical ordnance! To return to the actual situation of utterance, it is exactly Gibson’s aforementioned contextual ‘affordances’ that carry the day here. Interdisciplinary thinking confers on these affordances their status of participatory elements in the communicative process, in addition to, and sometimes subsuming or superseding, the purely verbal manifestation of, for example, a particular speech act such as a ‘promise’ (see Mey 2001: chapter 8 on this). Carlyle’s ‘wonder’ does not demand the abrogation of the scientific methods of ‘Mensuration and Numeration’, themselves duly venerated by his Herr Teufelsdröckh (Carlyle 1930: 57); however, it opposes the unique and exclusionary attention paid to these methods, and their defenders’ mostly shallow philosophy and ‘Attorney-Logic’ (Carlyle 1930). Science should not ‘proceed in the small chink-lighted, or even oil-lighted underground workshop of Logic alone’; we should beware of the ‘Logicchoppers, and treble-pipe Scoffers, and professed Enemies to Wonder’ (Carlyle 1930: 58). The ‘wonder’ of the situation and its context consists in integrating the various aspects of human behavior into the communicative act through a true ‘inter-’, rather than a misconceived ‘cross’disciplinary movement. As noted for the case of visual perception (Gibson 1979), what we perceive depends as much on what we can (‘afford’) expect to see, as on what we physically can be ‘tested on’ with regard to our seeing. Incorporating this dialectic (including Keller’s and Carlyle’s views) into our treatment of language-in-communication leads to greater insights in language’s social functions, and not just to more detailed and accurate, or even formalized, descriptions of selective portions of Carlyle’s ‘Garment of Thought’ by elevating it to what he thought should rather be called ‘the Flesh-Garment, the Body, of Thought’ (1930: 62). We can only achieve these insights by paying attention to the fact that language always is embodied in a concrete situation and derives its full meaning from there. Our interdisciplinary efforts at coping with Carlyle’s ‘wonder’ will produce worthwhile results only if we take the whole situation into account; this also denies other disciplines the right to impose on us a methodology not adapted to our needs, while excluding features that, though not ‘immanent’ to language considered as an abstract system, yet are indispensable in our dealing with language-in-use. It is our pragmatic duty to try and capture the various ways the users deal with language, even by transcending the narrow confines of their own context(s) along various interdisciplinary dimensions. Interestingly, and again interdisciplinarily, what Carlyle’s ‘clothing project’ implies is something that only re-emerged a century and a half later, in studies of impersonation, or ‘performance’ as a way of creating identities. If putting clothes on a human makes him into a ‘Man’ (and only uninterestingly and occasionally a ‘clothes-screen’), the natural question is: ‘So what if a woman puts on “her” clothes, does that make her into a “Woman”?’ And we could add: ‘How about a man performing as a woman by putting on “drag”?’ By putting on a woman’s clothes, by dressing up and performing as one, the drag queen becomes a woman in his own and others’ eyes, 15

Jacob L. Mey

irrespective of his ‘underlying’ sex. We live our lives in ‘performed identities’; it is the act of performing our sex that makes us gendered. (Which is precisely what Judith Butler has argued in many of her works; e.g. 1990). But such ‘impersonation’, as Douglas Robinson (2013: 68) points out, is typical of all performance, not just the sexual kind. Clothing oneself is performing an act of identity; but so is participating in a conversation, playing a game of tennis, singing in a choir or even carrying out menial tasks in the office. In society, we are always performing: ‘putting on acts’, as it is called, misleadingly and a bit downputtingly (as if there were something bad about the act); yet it is the most natural thing in the world, and the sooner a child acquires this technique, the better. We are not what we are, but what we perform as, and typically can get away with performing; as we go around on the world’s scene, ‘acting up’ or ‘acting out,’ we create our identities – provided always that we can get our audience to cooperatively confirm our acts (just as we saw it was necessary in the case of what we usually limit ourselves to, when dealing with these matters: our ‘speech acts’).

Conclusion Rather than using pieces of language material as help in constructing what is basically a mathematical, formal logic or computational discourse, whose ‘grammars’ are directly compatible with (and often derived from) systems of formal logic or automata theory, the general pragmatic ‘perspective’ (Verschueren 1999: 7) opens the field wide and takes in the total context of the situation, understood historically as not just what is present here and now, but including that which came before, as well as that which may be expected or predicted (or post-factually ratified) in an interdisciplinary (not to say ‘intergenerational’) view of the situational context and its affordances. This, at the end of the day, is the original object of linguistic pragmatics, defined as ‘the study of the use of language in human communication as determined by the conditions of society’ (Mey 2001: 6).

Suggestions for further reading Watzlawick, P., Beavin, J. H. and Jackson D. D. (1967) Pragmatics of Human Communication: A Study of Interactional Patterns, Pathologies, and Paradoxes. New York: Norton. A seminal, but currently mostly forgotten work in early US pragmatics, which for the first time effectively established the discipline as interdisciplinary (the authors are a psychologist, a linguist, and a medical doctor, respectively), also linking pragmatics to fields such as interaction studies, literary criticism, philosophy, and psychological/psychiatric practice, all richly illustrated with excerpts from their own practice, domestic and foreign literature, and professional interviews. (The famous tenet that ‘One cannot possibly not communicate’ is due to Watzlawick; chapter 3 of the book). Thomas, J. (1995) Meaning in Interaction: An Introduction to Pragmatics. London/New York: Longman. Easily the best and most reader-friendly short introduction to the field, also richly illustrated with wellchosen, often highly entertaining examples. The book reaches out to an interdisciplinary readership because of its reliance on making things understood in different frameworks, such as literary studies, police practices, interviewing, serendipitous encounters etc. Robinson, D. (2013) Introducing Performative Pragmatics. London: Routledge. This introduction to a new ‘performative’ methodology in linguistic pragmatics breaks away from the traditional approach which understands language as a processing device, and the brain as an internal computer. Instead, the author focuses on human agency, which is not restricted to language as we (think we) know it, but includes all sorts of ‘performative’ practices, easily understood and put into relation with language use in the pragmatic sense. Drawing on a wide spectrum of research and theory from the past thirty years in particular, the author, appealing to the interdisciplinary-oriented reader by invoking cases and examples from a wide variety of disciplines, presents a combination of ‘action-oriented approaches’ from sources such as John L. Austin, H. Paul Grice, Harold Garfinkel and Erving Goffman.

16

Interdisciplinarity Austin, J. L. (1962) How to Do Things with Words. Oxford: Oxford University Press. Austin’s posthumous legacy to the pragmatic community, which set in motion the theoretical interest in what he called ‘speech acts’. Performativity, the idea that words do something to the world, is a central theme, leading directly into the pragmatic conception of language as a ubiquitous tool in a multifarious societal perspective. Undoubtedly the work that first made linguistics into a societally relevant, interdisciplinary affair. Nerlich B. and Clarke D. D. (1996) Language, Action and Context:The Early History of Pragmatics in Europe and America, 1780–1930. Amsterdam/Philadelphia: John Benjamins. An indispensable read for anyone who wants to understand how pragmatics came to be the way it is practiced today – in fact, one could say: how it came into being in the first place. The book is a treasure trove for those interested in the interdisciplinary origins of pragmatics (often still thought of as a linguistic development, with semantics being the preferred/disputed module).The authors manage to provide a lively and extensively documented account of the ‘early years’ before most people even dreamt of ‘pragmatics’ as an independent, yet interdisciplinary discipline.

Note 1 I want to thank my clever sister-in-law, the late Ms. Karin Bangor of Seljord, Norway, for alerting me to the various facets of Thomas Carlyle’s inimitable and merciless wit.   A heartfelt thanks goes also to the editors of this volume for incisive remarks on earlier versions of the chapter which made me reconsider and rectify a number of my initial assumptions.

References Austin, J. L. (1962) How to Do Things with Words. Oxford: Oxford University Press. Bakhtin, M. M. (1994) Speech Genres and Other Late Essays, C. Emerson and M. Holquist (eds.), translated by V. W. McGee. Austin: University of Texas Press. Brown, P. and Levinson, S. C. (1978) ‘Universals in language usage: Politeness phenomena’, in E. Goody (ed.) Questions and Politeness: Strategies in Social Interaction. Cambridge: Cambridge University Press. 56–311. Butler, J. (1990) Gender Trouble: Feminism and the Subversion of Identity. London: Routledge. Carlyle, T. (1930) Sartor Resartus:The Life and Opinions of Herr Teufelsdröckh. London: Humphrey Milford for Oxford University Press. Chomsky, N. (1957) Syntactic Structures. The Hague: Mouton. Gibson, J. J. (1979) The Ecological Approach to Visual Perception. Boston: Houghton Mifflin. Ginsburg, S. (1975) Algebraic and Automata Theoretic Properties of Formal Languages.Amsterdam: North-Holland. Grice, H. P. (1989) Studies in the Way of Words. Cambridge, MA: Harvard University Press. Hjelmslev, L. (1953) Prolegomena to a Theory of Language, translated by F. Whitfield. Madison: University of Wisconsin Press. Hjelmslev, L. [1943] (1993) Omkring Sprogeoriens Grundlaeggelse. Copenhagen: The Linguistic Circle of Copenhagen. Kamp, H. and Reyle, U. (1993) From Discourse to Logic: Introduction to Model-Theoretic Semantics of Natural Language, Formal Logic, and Discourse Representation Theory. Berlin/New York: Springer. Keller, G. (1874) ‘Kleider machen Leute’, in G. Keller (ed.) Die Leute von Seldwyla, 2. Braunschweig:Vieweg. Lakoff, G. and Johnson, M. (1980) Metaphors We Live By. Chicago: University of Chicago Press. Levinson, S. C. (1979) ‘Activity types and language’, Linguistics, 17: 365–399. Levinson, S. C. (1983) Pragmatics. Cambridge: Cambridge University Press. Levinson, S. C. (2000) Presumptive Meanings:The Theory of Generalized Conversational Implicatures. Cambridge, MA: MIT Press. Malinowski, B. (1923) ‘The problem of meaning in primitive languages’, in C. K. Ogden and I. A. Richards (eds.) The Meaning of Meaning. New York: Harcourt, Brace and World. 296–336. Mey, J. L. (1951) ‘On the criterion of intersubjectivity in Moritz Schlick’s works’, unpublished licentiate thesis, Nijmegen: Berchmanianum Philosophical Faculty. Mey, J. L. (2000) When Voices Clash: A Study in Literary Pragmatics. Berlin/New York: Mouton de Gruyter. Mey, J. L. (2001) Pragmatics: An Introduction. Oxford/Malden, MA: Blackwell. Mey, J. L. (2012) ‘A brief sketch of the historic development of pragmatics’, in K. Allan (ed.) The Oxford Handbook of the History of Linguistics. Oxford: Oxford University Press. 587–612.

17

Jacob L. Mey Mey, J. L. (2013a) ‘Across the abyss: The semantics-pragmatics interface revisited’, Intercultural Pragmatics, 10(3): 487–494. Mey, J. L. (2013b) ‘The well-timed speech act: Bourrée for Bache’, in S. Breunig, A. Holsting, C. A. Maegaard and N. Nørgaard (eds.) A Rule of Thumb: For Carl Bache. Odense: University Press of Southern Denmark. 265–278. Mey, J. L. (2015) ‘Pragmatics through the prism of society’, in A. Capone and J. L. Mey (eds.) Interdisciplinary Studies in Pragmatics, Culture and Society. Berlin: Springer. 15–41. Morson, G. S. and Emerson, C. (1990) Mikhail Bakhtin: Creation of a Prosaics. Stanford: Stanford University Press. Nevile, M. (2006) ‘Making sequentiality salient: and-prefacing in the talk of airline pilots’, Discourse Studies, 8(2): 309–332. Offenbach, J. (1880) Les Contes d’Hoffmann. Paris: Choudens Fils. Pedersen, H. (1924) Språkvetenskapen under Nittonde Århundradet: Metoder och Resultat [Linguistic Science in the 19th Century: Methods and Results]. Stockholm: Norstedt and Söner. Recanati, F. (2004) ‘Pragmatics and semantics’, in: L. R. Horn and G.Ward (eds.) The Handbook of Pragmatics. Oxford: Wiley-Blackwell. 442–462. Recanati, F. (2010) Truth-Conditional Pragmatics. Oxford: Clarendon Press. Restle, F. (1980) ‘Review of Gibson 1979’, Contemporary Psychology, 25: 291–293. Robinson, D. (2013) Introducing Performative Pragmatics. New York/London: Routledge. Saussure, F. de. (1949) Cours de Linguistique Générale, C. Bally and A. Sechehaye (eds.). Paris: Payot. Schlick, M. (1979) ‘Meaning and verification’, in H. L. Mulder and B. F. Van de Velde-Schlick (eds.), Philosophical Papers, 2. Dordrecht: Reidel. 456–481. Searle, J. R. (1969) Speech Acts: An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Stalnaker, R. C. (1977) ‘Pragmatic presuppositions’, in A. Rogers, B. Wall and J. P. Murphy (eds.) Proceedings of the Texas Conference on Performatives, Presuppositions, and Implicatures. Washington, DC: Center for Applied Linguistics. 135–147. Van Orman Quine, W. (1958) Methods of Logic. London: Routledge and Kegan Paul. Van Orman Quine, W. (1960) Word and Object. Cambridge, MA: MIT Press. Verschueren, J. (1999) Understanding Pragmatics. London: Arnold.

18

PART I

Methods and modalities Data collection

3 NATURALLY OCCURRING DATA Andrea Golato

Introduction All pragmatic research is based on some form of linguistic data, be it invented, elicited, observed or recorded (for a similar argument, see Jucker 2009). Of these, no one method of data collection and analysis is intrinsically better than any other. Rather, the data type, method of data collection and method of analysis adopted in a particular study should depend solely on the research question at hand (for similar arguments, see also Grotjahn and Kasper 1991; Bardovi-Harlig 1999; Turnbull 2001; Barron 2003; Félix-Brasdefer 2007: 159; Golato and Golato 2013). Thus, a researcher interested in language intuition, for instance, goes about data collection and analysis differently than a researcher interested in actual language use. This chapter discusses naturally occurring data and their use in pragmatics research that focuses on spoken language. It starts out by providing a definition of the term and a rationale for the use of such data. It then discusses the two ways in which naturally occurring data are typically collected, namely using field notes and recordings.

Naturally occurring data – what are they and why should they be used? A useful definition of what counts as naturally occurring data is provided by Potter, who suggests using the ‘(conceptual) dead social scientist’s test’. Potter asks: Would the data be the same, or be there at all, if the researcher got run over on the way to work? An interview would not take place without the researcher there to ask the questions; a counselling session would take place whether the researcher turns up to collect the recording or not. (Potter 2002: 541) As Potter clearly highlights, naturally occurring data are data that are not directly elicited by the researcher; instead, they are data that are observed. Each and every feature of the interaction (e.g., the context, underlying assumptions, linguistic realisations of a given speech event) is talked into being (Heritage 1984) by the participants of the interaction, whereas in elicited data 21

Andrea Golato

these features are predetermined by the elicitation instrument designed by the researcher (FélixBrasdefer 2007; Félix-Brasdefer and Hasler-Barker, Chapter 4, this volume). One important argument for using naturally occurring data is that experimenters may not predict all (or even the most common) situations in which a given speech event (e.g., a compliment or an apology) may be produced (Kasper 2000). Also, different situations may lead to different realisations of a speech event: for instance, a compliment used at the beginning of a meal leads to a different response than a compliment produced at the end of a meal (Golato 2005). Moreover, as Atkinson and Heritage (1984: 3) note: ‘The experimenter is unlikely to anticipate the range, scope, and variety of behavioural variation that might be responsive to experimental manipulation’. Several studies have compared naturally occurring data with data obtained through other elicitation methods, such as role plays, written or oral discourse completion tasks (DCTs) or questionnaires. Comparisons of naturally occurring data and DCTs have shown that the latter fail to capture how sequences are initiated and then unfold in the interaction (Beebe and Cummings 1985;Wolfson 1989; Beebe and Cummings 1996). Moreover, subjects produce utterances on DCTs that are different in terms of length, complexity, and linguistic realisation than those in naturally occurring data (Bodman and Eisenstein 1988; Hartford and Bardovi-Harlig 1992; Hassall 1997; Turnbull 2001; Golato 2003; Bou-Franch and Lorenzo-Dus 2008; Félix-Brasdefer 2010; Economidou-Kogetsidis 2013).This was also found when DCT data were compared with large-scale corpus data, specifically, the 5 million-word Cambridge and Nottingham Corpus of Discourse in English (CANCODE) collected between 1991 and 2001 (Schauer and Adolphs 2006). However, in a study of requests, Economidou-Kogetsidis (2013) has shown that data from written DCTs and naturally occuring data display similar trends in terms of directness and ­lexical/phrasal modification, thus indicating that ‘to a certain extent, the WDCT requests represent an approximation to the naturally occurring requests’ (Economidou-Kogetsidis 2013: 33, emphasis in the original). Comparisons between role plays and naturally occurring data have shown that role plays capture interactional details (e.g., turn-taking features, on-line planning, hesitations, hedges) and thus yield more realistic data than other data elicitation methods (e.g., Bodman and Eisenstein 1988; Kasper and Dahl 1991; Turnbull 2001; Félix-Brasdefer 2007; Huth 2010; Bataller 2013). However, role-play data are still only an approximation of actual conversation as they differ in significant ways from naturally occurring data in terms of length, complexity, and prosodic and other linguistic features (Hassall 1997; Félix-Brasdefer 2007; Bataller and Shively 2011; Baller 2013). For these reasons, researchers generally recommend naturally occurring data for analyses of actual language use (Turnbull 2001; Golato 2003; Félix-Brasdefer 2007; Bou-Franch and Lorenzo-Dus 2008; Bataller and Shively 2011). We turn now to the two methods of collecting naturally occurring data, namely to field notes and audio and/or video recordings.

Field notes Researchers take field notes on site either as observers or as participant observers. When they hear or observe a linguistic exchange of interest to their study, they note it down as soon as possible after it has occurred. Typically, information is also recorded about the context of the interaction (e.g., the participants, the location, intonation, any other noteworthy elements). Such contextual information may be indispensable in data analysis, depending on the research question of the investigator (Kasper 2000). A researcher may conduct the field observations on his or her own, but might also send a number of research assistants into the field who are trained with regard to both the phenomenon of interest and data collection methods. Overall, this method allows for collecting naturalistic data quickly and in a wide range of conversational contexts, and 22

Naturally occurring data

from a variety of sources (Kasper 2000). However, this data collection method is not without limitations, the most obvious one being that researchers have to rely on their memory when noting down the observed exchanges. Research (Labov 1984; Lehrer 1989) has shown that a person’s recall of prior utterances is limited in terms of both quantity and quality, and frequently does not include elements which were not tightly integrated into the original sentence structure (e.g., modifiers, intensifiers, hedges). In addition, the lack of recordings means that there is no opportunity for replay and consequently no possibility for other researchers to independently verify the occurrence of the utterance. A second problem associated with this method of data collection is that the researcher may note down only the most common or expected linguistic forms of a given speech act, and may fail to recognise (or recognise only later in the interaction) that a speech event of interest has taken place. As Kasper (2000: 319) states: ‘[T]here is thus a real danger that memorisation and taking field notes will result in recording salient and expected (or particularly unexpected) facets of the interaction, at the expense of less salient but perhaps decisive (often indexical) material’. Given these limitations, it seems that field notes might be appropriate to use when researchers are interested in the semantic content or situational context of a given speech event, but less appropriate when actual language use is the object of investigation.

Recordings Recorded naturally occurring data allows for exact and repeated analysis of the linguistic material, and for data verification by other researchers. Whenever interactants have visual access to one another, as in most face-to-face interactions and through web conferencing services, conversations are typically video recorded so that the researcher has the same access to utterances, gestures, gaze, and other embodiments as the conversational partners themselves. In rare situations, video recordings are not feasible, for instance in certain institutional settings where subjects fear that video recordings may reveal industry secrets (Golato 2003). In these settings, audio recordings are made. Otherwise, audio recordings are used for capturing telephone interactions or non-video web conferencing interactions.Typically, the researcher sets up the recording equipment (at times multiple recording devices are necessary to capture the entire setting) and then leaves the room. Note that in many countries, ethical practices suggest and/or law requires that the researcher seek written consent from subjects prior to any recording. Working with recordings of naturally occurring data is not without its challenges. As Labov (1972: 209) stated: ‘The aim of linguistic research in the community must be to find out how people talk when they are not being systematically observed; yet we can only obtain this data by systematic observation’. With respect to the present discussion, this means that the very presence of recording equipment may alter subjects’ speech production. However, if given even a short time to adjust, subjects tend to forget that they are being recorded (Kasper 2000). Depending on the speech event, it can sometimes take longer to amass a corpus of an appropriate size (Kasper 2000), and the time it takes to transcribe the recordings can be considerable (as can be the costs if one pays for transcription services). However, the increasing availability of public corpora for a variety of languages (for English corpora, see Lee 2010) allows for examining data from a diverse set of speakers hailing from a variety of backgrounds (cf. Weisser, Chapter 5, this volume). A further challenge is that certain speech acts may not occur frequently in interaction or may be difficult to obtain (Kasper and Dahl 1991; Billmyer and Varghese 2000). In the latter situation, it may take some extra effort to secure permission to audio-video record in certain settings. However, recordings of interactions in cars at car dealerships in France (Mondada 2009), in families dealing with terminally ill patients (Beach 2009) and of arguments between individuals (Jackson and Jacobs 1980) are just a few of the studies that have investigated data that might at 23

Andrea Golato

first glance be considered difficult to obtain. Beebe and Cummings (1996) have stated that it is difficult to systematically compare speech samples from different groups and different points in time. A number of studies since, however, have successfully compared speech events across cultures and groups, even at different points in time (Placencia 1997; Pavlidou 2000; Golato 2002; Taleghani-Nikazm 2002). Similarly, Shively (2011) has demonstrated that naturally occurring data can successfully be used to demonstrate language development over time in study abroad contexts.

Conclusion A major challenge facing researchers who work with naturally occurring data is a lack of control of speaker and context variables, such as age and/or social backgrounds of the speakers (Yuan 2001). Researchers for whom such controls are important may follow Turnbull’s (2001) approach to data collection: here, subjects are selected according to speaker variables and receive certain tasks to perform that likely involve the speech act of interest. Crucially, however, subjects are not made aware that their talk is the object of the investigation. For instance, when studying negotiation strategies, teams of subjects engaged in a computer game in competition with other teams, subjects were under the impression that their game moves were the focus of the experiment, whereas in reality the researcher was interested in their language use. While it may seem that obtaining data in the form of field notes and recordings is more involved than collecting data via elicitation techniques, these two methods of data collection yield the most naturalistic forms of data as defined earlier. Such data are vital when researchers are interested in actual language use, whereas other methods of data collection may be useful for different research purposes (cf. Golato 2003; Jucker 2009).

Suggestions for further reading Sidnell, J. and Stivers, T. (eds.) (2014) The Handbook of Conversation Analysis. Malden, MA: Blackwell. Chapter 3 presents factors to consider during the actual recording of naturally occurring data, while chapter 4 discusses transcription techniques. Tracy, S. J. (2013) Qualitative Research Methods: Collecting Evidence, Crafting Analysis, Communicating Impact. Malden, MA: Blackwell. This book provides a very detailed overview of the factors to consider when conducting field work. The following books provide overviews of various research methods in (applied) linguistics: Dörnyei, Z. (2007) Research Methods in Applied Linguistics: Quantitative, Qualitative, and Mixed Methodologies. Oxford: Oxford University Press. Litosseliti, L. (2010) Research Methods in Linguistics. London/New York: Continuum International.

References Atkinson, J. M. and Heritage, J. (eds.) (1984) Structures of Social Action. Studies in Conversation Analysis. Cambridge: Cambridge University Press. Bardovi-Harlig, K. (1999) ‘Researching method’, in L. Bouton (ed.) Pragmatics and Language Learning (Monograph Series, vol. 9). Urbana: University of Illinois at Urbana-Champaign. 237–264. Barron, A. (2003) Acquisition in Interlanguage Pragmatics: Learning How to Do Things with Words in a Study Abroad Context. Amsterdam/Philadelphia: John Benjamins. Bataller, R. (2013) ‘Role-plays vs. natural data: Asking for a drink at a cafeteria in peninsular Spanish’, Íkala, Revista de Lenguaje y Cultura, 18(2): 111–126. Bataller, R. and Shively, R. L. (2011) ‘Role plays and naturalistic data in pragmatics research: Service encounters during study abroad’, Journal of Linguistics and Language Teaching, 2: 15–50. Beach, W. (2009) A Natural History of Family Cancer: Interactional Resources for Managing Illness. New York: Hampton Press.

24

Naturally occurring data Beebe, L. M. and Cummings, M. C. (1985) ‘Speech act performance: A function of the data collection procedure?’ Paper presented at the TESOL Convention, New York. Beebe, L. M. and Cummings, M. C. (1996) ‘Natural speech act data versus written questionnaire data: How data collection method affects speech act performance’, in S. M. Gass and J. Neu (eds.) Speech Act Across Cultures: Challenges to Communication in a Second Language. Berlin: Mouton de Gruyter. 65–88. Billmyer, K. and Varghese, M. (2000) ‘Investigating instrument-based pragmatic variability: Effects of enhancing discourse completion tasks’, Applied Linguistics, 21: 517–552. Bodman, J. and Eisenstein, M. (1988) ‘May God increase your bounty: The expression of gratitude in English by native and non-native speakers’, Cross Currents, 15: 1–21. Bou-Franch, P. and Lorenzo-Dus, N. (2008) ‘Natural versus elicited data in cross-cultural speech act realisation: The case of requests in Peninsular Spanish and British English’, Spanish in Context, 5: 246–277. Economidou-Kogetsidis, M. (2013) ‘Strategies, modification and perspective in native speakers’ requests: A comparison of WDCT and naturally occurring requests’, Journal of Pragmatics, 53: 21–38. Félix-Brasdefer, J. C. (2007) ‘Natural speech vs. elicited data: A comparison of natural and role play requests in Mexican Spanish’, Spanish in Context, 4: 159–185. Félix-Brasdefer, J. C. (2010) ‘Data collection methods in speech act performance. DCTs, role plays, and verbal reports’, in A. Martínez-Flor and E. Usó-Juan (eds.) Speech Act Performance: Theoretical, Empirical and Methodological Issues. Amsterdam/Philadelphia: John Benjamins. 41–56. Golato, A. (2002) ‘German compliment responses’, Journal of Pragmatics, 34: 547–571. Golato, A. (2003) ‘Studying compliment responses: A comparison of DCTs and recordings of naturally occurring talk’, Applied Linguistics, 24: 90–121. Golato, A. (2005) Compliments and Compliment Responses: Grammatical Structure and Sequential Organization. Amsterdam/Philadelphia: John Benjamins. Golato, A. and Golato, P. (2013) ‘Pragmatics research methods’, in C. A. Chapelle (ed.) The Encyclopedia of Applied Linguistics. Oxford, UK: Wiley-Blackwell. Published Online: 5 November 2012. Grotjahn, R. and Kasper, G. (1991) ‘Methods in second language research: Introduction’, Studies in Second Language Acquisition, 12: 109–112. Hartford, B. S. and Bardovi-Harlig, K. (1992) ‘Experimental and observational data in the study of interlanguage pragmatics’, in L. Bouton and Y. Kachru (eds.) Pragmatics and Language Learning (Monograph Series, vol. 3). Urbana: University of Illinois at Urbana-Champaign. 33–52. Hassall, T. (1997) ‘Requests by Australian learners of Indonesian’, unpublished doctoral dissertation, Australia National University, Canberra. Heritage, J. (1984) Garfinkel and Ethnomethodology. Cambridge: Polity Press in association with Blackwell Publishers. Huth, T. (2010) ‘Can talk be inconsequential? Social and interactional aspects of elicited second language interaction’, Modern Language Journal, 94: 537–553. Jackson, S. and Jacobs, S. (1980) ‘Structure of conversational argument: Pragmatic bases for the enthymeme’, The Quarterly Journal of Speech, 66: 251–265. Jucker, A. H. (2009). ‘Speech act research between armchair, field and laboratory. The case of compliments’, Journal of Pragmatics, 41: 1611–1635. Kasper, G. (2000) ‘Data collection in pragmatics research’, in H. Spencer-Oatey (ed.) Culturally Speaking: Managing Rapport through Talk across Cultures. London/New York: Continuum. 316–341. Kasper, G. and Dahl, M. (1991) ‘Research methods in interlanguage pragmatics’, Studies in Second Language Acquisition, 13: 215–247. Labov, W. (1972) Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Labov,W. (1984) ‘Field methods of the project on linguistic change and variation’, in J. Baugh and J. Sherzer (eds.) Language in Use: Readings in Sociolinguistics. Englewood Cliffs, NJ: Prentice Hall. 28–66. Lee, D. Y. W. (2010) ‘What corpora are available?’, in M. McCarthy and A. O’Keeffe (eds.) The Routledge Handbook of Corpus Linguistics. London: Routledge. 107–121. Lehrer, A. (1989) ‘Remembering and representing prose: Quoted speech as a data source’, Discourse Processes, 12: 105–125. Mondada, L. (2009) ‘The embodied and negotiated production of assessments in instructed actions’, Research on Language and Social Interaction, 42: 329–361. Pavlidou,T.-S. (2000) ‘Telephone conversations in Greek and German: Attending to the relationship aspect of communication’, in H. Spencer-Oatey (ed.) Culturally Speaking: Managing Rapport through Talk across Cultures. London/New York: Continuum. 121–142. Placencia, M. E. (1997) ‘Opening up closings: The Ecuadorian way’, Text, 17: 53–81.

25

Andrea Golato Potter, J. (2002) ‘Two kinds of natural’, Discourse Studies, 4: 543–548. Schauer, G. A. and Adolphs, S. (2006) ‘Expressions of gratitude in corpus and DCT data:Vocabulary, formulaic sequences, and pedagogy’, System, 34: 119–134. Shively, R. L. (2011) ‘L2 pragmatic development in study abroad: A longitudinal study of Spanish service encounters’, Journal of Pragmatics, 43: 1818–1835. Taleghani-Nikazm, C. (2002) ‘A conversation analytical study of telephone conversation openings between native and nonnative speakers’, Journal of Pragmatics, 34: 1807–1832. Turnbull, W. (2001) ‘An appraisal of pragmatic elicitation techniques for the social psychological study of talk: The case of request refusals’, Pragmatics, 11: 31–61. Wolfson, N. (1989) Perspectives. Sociolinguistics and TESOL. Boston, MA: Heinle & Heinle Publishers. Yuan,Y. (2001) ‘An inquiry into empirical pragmatics data-gathering methods: Written DCTs, oral DCTs, field notes, and natural conversations’, Journal of Pragmatics, 33: 271–292.

26

4 ELICITED DATA J. César Félix-Brasdefer and Maria Hasler-Barker

Introduction The Oxford Dictionary of English defines the verb elicit as an action that evokes or draws out (a reaction, answer, or fact) from someone, as in eliciting a response with a question. Pragmatics researchers often elicit a response from participants with predesigned tasks that are intended to examine particular aspects of speakers’ pragmatic knowledge. Pragmatic knowledge, according to Leech (1983) and Thomas (1983), is comprised of two components: (1) pragmalinguistic competence – knowledge about and performance of the conventions of language use or the linguistic resources available in a given language that convey ‘particular illocutions’ in contextually appropriate situations (Leech 1983: 11), and (2) sociopragmatic competence – knowledge about and performance consistent with the social norms in specific situations in a given society, as well as familiarity with variables of social power and social distance. To analyse data from a pragmatic perspective, numerous researchers have drawn on methodology from the fields of linguistic anthropology, psychology, neurolinguistics, sociolinguistics, discourse analysis, second language acquisition, and sociology (Ericsson and Simon 1993; Duranti 1997; Cohen 1998, 2012; Davis and Henze 1998; Kasper 2000; Noveck and Sperber 2004; Mackey and Gass 2005; Cohen and Macaro 2010; Félix-Brasdefer 2010, 2015; Meibauer and Steinbach 2011). In this chapter, we examine the techniques used in experimental pragmatics to elicit production, perception, and comprehension data that are reliable and valid to a degree. Specifically, we provide a selective account of methods for pragmatics research that elicit data using a variety of tasks, with particular attention to issues of reliability and validity. Since the instrument chosen to analyse pragmatic meaning may affect the credibility and interpretation of the data (e.g. speech act, (im-)politeness, implicature, irony, indirectness), researchers should be aware of the reliability and validity of the particular instrument they employ. The degree of reliability of an instrument refers to whether an instrument administered to the same respondent on a different occasion would yield similar results. Specifically, reliability regulates the quality of the instrument.Validity, on the other hand, refers to the degree to which an instrument measures that which it is intended to measure, and consequently the degree to which it allows adequate interpretation of the results; that is, validity refers to ‘the degree to which a test measures what it claims, or purports, to be measuring’ (Brown 1996: 231).Validity includes two types: construct validity refers to the internal structure of the instrument and the concept it measures, while content validity alludes 27

J. César Félix-Brasdefer and Maria Hasler-Barker

to whether the content of the variables (e.g. questions of a questionnaire or items on a test) is suitable to measure the latent concept (for additional information on validity and reliability, see Brown 2001; Cohen 2004; Muijs 2004; Cohen and Macaro 2010). In addition, face validity refers to asking participants what they think of specific sections of the content of the test. Researchers in pragmatics need to be aware of issues of reliability and validity to determine the degree to which a method is reliable, and whether the elicited data is credible for answering the question at hand. This chapter is organised as follows. First, we describe methods used to elicit pragmatic data through structured interviews, followed by production instruments, methods that elicit pragmatic perception, and methods of evaluating pragmatic language comprehension, including electrophysiological methods.We also discuss issues of validity and reliability and the importance of appropriately matching tasks and measurements to research questions.

Structured interviews It is important that the researcher be concerned with the collection of natural data reflecting spontaneous speech to account for general patterns of communicative behavior under natural conditions. Labov observed that ‘our goal is then to observe the way that people use language when they are not being observed,’ otherwise known as the Observer’s Paradox (Labov 1972: 209). Unable to achieve this ideal, Labov resorted to other techniques such as the sociolinguistic/structured interview. In a structured interview, the researcher guides the direction of the conversation by means of preplanned questions (in the framework of a sociolinguistic interview) or via a conversational task. The sociolinguistic interview, as commonly used in variationist sociolinguistics research (Labov 1984; Tagliamonte 2012), is defined as a series of hierarchically structured sets of questions with a conversational format. In the pragmatic context, researchers used this interview type to elicit narratives of personal experience or stories that are told in an open-ended format. Given the presence of the interviewer, there is a degree of control in the selection of the topics and questions asked. Conversational modules which focus on a group of questions with a particular topic (e.g. children’s games, premonitions, the danger of death, aspirations) are used to this aim (Labov 1984: 33). Data from sociolinguistic interviews have been used to examine pragmatic/discourse features such as quotative be like and general extenders (e.g. . . . . or something like that) (Tagliamonte 2012). In conversational tasks, in contrast, the researcher is not present, but participants are asked to engage in free conversation in dyads (e.g. Félix-Brasdefer and Lavin 2009). Both structured interview types, whether conversationally elicited narratives (sociolinguistic interview) or conversational data, allow the researcher to observe a wide range of variables at the pragmatic and discourse levels, such as narratives, discourse markers (e.g. well), or turn-taking/ overlap patterns. In addition to linguistic features, these types of data allow the researcher to examine nonverbal actions (in videotaped interactions) or prosodic resources that convey pragmatic meanings, such as low or high terminals, duration, pauses, or stress. Given the interactional style of the interview, the pragmatic features analysed are the result of social interaction in situated contexts with some degree of control on the part of the researcher.

Production instruments This section describes two instruments that elicit production data under controlled conditions, written discourse completion tasks (WDCTs) (also referred to as DCTs) and role plays, both of which elicit simulated data under controlled experimental conditions. WDCTs measure 28

Elicited data

non-interactive or offline (pragmatic) knowledge while role plays measure interactive or online knowledge. These instruments have been used in interlanguage pragmatics (ILP) and in crosscultural pragmatics research to elicit (simulated) speech act data in written and oral modalities, respectively (see Kasper and Dahl 1991; Bardovi-Harlig 2010; Félix-Brasdefer and Koike 2014).

Written discourse completion tasks (WDCTs) WDCTs elicit experimental speech act data under controlled conditions with the aim of measuring offline pragmalinguistic or sociopragmatic knowledge in a non-interactive format. That is, WDCTs measure what the participants know, rather than how they use their ability to interact with an interlocutor. These instruments are typically administered in paper-and-pencil or computer format (e.g. Levenston and Blum 1978; Billmyer and Varghese 2000; Yuan 2001). (1) shows a situation taken from the Cross-Cultural Speech Act Realization Project (CCSARP) (Blum-Kulka et al. 1989), a project in which a WDCT was used to collect written speech act data among native and nonnative speakers in seven countries. (1) At the university Ann missed a lecture yesterday and would like to borrow Judith’s notes. Ann: ________________________________________________________ Judith: Sure, but let me have them back before the lecture next week. (Blum-Kulka et al. 1989: 14) With this instrument, participants are asked to imagine the situation and to respond in writing with how they believe they would respond in a real-life situation. This task can be used to elicit reactive speech acts (e.g. a refusal to an invitation) or initiative acts, as the request in the foregoing example. Other formats of the WDCT include the enhanced DCT which adds detailed information about the interlocutors, including social power and distance (Billmyer and Varghese 2000). Finally, another form is the free discourse completion task (FDCT) (Barron 2003), also referred to as the dialogue production task (DPT) in pragmatic variation research (Schneider 2008). Unlike the WDCT, where each participant produces one turn, in the FDCT or DPT, each participant has ‘to produce a short dialogue involving two participants’ (Schneider 2008: 106). In these instruments, written data are elicited at either the speech act or the discourse level. Although these instruments have proven to elicit reliable data, the validity of the data needs to be interpreted with caution, as they elicit simulated data under highly controlled conditions (Hudson et al. 1995; Brown 2001; Cohen 2004, 2012; Roever 2004; Woodfield 2008).

Role plays Role plays are simulations of communicative face-to-face encounters that elicit spoken data in which two interlocutors assume roles under predefined experimental conditions. According to Crookall and Saunders, a role play can be defined as ‘a social or human activity in which participants “take on” and “act out” specified “roles”, often within a predefined social framework or situational blueprint (a “scenario”)’ (1989: 15–16). Role-play data allow researchers to examine interactional practices (e.g. speech act sequences, conventional routines, openings and closings), discourse markers (e.g. okay, well), mitigators (e.g. I think, I was wondering, probably), turn-taking patterns, and polite and impolite practices. Role plays allow control over a series of contextual parameters: the situation, the degree of social distance and social power between the 29

J. César Félix-Brasdefer and Maria Hasler-Barker

interlocutors, the weight of imposition, gender and age of the participants, as well as learning environment (foreign [e.g. Americans learning Spanish in Indiana] or second [e.g., Koreans learning English in Chicago]) and proficiency level, where applicable. Role plays can be of two types: closed and open (Kasper and Dahl 1991). In the closed role play (also called oral DCTs or oral production tasks), the participant responds to a role-play situation orally without a reply from an interlocutor.Various elicitation formats involving elicitation of an oral response under highly controlled conditions have developed from this oral response format. Rose’s (2000) cartoon oral production task (COPT) has the advantage of eliciting oral data within a particular visual context by means of pictures representing different social situations and can be audio recorded (as an oral DCT). In her study of L2 pragmalinguistic and sociopragmatic development, Flores-Salgado (2009) adapted the COPT to elicit requests and apologies among adult learners of English as an FL in Mexico. Hasler-Barker (2013) created a similar task using PowerPoint to deliver written prompts for compliments and compliment responses. The written descriptions of the scenarios were supplemented with photos of interlocutors. L2 Spanish participants were instructed to imagine that they were responding to the person whose image they saw. The time for both reading and responding was controlled and all responses were audio recorded. In an attempt to further increase the degree of validity of the WDCT, Schauer (2004) designed the computer-based Multimedia Elicitation Task (MET) with rich audiovisual and contextual information in the situation prompt. She examined ILP development of requests in German-speaking learners of English during a year-long study abroad program in Great Britain. The MET controls the time and the nature of the audio and visual input, guarantees equal conditions for every participant, elicits oral data, and is delivered by means of a computerised presentation format with visual (photographic images) and audio input (description of the situation). This task is also delivered by computer: participants simultaneously read the scenario on a screen and hear it over individual headsets, and all responses are audio recorded.The aural/oral production task has also been used to elicit conventional formulas among native and nonnative speakers (Bardovi-Harlig 2012). These oral production tasks are especially useful for identifying what participants know they should do and what they are able to produce under constrained conditions. They allow researchers to collect comparable data (across participants, testing conditions, or other experimental constraints). Open role plays, on the other hand, specify at least one actor’s role, but the course and outcome of the conversation are not predetermined (cf. Kasper and Dahl 1991: 228). During a role play, participants are asked to read a situational description and to interact orally as they would in a real situation with an interlocutor in face-to-face interaction. Most role plays employed explicitly instruct the interlocutor to either initiate (e.g. to make a request) or to react to (e.g. to refuse) a particular speech act, thus anticipating the conversational outcome (cf. [2]; HaslerBarker 2013). Other role-play formats do not prescribe the conversational outcome; that is, participants do not receive explicit instructions on what speech act to perform (e.g. Márquez Reiter et al. 2005). (2) Shows the description of a role play that includes contextual information about the setting, the relationship between the participants, gender, and the time of the event. In this situation, participants are asked to act out the situation based on the contextual information and roles provided in the scenario. (2) Refusing a friend’s invitation to attend a birthday party You are walking across campus when you run into a good friend of yours whom you haven’t seen for about a month.You and he have been studying in the same program at the 30

Elicited data

university for three years, and have studied and written papers together in the past, but you don’t have any classes together this semester since you have been doing an internship offcampus. He invites you to his 21st birthday party at his house next Friday night at 8:00 p.m. He tells you that a group of mutual friends that you both used to hang out with and whom you haven’t seen since the semester started will also be there.You know that this would be a good opportunity to see everyone again and to celebrate this special occasion with him. Unfortunately, you cannot make it. (Félix-Brasdefer 2008a: 186–187) With role-play data researchers can examine the pragmatic effect of a variety of features, including prosodic elements employed in an interaction to express tentativeness, politeness, or degrees of directness or indirectness (e.g. high and low terminals, stress, duration) (Selting 2010).Validity is relatively high with role plays because researchers can capture interactive data with turntaking, which more closely represents a natural conversation than an oral or written task (FélixBrasdefer 2007, 2010; Al-Gahtani and Roever 2012). This type of task allows researchers to use conversation analytic tools to understand language use. Still, these tasks rely on interlocutors to co-construct interaction, a fact which affects reliability and limits researcher control and the comparability of role plays.

Eliciting data related to participants’ perceptions Introspection or retrospection techniques are used to elicit data that trigger either a previous experience (or schema) or one aspect of participants’ short-term memory after completion of a task. Perceptual data can be elicited in at least three ways: politeness events, rating scales (e.g. Likert scale), and verbal reporting.

Politeness events This method is grounded in psychological research and applied in politeness and impoliteness research in cross-cultural pragmatics (Spencer-Oatey 2002; Culpeper et al. 2010). In this method, participants are asked to write about events which had a positive or negative effect on them. For example, in Culpeper and colleagues (2010), the researchers asked participants from five countries to report a conversation that had a negative effect on them, yielding an impoliteness event (see also Culpeper 2011, 195–219). To help participants in their reconstruction, the instructions specify: What was actually said, implied, or done, how/where was it said, implied, or done? By whom was it said, implied, or done? Did others hear it? What were your reactions? (Culpeper et al. 2010: 601). This method produces lengthy discourse related to impoliteness events that are based on the participants’ previous experience. Excerpt (3) shows an impoliteness event produced by a second language learner of Spanish (example from corpus collected by the first author). (3) When I was visiting a friend in Barcelona, we went out to eat one night. It was both of our second-to-last week in Spain, so we felt very confident in our ability to speak Spanish with our waiter. When he came to the table for the first time, he addressed us in Spanish. We responded, and from then on he would only speak to us in English. We were both very frustrated as we were trying to get in as much speaking practice as possible before we went home. Knowing we would never see him again, at the end of our meal we asked him why he would not converse with us in Spanish, and his response was that he doesn’t 31

J. César Félix-Brasdefer and Maria Hasler-Barker

like speaking broken Spanish with Americans. His tone of voice throughout the meal, but especially in his response to our question, was very condescending. This event occurred in a busy restaurant in downtown Barcelona. I don’t know if others in the restaurant heard it, but there was a very solid mix of English and Spanish speakers around us and we were seated relatively close together so it is very possible they did. We tried to be nice throughout the entire meal, even after his rude response, so we thanked him for serving us and left immediately after. We didn’t do anything out of the ordinary other than ask him directly why he would not speak Spanish with us. While the degree of validity of this method may be subject to the criticism that respondents may not always recall specific information, this method has proven reliable and valid to a degree because participants activate existing mental frames based on previous experiences. Rating scales can complement this method, as described in the next section.

Rating scales Eliciting data by means of rating scales allows researchers to observe degrees of perception of a particular variable in terms of a numeric scale, ranging from 1 to 9, 1 to 7, or 1 to 5. For example, participants can rate the degree of formality of a situation (e.g. 1 = too formal, 5 = too informal). Participants can also rate the degree of appropriateness of behaviour in a given conversation (e.g. 1 = inappropriate, 4 = appropriate). Rating scales are often used to obtain complementary information to triangulate primary data, as in Culpeper and colleagues’ (2010) research on impoliteness cited earlier. After a participant reported an impoliteness event, Culpeper and colleagues gauged the gravity of the offense by asking participants to select a rating on a five-point Likert scale in response to the following question: ‘How bad did the behaviour in the conversation make you feel at the time it occurred?’ (2010: 602). Triangulating data with rating scales increases the data’s validity and credibility. Numeric data are also relatively easy to quantify with means comparisons (paired-sample or independent-sample t-tests) or frequencies (Chi-square tests by computing the frequency counts in each category in the case of two or more groups). Although the responses to this indirect measure of pragmatic knowledge are the result of interpretation of the data, the data are valid to a degree if the instrument is piloted to ensure consistency of the responses. Researchers also need to be aware of individual variation in intra-cultural and intercultural contexts.

Verbal reports This section examines two types of introspection methods: think-aloud protocols and retrospective verbal reports in speech act research. Mackey and Gass (2005: 77) describe verbal reporting as ‘a special type of introspection and consists of gathering protocols, or reports, by asking individuals what is going through their minds as they are solving or completing a task’. Verbal reporting is grounded in psychological research within the Processing Model which interprets human cognition as information processing with limited capacity, namely, short-term (STM) and long-term memory (LTM) (Ericsson and Simon 1993). Under this framework, the authors (1993: 11) indicate that ‘it is assumed that information recently acquired (attended to or heeded) by the central processor is kept in STM, and is directly accessible for further processing (e.g. for producing verbal reports), whereas information from LTM must be first transferred to STM before it can be reported’. In think-aloud protocols, participants are instructed to think aloud

32

Elicited data

or to voice their thoughts while performing a task (e.g. WDCTs). On the other hand, retrospective reporting consists of verbal reports obtained from the participants immediately after completion of a task (e.g. role plays or WDCTs), while much information is still in STM and can be directly reported. In speech act research, both types of verbal reports have been used to draw inferences during production or after completion of WDCTs or role plays. Verbal reports are instrumental in understanding speech act performance because ‘one may learn what the respondents actually perceived about each situation (e.g. what they perceived about the relative role status of the interlocutors) and how their perceptions influenced their responses’ (Cohen 2004: 321). For instance, using a WDCT, Robinson (1992) examined the effects that concurrent (single-subject think-aloud) and retrospective verbal reporting had on recalling refusals among learners of English as an L2. Results showed that while concurrent reporting elicits specific information about the planning of the speech act, retrospective reporting aids in generating and investigating hypotheses. Widjaja (1997) examined the sociocultural perceptions that female EFL Taiwanese learners have of their L1 when refusing a date in English, and concluded that retrospective reporting draws inferences on the learners’ pragmatic knowledge. In order to examine sociocultural perceptions of L2 target norms, retrospective verbal reports collected by Félix-Brasdefer (2003, 2008a) showed that although advanced US learners of Spanish as an FL showed high levels of pragmalinguistic competence when declining an invitation in formal and informal situations, their perceptions of sociocultural target norms often differed from those of the L1 culture. Finally, using retrospective reporting immediately after the completion of two open role plays, Félix-Brasdefer (2008b) examined the perceptions of refusals to two invitations (friend–friend and employee– boss) among twenty advanced learners of Spanish as an FL. Results showed that retrospective reporting was instrumental in accessing information about cognition (i.e. the planning and execution of the refusal and the selection of the language of thought) and perception of sociocultural knowledge (social status differences in L1 and L2 and the perception of an insistence in Spanish and English). Both concurrent and retrospective reporting serve to validate instruments, thus increasing both content and construct validity.Verbal reporting can be used to increase the degree of content validity (or face validity) of an instrument, and thus ensure trustworthiness of the results (cf. also Woodfield 2008). Most important, if carried out with care, verbal reporting can be instrumental in analysing learners’ sociocultural perceptions of the target culture such as notions of politeness and respect, directness and indirectness, and speech act perception in formal and informal contexts.

Elicited data for pragmatic language comprehension As the field of pragmatics has evolved, so have methods for eliciting experimental data using multiple-choice questionnaires, methods from psycholinguistics (e.g. forced choice tasks such as truth value judgments, felicity judgments, picture selection tasks), and physiological measures (e.g. self-paced reading, eye-tracking, ERP, and fMRI), to test a variety of pragmatic phenomena.

Written multiple-choice completion tasks (WMCTs) Written multiple-choice completion tasks (WMCTs) aim to measure pragmatic knowledge comprehension. Participants choose the best or most appropriate option in response to a particular situation. WMCTs have been used to test pragmatic knowledge of implicatures among

33

J. César Félix-Brasdefer and Maria Hasler-Barker

second language learners, as in (4) taken from Roever (2006: 238). In this example, the most likely response to the exchange in (4) is (c), which is the result of a pragmatic inference. (4) Testing pragmatic knowledge of L2 implicature Jack is talking to his housemate Sarah about another housemate, Frank. Jack: ‘Do you know where Frank is, Sarah?’ Sarah:  ‘Well, I heard music from his room earlier.’

What does Sarah probably mean? a. b. c. d.

Frank forgot to turn the music off. Frank’s loud music bothers Sarah. Frank is probably in his room. Sarah doesn’t know where Frank is. (Roever 2006: 238)

The WMCT is an indirect way of assessing pragmalinguistic and sociopragmatic knowledge. When designing this instrument, NS norms and regional variation should be considered, as well as the design of distracter items to balance the selection of responses. As with any other data collection instrument,WMCT situations need to be pilot tested to ensure the description of the situations represent authentic scenarios and that there is a high degree of reliability and validity in the representativeness of the responses (Roever 2004; Liu 2007).

Forced judgment tasks In forced choice tasks, participants must choose between two or more options. The literature in experimental pragmatics has employed at least three types of forced judgment tasks to test different aspects of pragmatic comprehension among adults and children, such as implicature, speech acts, and prosodic cues. These are truth value judgment tasks, felicity judgment tasks, and picture-selection tasks. In a truth value judgment, participants determine whether an utterance is true or false in a given context. In this task, participants view a scenario, which can be a picture or even a tableau containing one or more items, and judge whether an utterance is true or false for that scenario. When presenting this task to children, researchers frequently use a puppet to deliver the utterances and the participants are asked whether the puppet is telling the truth. For example, Noveck (2001) tested French-speaking children and adults with this type of task and found that children were more literal in their interpretations (i.e. they interpreted some as including all) while adults were equivocally pragmatic (i.e. they interpreted some as excluding all) (cf. also Bott and Noveck 2004; Barner et al. 2011). Truth value judgment tasks are also used in conjunction with other types of tasks to evaluate additional cognitive factors. Bezuidenhout and Cutting (2002) used self-paced reading (see more later in this chapter about this type of task) and truth value judgment to test English-speaking participants’ processing times for literal and pragmatic implicatures. They found no significant difference in processing time, which supports their claim that both types of implicature are processed simultaneously. DeNeys and Schaeken (2007) combined a truth value task with a dot memory task to analyse the effect of cognitive load on pragmatic interpretations. They found that an increased cognitive load (i.e. the memory task) meant that Dutch-speaking participants took longer to interpret utterances and were less likely to accept a pragmatic interpretation. 34

Elicited data

Researchers have also used felicity judgment tasks to better understand comprehension of scalar implicatures. Like the truth value judgment task, participants are presented with pictures or tableaux. However, they are then asked to choose which of two utterances best matches what they see rather than if the utterance is true or false. In fact, Foppolo and colleagues (2012) showed that sensitivity to scalar implicatures is more easily captured in a felicity judgment task than in a truth value task, thus highlighting the importance of carefully matching tasks to research goals. (cf. also Chierchia et al. 2001; Gualmini et al. 2001; Foppolo et al. 2012). Instead of judging the fit of an utterance, picture selection tasks require participants to choose the picture that best fits an utterance. In addition to testing scalar implicatures (e.g. Katsos and Bishop 2011), researchers have tested other pragmatic features such as prosody. Kurumada and colleagues (2012) tested the influence of prosodic cues on picture selection by NS of American English, finding that hearers adapt quickly to prosody cues of different speakers. Forced choice instruments are especially useful for understanding comprehension and can be both highly reliable and valid for these purposes, though it is crucial to pilot test the stimuli for content validity. It is also important to recall that these types of instruments are appropriate only for measuring receptive aspects of pragmatics.

Physiological measures This section reviews data elicitation procedures through four measurements of physiological responses to pragmatic stimuli: self-paced reading, eye-tracking experiments, event-related potential (ERP), and functional magnetic resonance imaging (fMRI). Self-paced reading tasks allow participants to control the rate at which words or phrases appear on a computer screen. Reading times are correlated with processing times. For example, Breheny and colleagues (2006) used self-paced reading to better understand context-driving vs. default implicature generation approaches. For example, in the exchange in (5), both (a) and (b) are implied: (5) John: Was the exam easy? Mary:  Some of the students failed. a. Not all the students failed. b. The exam was not easy. According to the default approach, both implicatures are always generated, regardless of context, while in the context-driven approach, implicature (b), a particularised implicature, is generated only when the context requires it. Greek-speaking participants in Breheny and colleagues (2006) completed both a self-paced reading task and a forced judgment task about sentence fit. Their findings support the context-driven approach, as participants took significantly longer to read contexts that triggered particularised implicatures. Eye-tracking measurement uses a specialised device to measure eye positions and eye movement while participants complete a task. For instance, Bezuidenhout and Morris (2004) used eye tracking to measure English speakers’ processing of generalised conversational implicatures followed by a cancellation sentence (e.g. Some books had colour pictures. In fact all of them did, which is why the teachers liked them.). Their findings support the context-driven approach to implicatures (cf. also Huang and Snedeker 2009, 2011; Panizza and Chierchia 2011). Furthermore, Filik (2008) showed with eye tracking that, for English speakers, pragmatically anomalous words were integrated when appropriate contextual/cultural knowledge was accessed (e.g. ‘the mouse 35

J. César Félix-Brasdefer and Maria Hasler-Barker

picked up the dynamite’ is anomalous under normal circumstances, but acceptable in the context of the cartoon Tom & Jerry). Event-related potential (ERP) measures components of brain activity waves in response to stimuli. Coulson (2004) offers an overview of this type of research and its potential for understanding pragmatic comprehension. Researchers may use truth value judgment tasks for comprehension. For instance, Ditman and colleagues (2007) combined ERP and self-paced reading methodologies to access behavioural and neural correlates for processing pragmatic and syntactic violations. Their results indicated that combining the two experiment types offers a more comprehensive way to understand language processing. Van Berkum’s (2009) review of ERP research underscores why this type of measurement is well suited to understanding the connections between pragmatic processing and broader aspects of language processing. He suggests that listeners and readers likely anticipate what speakers and writers will do because processing of underspecified utterances in context happens so quickly, highlighting the need to access more information about the process of comprehension. fMRI measures brain activity by measuring blood flow to brain regions activated by stimuli. For example, participants in Kuperberg and colleagues’ (2000) study listened to sentences with pragmatic, semantic, or syntactic violations. Their data showed variations in brain activity depending on the type of violation, though they identified brain activity common to all three types. Eviatar and Just (2006) used fMRI with reading comprehension to test English speakers’ sensitivity to irony and conventional metaphor, finding that higher levels of discourse processing require brain activation beyond literal interpretations (cf. also Mashal et al. 2007). fMRI has also been used to understand pragmatic language processing problems. For instance, Tesink and colleagues (2009) studied native speakers of Dutch with autism spectrum disorders. They found problems at both the most basic levels and in high-level inferential processing for the participants in their study. fMRI measures have demonstrated success in understanding language processing disorders as well as in understanding sensitivity to pragmatic features. Physiological measures of language processing and comprehension have been used to measure reactions to various pragmatic targets. Prior to implementing these types of measurement, it is crucial to measure the content validity of stimuli through careful pilot testing.These measurements are reliable for testing comprehension under experimental conditions, which may affect their overall validity. It is crucial that instruments and means of measurement be chosen carefully for the questions that researchers wish to answer.

Conclusion This chapter examined methods of eliciting oral and written data in various pragmatic contexts. In pragmatics research, elicited data are often used to assess or measure a particular aspect of pragmatic knowledge (pragmalinguistic or sociopragmatic). In addition, in contrast to sociolinguistics, which shows a preference for the sociolinguistic interview (Labov 1984; Tagliamonte 2012), research in pragmatics may include a variety of data elicited from diverse methods under controlled conditions (e.g. role plays, written questionnaires, verbal reports). We also described methods used to elicit data to measure pragmatic language comprehension, including brain activity, eye tracking, and reaction times.These measures have provided information about comprehension of implicatures and other pragmatic phenomena. Note that the aforementioned elicitation methods need to be constantly refined and adapted to fit each research question, and ultimately triangulated with natural data. Issues of reliability and validity must be considered when designing elicitation measures and can be enhanced through careful pilot testing. More important, as researchers apply various methods for measurement 36

Elicited data

and use a variety of methods of analysis, we can come to a fuller understanding of pragmatic production, perception, and comprehension. On a final note about ethics, note that collecting elicited data may require approval of the Institutional Review Board (or similar group) from the institution where the data are collected in order to protect the rights of human subjects; this information should be reported in the study.

Suggestions for further reading Coulson, S. (2004) ‘Electrophysiology and pragmatic language comprehension’, in I. Noveck and D. Sperber (eds.) Experimental Pragmatics. San Diego, CA: Palgrave MacMillan. 187–206. Gives an overview of selected electrophysiological measurements and their suitability for assessing pragmatic language comprehension. Cohen, A. D. and Macaro, E. (2010) ‘Research methods in second language acquisition’, in E. Macaro (ed.) Continuum Companion to Second Language Acquisition. London: Continuum. 107–133. Offers an overview of traditional research methods and research instruments used in second language acquisition and interlanguage pragmatics. Holtgraves, T. (2002) Language as Social Action: Social Psychology and Language Use. Mahwah, NJ: Lawrence Erlbaum. Gives an overview of the intersection of psychological research and interactional linguistics with special focus on speech acts and politeness. Kasper, G. (2000) ‘Data collection in pragmatics research’, in H. Spencer-Oatey (ed.) Culturally Speaking: Managing Rapport through Talk across Cultures. London: Continuum. 316–369. Provides a description and assessment of selective methods used in pragmatics research with particular attention to interlanguage pragmatics research. Ericsson, K. A. and Simon, H. A. (1993) Protocol Analysis:Verbal Reports as Data. Cambridge, MA: MIT Press. Offers a comprehensive overview of introspective methodology (verbal reports) in psychology research.

References Al-Gahtani, S. and Roever, C. (2012) ‘Proficiency and sequential organization of L2 requests’, Applied Linguistics, 33(1): 42–65. Bardovi-Harlig, K. (2010) ‘Exploring the pragmatics of interlanguage pragmatics: Definition by design’, in A. Trosborg (ed.) Pragmatics across Languages and Cultures. Berlin: De Gruyter Mouton. 219–259. Bardovi-Harlig, K. (2012) ‘Variation in the pragmatic use of conventional expressions’, in J. C. Félix-Brasdefer and D. Koike (eds.) Pragmatic Variation in First and Second Language Contexts: Methodological Issues. Amsterdam/Philadelphia: Benjamins. 141–173. Barner, D., Brooks, N., and Bale, A. (2011) ‘Accessing the unsaid: The role of scalar alternatives in children’s pragmatic inference’, Cognition, 118(1): 84–93. Barron, A. (2003) Acquisition in Interlanguage Pragmatics: Learning How to Do Things with Words in a Study Abroad Context. Amsterdam/Philadelphia: Benjamins. Bezuidenhout, A. and Cutting, J. C. (2002) ‘Literal meaning, minimal propositions, and pragmatic processing’, Journal of Pragmatics, 34(4): 466–456. Bezuidenhout, A. and Morris, R. (2004) ‘Implicature, relevance, and default pragmatic inference’, in I. Noveck and D. Sperber (eds.) Experimental Pragmatics. San Diego, CA: Palgrave MacMillan. 257–282. Billmyer, K. and Varghese, M. (2000) ‘Investigating instrument-based pragmatic variability: Effects of enhancing discourse completion tests’, Applied Linguistics, 21(4): 517–552. Blum-Kulka, S., House, J., and Kasper, G. (1989) ‘Investigating cross-cultural pragmatics: An introductory overview’, in S. Blum-Kulka, J. House, and G. Kasper (eds.) Cross-cultural Pragmatics: Requests and Apologies. Norwood, NJ: Ablex. 1–34. Bott, L. and Noveck, I. (2004) ‘Some utterances are underinformative: The onset and time course of scalar inferences’, Journal of Memory and Language, 51(3): 437–457. Breheny, R., Katsos, N., and Williams, J. (2006) ‘Are generalised scalar implicatures generated by default? An on-line investigation into the role of context in generating pragmatic inferences’, Cognition, 100(3): 434–463. Brown, J. D. (1996) Testing in Language Programs. Upper Saddle River, NJ: Prentice Hall Regents.

37

J. César Félix-Brasdefer and Maria Hasler-Barker Brown, J. D. (2001) ‘Six types of pragmatics tests in two different contexts’, in K. Rose and G. Kasper (eds.) Pragmatics in Language Teaching. New York: Cambridge University Press. 301–325. Chierchia, G., Crain, S., Guasti, M. T., Gualmini, A., and Meroni, L. (2001) ‘The acquisition of disjunction: Evidence for a grammatical view of scalar implicatures’, in A. H.-J. Do, L. Domínguez, and A. Johansen (eds.) BUCLD 25: Proceedings of the 25th Annual Boston University Conference on Language Development. Somerville, MA: Cascadilla Press. 157–168. Cohen, A. D. (1998) Strategies in Learning and Using a Second Language. London: Longman. Cohen, A. D. (2004) ‘Assessing speech acts in a second language’, in D. Boxer and A. D. Cohen (eds.) Studying Speaking to Inform Second Language Learning. Clevedon: Multilingual Matters. 302–327. Cohen, A. D. (2012) ‘Research methods for describing variation in intercultural pragmatics for cultures in contact and conflict’, in J. C. Félix-Brasdefer and D. A. Koike (eds.), Pragmatic Variation in First and Second Language Contexts: Methodological Issues. Amsterdam: Benjamins. 17–48. Cohen, A. D. and Macaro, E. (2010) ‘Research methods in second language acquisition’, in E. Macaro (ed.) Continuum Companion to Second Language Acquisition. London: Continuum. 107–133. Coulson, S. (2004) ‘Electrophysiology and pragmatic language comprehension’, in I. Noveck and D. Sperber (eds.) Experimental Pragmatics. San Diego, CA: Palgrave MacMillan. 187–206. Crookall, D. and Saunders, D. (1989) Communication and Simulation. Clevedon: Multilingual Matters. Culpeper, J. (2011) Impoliteness: Using Language to Cause Offence. Cambridge: Cambridge University Press. Culpeper, J., Marti, L., Mei, M., Nevala, M., and Schauer, G. (2010) ‘Cross-cultural variation in the perception of impoliteness: A study of impoliteness events reported by students in England, China, Finland, Germany, and Turkey’, Intercultural Pragmatics, 7(4): 597–624. Davis, K. A. and Henze, R. C. (1998) ‘Applying ethnographic perspectives to issues in cross-cultural pragmatics’, Journal of Pragmatics, 30(4): 399–419. DeNeys, W. and Schaeken, W. (2007) ‘When people are more logical under cognitive load: dual task impact on scalar implicature’, Experimental Psychology, 54(2): 128–133. Ditman, T., Holcomb, P. J., and Kuperberg, G. R. (2007) ‘An investigation of concurrent ERP and selfpaced reading methodologies’, Psychophysiology, 44(6): 927–935. Duranti, A. (1997) Linguistic Anthropology. Cambridge: Cambridge University Press. Ericsson, K. A. and Simon, H. A. (1993) Protocol Analysis. Verbal Reports as Data, rev. ed. Cambridge, MA: MIT Press. Eviatar, Z. and Just, M. A. (2006) ‘Brain correlates of discourse processing: An fMRI investigation of irony and conventional metaphor comprehension’, Neuropsychologia, 44(12): 2348–2359. Félix-Brasdefer, J. C. (2003) ‘Validity in data collection methods in pragmatics research’, in P. Kempchinsky and C. E. Piñeros (eds.) Theory, Practice, and Acquisition: Papers from the Sixth Hispanic Linguistics Symposium and the Fifth Conference on the Acquisition of Spanish and Portuguese. Somerville, MA: Cascadilla Press. 239–257. Félix-Brasdefer, J. C. (2007) ‘Natural speech vs. elicited data: A comparison of natural and role-play requests in Mexican Spanish’, Spanish in Context, 4(2): 159–185. Félix-Brasdefer, J. C. (2008a) Politeness in Mexico and the United States: A Contrastive Study of the Realization and Perception of Refusals. Amsterdam: Benjamins. Félix-Brasdefer, J. C. (2008b) ‘Perceptions of refusals to invitation: Exploring the minds of foreign language learners’, Language Awareness, 17(3): 195–211. Félix-Brasdefer, J. C. (2010) ‘Data collection methods in speech act performance: DCTs, role-plays, and verbal reports’, in E. Usó Juán and A. Martínez-Flor (eds.) Speech Act Performance: Theoretical, Empirical, and Methodological Issues. Amsterdam/Philadelphia: Benjamins. 41–56. Félix-Brasdefer, J. C. (2015) The Language of Service Encounters: A Pragmatic-Discursive Approach. Cambridge: Cambridge University Press. Félix-Brasdefer, J. C. and Koike, D. (2014) ‘Perspectives on Spanish SLA from pragmatics and discourse’, in M. Lacorte (ed.) The Routledge Handbook of Hispanic Applied Linguistics. 25–43. Félix-Brasdefer, J. C. and Lavin, E. (2009) ‘Grammar and turn expansion in second language conversations’, in J. Collentine, M. García, B. Lafford, and F. Marcos Marín (eds.) Selected Proceedings of the 11th Hispanic Linguistics Symposium. Somerville, MA: Cascadilla Proceedings Project. 53–67. Filik, R. (2008) ‘Contextual override of pragmatic anomalies: Evidence from eye movements’, Cognition, 106(2): 1038–1046. Flores-Salgado, E. (2009) ‘A pragmatic study of developmental patterns in Mexican students making English requests and apologies’, in E. Alcón and A. Martínez-Flor (eds.) Speech Act Performance: Theoretical, Empirical and Methodological Issues. Amsterdam: Benjamins. 224–248.

38

Elicited data Foppolo, F., Guasti, M. T., and Chierchia, G. (2012) ‘Scalar implicatures in child language: Give children a chance’, Language Learning and Development, 8(4): 365–394. Gualmini, A., Crain, S., Meroni, L., Chierchia, G., and Guasti, M. T. (2001) ‘At the semantics/pragmatics interface in child language’, in R. Hastings, B. Jackson, and Z. Zvolensky (eds.) Semantics and Linguistic Theory. Ithaca, NY: Cornell University. 231–247. Hasler-Barker, M. (2013) ‘Effects of pedagogical intervention on the production of the compliment and compliment response sequence by second language learners of Spanish’, unpublished doctoral dissertation, Indiana University, Bloomington. Holtgraves, T. (2002) Language as Social Action: Social Psychology and Language Use. Mahwah, NJ: Lawrence Erlbaum. Hudson, T., Detmer, E., and Brown, J. D. (1995) Developing Prototypic Measures of Cross-cultural Pragmatics (Technical Report #7). Honolulu, Hawai’i: Second Language Teaching & Curriculum Center, University of Hawai’i. Huyang,Y. T. and Snedeker, J. (2009) ‘Online interpretation of scalar quantifiers: Insight into the semanticspragmatics interface’, Cognitive Psychology, 58: 376–415. Huyang, Y. T. and Snedeker, J. (2011) ‘Logic and conversation revisited: evidence for a division between semantic and pragmatic content in real time language comprehension’, Language and Cognitive Processes, 26(8): 1161–1172. Kasper, G. (2000) ‘Data collection in pragmatics research’, in H. Spencer-Oatey (ed.) Culturally Speaking: Managing Rapport through Talk across Cultures. London: Continuum. 316–341. Kasper, G. and Dahl, M. (1991) ‘Research methods in interlanguage pragmatics’, Studies in Second Language Acquisition, 13(2): 215–247. Katsos, N. and Bishop, D.V. (2011) ‘Pragmatic tolerance: Implications for the acquisition of informativeness and implicature’, Cognition, 120(1): 67–81. Kuperberg, G., McGuire, P. K., Bullmore, E. T., Brammer, M., Rabe-Hesketh, S., Lythgoe, D. J., and Williams, S. C. (2000) ‘Common and distinct neural substrates for pragmatic, semantic, and syntactic processing of spoken sentences: An fMRI study,’ Journal of Cognitive Neuroscience, 12(2): 321–341. Kurumada, C., Brown, M., and Tanenhaus, M. K. (2012) ‘Pragmatic interpretation of contrastive prosody: It looks like speech adaptation’, in N. Miyake, D. Peebles, and R. P. Cooper (eds.) Proceedings of the 34th Annual Conference of the Cognitive Science Society. Austin, TX: Cognitive Science Society. 647–652. Labov, W. (1972) Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Labov.W. (1984) ‘Field methods of the project on linguistic change and variation’, in J. Baugh and J. Sherzer (eds.) Language in Use: Readings in Sociolinguistics. Englewood Cliffs, NJ: Prentice-Hall. 28–54. Leech, G. N. (1983) Principles of Pragmatics. New York: Longman. Levenston, E. and Blum, S. (1978) ‘Discourse completion as a technique for studying lexical features of interlanguage’, Working Papers in Bilingualism, 15: 13–21. Liu, J. (2007) ‘Developing a pragmatics test for Chinese EFL learners’, Language Testing, 24(3): 391–415. Mackey, A. and Gass, S. (2005) Second Language Research: Methodology and Design. Mahwah, NJ: Lawrence Erlbaum. Márquez Reiter, R., Rainey, I., and Fulcher, G. (2005) ‘A comparative study of certainty and conventional indirectness: Evidence from British English and Peninsular Spanish’, Applied Linguistics, 26: 1–31. Mashal, N., Faust, M., Hendler, T., and Jung-Beeman, M. (2007) ‘An fMRI investigation of the neural correlates underlying the processing of novel metaphoric expressions’, Brain and Language, 100(2): 115–126. Meibauer, J. and Steinbach, M. (eds.) (2011) Experimental Pragmatics/Semantics. Amsterdam: Benjamins. Muijs, D. (2004) Doing Quantitative Research in Education with SPSS. London: Sage. Noveck, I. (2001) ‘When children are more logical than adults: Experimental investigations of scalar implicature’, Cognition, 78(2): 165–188. Noveck, I. and Sperber, D. (eds.) (2004) Experimental Pragmatics. San Diego, CA: Palgrave Macmillan. Panizza, D. and Chierchia, G. (2011) ‘Numerals and scalar implicatures’, in J. Meibauer and M. Steinbach (eds.) Experimental Pragmatics/Semantics. Amsterdam: Benjamins. 129–150. Robinson, M. (1992) ‘Introspective methodology in interlanguage pragmatics research’, in G. Kasper (ed.) Pragmatics of Japanese as Native and Target Language (Technical Report #3). Honolulu, Hawai’i: Second Language Teaching & Curriculum Center, University of Hawai’i. 27–82. Roever, C. (2004) ‘Difficulty and practicality in tests of interlanguage pragmatics’, in D. Boxer and A. D. Cohen (eds.) Studying Speaking to Inform Second Language Learning. Clevedon: Multilingual Matters. 283–301. Roever, C. (2006) ‘Validation of a web-based test of ESL pragmalinguistics’, Language Testing, 23(2): 229–256.

39

J. César Félix-Brasdefer and Maria Hasler-Barker Rose, K. R. (2000) ‘An exploratory cross-sectional study of interlanguage pragmatic development’, Studies in Second Language Acquisition, 22(1): 27–67. Schauer, G. (2004) ‘May you speak louder maybe? Interlanguage pragmatic development in requests’, EUROSLA Yearbook, 4: 253–272. Schneider, K. (2008) ‘Small talk in England, Ireland, and the U.S.A.’, in K. Schneider and A. Barron (eds.) Variational Pragmatics: A Focus on Regional Varieties in Pluricentric Languages. Amsterdam: Benjamins. 97–139. Selting, M. (2010) ‘Prosody in interaction: State of the art’, in D. Barth Weingarten, E. Reber, and M. Selting (eds.) Prosody in Interaction. Amsterdam: Benjamins. 3–40. Spencer-Oatey, H. (2002) ‘Managing rapport in talk: Using rapport sensitive incidents to explore the motivational concerns underlying the management of relations’, Journal of Pragmatics, 34: 529–545. Tagliamonte, S. A. (2012) Variationist Sociolinguistics: Change, Observation, Interpretation. Malden, MA: Wiley-Blackwell. Tesink, C. M., Buitelaar, J. K., Petersson, K. M.,Van Der Gaag, R. J., Kan, C. C., Tendolkar, I., and Hagoort, P. (2009) ‘Neural correlates of pragmatic language comprehension in autism spectrum disorders’, Brain, 132: 1941–1952. Thomas, J. (1983) ‘Cross-cultural pragmatic failure’, Applied Linguistics, 4: 91–112. Van Berkum, J. (2009) ‘The neuropragmatics of “simple” utterance comprehension: An ERP review’, in U. Sauerland and K.Yatsushiro (eds.) Semantics and Pragmatics: From Experiment to Theory. London: Palgrave MacMillan. 276–316. Widjaja, C. (1997) ‘A study of date refusal: Taiwanese females vs. American females’, University of Hawai’i Working Papers in ESL, 15(2):1–43. Woodfield, H. (2008) ‘Problematising discourse completion tasks:Voices from verbal report’, Evaluation and Research in Education, 21(1): 43–69. Yuan, Y. (2001) ‘An inquiry into empirical pragmatics and data-gathering methods: Written DCTs, oral DCTs, field notes, and natural conversations’, Journal of Pragmatics, 33(2): 272–292.

40

5 CORPORA Martin Weisser

Introduction Corpora (sing. corpus) are collections of naturally occurring (electronic) texts that are compiled according to specific design principles. Corpora are constructed either to reflect a language in general terms in all its diversity and aspects of use, or one of its specific sub-domains, such as Business English, English for Academic Purposes, Learner English, the language of blogs, or service encounters. The former type of corpus is usually referred to as a general corpus, and the latter as a specialised one. General corpora are assumed to fulfil two important criteria underlying their design, those of balance and representativeness. A corpus can be seen as balanced if its contents, as far as possible, represent as many different genres of the language to an appropriate degree in terms of selection and size. Its representativeness is judged by how well it reflects the typical aspects of the language. For specialised corpora, the issue of balance is frequently seen as somewhat less important, while representativeness is a more important factor. Corpora can essentially further be subdivided into written and spoken corpora, although there are also categories such as scripted speech, that is, texts written to be spoken (e.g., news broadcasts) that fall in between. And of course we should not forget that literary works, such as novels, often do contain ‘simulated’ spoken language in the form of dialogues between their characters. For the area of pragmatics, genuine spoken corpora are arguably most relevant, so the present chapter will focus on these. This does not mean, though, that the communicative strategies that are relevant to the study of pragmatics are not employed in writing, only that they may occur there in somewhat reduced form. This smaller repertoire is, for instance, due to the relative absence of questions or other markers of direct interaction. Corpus linguistics – as a method of linguistic investigation – generally uses corpora for developing or testing linguistic hypotheses based on real-life data, thereby complementing and/ or corroborating the researcher’s intuition, using relatively large samples of data that can be processed quickly and efficiently by means of the computer. Perhaps the three most important techniques employed to achieve this are: 1) keyword-in-context (KWIC) concordancing 2) the creation of word or n-gram frequency lists, and 3) the investigation of common collocations. 41

Martin Weisser Concordance Hits 477 Hit 1 2 3 4 5

KWIC all to this room." "Frequently." "How often?" "Well, some hundreds of times." "Then how many . "The man who wrote it was presumably well to do, ” I remarked, endeavouring to imitate fire and laughed heartily for some minutes. "Well, really!" he cried, and then he choked door. Large sitting-room on the right side, well furnished, with long windows almost to the just wondering whether I should not do well to follow them when up the lane

6 7 8 9 10

of affairs," said I; "and what then?" " Well, I found my plans very seriously menaced. mouth than the whole crowd of spectators, well dressed and ill—gentlemen, ostlers, and serv . SHERLOCK HOLMES,—You really did it very well. You took me in completely. Until after , and came down just as you departed. "Well, I followed you to your door, and it, and the muscles are more developed," " Well, the snuff, then, and the Freemasonry?" "I

11 12 13 14

where you rest it upon the desk?" "Well, but China?" "The fish that you have More simple." Mr. Jabez Wilson laughed heavily. "Well, I never!" said he. “I thought at months ago." "Very good. Now, Mr. Wilson?" "Well, it is just as I have been assistant, Mr. Holmes; and I know very well that he could better himself and earn

Figure 5.1 The first fourteen results (hits) of a KWIC concordance of the word well, nicely illustrating the distinction between its uses as a discourse marker (henceforth DM) and an ordinary adverb.

In concordancing, a number of examples are identified in a corpus through a search for individual word forms or more complex linguistic patterns, for example, verb paradigms.The results, or hits in technical terms, are generally presented in a context of a variable number of characters or words on either side, as shown in Figure 5.1. Single word frequency or n-gram lists (where n refers to any number of words occurring in sequence) allow the researcher to identify salient topics, recurring lexical items or chunks, or relevant grammatical constructions in corpora. Figure 5.2 shows an example of a four-gram list from a small literary corpus. Taking a closer look at this, we can see that the characters in the books are essentially in most cases reporting on their perceptions, understanding of, or thoughts on matters being discussed. In combination with the recurrent use of first-person pronouns, this feature indicates that we are dealing with

Total No. of N-Gram Types 80 Rank 1 2 3 4 5 6 7 8 9 10 11 12

Freq 20 14 12 12 11 11 10 10 9 8 8 8

Range 5 2 4 2 3 2 2 2 2 2 3 2

Total No. of N-Gram Tokens 536 N-gram i do not know i have no doubt i do not think to me to be but i did not it seemed to me i thought of the of a man who that i could not do not think that i have heard of i was unable to

Figure 5.2 N-Gram list

42

Corpora

non-argumentative texts. In contrast, argumentative texts, such as research papers, would use different communicative strategies involving more third-person pronouns and nouns, as well as different, ‘fact-reporting’, verbs, especially in the presentation/discussion of their own findings and in making reference to the findings or claims of others (cf. Hyland 1999). Collocation analysis employs statistics in order to identify the words that most frequently cooccur with a given word, referred to as the node, and within a certain span of it, that is, n many words to the left or right. Ordinarily, the span tends to be limited to a window of maximally five words. In addition to the statistical measures, collocations may also be identified through n-gram analyses, but then generally the node and its collocate have to occur in sequence. In corpus-based pragmatics, this type of analysis allows the researcher to identify what Partington (2014: 279) calls evaluative prosody, but has also been called semantic, pragmatic, discourse, or emotive prosody (Partington 2014: 279). The term itself mainly refers to a kind of connotative meaning that words may possess, and which predisposes them to co-occur with other words that have similar connotations. This type of priming (Hoey 2005) can be detected through collocation analysis and used to investigate the attitude of speakers/writers towards particular topics. Classic examples of such primings would be the pair of intensifying adverbs utterly and perfectly, where ‘the first. . . . tends to co-occur with negative items, [and] the second, more predictably, with good ones’ (Partington 2014: 283). While the insights gained from concordancing and collocation analysis can be seen as what has been described as corpus-based (Tognini-Bonelli 2001: 84–85) because a particular feature has to be known prior to conducting the analysis, those of n-gram analysis may be defined as corpus-driven (Tognini-Bonelli 2001: 84–85) as the information gleaned is more or less purely derived directly from the data.

Historical perspective and current work Corpora have been used for analysing pragmatic phenomena since at least the mid-1990s, but most of the earlier corpus-based work in linguistics (e.g., Stenström 1994 or Aijmer 1996) was carried out on corpora that were not specifically designed to examine pragmatic features, such as the London-Lund Corpus (see Svartvik 1990) or the Bergen Corpus of London Teenage Language (COLT; see Stenström et al. 2002). However, basic un-annotated corpora, that is, those that have not been enriched with additional linguistic information, only allow the researcher to look for relatively known expressions. Thus, in the past, research into pragmatic phenomena has mainly been corpus-based and limited to identifying patterns, such as well-known speech act expressions or pragmatic markers, extracting examples of these through concordances, and investigating their contexts of use. This research has already yielded valuable insights into interactional strategies like adjacency pairs or initiation-response-feedback (IRF) sequences (cf. Stenström 1994; Tsui 1994), detailed descriptions of individual speech acts such as thanking, apologising, requesting and offering, and demonstrated the importance of DMs in conversation (cf. Aijmer 1996). However, a substantial part of the pragmatics research using corpora still focuses only on a very limited picture of the many diverse functions of language use.This is perhaps not surprising given that there is – due to the influence of context and to the frequent use of indirectness in human language – by-and-large no direct mapping between linguistic form and illocutionary force. No doubt much of what we want to communicate in everyday language can be expressed in a multitude of different forms, and with an equal multitude of subtle differences in meaning.The main challenge for corpus pragmatics is how to identify and capture these differences in order to enable researchers in pragmatics to investigate associated illocutionary functions in different social and situational contexts. 43

Martin Weisser

For general corpora, the advent of morpho-syntactic annotation in late 1959/early 1960s (Van Halteren 1999: 10) provided a major advance in the exploitability of corpus data because it helped disambiguate between ‘grammatically polysemous’ word forms, that is, those where one word form may belong to different word classes, and thus perform different lexico-grammatical functions. A classic example of this is the word form round, which may represent ‘a preposition, adverb, noun, verb or adjective’ (Van Halteren 1999: 5). A similar ‘revolution’ was long needed in pragmatics, and has hopefully now begun with the introduction of annotated corpora that contain explicit information about pragmatic phenomena, such as speech acts, as well as other contextually relevant information. To illustrate the potential for such corpora to inform pragmatics research, I first want to describe where some of these corpora originated, how they have evolved over the years, and the different types of content they may encompass. I will then illustrate how these corpora can be exploited in improving our understanding of spoken – and perhaps also written – language. Earlier pragmatically enriched corpora were predominantly concerned with investigating and annotating transactional and/or problem-solving interactions. The motivation for creating such corpora was frequently less of a linguistic nature, rather than to foster advances in spoken language technology for human–computer interaction. Many of these corpora, for example, the HCRC Maptask Corpus (Carletta et al. 1996), the COCONUT Corpus (Di Eugenio et al. 1998), the Monroe Corpus (Stent 2000), and Trains (Heeman and Allen 1995), were created mainly by computer scientists for the investigation of dialogue structure and (conversational) moves/dialogue acts (roughly equivalent to speech acts) or for modelling conversational agents. The pragmatic annotation they contain often reflects this limited, purpose-oriented scope. An example is the set of twelve move codes used in the Maptask Corpus (Kowtko et al. 1993): a) INSTRUCT: b) CHECK: c) QUERY-YN: d) QUERY-W: e) EXPLAIN: f) ALIGN: g) CLARIFY: h) REPLY-Y: i) REPLY-N: j) REPLY-W: k) ACKNOWLEDGE: l) READY:

direct or indirect request/instruction checking one’s understanding/request for confirmation yes-no question wh-question stating new information (unsolicited) verifying agreement or understanding clarifying prior information upon request positive response to yes-no question, check, or align negative response to yes-no question, check, or align reply to wh-question or check backchannelling acknowledgement indicating the beginning of a new interaction sequence.

In this list, we can observe a mix of speech act-like (a, b, e, f, g, k, l) and syntactic labels (c, d, h, i, j). These categories are clearly not suitable for representing the wide range and different nuances of pragmatic meaning I referred to earlier, while they may well be sufficient to achieve the task at hand, finding a common route to a destination, based on two maps that contain different levels of detail. In order to understand the potential options for categorising or classifying speech acts, we first need to take a closer look at the historical development of descriptive taxonomies in speech act theory. The first two taxonomies were established by the ordinary-language philosophers Austin and Searle. Austin (1962), in developing his theory of performative verbs, distinguished between Verdictives (uttering a verdict, e.g. true or false, guilty or not guilty), Exercitives (exercising power verbally, e.g. through appointing, sentencing, warning), Commissives (making 44

Corpora

commitments), Expositives (explaining one’s verbal strategies, e.g. I assume. . . ., I concede. . . .), and Behabitives (expressing speaker attitudes or ‘(im)politeness formulae’, as in acts of apologising, congratulating, swearing). Searle later criticised this taxonomy and devised his own, comprising Representatives (‘commit[ting] the speaker. . . . to the truth of the expressed proposition’, expressing a belief), Directives (‘get[ting] the hearer to do something’, including providing information), Commissives (‘commit[ting] the speaker. . . . to some future course of action’), Expressives (cf. Austin’s behabitives), and Declarations (performing a verbal act + bringing about a change, e.g. firing someone) (1975: 354–358). Both taxonomies are predominantly speaker-oriented, and in essence ignore the collaborative nature of verbal interaction, as stressed by Grice (1989) in his Co-operative Principle (and already implicit in the Maptask scheme), as well as features of discourse that do not rely on performatives alone. To overcome some of these issues, and to devise a consensus model for annotating dialogues, the Discourse Resource Initiative (DRI), a consortium of academics and members of telecommunications companies from around the world, engaged in collaborative efforts in the late 1990s. To this end, the DRI held three workshops between 1995 and 1998, one of the outcomes being an annotation scheme called Dialogue Act Markup in Several Layers (DAMSL; Allen and Core 1997). This scheme classifies the communicative function of acts according to four dimensions: 1) 2) 3) 4)

Communicative status Information level and status Forward-looking communicative function Backward-looking communicative function

Within these dimensions, it defines the following set of dialogue act labels: (Re-)Assert, Abandoned/Uninterpretable, Accept(-part), Action-directive, Answer, Assert, Commit, Completion, Conventional Closing, Explicit-performative, Hold, Info-Request, Maybe, Offer, Open-option, Other-forward-function, Other-statement, Reject, Repeat-rephrase, Self-talk, and Signal-nonunderstanding. Even though these categories (or their labels) may not always be intuitively transparent, and may still not cover all the potential acts we can accomplish with speech, they already go far beyond the taxonomies discussed earlier. For instance, they provide better labels for indicating the interaction between speakers by trying to distinguish between forward-looking, that is, initiating (e.g. Offer, Open-option, Info-Request), and backward-looking, that is, responding (e.g. Accept, Reject, Answer), elements in interaction, as well as acknowledging that not everything that occurs in a dialogue necessarily represents successful interaction (e.g. Abandoned/Uninterpretable, Repeat-rephrase, Self-talk). Furthermore, they have already been applied, albeit with limited success, in annotating task-oriented corpora such as Coconut, Monroe, and Trains manually (see Weisser 2014 for an in-depth discussion). However, based on these attempts, and the feedback of participants of the third DRI workshop, we can conclude that, overall, the DAMSL scheme has proved too unwieldy, and it is perhaps the inherent difficulties in categorising speech/dialogue acts according to the different dimensions and labels which led the creators of the pragmatically annotated version of the Switchboard Corpus (see Jurafsky et al. 1997) to create the simplified SWBD-DAMSL scheme. Despite some inconsistencies in the annotations and application of the scheme, this version of the Switchboard Corpus, a collection of 2,430 telephone conversations by native speakers of American English (Godfrey and Holliman 1997), is particularly noteworthy because it is the first pragmatically annotated corpus of unconstrained natural dialogue, revolving around fifty-two topics that range from AIDS, drug testing, gun control and music, to woodworking, thus covering a wide variety of different everyday interests. The inconsistencies referred to earlier mainly consist in the application of 45

Martin Weisser

speech-act labels such as sv, which is meant to indicate opinions, but is often used in cases where no expression of an opinion is found, or ad, which indicates an action directive, where such directives may also include imperatives that start with Let’s, which would be better classified as suggestions and so forth (Weisser 2014). More recent pragmatically annotated corpora, which were created with (more) linguisticsoriented objectives in mind (rather than language engineering), are the SPICE-Ireland (Kirk 2013) and the SPAADIA corpus (Leech and Weisser 2013). The former is based on the spoken part of the Irish sub-corpus of the International Corpus of English (ICE), contains 626,597 words from ‘15 spoken registers and 300 texts’ (Kirk 2013), and has been annotated pragmatically and, to a fairly large extent, also prosodically. However, as laudable as the project is, it has two major drawbacks concerning these annotations (see Figure 5.3). 1LOOk at your 8vEgetables% 1ThInk 1flAvour% while you 're 1Actually 1chOOsing them% and if you 1cAn 1fEE1 them% and 1pIck up% and get the 1crIspest% and 1frEshest ones you 1cAn%

Figure 5.3  SPICE-Ireland sample

The first of these issues is that, in an effort to keep the text more readable and avoid the use of heavy coding, the annotation itself mixes prosody, pragmatic codes, and orthography. Ironically, however, this actually makes the text less readable, and hence also less easily searchable.The prosodic markup that can be seen in Figure 5.3 includes pauses (marked or ), intonational phrase boundaries (%), and emphasis (marked through capital letters in places other than in sentence-initials or proper nouns), as well as eleven tone tunes (marked by numbers 1–11; Kallen and Kirk 2012: 35–38), which are directly interspersed with the text. The second issue relates to the low number of speech-act categories, which does not allow researchers to conduct any fine-grained pragmatic analyses. These categories comprise Searle’s five original ones (marked . . ., . . ., . . ., . . ., and . . .), plus four extensions (Kallen and Kirk 2012: 28–35): 1) indeterminate conversationally relevant utterances (. . .): backchannels and minimal responses 2) social expressions (. . . ): greetings, farewells, etc. 3) pragmatically non-analysable utterances (. . . . ): units that are incomplete or not understandable 4) keyed utterances (. . . . ): units that suggest a particular speech act, due to their form, but are meant as another. As far as 1) is concerned, these definitely constitute important speech-act units, but are by no means ‘indeterminate’, as they generally have an acknowledging function. The items under 2) clearly also represent speech acts, but were in fact already subsumed under Austin’s behabitives and Searle’s expressives, so it is interesting to find them here as a separate category. The addition of 3) reflects the fact that the scheme is being used with real-life data, where unintelligible material is not uncommon. Extension 4) is special in the sense that it does not in itself represent a speech-act category, but indicates a special case of ‘indirectness’, for instance the use of irony 46

Corpora

or humour, where what is supposed to be conveyed may actually mean the opposite of what is said. Hence, the K is appended to another speech-act category label to indicate this divergence. In addition to marking these categories, the corpus contains other pragmatics-relevant markup, namely for DMs (marked with *, and hyphenated if the marker contains multiple words) and tags (marked with @). However, DMs are not treated as constituting separate speech acts, even if they normally realise either acknowledgements or function to ‘initiate’ new subtopics/transactions. This is, unfortunately, a common problem in almost all annotation schemes to date, despite the fact that DMs and their functions have received extensive treatment in the pragmatics literature, including whole books such as Schiffrin (1987) or Jucker and Ziv (1998). Regarding tags, we encounter the opposite issue, as question tags are generally marked as separate speech acts in SPICE-Ireland, while in fact they generally function as ‘postmodifiers’: they either turn the ‘default behaviour’ of declarative structures from expressing a statement into becoming a request for confirmation; or they may represent an attempt to involve the listener more (Tottie and Hoffmann 2006: 302). Their use as genuine directives is relatively rare in comparison. As stated in the discussion of prior attempts at defining speech/dialogue act taxonomies, the variety of features pertaining to spoken (and written) interaction that may be of interest to researchers in pragmatics cannot really be covered by such a relatively coarse scheme as that adopted in the SPICE. For instance, researchers are often interested in analysing question– response patterns, but as both direct imperatives and requests for information, for instance, are conflated under directives in Searle’s original taxonomy, and thus also in the corpus, it becomes a non-trivial task to extract such patterns from the corpus for investigation. Conversely, it would take an equal effort to extract all pure imperatives from the data, as well as particular instances of any of the other broad categories. This is why a more fine-grained taxonomy, using labels that are closer to what researchers are in fact interested in, rather than relatively abstract hypernyms, would be more appropriate for pragmatics research. One corpus that does make use of such a taxonomy is the SPAADIA corpus (ver. 1 released in 2013), which originated from the considerably larger SPAAC corpus (Leech and Weisser 2003), and contains thirty-five transactional dialogues (some 21,265 words) of Trainline timetable and booking exchanges between an operator and callers. The original (SPAAC) taxonomy used to annotate this corpus comprised forty-one speech acts, but has since been revised and expanded to fifty-seven, which allow for the automatic identification and annotation of around 200 combinations automatically within the Dialogue Annotation and Research Tool (DART; Weisser 2016a). As the new DART taxonomy is too extensive to discuss in detail here, and also likely to be further updated in future, please refer to http://martinweisser.org/publications/ DART_scheme.html for full details of all the labels and approximate functions they represent. Despite its fine-grained nature, which already makes it more useful for many types of pragmatics research than the coarse SPICE-Ireland, this taxonomy is designed to be generically applicable to any type of dialogue, and is therefore more flexible than those that were primarily created for language engineering purposes, such as the DAMSL and SWBD-DAMSL, even if the latter has also been applied to the Switchboard corpus. For a more detailed discussion of this, see Weisser (2014). The SPAADIA corpus uses the DART annotation format, a very simple – and, in contrast to the SPICE format, highly readable – form of XML, depicted in Figure 5.4. The first two lines of Figure 5.4 contain XML-specific information that is unrelated to the linguistic content.The first line contains the so-called XML declaration, which simply identifies the file as being in XML format, while the second, optional, one contains a link to an external style sheet that specifies how the content should be displayed in an application such as a web 47

Martin Weisser

Figure 5.4  DART XML format

browser. The XML elements – often also referred to as ‘XML tags’ – that mark up the content appear in angle brackets (), with the name of the tag following the opening angle bracket. If these tags are paired, they may either contain textual content or other tags within them, and then the closing element has a forward slash (/) before the name to indicate the end of each level within the hierarchical structure. The linguistic content in Figure 5.4 is contained in what I have labelled ‘dialogue container’, an XML element that identifies the content as a dialogue. This element has three attributes, an ‘id’, which specifies the name of the dialogue file to distinguish it from other dialogues within the corpus, ‘corpus’, the name of the corpus the dialogue is a part of, and ‘lang’, which identifies the (main) language used in the corpus in abbreviated form, in our example ‘en’ for ‘English’. Each dialogue is separated into turns (), with the attributes ‘n’ for ‘number’ and ‘speaker’ to identify the current speaker. Each turn is subdivided into communicative c-units that express a single or combined speech act, as can be seen in the example in Figure 5.5. As Figure 5.5 shows, the annotation format (see Leech and Weisser 2013 for details) for each unit contains syntactic information in the opening tag (, , , ), an identifying number (10–14), a speech-act label (init, reqInfo, answ[er], answ[er]Elab[oration]), semantic (topic; number, credit card) and semantico-pragmatic (mode; closed, closure) information, as well as polarity, if applicable. 48

Corpora

Figure 5.5  SPAADIA (ver. 1) coding sample

Corpora annotated in this multilevel annotation format can easily be analysed by extracting formal (syntactic) and functional (speech-act) properties for individual speakers or speaker populations, in order to compare their interactional behaviour, as illustrated in Weisser (2016b). Here, the communicative strategies used by a British and an American call centre agent are profiled and compared to one another, as well as to those employed by their respective callers, by investigating the four areas of initiative and information-management, efficiency, (im)politeness and deference and (in)directness. This research was carried out on an updated version of the SPAADIA corpus and a subset of SRI’s Amex Travel Agent Data (Kowtko and Price 1989), both of which were annotated in DART automatically and then post-edited to ensure accuracy at all levels of annotation. A revised version of DART (to be released in 2016) was then used to filter and extract the relevant speech-act statistics, and further analyses were conducted using DART’s built-in special concordancing facility.

Critical issues and further directions One major critical issue in annotating corpora pragmatically has always been the depth and precision of the speech-act taxonomies employed. As I have tried to point out, the many diverse functions of language use certainly require us to describe speech acts with a certain degree of precision, a challenge which demands a relatively fine-grained taxonomy. However, there is a trade-off between this precision and what can practically be annotated with an equally high degree of consistency, at least if the annotation is carried out entirely manually, as was done for most of the projects referred to here. In manual annotation, it is necessary for annotators to pay close attention to a variety of features pertaining to both form and function, something that is highly time-consuming and tiring, so that it becomes extremely difficult to achieve any consistency, especially working with a team of annotators. This is probably at least one of the reasons 49

Martin Weisser

the annotators of SPICE-Ireland opted for their limited set of speech-act labels, which made it possible to annotate a sizeable amount of data using a limited budget. However, as DART has been publicly and freely available to the research community since 2015, it should no longer be necessary to sacrifice breadth for quantity, and more and larger pragmatically annotated corpora should become available soon. Some of my own ongoing, longer-term projects that use DART for this purpose are the creation of a new annotated version of the Switchboard corpus, as well as of the Hong Kong Corpus of Spoken English (Cheng et al. 2008). One feature all current annotations schemes/corpora lack, though, is a way to encode the participants’ relative status in a spoken interaction. Incorporating this would probably provide even deeper insights into verbal interaction. An extension of speech-act annotation to written language also appears realistic and desirable. To this end, I have recently started developing the Text Annotation and Research Tool (TART), DART’s ‘big brother’, in order to attempt performing automated annotation of written texts, and to see whether, or to what extent, the categories devised for DART can be applied to the annotation and analysis of written language.

Conclusion In this chapter, I have tried to achieve two objectives: to trace the development and rough timeline of the use of corpora for pragmatic purposes, and to point out where some of the main issues and opportunities in doing corpus pragmatics lie. A considerable amount of the ‘research’ in pragmatics, however, especially as conducted by members of the so-called Anglo-American School (Horn and Ward 2004: x), is still based on invented examples. And even if proponents of the ‘European Continental School’ (Horn and Ward 2004: x), particularly those involved in corpus linguistics, have been working with real-life data for a considerable time, they have generally been conducting in-depth qualitative analyses on relatively small quantities of data. However, the advent of larger-scale annotated corpora and associated technologies for producing them, such as DART, have provided a way forward that could liberate pragmatics from these shortcomings and allow us to gain even deeper insights into how spoken – and perhaps also written – language works.

Suggestions for further reading Adolphs, S. (2008) Corpus and Context: Investigating Pragmatic Functions in Spoken Discourse. Amsterdam/ Philadelphia: John Benjamins. This book illustrates how more ‘traditional’ corpus linguistics techniques, such as concordancing, can be applied to investigating speech acts in context, based on non-annotated corpora. Aijmer, K. and Rühlemann, C. (2014) Corpus Pragmatics: A Handbook. Cambridge: Cambridge University Press. The first, and so far only, comprehensive handbook on the emerging subdiscipline of corpus pragmatics, covering a variety of different topics and techniques in the area. Archer, D., Aijmer, K. and Wichmann, A. (2012) Pragmatics: An Advanced Resource Book for Students. London/ New York: Routledge. Section A2.5 provides a brief summary of some of the basic work that has been done in annotating and investigating corpora pragmatically. Weisser, M. (2014) ‘Speech act annotation’, in K. Aijmer and C. Rühlemann (eds.) Corpus Pragmatics: A Handbook. Cambridge: Cambridge University Press. 84–113. This chapter provides an overview of different annotation formats, their application and their relative merits and demerits, focussing mainly on a comparison of DAMSL-based dialogue schemes vs. the DART scheme.

50

Corpora

References Aijmer, K. (1996) Conversational Routines in English: Convention and Creativity. London: Longman. Aijmer, K. and Rühlemann, C. (2014) Corpus Pragmatics: A Handbook. Cambridge: Cambridge University Press. Allen, J. and Core, M. (1997) Draft of DAMSL: Dialog Act Markup in Several Layers. Available from: ftp://ftp. cs.rochester.edu/pub/packages/dialog-annotation/manual.ps.gz (accessed 7 May 2015). Austin, J. (1962) How to Do Things with Words. Oxford: Oxford University Press. Carletta, J., Isard, A., Isard, S., Kowtko, J., Doherty-Sneddon, G. and Anderson, A. (1996) HCRC Dialogue Structure Coding Manual, (HCRC Technical Report 82). Edinburgh. University of Edinburgh. Cheng, W., Greaves, C. and Warren, M. (2008) A Corpus-Driven Study of Discourse Intonation. Amsterdam: John Benjamins. Di Eugenio, B., Jordan, P. and Pylkkianen, L. (1998) The COCONUT Project: Dialogue Annotation Manual, (ISP Technical Report 98–1). Pittsburgh, PA: University of Pittsburgh. Godfrey, J. and Holliman, E. (1997) Switchboard-1 Release 2. Philadelphia: Linguistic Data Consortium. Available from https://catalog.ldc.upenn.edu/LDC97S62 (accessed 1 August 2015). Grice, H. P. (1989) Studies in the Way of Words. Cambridge, MA: Harvard University Press. Heeman, P. A. and Allen, J. (1995) The Trains 93 Dialogues, (TRAINS Technical Note 94–2). Rochester, NY: University of Rochester. Hoey, M. (2005) Lexical Priming: A New Theory of Words and Language. London/New York: Routledge. Horn, L. and Ward, G. (2004) The Handbook of Pragmatics. Oxford: Blackwell. Hyland, K. (1999) ‘Academic attribution: Citation and the construction of disciplinary knowledge’, Applied Linguistics, 20(3): 341–367. Jucker, A. and Ziv,Y. (eds.) (1998) Discourse Markers: Description and Theory. Amsterdam: John Benjamins. Jurafsky, D., Shriberg, E. and Biasca, D. (1997) Switchboard SWBD-DAMSL Shallow-Discourse-Function Annotation Coder Manual. Available from: www.stanford.edu/~jurafsky/ws97/ics-tr-97–02.ps (accessed 7 May 2015). Kallen, J. and Kirk, J. (2012) SPICE-Ireland: A User’s Guide. Belfast: Cló Ollscoil na Banríona. Kirk, J. (2013) ‘Beyond the structural levels of language: An introduction to the SPICE-Ireland corpus and its uses’, in J. Cruickshank and R. McColl Millar (eds.) After the Storm: Papers from the Forum for Research on the Languages of Scotland and Ulster Triennial Meeting, Aberdeen 2012. Aberdeen: Forum for Research on the Languages of Scotland and Ireland. 207–232. Kowtko, J., Isard, S. and Doherty, G. (1993) Conversational Games within Dialogue. Edinburgh: Human Communication Research Centre. Kowtko, J. and Price, P. (1989) ‘Data collection and analysis in the air travel planning domain’, in Proceedings of DARPA Speech and Natural Language Workshop, October 1989. Leech, G. and Weisser, M. (2003) ‘Generic speech act annotation for task-oriented dialogues’, in D. Archer, P. Rayson, A. Wilson and T. McEnery (eds.) Proceedings of the Corpus Linguistics 2003 Conference. Lancaster: Centre for Computer Corpus Research on Language Technical Papers, Lancaster University. 441–446. Leech, G. and Weisser, M. (2013) The SPAADIA Annotation Scheme. Available from: http://martinweisser. org/publications/SPAADIA_Annotation_Scheme.pdf (accessed 27 January 2016). Partington, A. (2014) ‘Evaluative prosody’, in K. Aijmer and C. Rühlemann (eds.) Corpus Pragmatics: A Handbook. Cambridge: Cambridge University Press. 279–303. Schiffrin, D. (1987) Discourse Markers. Cambridge: Cambridge University Press. Searle, J. (1975) ‘A taxonomy of illocutionary acts’, in K. Gunderson (ed.) Language, Mind and Knowledge, (Minnesota Studies in the Philosophy of Science, vol. III). Minneapolis: University of Minnesota Press. 344–369. SPAADIA Corpus (plus concordancer). Available from: http://martinweisser.org/#spaadia (accessed 27 January 2016). Stenström, A. (1994) An Introduction to Spoken Interaction. London: Longman. Stenström, A., Andersen, G. and Hasund, I. (2002) Trends in Teenage Talk. Amsterdam: John Benjamins. Stent, A. (2000) The Monroe Corpus (Technical Report 728/TN 99–2). Rochester, NY: University of Rochester. Svartvik, J. (ed.) (1990) The London-Lund Corpus of Spoken English: Description and Research. Lund: Lund University Press. Switchboard annotated pre-release version. Available from: www.stanford.edu/~jurafsky/swb1_dialogact_ annot.tar.gz (accessed 7 June 2015).

51

Martin Weisser Tognini-Bonelli, T. (2001) Corpus Linguistics at Work. Amsterdam: John Benjamins. Tottie, G. and Hoffmann, S. (2006) ‘Tag questions in British and American English’, Journal of English Linguistics, 34(4): 283–311. Tsui, A. (1994) English Conversation. Oxford: Oxford University Press. Van Halteren, H. (1999) Syntactic Wordclass Tagging. Dordrecht: Kluwer Academic Publishers. Weisser, M. (2014) ‘Speech act annotation’, in K. Aijmer and C. Rühlemann (eds.) Corpus Pragmatics: A Handbook. Cambridge: Cambridge University Press. 84–114. Weisser, M. (2016a) ‘DART – the Dialogue Annotation and Research Tool’, Corpus Linguistics and Linguistic Theory. Weisser, M. (2016b) ‘Profiling agents and callers: A dual comparison across speaker roles and British vs. American English’, in L. Pickering, E. Friginal and S. Staples (eds.) Talking at Work: Corpus-Based Explorations of Workplace Discourse. London: Palgrave Macmillan.

52

Nonverbal communication

6 BRITISH SIGN LANGUAGE (BSL) Gary Quinn

Introduction Many people believe that the hands are the focal point of sign language. However, many elements of sign languages such as British Sign Language (BSL) are conveyed through the nonmanual features of the language. In other words, much of BSL is expressed in ways other than the use of the hands, for example, through facial expression and body language (Sutton-Spence and Woll 1999: 81), both of which make significant contributions to the pragmatics of signed languages.This is because signed languages use a markedly different modality to spoken ones. In a spoken language such as English, the modality is oral/aural, and this imposes certain constraints on how the language functions. By contrast, the visual-spatial nature of BSL allows for language to be expressed very differently (Stokoe 1960; Deuchar 1984; Brennan 1992). BSL is an indigenous British language that many Deaf1 people use as their native language. The term ‘visual-spatial’ refers to both the perception and production of BSL: it is produced in a medium perceived visually, using movement of the hands, arms, head and the upper part of body, including the face (Brennan et al. 1984; Deuchar 1984). BSL is produced in what can be described as a ‘controlled’ spatial area. This area of signing space ‘extends from approximately just above the head to the waist and in width from elbow to elbow when the arms are held loosely bent’ (Brennan 1992: 22). By contrast, spoken languages are particularly suited to the auditory medium, expressed by the mouth and received by the ear, and we know they have evolved over thousands of years. In comparison with spoken languages much less is known about sign languages, and their study is still relatively recent, but it is clear that sign languages may have also evolved over time. Different varieties of BSL exist throughout the United Kingdom that can be considered similar to different dialects of spoken English. There is considerable regional variation in BSL and in Scotland alone, there are an estimated five signs for the colour green, used by the Deaf communities in Glasgow, Edinburgh, Aberdeen, Dundee and the Catholic Deaf community in the west of Scotland. Much of this variation is attributable to a process described as ‘schoolization’ (Quinn 2010). Previously, regional dialects in BSL developed in the residential schools for the deaf at different locations throughout the United Kingdom. It is difficult to evidence exactly when BSL appeared or developed. This is partly because it is not possible to document BSL in a written form in the same way as English and also because, until recently, there was no way of recording it in a visual format such as video. However, there is some evidence that sign language was in usage in the sixteenth century (Sutton-Spence and 55

Gary Quinn

Woll 1999: 34). One historic source of information on BSL is the Chirologia or the National Language of the Hand, a book published by John Bulwer in 1644 (Deuchar 1984; Kyle and Woll 1985; Miles 1988). Of the signs described within this work some are still in use in the present whilst others are not (Miles 1988: 15). Sign language was first brought to the attention of the linguistics world by William Stokoe, in the United States, during the 1960s. This had positive effects for sign language communities across the world. In Britain, inspired by Stokoe’s work, Mary Brennan established a research group known as the Edinburgh British Sign Language Project, in 1979, culminating in the publication of the influential Words in Hand (Brennan et al. 1984). These early pieces of research proved that sign languages are languages in the same way that spoken languages are considered to be (Sutton-Spence and Woll 1999: 9). Deuchar’s study highlights the distinctively different structure that BSL has in comparison with English, though both languages are capable of conveying the same meaning and emphasises that signs represent neither the sounds nor meanings of English words (Deuchar 1984: 8).

Pragmatics of BSL and other sign languages Politeness Whilst there are papers about politeness forms in sign language (Brennan 1994, Sutton-Spence and Woll 1999), these have mostly focused on the social-norm view of politeness (first-order politeness), in other words, what people perceive as polite behaviour and use of language. Sutton-Spence and Woll (1999: 241–253) devote an entire chapter to the use of unacceptable signs in BSL and a discussion of how the lexicon has been modified for political correctness. Second-order politeness, which forms part of the study of pragmatics, is a theoretical construct used within theory of language usage (Watts et al. 1992: 3), and has been described in the works of Leech, the linguist responsible for devising the politeness principle: Minimize (all things being equal) the expression of impolite beliefs; maximize (all things being equal) the expression of polite beliefs. (Leech 1983: 81) The politeness principle, in turn, is related to the cooperative principle devised by Grice (1975). Issues relating to politeness and pragmatics are not concepts exclusive to spoken languages; meaning in interaction is equally important in the use of BSL. In a similar way to other-initiated repair in English, when, for example, an addressee may indicate that they have not understood the speaker, in sign languages it is equally possible for the addressee to accept ‘blame’ for misunderstanding as a polite way of indicating that the signer has not been understood (Quinn 2004). In BSL, the equivalent other-initiated repair may consist of the addressee tilting their head to the side and furrowing their brows. Quinn’s study also indicates that non-manual features, such as head tilting and eyebrow movements, play an important role in second-order politeness in BSL, and research into American Sign Language (ASL) by Roush (2007) and Hoza (2007) supports this.

Reference in sign language Compared to spoken languages, the way that reference may be present can be quite different in a language that uses a visual and spatial modality. Therefore, in BSL as in other signed languages, reference can work in a very different way. BSL naturally uses the hands and the 56

British Sign Language

signing space, the area in front of the signer’s body, in producing references. Objects, concepts and ideas can be established in a certain location, within the signer’s signing space and later referred back to in the sign language equivalent of pronoun use. In comparison to spoken languages, these references can be more detailed and incorporate information on location through the use of what are referred to as ‘classifier predicates’ and ‘constructed action’ (Dudis 2002: 53). To illustrate this, it is worth considering the concept of ‘buoys’ (Liddell 2003: 223–260). The use of buoys occurs when a signer has produced a sign using two hands and then retains one in the finishing place of that sign, while simultaneously using the other hand to produce subsequent signs. The buoy is the term given to the hand retaining the original sign, which is held there as a marker and may be referred back to by use of the active hand in a form of demonstrative reference. Liddell (2003) describes different types of buoy, one of which is referred to as a ‘pointer buoy’, using the index finger to indicate the referent. An additional type of buoy, not present in Liddell’s study of ASL, was identified by VogtSvendsen and Bergman (2007) in their work on Norwegian and Swedish Sign Languages. This new form of buoy is called a ‘point’ buoy as it is used to indicate a point in either space or time. They identified just two handshapes used for this type of buoy and discovered that one distinguishing feature of the point buoy is its use only within a two-dimensional plane, rather than a three-dimensional representation. Vogt-Svendsen and Bergman also found the plane used to be, generally, the horizontal one (2007). BSL has examples of all of these types of buoys.

Turn-taking in sign language Turn-taking in sign language works differently to the descriptions given of the process in spoken languages. During a signed conversation the addressees usually maintain eye gaze on the signer, and wait for their turn to sign. They may use back channels such as head nodding and other non-manual features. The signer signals that his or her turn is close to completion by changing the pace of signing, either by speeding up or slowing down before stopping completely (Johnston and Schembri 2007: 262).The signer then uses his or her eye gaze to direct the turn to one of the ‘listeners’ and may supplement this with the use of a pointed index finger or upturned palm in the direction of the listener (Van Herreweghe 2002: 154; Johnston and Schembri 2007: 263), who has indicated readiness to take the turn by raising his or her hands. Baker and Cokely (1980) describe how conversation is regulated with strategies like hand waving and tapping someone to get attention. Addressees also use regulators to show they are ready to accept someone else’s turn (Smith 1999; Smith and Sutton-Spence 2005). The norm is for addressees to maintain eye gaze on the signer’s face (Siple 1978; Swisher et al 1989). This is supported by research on turn-taking in more formal situations which found that a signer cannot begin to sign unless he or she has eye contact with other participants (Baker 1977; Mather 1996). Baker’s research (1977) has been supported by Van Herreweghe (2002), who also studied the strategies signers use to claim a turn. Baker (1977) states a number of strategies used to signal turn-claiming by the addressee, and these include an increase in size and quality of head nodding, holding the hand with the palm upwards and higher than the fingertips, waving a hand to the speaker, moving their eye gaze away from the speaker, and interrupting to claim the turn. What Baker did not cover was an account of trouble-source turns, or of grounding and repairs. Research is ongoing in the area of sign pragmatics, and although relatively little has been published to date, the body of work is slowly growing. 57

Gary Quinn

Overlap Overlap is another issue that can occur in signed conversation. Coates and Sutton-Spence (2001) conducted research to examine whether the two modes of conversational organisation, single floor and collaborative floor, apply to sign language interactions and observe potential differences for gender. Evidence suggests that both single floor and collaborative floor are used within BSL (Coates and Sutton-Spence 2001), supporting the spoken-language models developed by Edelsky (1981). Smith and Sutton-Spence (2005) found that Deaf people have advantages in collaborative floor as signers can see more in their peripheral vision.This means that signs can be ‘held’ in a way that words cannot and, whereas overlapping speech can create interfering noise, signs produced simultaneously are less interfering.

Repairs in signed language Dively (1998) has studied conversational repairs in ASL. Her studies found evidence of a variety of repair types that mirror those used in spoken English, such as self-repairs, other repairs, replacement repairs and word-search repairs. This study also confirmed that, as in spoken English, the most common form of repair was self-repair. However, she also uncovered some types of repair that differed from spoken English. These included the use of ‘nonhanded’ signs, or non-manual features, which involved adding extra meaning to a manual sign by the use of particular facial expressions or upper body posture. An example of this might be when an addressee moves his or her upper body back and away from the signer to indicate doubt about the information he or she is receiving. Another repair technique peculiar to sign language is the use of space, which is not possible in the linear delivery of a spoken language.This corrective use of space might be apparent if, for example, a signer uses a particular location in space to set up a concept and later refers back to the concept by pointing at a different location, but, realising the mistake, reverts to the original, correct location. Imagine a signer talking about a recent visit to Edinburgh who establishes ‘Edinburgh’ in the conversation on his right-hand side. Later during the same conversation, he refers to Edinburgh again, but this time points to his left. The repair occurs when, upon realising the mistake, he corrects the location of the sign. Dively (1998) also found that other-initiated repair often related to use of regional signs that might be unfamiliar to the listener.

Prosody in sign language As with any spoken language, in order to use BSL proficiently, it is not sufficient merely to know the vocabulary and grammatical structures of the language; prosody is an important part of all languages, including signed languages. Similar to spoken languages, signed prosody involves rhythm, emphasis, pausing and rate, as well as intonation. The prosodic elements are not only used to convey paralinguistic information about the signer’s mood and emotions or just to support the morphosyntactic structures of the language, but play an important role in conversation management. Research by Nespor and Sandler (1999) shows that the equivalent of intonation in sign languages is facial expression and that signed and spoken intonation systems are remarkably similar in function and complexity although delivered via different modalities.

Conclusion While scholars have conducted considerable research into the area of pragmatics in spoken languages, comparatively little research has been done into the pragmatics of signed languages. Linguists started researching sign languages only in the 1980s and began with the more structural 58

British Sign Language

features of phonology, semantics and grammar. More recent research, as discussed earlier, has broadened its scope to include pragmatics and discourse analysis. Most striking here has been the number of similarities of pragmatic features and functions particularly given the inherent differences between signed and spoken languages with regard to their modality (visual-spatial as opposed to oral-aural), their simultaneity (as opposed to linear presentation of meaning) and their grammatical and phonological systems.

Suggestions for further reading Stokoe, W. C. (1960) ‘Sign language structure: An outline of the visual communication system of the American deaf ’, Studies in Linguistics, Occasional Papers 8, Buffalo, NY: University of Buffalo Press. A seminal text in the field of sign language linguistics, this work by Stokoe is essential reading for any interested in the field. Stokoe’s publication was the first to challenge accepted beliefs that sign languages were just gesture systems and paved the way for much of the subsequent research in sign linguistics. Sutton-Spence, R. and Woll, B. (1999) The Linguistics of British Sign Language: An Introduction. Cambridge: Cambridge University Press. This book provides an accessible and comprehensive overview of the linguistics of BSL. Johnston,T. and Schembri, A. (2007) Australian Sign Language: An Introduction to Sign Linguistics. Cambridge: Cambridge University Press. Although about Australian sign language (Auslan), this publication is a natural sequel to the work by Sutton-Spence and Woll and updates many of the ideas introduced in their book. Liddell, S. K. (2003) Grammar, Gesture, and Meaning in American Sign Language. Cambridge: Cambridge University Press. For further information about referencing and the different ways it can be used in sign languages, Liddell’s book provides detail about buoys and tokens in ASL. Nespor, M. and Sandler,W. (1999) ‘Prosody in Israeli Sign Language’, Language and Speech, 42(2–3): 143–176. Although rather technical, this article is one of several published by this group of researchers which examine prosodic features, particularly intonation, in Israeli sign language. Their general findings and premises are applicable to many other, if not all signed languages.

Note 1 Following the convention suggested by Woodward (1972) and now commonly accepted, the use of the uppercase ‘Deaf ’ denotes deaf people who use sign language and regard themselves as members of a linguistic and cultural minority, while the lowercase ‘deaf ’ refers to those with hearing loss.

References Baker, C. (1977) ‘Regulators and turn-taking in American Sign Language discourse’, in L. A. Friedman (ed.) On the Other Hand: New Perspectives on American Sign Language. New York: Academic Press. 215–236. Baker-Shenk, C., and Cokely, D. (1980) American Sign Language: A Teacher’s Resource Text on Grammar and Culture. Silver Spring, MD: T. J. Publishers. Brennan, M. (1992) ‘The visual world of British Sign Language: An introduction’, in D. Brien (ed.) Dictionary of British Sign Language/English, 1–133. London: Faber and Faber Limited. Brennan, M. (1994) ‘Pragmatics and productivity’, in I. Ahlgren, B. Bergman and M. Brennan (eds.) Perspectives on Sign Language Usage, 371–390. Durham: Immersion for Spanish Language Acquisition. Brennan, M., Colville, M. D. and Lawson, L. K. (1984) Words in Hand. Wembley: Adept Press Ltd. Coates, J. and Sutton-Spence, R. (2001) ‘Turn-taking patterns in deaf conversation’, Journal of Sociolinguistics, 5: 507–529. Deuchar, M. (1984) British Sign Language. London: Routledge & Kegan Paul. Dively, V. (1998) ‘Conversational repairs in ASL’, in C. Lucas (ed.) Pinky, Extension & Eye Gaze, 137–169. Washington, DC: Gallaudet University Press. Dudis, P. (2002) ‘Grounded blend maintenance as a discourse strategy’, in C. Lucas (ed.) Turn-Taking, Fingerspelling, and Contact in Signed Languages. Washington, DC: Gallaudet University Press. Edelsky, C. (1981) ‘Who’s got the floor?’ Language in Society, 10: 383–421.

59

Gary Quinn Grice, H. P. (1975) ‘Logic and conversation’, in P. Cole and J. L. Morgan (eds.) Syntax and Semantics, 3: Speech Acts, 41–58. Academic: New York. Hoza, J. (2007) It’s Not What You Sign, It’s How You Sign It: Politeness in American Sign Language. Washington, DC: Gallaudet University Press. Johnston,T. and Schembri, A. (2007) Australian Sign Language: An Introduction to Sign Linguistics. Cambridge: Cambridge University Press. Kyle, J. G. and Woll, B. (1985) Sign Language:The Study of Deaf People and Their Language. Cambridge: Cambridge University Press. Leech, G. (1983) Principles of Pragmatics. Harlow: Longman. Liddell, S. K. (2003) Grammar, Gesture, and Meaning in American Sign Language. Cambridge: Cambridge University Press. Mather, S. M. (1996) ‘Initiation in visually constructed dialogue: Reading books with three- to eight-yearold students who are deaf and hard of hearing’, in C. Lucas (ed.) Multicultural Aspects of Sociolinguistics in Deaf Communities, 109–131. Washington, DC: Gallaudet University Press. Miles, D. (1988) British Sign Language: A Beginner’s Guide. Bath: The Bath Press. Nespor, M. and Sandler,W. (1999) ‘Prosody in Israeli Sign Language’, Language and Speech, 42(2–3): 143–176. Quinn, G. (2004) Pragmatics and Politeness in British Sign Language. Lancaster: Lancaster University. Quinn, G. (2010) ‘Schoolization: An account of the origins of regional variation in British Sign Language’, Sign Language Studies, 10(4): 476–501. Roush, D. (2007) ‘Indirectness strategies in American Sign Language requests and refusals: Deconstructing the deaf-as-direct stereotype’, in M. Metzger and E. Fleetwood (eds.) Translating Sociolinguistic and Consumer Issues in Interpreting, 103–156. Washington, DC: Gallaudet University Press. Siple, P. (1978) ‘Visual constraints for sign language communication’, Sign Language and Linguistics, 19: 95–110. Smith, S. (1999) ‘Adult–child interaction in a BSL nursery: Getting children’s attention’, paper presented at European Science Foundation ‘Intersign’ workshop on Sign Language Acquisition, London. Smith, S. and Sutton-Spence, R. (2005) ‘Adult–children interaction in a BSL nursery – getting their attention!’, Sign Language & Linguistics, 8: 131–152. Stokoe, W. C. (1960) ‘Sign language structure: An outline of the visual communication system of the American deaf ’, Studies in Linguistics, Occasional Papers, 8. Buffalo, NY: University of Buffalo Press. Sutton-Spence, R. and Woll, B. (1999) The Linguistics of British Sign Language: An Introduction. Cambridge: Cambridge University Press. Swisher,V., Christie, K. and Miller, S. (1989) ‘The reception of signs in peripheral vision by deaf persons’, Sign Language Studies, 63: 99–125. Van Herreweghe, M. (2002) ‘Turn-taking mechanisms and active participant in meetings with deaf and hearing participants in Flanders’, in C. Lucas (ed.) Turn-Taking, Fingerspelling and Contact in Signed Languages, 73–106. Washington, DC: Gallaudet University Press. Vogt-Svendsen, M. and Bergman, B. (2007) ‘Point buoys: The weak hand as a point of reference for time and space’, in M.Vermeerbergen, L. Leeson and O. Crasborn (eds.) Simultaneity in Signed Language: Form and Function. Amsterdam: John Benjamins. 217–235. Woodward, J. (1972) ‘Implications for Sociolinguistic Research among the Deaf ’. Sign Language Studies, 1: 1-7. Watts, R., Ide, S. and Ehlich, K. (1992) ‘Introduction’, in Watts, R., Ide, S. and Ehlich, K. (eds) Politeness in Language: Studies in its History,Theory and Practice (1-17). Berlin: Mouton de Gruyter.

60

7 GESTURE AND PRAGMATICS From paralinguistic to variably linguistic Alan Cienki

Introduction From the point of view of mainstream Western linguistic theory in the second half of the twentieth century, gesture was primarily seen as a matter of pragmatics in the sense that it was not considered to relate to the ‘core’ fields in linguistics of phonetics, phonology, morphology, syntax, and semantics. However, in the past few decades, several changes have occurred, placing gesture in a different position in the scientific landscape. Gesture studies has grown as an interdisciplinary field of inquiry, and further developments in approaches focusing on language use – including ones as diverse as cognitive linguistics and conversation/interaction analysis – are giving rise to a variably multimodal approach to language (as audiovisual in nature). The first part of this chapter concerns the earlier view of gesture within pragmatics, then we will turn to more recent ways of thinking, whereby gesture use is not viewed in terms of the pragmatic versus the semantic, but as located on a continuum of the semantic to the pragmatic. While any visible, effortful, bodily movement can be studied as gesture (Kendon 2004), we will narrow the topic here by focusing on movement of the hands and forearms by speakers when the movement is not part of an instrumental action (such as holding a pen and writing) and does not involve touching oneself or another (as in scratching one’s head or patting someone on the back), even though such actions can also clearly have pragmatic import (see, e.g., Kreidlin 2000).

Earlier views The traditional view, placing pragmatics outside the core areas of linguistics, arose from the division made in generative linguistics between linguistic competence and linguistic performance (Chomsky 1965), the former considered the purview of linguistics proper. One other consequence of that theoretical position was the separation of what constituted linguistic signals (sound segments for spoken languages that fell inside the boundary of the classical category of Language) versus other signals that were considered outside of it and thus paralinguistic. A related way of thinking about human communication has been in terms of dividing all relevant behaviors into two categories: the verbal versus the nonverbal. These divisions align with the separation of what kinds of meaning were considered semantic (such as reference to states 61

Alan Cienki

of affairs in the world) versus those that had pragmatic import (how one should interpret the semantic message, given the context). These pairs of epistemological divisions in linguistics between competence versus performance and semantics versus pragmatics, and the ontological division between verbal versus nonverbal behaviors, were (and for many, still are) aligned with each other. This has made it natural to consider gesture, as one nonverbal kind of performance behavior, from the point of view of pragmatics. Witness, for example, experimental research where the condition in which utterances are accompanied by pointing is called the ‘pragmatic condition’, as opposed to the ‘semantic condition’, which is the one not involving gesture (e.g., Tfouni and Klatzky 1983). This division also carries over implicitly in the application of categories that were the mainstay of traditional pragmatics in the late twentieth century, such as the types of possible speech acts. Studies on different languages (e.g., Kendon 1995 on Southern Italian; Müller 1998, 2004 on German; Payrató 1993 on Catalan; Streeck 2009 on English and Ilokano) have demonstrated how certain use of gestures can have illocutionary force, indicating an utterance as being assertive, directive, expressive, commissive, and so forth. Grishina (2014), for example, points out quantitatively significant differences in her Russian data between how speakers utter questions, imperatives, and statements according to the path of motion followed by the gesturing hand, and whether speakers use one or two hands when gesturing. The role of deixis with pointing gestures has played a prominent role in this research in terms of adding to the contextual understanding of speech. For example, Kelly and colleagues make this argument based on their findings that ‘people were more likely to interpret an utterance as an indirect request when speech was accompanied by a relevant pointing gesture than when speech or gesture was presented alone’ (1999: 577). Furthermore, Enfield and colleagues (2007) observe how big versus small pointing gestures allow speakers to place the accompanying verbal information in the foreground or background (respectively), without being on record as having done so, thus not giving excessive information or too little information. In this sense, the use of gesture with speech works within Grice’s (1975) conversational parameters of the maxims of quantity and manner (not over- or under-telling). Another context in which gesture research has been connected to pragmatics is when linguists look at ‘border phenomena’ in their field. An example of a linguistic category is that of interjections, a lexical category that arguably has not played a very prominent role in syntactic theory. However, Eastman (1992), for example, shows how interjections in her Swahili data blur the language/gesture boundary by involving both behaviors in systematic ways. In a similar vein, some psycholinguists have turned to the study of gesture in populations beyond neurotypical adult native speakers of a language. Groups of participants beyond this prototypical scenario that have been researched in terms of pragmatic use of gestures have been children; people with autism or aphasia, and/or people who have had a stroke; and even computers in terms of how people might communicate with them. We will consider each of these briefly. Many studies have been conducted on the pragmatics of children’s use of speech and gesture. A number of Tomasello’s works (e.g., 1998) concern the variety of gesture types children use in connection with joint attention during early word learning. Liebal and colleagues (2009) couch this issue in terms of the pragmatics of cooperative communication: shared experience is key for infants to interpret pointing gestures. A number of studies concern children’s use of pointing and other conventional gestures to accomplish their own or to comprehend others’ types of speech acts (Tomasello and Farrar 1986; Kelly 2001; Guidetti 2002, 2005; Bucciarelli et al. 2003). So and colleagues (2010) consider gesture and its relation to discourse-pragmatic principles of determining reference. Kirk and colleagues (2011) found that while all the children they tested answered more comprehension questions correctly when gesture accompanied 62

Gesture and pragmatics

speech, the difference between the conditions of speech + iconic representational gesture versus just speech as input was only significant in their results for children with language impairments. This last study points the way to considering how gesture has also been examined in relation to the pragmatics of communication by and with neuro-atypical adult populations. As mentioned earlier, gesture is often considered to play a pragmatic role in such research because of the assumed dichotomy between verbal and nonverbal communication. Groups studied have included people with autism or those who have had a stroke and have aphasia. For example, Baron-Cohen (1988) presents an overview of social and pragmatic deficits found with autism, which include lack of use – and lack of recognition of use – of communicative gestures in social contexts. Early studies on aphasia claimed a link between the severity of the aphasia and the ability to use and comprehend symbolic gestures (Goodglass and Kaplan 1963; Gainotti and Lemmo 1976). However, later research reveals a more complex picture. Cicone and colleagues (1979) found that the gestures of Wernicke’s aphasics generally resembled those of normal control ­participants, while the gestures of Broca’s aphasics equaled or surpassed those of the normal control participants in the clarity of their communication. Rousseaux and colleagues (2010) found that patients with stroke in the right hemisphere had difficulties with the pragmatics of gesture production and comprehension, while those with left-hemisphere strokes showed an increase in gesture production, compensating for aphasia. This phenomenon with left-hemisphere aphasia is also documented in Donkers and colleagues (1998) and in Goodwin’s (2008, 2012, and elsewhere) studies of a man whose left-hemisphere aphasia left him with only three words in English, yet who compensated skillfully through the use of gestures and intonation contours. Another area in which the pragmatic implications of gesture use have been examined is in various contexts of human–computer interaction, for example, trying to get robots and virtual agents to behave in ways that humans find pragmatically appropriate. One aspect of concern here is the sometimes complementary natures of the information expressed in the ‘linear’ structure of words versus the holistic, imagistic nature of gesture, as discussed in McNeill (1992). For example, sometimes words convey semantic information and the accompanying gestures convey pragmatic information, such as when gestures indicate a particular stance for interpreting the speech (such as a palm-up open hand to indicate doubt). Cassell and Stone (1999) compare the situation to the relation of words and graphics within one presentation; in both multimodal contexts, similar problems arise for computer ‘comprehension’ of the different modes of communication. Another issue in such research is how to operationalize the role of pragmatics in communication, for example, in terms of pragmatic constraints, to guide gesture use by robots appropriately (Cassell et al. 2007). Furthermore, the variety of forms and the multifunctionality of functions of human gestures pose extremely difficult problems for designing computer recognition systems of gestures that can infer the intentions of the speaker/gesturer (Nehaniv 2005). The papers in Kopp and Wachsmuth (2010) provide a comprehensive overview of many relevant issues. A more recently developing area is that of communication with computers via touch devices, involving gestural interfaces. Here pragmatics is relevant in several ways, including the ergonomic properties of gestures (what gestures ‘mean’ to human users if they are more versus less effortful to produce) and emotional/affective associations with gestural forms and movements (Kammer et al. 2010).

Recent developments What many of these studies share is that gesture is considered from the point of view of pragmatics because gesture is considered ‘extralinguistic’. However, this view has been argued against since at least the 1980s, most notably by Kendon (1980, 1994, 2004) and McNeill (1985, 1992, 63

Alan Cienki

2005). McNeill’s position is that gesture is part of language, and Kendon’s is that speech and gesture are partners in communication at a higher category level, that of ‘utterance’. Goodwin’s (2012 and elsewhere) view of gesture as one of several semiotic resources that speakers have at their disposal is therefore similar to that of Kendon’s in taking a higher-level perspective on communication. One consequence of these perspectives is a consideration of gesture as coverbal rather than as nonverbal, or, avoiding this terminology altogether, Kendon (personal communication) prefers to refer to speech as ‘utterance dedicated audible action’ and to gesture as ‘utterance dedicated visible action’. This contemporary approach in gesture studies problematizes the issue of what behaviors fall into the domain of linguistics, a situation that also calls into question the boundary between the semantic and the pragmatic. At the same time, independent theorizing in the field of cognitive linguistics has been supporting the position that it is more appropriate to view semantics and pragmatics on two ends of a continuum for characterizing meaning: from using spoken words as cues to prompt others to construct meanings to employing other behaviors specific to the local context as cues to prompt others to construct meanings (see, e.g., explicit views in Langacker 1987 and implicit arguments in Traugott 1988). Recent research in gesture studies is consonant with these views, showing a range of ways exists in which gesture is meaningful in communicative contexts, spanning what are traditionally considered semantic and pragmatic functions, along a continuum. According to this view every gesture can be analyzed as having pragmatic import to some degree (Payrató and Teßendorf 2014). A specific area of research to mention here concerns what have been called gesture families or recurrent gesture types. Gesture families consist of groups of gestures that share (a) similarities in form features (hand shape and spatial orientation of the hand, the location in which the gesture is produced relative to the gesturer’s body, and the movement of the hand) and (b) a set of related functions. Some of these functions involve reference to the ideas the speaker is communicating (more semantic) and others involve meaning related more to the context of use (more pragmatic). An example is gestures involving a relatively flat, open-hand produced palm up in front of the speaker. Palm-up open-hand (PUOH) gestures serve a range of functions (Kendon 2004; Müller 2004), at least for speakers of European languages, from displaying a small object to an addressee on the open palm, to using the extended fingers of the flat palm-up hand to point to a visible referent, to holding out the open hand when introducing an idea or a question, to indicating that one does not know something. The first two uses listed may be considered more semantic in nature because of how they relate more to the ‘propositional content’ of the utterance, and the latter are more pragmatic by virtue of their metadiscursive functions. However, since gestures are multifunctional by nature, even more ‘semantic’ uses of PUOH gestures can be overlaid with pragmatic functionality. Gestures share with verbal utterances a potential for multifunctionality, but in gesture, multifunctionality appears to be more of a ubiquitous norm. Since within a given culture such families of gestures recur across various contexts, scholars within the project ‘Towards a Grammar of Gesture’ (www.togog.org) began to refer to them as ‘recurrent gestures’ (see Bressem and Müller 2014; Ladewig 2014b for overviews). Recurrent gestures show a range of functions crossing from the more semantic to the more pragmatic, as the foregoing discussion of the PUOH illustrates. Other examples include: •



The cyclic gesture (Ladewig 2011, 2014a), involving the outward rotation of one’s index finger or whole hand in contexts of describing continuous actions (semantic function) or while searching for an idea or word while talking (thus showing this process pragmatically); The raised flat hand or index finger with the palm of the hand facing away from the speaker, oscillating with a left–right motion, often used by speakers of French (Calbris 64

Gesture and pragmatics



1990), English (Harrison 2009), and other languages in contexts of negation, ranging from explicitly verbalized, grammatical (semantic) negation to contexts where the gesture can pragmatically add negative import to the otherwise not explicitly negative verbal utterance; The ‘brushing away’ gesture (Teßendorf 2005, 2014) by speakers of (Iberian) Spanish in which the bent fingers of the speaker’s hand open up as the hand sweeps away from the body, used with explicit verbal dismissal (semantic) or implicit dismissal of an idea uttered by the speaker or interlocutor (pragmatic).

Cienki (2013b) suggests that we see here in gesture a phenomenon parallel to the one Traugott (1988) described for diachronic semantic change in verbal language, namely: semantic ‘bleaching’ being a partner of pragmatic ‘strengthening’, as an expression used for physical or metaphoric reference comes to be used in relation to the discourse content itself (compare the discussion in Ladewig 2014b).

Conclusion Looking at gesture with speech from the point of view of pragmatics thus raises important theoretical and empirical questions. In part, this is a natural consequence of turning to linguistic data in the form of video recorded talk – seeing and hearing the form of language use that is at the heart of most humans’ face-to-face ‘canonical encounter’, to use Clark’s (1973: 35) term. This perspective on language can incline one to reach different conclusions about what is involved in pragmatics and in language itself than if one focuses just on written language, which is implicitly what so much of modern mainstream linguistics has done, resulting in a theoretical bias toward written language (Linell 2005). Consequently, Cienki (2012: 155; 2013a: 195) proposes that linguists view language itself in a different way: as having a flexible boundary that overlaps to varying degrees and in a dynamic fashion with other semiotic systems. Another important issue is the temporal contour of spoken language and gesture use. Speech and gesture are tied to the dimension of time. Given that gestures often slightly precede the speech that semantically and pragmatically ‘goes’ with them (McNeill 1992), gestures ‘prefigure’ the meanings and pragmatic functions of utterances before they are verbalized (Streeck 1993), if they are verbalized at all. To what degree are those seeing a speaker perceiving (and cognitively processing) embodied cues for contextualizing utterances before the utterances are even verbalized for them to hear? The profound implications this has in terms of what we might call our ‘pragmatic cognition’ in real-time spoken discourse remains to be explored. To complicate the picture, there is increasing recognition that sign languages include gestural elements (Okrent 2002; Liddell 2003), bringing additional consideration of how graded categories can be handled theoretically as part of a language system (Wilcox and Xavier 2013). In various ways, linguists’ turn to manual behavior goes hand in hand with the questioning of a strict boundary between the semantic and the pragmatic.

Suggestions for further reading Bressem, J. and Müller, C. (2014) ‘A repertoire of German recurrent gestures with pragmatic functions’, in C. Müller, A. Cienki, E. Fricke, S. H. Ladewig, D. McNeill, and J. Bressem (eds.) Body – Language – Communication. An International Handbook on Multimodality in Human Interaction (vol. II). Berlin: De Gruyter Mouton. 1575–1592. This chapter explores how sets of gestures, each of which has an identifiable semantic core, are used in contexts in which they have different kinds of illocutionary force and thereby serve various kinds of pragmatic functions.

65

Alan Cienki Kendon, A. (2004) Gesture:Visible Action as Utterance. Cambridge: Cambridge University Press. Kendon is one of the founders of modern gesture studies, and this is the culmination of his work on gesture to date. Since communication in interaction is central to his research, issues in pragmatics play both explicit and implicit roles throughout his work. Müller, C. (1998) Redebegleitende Gesten. Kulturgeschichte – Theorie – Sprachvergleich. Berlin: Berlin Verlag A. Spitz. The book is of note here for its treatment of the pragmatic function of gestures in relation to other functions they can serve (e.g., accomplishing reference, or relating to the discourse level). In addition, it provides a valuable historical survey of the field of gesture studies. Payrató, L. and Teßendorf, S. (2014) ‘Pragmatic gestures’, in C. Müller, A. Cienki, E. Fricke, S. H. Ladewig, D. McNeill, and J. Bressem (eds.) Body – Language – Communication. An International Handbook on Multimodality in Human Interaction (vol. II). Berlin: De Gruyter Mouton. 1531–1540. This chapter provides a useful overview of recent research on pragmatic uses of gesture, moving beyond the platitude that ‘all gestures are pragmatic’ to delve into finer distinctions. Streeck, J. (2009) Gesturecraft:The Manu-facture of Meaning. Amsterdam: John Benjamins. Since Streeck’s work on gesture builds on the field of interaction analysis, pragmatics inherently plays an important role in it.The chapter in this book on ‘speech-handling’ gives particular attention to gesture’s discourse-pragmatic functions.

References Baron-Cohen, S. (1988) ‘Social and pragmatic deficits in autism: Cognitive or affective?’, Journal of Autism and Developmental Disorders, 18(3): 379–402. Bressem, J. and Müller, C. (2014) ‘A repertoire of German recurrent gestures with pragmatic functions’, in C. Müller, A. Cienki, E. Fricke, S. H. Ladewig, D. McNeill, and J. Bressem (eds.) Body – Language – Communication. An International Handbook on Multimodality in Human Interaction. Berlin: De Gruyter Mouton. 1575–1592. Bucciarelli, M., Colle, L., and Bara, B. G. (2003) ‘How children comprehend speech acts and communicative gestures’, Journal of Pragmatics, 35: 207–241. Calbris, G. (1990) The Semiotics of French Gestures. Bloomington: Indiana University Press. Cassell, J., Kopp, S., Tepper, P., Ferriman, K., and Striegnitz, K. (2007) ‘Trading spaces: How humans and humanoids use speech and gesture to give directions’, in T. Nishida (ed.) Conversational Informatics. Chichester: John Wiley and Sons. 133–160. Cassell, J. and Stone, M. (1999) ‘Living hand to mouth: Psychological theories about speech and gesture in interactive dialogue systems’, in S. E., Brennan, A. Giboin, and D. Traum (eds.) Proceedings of the AAAI Fall Symposium on Psychological Models of Communication in Collaborative Systems. North Falmouth, MA: AAAI. 34–42. Available from: www.aaai.org/Press/Reports/Symposia/Fall/fs-99–03.php (accessed May 2015). Chomsky, N. (1965) Aspects of the Theory of Syntax. Cambridge, MA: MIT Press. Cicone, M., Wapner, W., Foldi, N., Zurif, E., and Gardner, H. (1979) ‘The relation between gesture and language in aphasic communication’, Brain and Language, 8: 324–349. Cienki, A. (2012) ‘Usage events of spoken language and the symbolic units we (may) abstract from them’, in J. Badio and K. Kosecki (eds.) Cognitive Processes in Language. Bern: Peter Lang. 149–158. Cienki, A. (2013a) ‘Cognitive Linguistics: Spoken language and gesture as expressions of conceptualization’, in C. Müller, A. Cienki, S. Ladewig, D. McNeill, and S. Teßendorf (eds.) Body – Language – Communication. An International Handbook on Multimodality in Human Interaction. Berlin: De Gruyter Mouton. 182–201. Cienki, A. (2013b) ‘Image schemas and mimetic schemas in cognitive linguistics and gesture studies’, Review of Cognitive Linguistics, 11(2): 417–432. Clark, H. (1973) ‘Space, time, semantics, and the child’, in T. E. Moore (ed.) Cognitive Development and the Acquisition of Language. New York: Academic Press. 27–63. Donkers, N. F., Ludy, C. A., and Redfern, B. B. (1998) ‘Pragmatics in the absence of verbal language: Descriptions of a severe aphasic and a language-deprived adult’, Journal of Neurolinguistics, 11(1–2): 179–190. Eastman, C. M. (1992) ‘Swahili interjections: Blurring language-use/gesture-use boundaries’, Journal of Pragmatics, 18: 273–287.

66

Gesture and pragmatics Enfield, N. J., Kita, S., and De Ruiter, J. P. (2007) ‘Primary and secondary pragmatic functions of pointing gestures’, Journal of Pragmatics, 39: 1722–1741. Gainotti, G. and Lemmo, M. A. (1976) ‘Comprehension of symbolic gestures in aphasia’, Brain and Language, 3: 345–360. Goodglass, H. and Kaplan, E. (1963) ‘Disturbance of gesture and pantomime in aphasia’, Brain, 86: 703–720. Goodwin, C. (2008) ‘A competent speaker who can’t speak: The social life of aphasia’, Journal of Linguistic Anthropology, 14(2): 151–170. Goodwin, C. (2012) ‘The co-operative, transformative organization of human action and knowledge’, Journal of Pragmatics, 46: 8–23. Grice, P. (1975) ‘Logic and conversation’, in P. Cole and J. Morgan (eds.) Syntax and Semantics 3: Speech Acts. New York: Academic Press. 41–58. Grishina, Y. (2014) ‘Zhesty i pragmaticheskie kharakteristiki vyskzyvaniia’ [‘Gestures and pragmatic characteristics of the utterance’], in O.V. Fyodorova and A. A. Kibrik (eds.) Mul’timodal’naia Kommunikaciia: Teoreticheskie i Empiricheskie Issledovaniia [Multimodal Communication: Theoretical and Empirical Studies]. Moscow: Buki vedi, 25–47. Guidetti, M. (2002) ‘The emergence of pragmatics: Forms and functions of conventional gestures in young French children’, First Language, 22: 265–285. Guidetti, M. (2005) ‘Yes or no? How young French children combine gestures and speech to agree and refuse’, Journal of Child Language, 32: 911–924. Harrison, S. (2009) ‘Grammar, gesture, and cognition: The case of negation in English’, PhD dissertation, Université Michel de Montaigne Bordeaux 3. Kammer, D., Wojdziak, J., Keck, M., Groh, R., and Taranko, S. (2010) ‘Towards a formalization of multitouch gestures’, in Proceedings of the ACM International Conference on Interactive Tabletops and Surfaces. ITS. 49–58. Available from: http://dl.acm.org/citation.cfm?id=1936652 (accessed May 2015). Kelly, S. D. (2001) ‘Broadening the units of analysis in communication: Speech and nonverbal behaviors in pragmatic comprehension’, Journal of Child Language, 28: 325–349. Kelly, S. D., Barr, D. J., Church, R. B., and Lynch, K. (1999) ‘Offering a hand to pragmatic understanding: The role of speech and gesture in comprehension and memory’, Journal of Memory and Language, 40: 577–592. Kendon, A. (1980) ‘Gesticulation and speech: Two aspects of the process of utterance’, in M. R. Key (ed.) The Relation between Verbal and Nonverbal Communication. The Hague: Mouton. 207–227. Kendon, A. (1994) ‘Do gestures communicate? A review’, Research on Language and Social Interaction, 27(3): 175–200. Kendon, A. (1995) ‘Gestures as illocutionary and discourse structure markers in Southern Italian conversation’, Journal of Pragmatics, 23: 247–279. Kendon, A. (2004) Gesture:Visible Action as Utterance. Cambridge: Cambridge University Press. Kirk, E., Pine, K. J., and Ryder, N. (2011) ‘I hear what you say but I see what you mean:The role of gestures in children’s pragmatic comprehension’, Language and Cognitive Processes, 26(2): 149–170. Kopp, S. and Wachsmuth, I. (eds.) (2010) Gesture in Embodied Communication and Human–Computer Interaction. Berlin: Springer. Kreidlin, G.Y. (2000) ‘Neverbal’naia Semiotika v eë Sootnoshenii s Verbal’noi’ [‘Nonverbal semiotics in relation to verbal semiotics’], unpublished habilitation thesis, Russian State University for the Humanities. Ladewig, S. H. (2014a) ‘The cyclic gesture’, in C. Müller, A. Cienki, E. Fricke, S. H. Ladewig, D. McNeill, and J. Bressem (eds.) Body – Language – Communication. An International Handbook on Multimodality in Human Interaction. Berlin: De Gruyter Mouton. 1605–1618. Ladewig, S. H. (2014b) ‘Recurrent gestures’, in C. Müller, A. Cienki, E. Fricke, S. H. Ladewig, D. McNeill, and J. Bressem (eds.) Body – Language – Communication. An International Handbook on Multimodality in Human Interaction. Berlin: De Gruyter Mouton. 1558–1575. Ladewig, S. H. (2011) ‘Putting the cyclic gesture on a cognitive basis’, CogniTextes, 6. Available from: http:// cognitextes.revues.org/406#quotation (accessed May 2015). Langacker, R. W. (1987) Foundations of Cognitive Grammar. Theoretical Prerequisites. Stanford: Stanford University Press. Liddell, S. K. (2003) Grammar, Gesture, and Meaning in American Sign Language. Cambridge: Cambridge University Press. Liebal, K., Behne, T., Carpenter M., and Tomasello, M. (2009) ‘Infants use shared experience to interpret pointing gestures’, Developmental Science, 12(2): 264–271.

67

Alan Cienki Linell, P. (2005) The Written Language Bias in Linguistics: Its Nature, Origins, and Transformations. London: Routledge. McNeill, D. (1985) ‘So you think gestures are nonverbal?’, Psychological Review, 92(3): 350–371. McNeill, D. (1992) Hand and Mind: What Gestures Reveal about Thought. Chicago: University of Chicago Press. McNeill, D. (2005) Gesture and Thought. Chicago: University of Chicago Press. Müller, C. (1998) Redebegleitende Gesten. Kulturgeschichte – Theorie – Sprachvergleich. Berlin: Berlin Verlag A. Spitz. Müller, C. (2004) ‘Forms and uses of the Palm Up Open Hand: A case of a gesture family?’, in C. Müller and R. Posner (eds.) The Semantics and Pragmatics of Everyday Gestures: The Berlin Conference. Berlin: Weidler Buchverlag. 233–256. Nehaniv, C. L. (2005) ‘Classifying types of gesture and inferring intent’, in Proceedings of the AISB 05 Symposium on Robot Companions: Hard Problems and Open Challenges in Robot-Human Interaction. AISB. 74–81. Available from: http://aisb.org.uk/publications/proceedings/aisb2005/5_RoboComp_final.pdf (accessed October 2015). Okrent, A. (2002) ‘A modality-free notion of gesture and how it can help us with the morpheme vs. gesture question in sign language linguistics (or at least give us some criteria to work with)’, in R. P. Meier, K. Cormier, and D. Quinto-Pozos (eds.) Modality and Structure in Signed and Spoken Languages. Cambridge: Cambridge University Press. 175–198. Payrató, L. (1993) ‘A pragmatic view on autonomous gestures: A first repertoire of Catalan emblems’, Journal of Pragmatics, 20: 193–216. Payrató, L. and Teßendorf, S. (2014) ‘Pragmatic gestures’, in C. Müller, A. Cienki, E. Fricke, S. H. Ladewig, D. McNeill, and J. Bressem (eds.) Body – Language – Communication. An International Handbook on Multimodality in Human Interaction. Berlin: De Gruyter Mouton. 1531–1540. Rousseaux, M., Daveluy, W., and Kozlowski, O. (2010) ‘Communication in conversation in stroke patients’, Journal of Neurology, 257(7): 1099–1107. So, W. C., Demir, Ö. E., and Goldin-Meadow, S. (2010) ‘When speech is ambiguous, gesture steps in: Sensitivity to discourse-pragmatic principles in early childhood’, Applied Psycholinguistics, 31: 209–224. Streeck, J. (1993) ‘Gesture as communication I: Its coordination with gaze and speech’, Communication Monographs, 60: 275–299. Streeck, J. (2009) Gesturecraft:The Manu-facture of Meaning. Amsterdam: John Benjamins. Teßendorf, S. (2005) ‘Pragmatische Funktionen spanischer Gesten am Beispiel des “gesto de barrer” ’, unpublished MA thesis, Freie Universiteit Berlin. Teßendorf, S. (2014) ‘Pragmatic and metaphoric – Combining functional and cognitive approaches in the analysis of the “brushing aside gesture” ’, in C. Müller, A. Cienki, E. Fricke, S. H. Ladewig, D. McNeill, and J. Bressem (eds.) Body – Language – Communication. An International Handbook on Multimodality in Human Interaction. Berlin: De Gruyter Mouton. 1540–1558. Tfouni, L.V. and Klatzky, R. L. (1983). ‘A discourse analysis of deixis: Pragmatic, cognitive and semantic factors in the comprehension of “this”, “that”, “here” and “there” ’, Journal of Child Language, 10: 123–133. Tomasello, M. (1998) ‘Reference: Intending that others jointly attend’, Pragmatics and Cognition, 6(1–2): 1–2. Tomasello, M. and Farrar, M. J. (1986) ‘Joint attention and early language’, Child Development, 57: 1454–1463. Traugott, E. C. (1988) ‘Pragmatic strengthening and grammaticalization’, in Proceedings of the Fourteenth Annual Meeting of the Berkeley Linguistics Society. Berkeley, CA: Berkeley Linguistics Society. 406–416. Wilcox, S. and Xavier, A. N. (2013) ‘A framework for unifying spoken language, signed language, and gesture’, Revista Todas as Letras, 15(1): 88–110.

68

8 PARALANGUAGE Tim Wharton

Introduction According to Albert Mehrabian’s ‘7%-38%-55% Rule’ (1981), 55 per cent of the emotional information communicated in a given exchange is communicated via ‘body language’, 38 per cent via the words we say (intonation, affective tone of voice etc.) and only 7 per cent from the words themselves. While the accuracy of these claims has been criticised (Trimboli and Walker 1987), and, indeed, Mehrabian (2014) himself has gone on record as saying his results are often misrepresented, the so-called nonverbal dominance view of communication continues to play an important role in ethology, developmental psychology, Freudian psychoanalysis and other fields. The very existence of such a rule is at least suggestive. When it comes to the communication of affect, there is an assumption that the nonverbal dimension of human communication is at least as important as the verbal one, if not more so. All of which, rightly or wrongly, has been of virtually no concern to linguists.The nonverbal contributions to linguistic communication were left behind by the so-called cognitive revolution of the 1960s and 1970s (Hecht and Ambady 1999) as generative linguists abstracted away from paralinguistic phenomena in order to focus on the linguistic code. Some functionalist linguists, discourse analysts and sociologists continued to address aspects of nonverbal communicative behaviour (see e.g. Goffmann 1964; Garfinkel 1967; Hymes 1972; Goodwin 1981; Brown and Yule 1983; Bolinger 1983; Schiffrin 1994; Gumperz 2001), but distinctions that are important from a pragmatic view have been left relatively unexplored.The question of how that 93 per cent of behaviours might interact with the 7 per cent is largely ignored. Those of us who work in pragmatics can afford ourselves no such luxury. Firstly, the aim of a pragmatic theory is to explain how utterances are understood, and utterances, of course, have both linguistic and non-linguistic properties. Secondly, the emotional dimension to speaker meaning is at least as important, often more important, than those dimensions that tend to receive more attention. Any pragmatic theory worth its salt simply must have a view on nonverbal communicative behaviours and how they contribute to speakers’ meanings.

A problem of definition But when it comes to the analysis of ‘paralanguage’, there is a problem of definition. According to Poyatas (1993), paralanguage includes only the vocal aspects of language use that are not, 69

Tim Wharton

strictly speaking, part of language: affective tone of voice, the non-linguistic elements of prosody (effectively, the 38 per cent in the foregoing rule). Facial expression and gesture, which some researchers do analyse as paralinguistic (see later in this chapter), are not part of paralanguage. These, Poyatas claims, are better analysed as elements in a broader category of ‘kinesics’, which includes posture and ‘proxemics’. Proxemics concerns itself with the way communicators orient themselves to one another and the distance they maintain between themselves. (We become aware of this only when someone invades our ‘body space’.) Paralanguage is also sometimes defined so as to include ‘oculesics’ – the frequency of glances, patterns of eye fixation, pupil dilation – and ‘haptics’ – the communicative domain of touch. For Poyatas, paralanguage is: [those] nonverbal voice qualities, voice modifiers and independent utterances produced or conditioned in the areas covered by the supraglottal cavities (from the lips and the nares to the pharynx), the laryngeal cavity and the infraglottal cavities (lungs and esophagus), down to the abdominal muscles, as well as the intervening momentary silences, which we use consciously or unconsciously. (Potayas 1993: 6) Abercrombie (1968) takes a radically different stance. He defines paralanguage so as to include all those aspects of linguistic communication that are not part of language per se, but are nonetheless somehow involved with the message or meaning a communicator conveys. In order to avoid too much overlap with other sections of this handbook, the conceptualisation of paralanguage adopted here will necessarily be the one Poyatas proposed. But notice that, so defined, the set of paralinguistic behaviours is a very (very) small one – basically, all those aspects of prosody that are not deemed grammatical – and so before continuing I would like to add the following two caveats. Firstly, and perhaps most importantly, defining paralanguage to include vocal but not, say, facial expressions is to overlook the vitally important observation that they work closely in tandem with each other. Consider the following from Bolinger: This kind of working in parallel is easiest to demonstrate with exclamations. An ah! of surprise, with a high fall in pitch, is paralleled by a high fall on the part of the eyebrows. . . . . A similar coupling of pitch and head movement can be seen in the normal production of a conciliatory and acquiescent utterance such as ‘I will’ with the accent at the lowest pitch – we call this a bow when it involves the head, but the intonation bows at the same time. (Bolinger 1983: 98) The parallels are so strong that a single, homogeneous account of these para-/non-linguistic behaviours seems to be required, one that embraces the fact that they are, for the most part, closely interlinked. Secondly, in recent years there has been a huge amount of work in pragmatics suggesting that the gap between linguistic meaning and speaker meaning is much wider than the one Grice originally sketched in his distinction between what is said and what is implicated (1967, 1989). Work since his famous William James lectures suggests that speaker meaning is not just underdetermined by linguistic meaning, but massively so (Sperber and Wilson [1986]1995; Bach 2001; Blakemore 2002; Carston 2002; Recanati 2004). And that gap is bridged not by more coding, but by inference.

70

Paralanguage

The central aims of pragmatics are to explain how, by and large, hearers are accurate in their reasoning (in the sense that it broadly reflects the speaker’s communicative intentions), the principles behind that reasoning and the route it takes. Any attempt to characterise linguistic communication, then, must reflect the fact that it is an intelligent, inferential activity involving the expression and recognition of intentions. An inferential, not a coding-decoding, approach is required, and our analysis of paralanguage should be no different.

Towards a pragmatics of paralanguage Paralanguage and pragmatics: the state of play Many approaches to the broad category of nonverbal communicative behaviours, however, have been highly resistant to developments within pragmatics. Adam Kendon’s work has provided numerous insights into gestural communication, as well as a wide range of descriptive categories into which gestures fall (McNeill 1992; Kendon 2004). But he explicitly endorses a view under which the, as he terms them, invisible, ‘mysterious forces’ of intentionality are beyond the realm of scientific study (2004: 15). Cooperrider (2011), who largely endorses this view, outlines some of the arguments against adopting a more ‘pragmatic’ view on gestural communication. In his work on facial expression, Alan Fridlund (1994) also circumvents entirely the notion of intentionality. Research into paralanguage is much the same, and works on an assumption those working in pragmatics dismissed long ago: that if something is communicated, it must be because it is encoded (Ohala 1984; Ladd 1996; Gussenhoven 2004). Ohala’s ‘Frequency Code’ is a prime example. It is commonly known that the frequencies produced by small animals, with consequently small larynxes, are typically higher pitched than those produced by larger animals with large larynxes: according to Ohala, high-pitched vocalisations encode harmlessness, unassertiveness or submission, and low-pitched vocalisations encode danger, assertiveness and dominance. Inference, intentions, pragmatics are simply nowhere in the picture. Lacking these key concepts, an explanatory account of precisely how paralanguage contributes to human linguistic communication will remain incomplete.

MeaningNN and showing Paul Grice’s ‘Meaning’ (1957) is one of the most influential philosophical papers of the past fifty years, and remains a huge influence on linguists, pragmatists and cognitive scientists as well as philosophers. In it, he began by drawing a distinction between what he called natural, or indicator, meaning and non-natural meaning (meaningNN). The latter he attempted to characterise in terms of the expression and recognition of intentions. Consider the distinction between the kind of meaning inherent in (1) and (2): (1) That black smoke meansN the tyre factory is on fire. (2) That white smoke meansNN the Vatican conclave has elected a new pope.1 For Grice ‘what is meantNN’ is roughly coextensive with what is intentionally communicated, and his notion of non-natural meaning has had a major influence on the development of pragmatics. But a point often missed is that in the course of a communicative exchange, paralinguistic

71

Tim Wharton

phenomena (some of which are, in a sense, natural) often play a role in determining what has been meantNN. Consider the following examples: (3) (4) (5) (6) (7)

S (voice trembling): I feel fine, honestly. . . S (smiling to A, in a delighted tone of voice): It’s great! S (lying, deliberately faking a happy tone of voice): It’s wonderful news. . . S: Aaaaaargh! That flaming hurts! Ow! Ow! Ow! Ow! Ow! Ouch! S (shivering, exaggerating a bilabial trill): Can we go please go inside?

In (3), vocal manifestations of the (S)peaker’s illness indicate to his (A)ddressee that he is nowhere near as well as he would like the (A)ddressee to believe; in (4), S’s delighted tone of voice calibrates just how great S believes whatever she is referring to is; in (5), S fakes a natural behaviour, which indicates that he is being ironic, and means the opposite of what he has said; in (6), S’s natural paralinguistic expressions of pain say as much as the words he utters; in (7), S exaggerates a shiver, perhaps so as to provide evidence for her request. A controversial feature of Grice’s account of intentional communication is the line he draws between showing and meaningNN, where meaningNN typically involves a linguistic convention or code. This distinction has had important effects on the evolution of pragmatics since, following Grice, pragmatists have focused on the notion of meaningNN and tended to abstract away from cases of showing. However, I have argued (see Wharton 2009), that while there is room for disagreement on whether cases of ‘showing’ always amount to cases of meaningNN, there is little doubt that cases of ‘showing’ do qualify as cases of intentional communication of the kind a pragmatic theory should be able to handle: meaning and communicating do not always line up. Bearing this in mind, and looking once more at (3)–(7), but this time through the lens of an inferential, as opposed to a code, model of communication, a range of further questions comes to mind. In some cases, for example, paralinguistic behaviours betray our thoughts and feelings to others in a way that does not amount to intentionally communicating them (see [3]). In others, they may be deliberately produced in a way that clearly amounts to intentional communication (see [4]). In still further cases, they are involuntarily produced but may be deliberately (or overtly) shown (see [5]) and even exaggerated (see [6] and [7]). Intentional verbal communication, then, involves a mixture of natural and non-natural meaning, and an adequate pragmatic theory should take account of both. Many aspects of paralanguage are natural but also shown.

Signs and signals A further complication is that it appears not all phenomena that mean naturally work in the same way. In the study of the information transmission between non-human animals, a distinction is made between signs and signals. Signs carry information by providing evidence for it. Signals, on the other hand, are those natural behaviours that convey information and have been ‘moulded by natural selection to do so’ (Seeley 1989: 547). It is their adaptive function to carry information, where the adaptive function of a given trait or behaviour is the effect which is historically responsible for the reproduction and propagation of that trait or behaviour within a species (Millikan 1984; Sperber 2007). Whilst a sign may happen to carry information for an observer, it would go on being produced whether or not it carries this information. As a result of regular travel across dusty soil, a predatory species might leave traces of its presence. Certain species of prey might learn to associate such traces with the predator’s presence. The traces themselves, however, cannot be said to 72

Paralanguage

have a signalling function. Signals, by contrast, do have a communicative function. The function of the honeybee’s dance is to inform other honeybees about the location of nectar.The function of the bullfrog’s call is to alert female frogs to the fact that he is searching for a mate. If they did not carry this information, it would be hard to see why these behaviours survive. Most animal communication seems to be based on signalling systems of this type.2 In Wharton (2009), I argued that these distinctions apply in the whole range of human communicative behaviours. I illustrate the distinction between natural signs and signals in the human case by comparing shivering with smiling (compare [4] and [7]). Shivering is a natural behaviour whose function is to generate heat by rapid muscle movement. It may provide evidence to an observer that the individual is feeling cold. However, it is not its function to carry this information: it is not a signal but a sign. Smiling, by contrast, appears to have evolved as a signalling activity whose function is to convey information to others (Van Hooff 1972; Ekman 1989, 1992, 1999; Fridlund 1994). As Ekman puts it, smiling and other spontaneous facial expressions ‘have been selected and refined over the course of evolution for their role in social communication’ (1999: 51). Like the bee dance and the bullfrog calls, they are signals rather than signs. If some natural behaviours are coded signals, we would predict that they are interpreted by specialised, perhaps dedicated, neural machinery, and this prediction appears to be borne out. Both non-human primates and humans have neural mechanisms dedicated both to recognising faces and to processing facial expressions (Gazzaniga and Smiley 1991). It is not hard to think of counterparts to shivering and smiling in other domains of paralanguage. For instance, a speaker’s mental or physical state may affect the paralinguistic properties of her utterance, enabling a hearer with the appropriate experience or background knowledge to infer whether she is sober or drunk, healthy or ill, calm or anxious and so forth. As with shivering, these paralinguistic properties carry information about the speaker’s mental or physical state, but it is not their function to do so: they are natural signs, interpreted by inference rather than decoding. On the other hand, affective tones of voice, like affective facial expressions, may well be natural signals, interpreted by innately determined codes and via dedicated mechanisms rather than according to more general pragmatic principles.

Conclusion Natural codes are found in animals with no capacity for inferential intention recognition. Honeybees and frogs both lack the ability to infer the intentions of others, but they can still inform each other by means of their dance-based or vocal code. Communication among humans, by contrast, not only requires the capacity for inferential intention recognition, but may be achieved in the absence of any code at all – such as when I nudge my empty plate towards you and you infer that I’d like another slice of apple pie. Human linguistic communication exploits the human ability to understand the behaviour of others in terms of the intentions behind it – sometimes known as the ‘mindreading’ ability. It involves both coding and inferential intention recognition, and this observation is no less important when it comes to our analyses of how speakers and hearers express and interpret paralanguage than they are to our analyses of how they utter and understand words. Indeed, in a recent article (Vigliocco et al. 2014), the authors pointed out that most of our current understanding of language and cognition is based on the 7 per cent of Mehrabian’s rule (and, even more narrowly, on words from only a small subset of human natural languages). As this chapter has been at pains to point out, linguistic communicative acts are complex acts, composites of linguistic, non-linguistic and paralinguistic behaviours. I wonder how long it is before 73

Tim Wharton

someone asks whether drawing a line between the verbal and the nonverbal in the first place is more to do with convenient abstraction, than it is based on a sound research strategy.

Suggestions for further reading Cosmides, L. and Tooby J. (2000) ‘Evolutionary psychology and the emotions’, in M. Lewis and J. Haviland Jones (eds.) Handbook of Emotions. New York: Guilford. 91–115. This chapter explores the origins, function and evolution of emotions – often communicated by ­paralanguage – and situates them within a cognitive psychological framework. Grice, H. P. (1957) ‘Meaning’, in Grice, H. P. (1989) Studies in the Way of Words. Cambridge, MA: Harvard University Press. This seminal paper is required reading for anyone interested in inferential theories of communication. It also provided the inspiration for several of the distinctions presented in the above article. Kendon, A. (2004) Gesture:Visible Action as Utterance. Cambridge: Cambridge University Press. This book presents an exhaustive analysis of everyday conversations and, through this, explores the varied role of gesture in utterance interpretation. Wilson, D. and Sperber, D. (2004) ‘Relevance Theory’, in L. R. Horn and G. Ward (eds.) Handbook of Pragmatics. Oxford: Blackwell. 607–632. This chapter provides a concise summary of relevance theory, the theory of communication and cognition that provides the theoretical underpinning for much of the foregoing discussion. Wharton, T. (2009) Pragmatics and Non-verbal Communication. Cambridge: Cambridge University Press. This book explores how ‘natural’, non-linguistic behaviours – tone of voice, facial expressions, gesture – interact with the linguistic properties of utterances, and how all communicative behaviors can be integrated within an inferential theory of communication.

Notes 1 The reader is invited to consider whether high-pitched vocalisations ‘meaning’ harmless, unassertive or submissive and low-pitched vocalisations ‘meaning’ dangerous, assertive or dominant are cases of natural or non-natural meaning. 2 The reader is further invited to consider whether high-pitched vocalisations ‘meaning’ harmless, unassertive or submissive, and low-pitched vocalisations ‘meaning’ dangerous, assertive or dominant are signs or signals.

References Abercrombie, D. (1968) ‘Paralanguage’, British Journal of Disorders of Communication, 3: 55–59. Bach, K. (2001) ‘You don’t say?’, Synthese, 127: 11–31. Blakemore, D. (2002) Relevance and Linguistic Meaning: The Semantics and Pragmatics of Discourse Markers. Cambridge: Cambridge University Press. Bolinger, D. (1983) ‘The inherent iconism of intonation’, in J. Haiman (ed.) Iconicity in Syntax. Amsterdam: John Benjamins. 97–109. Brown, G. and Yule, G. (1983) Discourse Analysis. Cambridge: Cambridge University Press. Carston, R. (2002) Thoughts and Utterances:The Pragmatics of Explicit Communication. Oxford: Blackwell. Cooperrider, K. (2011) ‘Review of “Pragmatics and nonverbal communication” by T. Wharton (2009)’, Gesture, 11(1): 81–88. Ekman, P. (1989) ‘The argument and evidence about universals in facial expressions of emotion’, in H. Wagner and A. Manstead (eds.) Handbook of Social Psychophysiology. New York: Wiley. 143–164. Ekman, P. (1992) ‘An argument for basic emotion’, Cognition and Emotion, 6(3–4): 169–200. Ekman, P. (1999) ‘Emotional and conversational nonverbal signals’, in L. Messing and R. Campbell (eds.) Gesture, Speech and Sign. Oxford: Oxford University Press. 45–57. Fridlund, A. (1994) Human Facial Expression: An Evolutionary View. San Diego: Academic Press. Garfinkel, H. (1967) Studies in Ethnomethodology. Englewood Cliffs, NJ: Prentice Hall. Gazzaniga, M. and Smiley, C. (1991) ‘Hemispheric mechanisms controlling voluntary and spontaneous facial expressions’, Journal of Cognitive Neuroscience, 2: 239–245.

74

Paralanguage Goffman, E. (1964) ‘The neglected situation’, American Anthropologist, 66(6): 133–136. Goodwin, C. (1981) Conversational Organisation: Interaction between Speakers and Hearers. New York: Academic Press. Grice, H. P. (1957) ‘Meaning’, Philosophical Review, 66: 377–388. Grice, H. P. (1967) William James Lectures I–VII, unpublished typescripts. Grice, H. P. (1989) Studies in the Way of Words. Cambridge, MA: Harvard University Press. Gumperz, J. (2001) ‘Interactional linguistics: A personal perspective’, in D. Schiffrin, D.Tannen and H. Hamilton (eds.) The Handbook of Discourse Analysis. Oxford: Blackwell. 215–228. Gussenhoven, C. (2004) The Phonology of Tone and Intonation. Cambridge: Cambridge University Press. Hecht, M. and Ambady, N. (1999) ‘Non-verbal communication and psychology: Past and future’, The New Jersey Journal of Communication, 7(2): 1–14. Hymes, D. (1972) ‘On Communicative Competence’, in J. Pride and J. Holmes (eds.) Sociolinguistics. Harmondsworth: Penguin. 269–293. Kendon, A. (2004) Gesture:Visible Action as Utterance. Cambridge: Cambridge University Press. Ladd, R. (1996) Intonational Phonology. Cambridge: Cambridge University Press. McNeill, D. (1992) Hand and Mind: What Gestures Reveal about Thought. Chicago: University of Chicago Press. Mehrabian, A. (1981) Silent Messages: Implicit Communication of Emotions and Attitudes. Belmont, CA: Wadsworth. Mehrabian, A. (2014) Radio Interview: ‘More or Less’, BBC World Service, 2 December 2014. Available from: www.bbc.co.uk/programmes/p02clw50 (accessed 21 January 2015). Millikan, R. (1984) Language,Thought and Other Biological Categories. Cambridge, MA: MIT Press. Ohala, J. (1984) ‘An ethological perspective on common cross-language utilization of F0 of voice’, Phonetica, 41: 1–16. Poyatas, F. (1993) Paralanguage: A Linguistic and Interdisciplinary Approach to Interactive Speech and Sounds (Current Issues in Linguistic Theory, 92). Amsterdam: John Benjamins. Recanati, F. (2004) Literal Meaning. Cambridge: Cambridge University Press. Schiffrin, B. (1994) Approaches to Discourse. Oxford: Blackwell. Seeley, T. (1989) ‘The honey-bee colony as a superorganism’, American Scientist, 77: 546–553. Sperber, D. (2007) ‘Seedless grapes: Nature and culture’, in S. Laurence and E. Margolis (eds.) Creations of the Mind:Theories of Artifacts and Their Representation. Oxford: Oxford University Press. 124–137. Sperber, D. and Wilson, D. (1986) 1995 Relevance: Communication and Cognition. Oxford: Blackwell. Trimboli, A. and Walker, M. (1987) ‘Non-verbal domination in the communication of affect: A myth?’, Journal of Non-verbal Behaviour, 11(3): 180–190. Van Hooff, J. (1972) ‘A comparative approach to the phylogeny of laughter and smiling’, in R. Hinde (ed.) Non-verbal Communication. Cambridge: Cambridge University Press. 209–238. Vigliocco, G., Perniss, P. and Vinson, D. (2014) Philosophical Transactions of the Royal Society Series B, 369(1651): 1–7. Wharton, T. (2009) Pragmatics and Non-verbal Communication. Cambridge: Cambridge University Press.

75

PART II

Established fields Pragmatics and variation

9 VARIATION AND CHANGE Historical pragmatics Andreas H. Jucker and Daniela Landert

Introduction Languages, such as English, Russian, Japanese or Tzeltal, are not homogeneous entities. They are always subject to internal variation. Speakers make use of the resources of a language when they communicate with each other, but each speaker has his or her own small idiosyncrasies in the way they use these resources. They innovate consciously or unconsciously by modifying some aspect of their linguistic resources. They may pronounce an existing word in a slightly different way; they may use an existing word with a slightly different shade of meaning or in a different syntactic context; they may borrow a word from another language or invent a new word in order to meet new communicative demands, for example, to name a new or an existing concept in a creative – humorous or poetic – way. As a result of these modifications, languages cannot be described as uniform entities. Variability is an integral part of each living language. But this variability is not without regularity. Some variability may be idiosyncratic and pertain to individual speakers only. But to the extent that other speakers adopt the innovations and use them more regularly, idiosyncrasies become regular options, and if most speakers of a language adopt a particular option, it is no longer just an option but a regular part of this particular language or language variety. Thus individual idiosyncrasies can lead to regular variation in a language, and regular variation can lead to language change. Such variability exists on all levels of a language; on the level of its phonology, morphology, lexicon, syntax and semantics. And – crucially – such variability also exists on the level of language use, that is, the level of pragmatics. The research into such variability has a very long history for some of the core levels of language, in particular for phonology and the lexicon of specific languages. However, on the level of pragmatics, research into variability has only just started. The field of variational pragmatics, which looks at synchronic variation in different varieties of the same language, started officially in 2008 with the volume Variational Pragmatics (Schneider and Barron 2008; see also Barron and Schneider 2009; Schneider 2010; and Barron, Chapter 10, this volume). Research into diachronic variability on the level of language use is conducted in the field of historical pragmatics. This field was inaugurated with the volume Historical Pragmatics in 1995 (Jucker 1995), and it has become one of the most vibrant branches of historical linguistics, with

79

Andreas H. Jucker and Daniela Landert

a dedicated journal (Journal of Historical Pragmatics), a comprehensive handbook (Jucker and Taavitsainen 2010) and a recent textbook (Jucker and Taavitsainen 2013). In its entirety, historical pragmatics studies more than just diachronic variability. It also studies pragmatic features in earlier periods irrespective of whether they are subject to variation and/or change. This part of historical pragmatics has been called ‘pragmaphilology’ (Jacobs and Jucker 1995) or ‘historical discourse analysis proper’ (Brinton 2001). Variability at a particular stage of a language – as pointed out earlier – may lead to language change over time. The study of changes in pragmatic entities over time is called diachronic pragmatics (Jacobs and Jucker 1995). Brinton distinguishes between studies that adduce discourse factors to explain language change (‘discourse-oriented historical linguistics’) and studies that trace the development of discourse features across time (‘diachronic[ally oriented] discourse analysis’) (2001: 140). Historical pragmatics is predominantly based on the wider conceptualisation of pragmatics which includes the social and cultural context of language use. This approach has also been termed the Continental European or social approach to pragmatics in contrast to the more narrow conceptualisation of pragmatics as a study of utterance meaning in context (variously called Anglo-American, philosophical or theoretical pragmatics) (see Huang 2007: 4–5, 2012: 8; Jucker 2008: 894–895, 2012: 501–503; Chapman 2011: 5). The study of older stages of a language, and in particular the study of pragmatic entities at such older stages and the mechanisms of their change, usually requires a considerable amount of familiarity with the social and cultural context in which they were used, and thus the broader perspective of a social approach to pragmatics is often a necessity. In the following we shall review some of the relevant work in historical pragmatics. In Data and methodology, we shall first give a brief outline of some of the data problems and the problems of methodology in historical pragmatics. In Pragmaphilology, we focus briefly on historical pragmatic work in the pragmaphilological tradition, that is to say on work that studies pragmatic features in earlier periods without focusing on either synchronic or diachronic variation. Diachronic change will be devoted to diachronic variation proper, that is to say on studies of pragmatic entities that change over time. Synchronic variation in earlier periods turns to work on synchronic variation in former times. Investigating variation: pragmatic variables will introduce the concept of the ‘pragmatic variable’ and discuss some pertinent methodological and theoretical questions.

Data and methodology Historical pragmatics has to rely on data and on methodologies that differ somewhat from those used in other branches of pragmatics. Philosophical methods, for instance, are of very limited value for historical pragmatics since earlier stages of a language are not accessible through the introspection of the scholar. Likewise, it is impossible to interview native speakers who lived centuries ago and impossible to recruit such speakers for experimental research methods, such as discourse completion tasks, interaction tasks, role plays and the like. As a result the historical pragmaticist has to rely on empirical methods and in particular field methods of observation and analysis of existing data. Generally texts written in the past form the data source, although recent work in historical pragmatics has started to make use of the diachronic dimension of speech recordings. Such recordings have been around for almost a century now and already provide ample opportunities for tracing diachronic developments, even if the range of recordings for the early decades of the previous century are very limited. Political speeches and various radio programmes, for instance, are readily available, but it is more difficult to get access to less formal conversational material in natural settings (see Jucker and Landert 2015).

80

Historical pragmatics

In most cases, however, historical pragmatics relies on written data. Most of the early work in historical pragmatics endeavoured to find data that was as close as possible to the spoken language of the past because it was felt that this type of data was somehow more authentic and more worthy of pragmatic analysis. Play texts, for instance, were argued to be close to conversational language in spite of their fictional nature. Court records and witness depositions are transcripts of actual spoken interactions and were, therefore, also considered good candidates as data for historical pragmatics (see Taavitsainen and Jucker 2010: 7–11 for an overview). Koch and Oesterreicher in various publications (Koch and Oesterreicher 1985, 2011; Koch 1999) have introduced a framework that distinguishes between the dichotomy of graphic and phonic modality of language, and the scale extending from the language of immediacy to the language of distance. Within this model, the scholar can search more specifically for instances of the language of immediacy, characterised by informality, dialogicity and spontaneity, even if only sources in the graphic code are available. Letters, for instance, are graphically encoded, but in many cases they are instances of the language of immediacy. Culpeper and Kytö (2010) set out to specifically trace features of spoken language in the written data sources of the early modern English period. For this purpose, they compiled the 1.2 million-word Corpus of English Dialogues 1560–1760, containing data more or less directly related to spoken face-to-face interaction, that is, comedy drama, didactic works in dialogue form, prose fiction, trial proceedings and witness depositions, which all include either constructed or authentic dialogue. On this basis they investigate lexical bundles, repetitions and ‘pragmatic noise’, which includes items such as ah, ha, hah, o, oh, ho, um, hum. An alternative approach to the problem of data access starts from the presumption that all instances of language use – whether spoken or written – are communicative and, therefore, worthy of pragmatic investigations. This approach also maintains that language use in all its forms is always contextualised and subject to specific constraints. It is, therefore, not legitimate to analyse one type of language (e.g. fictional conversations in a drama) and generalise the results to other types (e.g. natural conversation among close friends). Each type of language is interesting in itself and should be analysed as such (Jucker and Taavitsainen 2010: 25–26). In the selection and analysis of naturally occurring data, the historical pragmaticist faces the dilemma of choosing between a microscopic and a macroscopic perspective. Adopting a microscopic perspective, the researcher focuses on small amounts of data with highly contextualised readings and background knowledge of the sociocultural situation. Such research is generally qualitative because there is too little material for any observations of statistical regularities and there is little scope for generalisations to larger diachronic trends. In a macroscopic perspective, on the other hand, the researcher relies on large, usually electronically readable, corpora, which allow the identification of trends and larger generalisations. Recent years have seen the publication of a large number of relevant corpora with historical data. Examples include the 1.6 million-word Helsinki Corpus of English Texts published some twenty-five years ago and the recent publication of the 400 million-word Corpus of Historical American English (for an overview, see Kytö 2010). In long lists of hits from such corpora, there is, however, a danger of losing the nuances of meaning. Search terms are no longer richly contextualised, and it is possible that the researcher may miss crucial differences in the significance of specific search strings because such differences manifest themselves only in the larger context in which they occur. Historical pragmatic research is, thus, often a search for the best possible compromise between the two perspectives, namely an attempt to discern the larger picture including diachronic trends while still remaining aware of the embeddedness of the linguistic item under investigation into larger linguistic, social and cultural contexts.

81

Andreas H. Jucker and Daniela Landert

Pragmaphilology Under this heading we subsume work on pragmatic entities as evidenced in earlier stages of a particular language. Such work may be devoted to the writing of one particular author or to other data that spans only a short period of time. Neither synchronic variation (i.e. variation according to such factors as region, social status, gender, genre and the specific communicative situation) nor diachronic variation is a particular focus of such research (cf. later in this chapter on synchronic variation and the pragmaphilological approach). Pragmaphilological analyses spanning several centuries may focus, for instance, on the Old English period, which – according to the standard textbooks – lasted some 700 years from 450 to 1150. In doing so, they abstract away from diachronic differences and analyse data synchronically without paying attention to diachronic variation. Strictly speaking, however, it is usually a simplification to ignore the diachronic dimension in whatever data a scholar chooses to analyse. Even the work of one single author may have a short-term diachronic dimension between earlier and later texts. A single play text or any other single text may appear to have no diachronic dimension at all, but even this may be deceptive as such a text may contain both archaisms and innovations, that is, earlier and later stages of language development may be represented in one and the same text. In the history of the English language, two authors have received an inordinate amount of attention in historical pragmatics – Geoffrey Chaucer and William Shakespeare (for overviews, see Pakkala-Weckström 2010 and Busse and Busse 2010). The numerous historical pragmatics studies of Chaucer’s work may serve as an illustration of some of the problems historical pragmatics has to deal with. Geoffrey Chaucer (1341–1400) wrote his works before Gutenberg had invented the art of printing with movable type, and none of Chaucer’s original handwriting has survived. We know his work through handwritten copies. His best-known work, The Canterbury Tales, for instance, has survived in eighty-two different manuscripts, and present-day editions are abstractions that try to give an account of the text that is both faithful to what the editors think the author must have intended and helpful to the present-day reader. The typeface is obviously different from the original handwriting. In addition, scholarly editions may choose to replace unfamiliar characters or abbreviations with appropriate present-day characters and extended versions of the abbreviations. They may also introduce line breaks and punctuations which were not part of the writing tradition of the time. Thus the historical linguist cannot be sure whether the available text is really what Chaucer had intended (Pakkala-Weckström 2010: 228). With authors who have received less attention from editors and historical linguists and for whom we know less about the textual histories and the communicative contexts of their work, the problems may be exacerbated. It is usually the scholarly editions which are used for the compilation of corpora, but in the compilation process the apparatus of explanatory notes and variant readings tends to disappear. What is left is a text that is to a large extent an idealisation far removed from the complexities of its textual histories. Often it may be impracticable to go back to the original manuscript, even though this could often lead to relevant insights (Caroll et al. 2013). The historical pragmaticists, therefore, must strive to find a balance between a reliance on the editorial work of the standard editions of these works and an awareness that these editions are idealisations. Historical pragmaticists have studied Chaucer’s work from many different angles. Some studies deal with Chaucer’s entire work, while some focus on the Canterbury Tales, on individual tales or even on specific passages from one of the tales (see Pakkala-Weckström 2010: 222 for a useful overview). Among these studies three topic areas have proved particularly popular with historical pragmaticists.These include firstly issues of politeness (e.g. Sell 1985 or Shimonomoto 2000). Related to this, the use of pronominal terms of address has also been analysed extensively. 82

Historical pragmatics

Chaucer’s characters have an option of using the singular pronoun thou (and its case forms) when addressing a single addressee or the plural pronoun ye (and its case forms).To some extent the system is similar to the situation in present-day French or German, but the rules appear to be both more complex and more flexible (see e.g. Mazzon 2000; Honegger 2003; Jucker 2006). And, finally, the use of specific speech acts in Chaucer’s work has received attention from various scholars. Examples include promises (Arnovick 1994; Pakkala-Weckström 2002), insults (Jucker 2000) and greetings and farewells (Jucker 2011). Such pragmaphilological work analyses the communicative behaviour of fictional characters, fictional characters, moreover, who interact with each other in perfectly crafted verse. These conversations, therefore, cannot be taken as representative of language use in different contexts by, for instance, everyday fourteenth-century speakers of English. However, they give an insight into one specific and fascinating area of language use.

Diachronic change Diachronic pragmatics deals with the changes of pragmatic features over time. Diachronic change can either be studied by comparing two (or more) different points in time, or by tracing the development of pragmatic features over a certain period. In the first case, data from different periods are compared with each other, similar to comparing data sets that differ with respect to other characteristics, for example different genres. Such parallels to comparative approaches can best be seen in studies that combine both perspectives. In a study of directive speech acts, Moessner (2010), for instance, combines a comparison of two periods, early modern English and present-day English, with a comparison across three genres, legal, religious and scientific discourse. Directives can be realised in several ways, with more direct or more indirect strategies. Moessner analyses the frequency and realisation of these different strategies in both periods and in all three genres and compares her findings across both dimensions in a similar way. The second option has sometimes been called the ‘evolutionary’ approach (e.g. Fritz 2012: 106). It is not based on a comparison of distinct points in time, but rather consists in a stepby-step tracing of developments which often span many centuries. An example is the study of how the present-day form goodbye (or short: bye) has developed from the blessing God be with you. Arnovick (1999) studied this development with the help of a large corpus of English plays from the fourteenth to the twentieth century. Her study shows the gradual loss of the function of blessing, which is replaced by a general closing function for partings. At the same time, the form undergoes contraction and change to a point that the original meaning of the parting is no longer transparent and God is replaced with good. The boundaries between synchronic and diachronic approaches are not always clear-cut. For instance, pragmaphilological research that only looks at one specific earlier period can still have a diachronic dimension by (more or less explicitly) relating the results to findings of presentday language use. An example of this is Culpeper and Archer’s (2008) study of requests and directness in early modern English. Their data come from the period between 1640 and 1760 and are treated as a synchronic data set, meaning that changes throughout this period are not investigated. However, Culpeper and Archer compare their findings with results from the large Cross Cultural Speech Act Realization Project carried out in the 1980s (Blum-Kulka et al. 1989). As a consequence, their synchronic study of requests in early modern English has a diachronic dimension, because it points to differences of this speech act between the early modern period and the late twentieth century. When investigating the change of pragmatic features over time, researchers have two basic options. Either they focus on a particular form and investigate how its pragmatic function 83

Andreas H. Jucker and Daniela Landert

changes over time, or they focus on a particular pragmatic function and investigate the various forms with which this function is realised over time.The first option is known as ‘form to function mapping’ (Jacobs and Jucker 1995: 13). It can be illustrated with the study of the development of Jesus, the name of a religious figure, into the primary interjection Jesus! (Gehweiler 2008). The expletive Jesus! is used as an interjection and no longer has a referential meaning. Thus, while the form is the same, the function undergoes change. However, as it is often the case (see earlier in the case of goodbye), the form also undergoes some degree of change, when the taboo expletive is later shortened into gee!, in this case with the additional motivation of obscuring its original taboo form. The second option is termed ‘function to form mapping’ and it is used, for instance, to investigate speech acts such as requests, apologies or compliments (Jacobs and Jucker 1995: 13). In this case, the pragmatic function of the speech act remains (more or less) constant over time, but the ways in which it is realised often change considerably.The fact that the form of a speech act changes over time poses a practical problem in its identification. With the historical corpora available today, retrieving all instances of a particular linguistic form is relatively easy. However, it is not possible to search these corpora automatically for pragmatic functions. A rather timeconsuming solution consists in reading through all texts and identifying each instance of a speech act ‘by hand’ (see Kohnen 2007: 139, 2015). Other approaches build on the fact that some speech acts have recurrent formal characteristics, such as the frequent use of the Illocutionary Force Indicating Devices (IFIDs) sorry, pardon, excuse and forgive in apologies (Jucker and Taavitsainen 2008). By searching for these devices, a large proportion of apologies can be found, even though some instances will still be missed. A third way of investigating speech acts diachronically apart from automatic searches for IFIDs and manual identification involves the analysis of metacommunicative expressions, that is, words or expressions used to talk about speech acts. For instance, by searching for the expression compliment (in its various forms), it is possible to identify passages in which people talk about compliments (Taavitsainen and Jucker 2008). Such passages provide important insights into what counted as a compliment at an earlier time, and sometimes instances of compliments can be found nearby (2008: 207–208).

Synchronic variation in earlier periods Pragmatic features do not only change over time, they can also vary synchronically, depending, for instance, on region, social status and gender of language users, and on contextual factors, such as genre and the specific communicative situation. As such, they can be investigated from a sociolinguistic point of view. Compared to other levels of language, such as phonology, morphology or lexicon, the variation of pragmatic features in earlier periods has so far received little attention. However, research on synchronic variation has been conducted from the very beginning of historical pragmatics. In particular the variation across different genres and text types has received considerable attention (see, for instance, Biber and Finegan 1992 and Taavitsainen 1993 for two early approaches). Research providing insight into synchronic variation has often been carried out within the scope of pragmaphilology (cf. earlier in this chapter). Jacobs and Jucker define pragmaphilology as the field of study that ‘describes the contextual aspects of historical texts, including the addressers and addressees, their social and personal relationship, the physical and social setting of text production and text reception, and the goal(s) of the text’ (1995: 11). This wide definition of context covers all factors that can lead to variation of pragmatic features. In contrast, the narrow conception of pragmatics does not include such aspects (see earlier in this chapter). It is in response to the narrow conception of pragmatics that a new branch of historical sociopragmatics 84

Historical pragmatics

has recently been suggested (see Culpeper 2010, 2011). The term sociopragmatics points to the relation of this research area with sociolinguistics, and in particular with interactional sociolinguistics. Historical sociopragmatics is mainly concerned with how pragmatic meaning is constructed in interaction. The variationist perspective, that is, the correlation of pragmatic and sociological features (see next section), is of secondary interest (Culpeper 2010: 74–75; see also Nevalainen and Raumolin-Brunberg 2003: 17–18). Among the earliest approaches to synchronic variation of pragmatic features is research on terms of address, in particular pronominal terms of address (see Mazzon 2010 for a recent overview). Due to the strong influence of social identity and the relation between the interactants on the choice of pronominal terms of address, social variation has been a central aspect of such investigations, for example in Brown and Gilman’s (1960) seminal paper. We briefly mentioned earlier that speakers of Middle English had the choice between two address terms, ye and thou. The study of these pronouns is concerned with the (mostly) language-external factors that determine which pronoun is used. Among the important factors influencing the choice are social status, age and the relationship between the interactants (see Burnley 2003: 29). In addition, situational factors play a role, so that even within a conversation, switches can take place for rhetorical or affective reasons. Jucker (2006) illustrates with examples from several of Chaucer’s Canterbury tales how switches between ye and thou often mark a turning point in the story, at which the power balance between two characters shifts. Research on pronominal terms of address in Middle English thus studies how the choice of address term varies depending on the identity of the interlocutors, their roles and relationships and the communicative situation. Lutzky’s (2012) study of discourse markers in early modern English is another example of a sociopragmatically oriented approach to historical pragmatics. One part of her analysis is based on a corpus of drama texts which are annotated for social status and gender of the characters. She shows that the discourse marker marry is preferred by characters of low status, while the discourse markers why and well tend to be preferred by characters of higher social status (2012: 244–247). Moreover, gender plays a role so that, for instance, male characters more frequently use the discourse marker well, and it is more frequent in dyads of the same gender, compared to dyads of opposite gender (2012: 261). Other types of variation that can be analysed are genre variation and regional variation. The first type has been investigated quite frequently, and an example was mentioned in the previous section. In contrast, regional variation is so far a rather underexplored area of historical pragmatics. We have presented approaches to historical pragmatics as falling into three general groups – pragmaphilology, which describes pragmatic features at an earlier point in time without a special focus on variation or change; approaches dealing with diachronic change; and studies of synchronic variation in earlier periods. As we have pointed out before, these are not strictly distinct categories. Many studies that are pragmaphilological in aim include aspects of synchronic variation or diachronic change, and similarly there is a certain degree of overlap between the other two categories. The difference between the approaches often depends on the focus of a study, that is, whether the main aim consists in describing a certain pragmatic feature from a pragmaphilological, diachronic or variationist perspective.

Investigating variation: pragmatic variables Variationist studies are based on the general principle of correlating linguistic features with language-external factors such as social status, gender or origin of the speaker. Jucker and Taavitsainen (2012) have recently proposed the term pragmatic variable in order to apply the variationist 85

Andreas H. Jucker and Daniela Landert

framework to pragmatic entities. How linguistic features are realised, for example whether a pronominal term of address is realised as thou or as ye, depends on sociodemographic factors, and therefore linguistic features are termed dependent variables. In contrast, the factors influencing the linguistic realisation are termed independent variables.Variationist sociolinguistics – and by implication variationist historical pragmatics – is thus concerned with investigating the influence of independent sociodemographic variables on dependent linguistic variables. Pragmatic studies of variation in earlier periods share the problems of other fields of historical sociolinguistics with respect to the limited availability of data (Auer et al. 2015: 5–7). If no texts have survived from a specific group of speakers, their language use can simply no longer be investigated. As a consequence of the general imbalance in education and literacy (Kiełkiewicz-Janowiak 2012: 308), women and members of lower social status are typically underrepresented in the surviving data from earlier periods. This means that much less is known about the language use of these groups and that the study of social variation has to be restricted to groups for which more data is available (see e.g. Nevalainen and RaumolinBrunberg 2003: 137). Moreover, information about the sociodemographic background of the authors of surviving texts is not always available, which further restricts the data that can be used for such studies. The difficulty of defining the dependent variable poses an additional problem for the analysis of variation of pragmatic features. Dependent variables are linguistic features that can be realised in two or more different ways, depending on context. In order to study the effect of sociodemographic variables, it is important that all possible realisations are equivalent in meaning and function. In their sociohistorical study of early modern English letters, Nevalainen and Raumolin-Brunberg (2003) investigate the variation between the third-person singular suffixes –th and –s, that is to say the replacement of old forms such as eateth, doth or hath with the present-day English forms eats, does and has. Between the beginning of the fifteenth and the end of the seventeenth century, the older –th forms were replaced with the newer –s forms (2003: 68). The function of the ending remained the same, and both endings equally expressed the thirdperson singular.The only difference between them was that one was the older and the other the newer form, and that different groups of speakers preferred one form over the other at various points in this development. Such constancy in function and meaning of the various realisations is much rarer for pragmatic variables (Jucker and Taavitsainen 2012: 296–297). We can say that the two variants thou and ye fulfil the same core function of addressing an interlocutor, but – as pointed out earlier – they clearly differ in their connotations and they are not fully equivalent in function. In addition, it is often difficult to define all possible variants with which a pragmatic variable can be realised (Jucker and Taavitsainen 2012: 303). For morphological variables, like the third-person singular ending, the number of possible variants is small and usually easy to determine. In contrast, it is more difficult to define a set of discourse markers that are possible realisations of the same functions. For some speech acts, such as compliments, it is even impossible to compile a complete list of possible realisations. Despite these difficulties, the concept of the pragmatic variable is promising for the study of language in earlier periods. A clear definition of the pragmatic variable under analysis is a prerequisite for the study of synchronic variation of pragmatic features. Heightened awareness of the difficulties for defining pragmatic variables might be a first step in creating new approaches to the problems mentioned previously. Moreover, pragmatic variables might take on a central role in the investigation of the relationship between synchronic variation and diachronic change of pragmatic features. Again, it is necessary to clearly define a pragmatic variable and its possible variants in order to trace the diffusion of pragmatic innovations across different groups of speakers. This is an area in which there is still a lot of potential for future research. 86

Historical pragmatics

Conclusion Diachronic language studies and variationist studies have a long and eminent history in the field of linguistics, but it is only fairly recently that diachronic language studies extended their scope from the language systems to more systematic investigations of how language was actually used in earlier periods, and, similarly, that variationist studies extended their scope from the traditional core areas of linguistics to include pragmatic variables.The former extension has now established itself as historical pragmatics, the latter as variational pragmatics.Within the context of historical pragmatics, diachronic variation, that is, the development of pragmatic entities over time, has always been an important aspect. At the same time, scholars working in the framework of sociohistorical linguistics have extended the research tools of present-day sociolinguistics to historical data. However, a more systematic application of a variationist framework to pragmatic entities in a historical context is still very much in its infancy and a lot of exciting work remains to be done. In principle, it would be important to explore all of the established levels of sociolinguistic variation – social class, gender, age, region and genre – to all sorts of pragmatic entities – speech acts, discourse markers, conversational styles and so on – in order to determine the interrelationship between synchronic variation at a given point in time and diachronic change. In reality many of these dimensions will remain very difficult to investigate because of the limited survival of relevant material and the impossibility of carrying out historical experiments.

Suggestions for further reading Jucker, A. H. and Taavitsainen, I. (eds.) (2010) Historical Pragmatics (Handbooks of Pragmatics 8). Berlin/ New York: De Gruyter Mouton. This handbook contains twenty-two survey articles that present detailed state-of-the-art accounts of the entire field of historical pragmatics, from data and methodology and grammaticalisation theory to pragmatic entities, such as discourse markers and speech acts, and to individual discourse domains, such as scientific writing and literary discourse. Jucker, A. H. and Taavitsainen, I. (2013) English Historical Pragmatics (Edinburgh Textbooks on the English Language). Edinburgh: Edinburgh University Press. This is the first textbook in the field of historical pragmatics. It is intended for advanced students of linguistics or pragmatics and provides a clear and concise introduction to the field. The chapters come with exercises and suggestions for further reading. Nevalainen, T. and Raumolin-Brunberg, H. (2003) Historical Sociolinguistics. Harlow/London: Pearson. This book provides an excellent overview of the potentials and the problems of sociolinguistic variability studies in an historical context. On the basis of an extensive corpus of personal letters from the Middle English and early modern English period, it discusses the factors that promoted linguistic changes, and it identifies the people leading these changes. Taavitsainen, I. and Jucker, A. H. (2015) ‘Twenty years of historical pragmatics: Origins, developments and changing thought styles’, Journal of Historical Pragmatics, 16(1): 1–25. This article surveys the changes in linguistics that led to the rise of historical pragmatics in the mid1990s, and the developments in the field within the past twenty years. The changes and developments are discussed in terms of shifting thought styles. Seven different turns are identified: the pragmatic turn, the sociocultural turn, the dispersive turn, the empirical turn, the digital turn, the discursive turn and the diachronic turn.

Corpora Corpus of Historical American English (COHA) (http://corpus.byu.edu/coha/). The Helsinki Corpus of English Texts (HC) (1991). Department of Modern Languages, University of Helsinki. Compiled by Matti Rissanen (Project leader), Merja Kytö (Project secretary); Leena Kahlas-Tarkka, Matti Kilpiö (Old English); Saara Nevanlinna, Irma Taavitsainen (Middle English); Terttu Nevalainen, Helena Raumolin-Brunberg (Early Modern English). In ICAME Collection of English Language Corpora (CD-ROM), 2nd ed.

87

Andreas H. Jucker and Daniela Landert

References Arnovick, L. K. (1994) ‘Dorigen’s promise and scholars’ premise: The orality of the speech acts in the Franklin’s Tale’, in M. C. Amodio (ed.) Oral Poetics in Middle English Poetry. New York: Garland. 125–147. Arnovick, L. K. (1999) Diachronic Pragmatics. Seven Case Studies in English Illocutionary Development (Pragmatics & Beyond New Series 68). Amsterdam/Philadelphia: John Benjamins. Auer, A., Peersman, C., Pickl, S., Rutten, G., and Vosters, R. (2015) ‘Historical sociolinguistics:The field and its future’, Journal of Historical Sociolinguistics, 1(1): 1–12. Barron, A. and Schneider, K. P. (2009) ‘Variational pragmatics: Studying the impact of social factors on language use in interaction’, Intercultural Pragmatics, 6(4): 425–442. Biber, D. and Finegan. E. (1992) ‘The linguistic evolution of five written and speech-based English genres from the 17th to the 20th centuries’, in M. Rissanen, O. Ihalainen, T. Nevalainen and I. Taavitsainen (eds.) History of Englishes. New Methods and Interpretations in Historical Linguistics. Berlin: Mouton de Gruyter. 688–704. Blum-Kulka, S., House J. and Kasper G. (eds.) (1989) Cross-cultural Pragmatics: Requests and Apologies (Advances in Discourse Processes 31). Norwood, NJ: Ablex. Brinton, L. J. (2001) ‘Historical discourse analysis’, in D. Schiffrin, D. Tannen and H. E. Hamilton (eds.) The Handbook of Discourse Analysis. Oxford: Blackwell. 138–160. Brown, R. and Gilman, A. (1960) ‘The pronouns of power and solidarity’, in T. A. Sebeok (ed.) Style in Language. Cambridge, MA: MIT Press. 253–276. Burnley, D. (2003) ‘The T/V pronouns in later Middle English Literature’, in I. Taavitsainen and A. H. Jucker (eds.) Diachronic Perspectives on Address Term Systems (Pragmatics & Beyond New Series 107). Amsterdam/Philadelphia: John Benjamins. 27–45. Busse, U. and Busse, B. (2010) ‘Shakespeare’, in A. H. Jucker and I. Taavitsainen (eds.) Historical Pragmatics (Handbooks of Pragmatics 8). Berlin/New York: De Gruyter Mouton. 247–281. Caroll, R., Peikola, M., Salmi, H.,Varila, M.-L., Skaffari, J. and Hiltunen, R. (2013) ‘Pragmatics on the page. Visual text in late medieval English books’, European Journal of English Studies, 17(1): 54–71. Chapman, S. (2011) Pragmatics (Palgrave Modern Linguistics). Houndsmills, Basingstoke: Palgrave Macmillan. Culpeper, J. (2010) ‘Historical sociopragmatics’, in A. H. Jucker and I.Taavitsainen (eds.) Historical Pragmatics (Handbooks of Pragmatics 8). Berlin/New York: De Gruyter Mouton. 69–94. Culpeper, J. (ed.) (2011) Historical Sociopragmatics (Benjamins Current Topics 31). Amsterdam/Philadelphia: John Benjamins. Culpeper, J. and Archer, D. (2008) ‘Requests and directness in Early Modern English trial proceedings and play texts, 1640–1760’, in A. H. Jucker and I.Taavitsainen (eds.) Speech Acts in the History of English (Pragmatics & Beyond New Series 176). Amsterdam/Philadelphia: John Benjamins. 45–84. Culpeper, J. and Kytö, M. (2010) Early Modern English Dialogues. Spoken Interaction as Writing (Studies in English Language). Cambridge: Cambridge University Press. Fritz, G. (2012) ‘Kontroversen – Ein Paradigma für die Geschichte von Kommunikationsformen’, Jahrbuch für germanistische Sprachgeschichte, 3(1): 105–126. Gehweiler, E. (2008) ‘From proper name to primary interjection. The case of “gee!” ’, Journal of Historical Pragmatics, 9(1): 71–93. Honegger, T. (2003) ‘ “And if ye wol nat so, my lady sweete, thanne preye I thee [. . .].”: Forms of address in Chaucer’s Knight’s Tale’, in I. Taavitsainen and A. H. Jucker (eds.) Diachronic Perspectives on Address Term Systems (Pragmatics & Beyond New Series 107). Amsterdam/Philadelphia: John Benjamins. 61–84. Huang,Y. (2007) Pragmatics. Oxford: Oxford University Press. Huang.Y. (2012) The Oxford Dictionary of Pragmatics. Oxford: Oxford University Press. Jacobs, A. and Jucker, A. H. (1995) ‘The historical perspective in pragmatics’, in A. H. Jucker (ed.) Historical Pragmatics. Pragmatic Developments in the History of English (Pragmatics & Beyond New Series 35). Amsterdam/Philadelphia: John Benjamins. 3–33. Jucker, A. H. (ed.) (1995) Historical Pragmatics. Pragmatic Developments in the History of English (Pragmatics & Beyond New Series 35). Amsterdam/Philadelphia: John Benjamins. Jucker, A. H. (2000) ‘Slanders, slurs and insults on the road to Canterbury. Forms of verbal aggression in Chaucer’s Canterbury Tales’, in I. Taavitsainen, T. Nevalainen, P. Pahta and M. Rissanen (eds.) Placing Middle English in Context (Topics in English Linguistics 35). Berlin: Mouton de Gruyter. 369–389. Jucker, A. H. (2006) ‘ “Thou art so loothly and so oold also”:The use of ye and thou in Chaucer’s Canterbury Tales’, Anglistik, 17(2): 57–72. Jucker, A. H. (2008) ‘Historical pragmatics’, Language and Linguistics Compass, 2(5): 894–906.

88

Historical pragmatics Jucker, A. H. (2011) ‘Greetings and farewells in Chaucer’s Canterbury Tales’, in P. Pahta and A. H. Jucker (eds.) Communicating Early English Manuscripts (Studies in English Language). Cambridge: Cambridge University Press. 229–240. Jucker, A. H. (2012) ‘Pragmatics in the history of linguistic thought’, in K. Allan and K. Jaszczolt (eds.) The Cambridge Handbook of Pragmatics. Cambridge: Cambridge University Press. 495–512. Jucker, A. H., and Landert, D. (2015) ‘Historical pragmatics and early speech recordings: Diachronic developments in turn-taking and narrative structure in radio talk shows’, Journal of Pragmatics, 79: 22–39. Jucker, A. H. and Taavitsainen, I. (2008) ‘Apologies in the history of English. Routinized and lexicalized expressions of responsibility and regret’, in A. H. Jucker and I. Taavitsainen (eds.) Speech Acts in the History of English (Pragmatics and Beyond New Series 176). Amsterdam/Philadelphia: John Benjamins. 229–244. Jucker, A. H. and Taavitsainen, I. (eds.) (2010) Historical Pragmatics (Handbooks of Pragmatics 8). Berlin/ New York: De Gruyter Mouton. Jucker, A. H. and Taavitsainen, I. (2012) ‘Pragmatic variables’, in J. M. Hernández-Campoy and J. C. CondeSilvestre (eds.) The Handbook of Historical Sociolinguistics. Oxford: Blackwell. 293–306. Jucker, A. H. and Taavitsainen, I. (2013) English Historical Pragmatics (Edinburgh Textbooks on the English Language). Edinburgh: Edinburgh University Press. Kiełkiewicz-Janowiak, A. (2012) ‘Class, age, and gender-based patterns’, in J. M. Hernández-Campoy and J. C. Conde-Silvestre (eds.) The Handbook of Historical Sociolinguistics. Oxford: Blackwell. 307–331. Koch, P. (1999) ‘Court records and cartoons: Reflections of spontaneous dialogue in Early Romance texts’, in A. H. Jucker, G. Fritz and F. Lebsanft (eds.) Historical Dialogue Analysis (Pragmatics & Beyond New Series 66). Amsterdam/Philadelphia: John Benjamins. 399–429. Koch, P. and Oesterreicher,W. (1985) ‘Sprache der Nähe – Sprache der Distanz: Mündlichkeit und Schriftlichkeit im Spannungsfeld von Sprachtheorie und Sprachgeschichte’, Romanistisches Jahrbuch, 36: 15–43. Koch, P. and Oesterreicher, W. (2011) Gesprochene Sprache in der Romania. Französisch, Italienisch, Spanisch, 2nd rev. and exp. edn. Berlin: De Gruyter. Kohnen, T. (2007) ‘Text types and the methodology of diachronic speech act analysis’, in S. M. Fitzmaurice and I. Taavitsainen (eds.) Methods in Historical Pragmatics. Berlin: Mouton de Gruyter. 139–166. Kohnen, T. (2015) ‘Speech acts: A diachronic perspective’, in K. Aijmer and Ch. Rühlemann (eds.) Corpus Pragmatics. A Handbook. Cambridge: Cambridge University Press. 52–83. Kytö, M. (2010) ‘Data in historical pragmatics’, in A. H. Jucker and I. Taavitsainen (eds.) Historical Pragmatics (Handbooks of Pragmatics 8). Berlin/New York: De Gruyter Mouton. 33–67. Lutzky, U. (2012) Discourse Markers in Early Modern English (Pragmatics & Beyond New Series 227). Amsterdam/Philadelphia: John Benjamins. Mazzon, G. (2000) ‘Social relations and form of address in the Canterbury Tales’, in D. Kastovsky and A. Mettinger (eds.) The History of English in a Social Context. A Contribution to Historical Sociolinguistics. Berlin: Mouton de Gruyter. 135–168. Mazzon, G. (2010) ‘Address terms’, in A. H. Jucker and I.Taavitsainen (eds.) Historical Pragmatics (Handbooks of Pragmatics 8). Berlin/New York: De Gruyter Mouton. 351–376. Moessner, L. (2010) ‘Directive speech acts. A cross-generic diachronic study’, Journal of Historical Pragmatics, 11(2): 219–249. Nevalainen, T. and Raumolin-Brunberg, H. (2003) Historical Sociolinguistics. Harlow/London: Pearson. Pakkala-Weckström, M. (2002) ‘‘Have heer my trouthe – til that myn herte breste’: Dorigen and the difficulty of keeping promises in the ‘Franklin’s Tale’’, in H. Raumolin-Brunberg, M. Nevala, A. Nurmi and M. Rissanen (eds.) Variation Past and Present. VARIENG Studies on English for Terttu Nevalainen, (Mémoires de la Société Néophilologique de Helsinki 61). Helsinki: Société Néophilologique. 287–300. Pakkala-Weckström, M. (2010) ‘Chaucer’, in A. H. Jucker and I. Taavitsainen (eds.) Historical Pragmatics (Handbooks of Pragmatics 8). Berlin/New York: De Gruyter Mouton. 219–245. Schneider, K. P. (2010) ‘Variational pragmatics’, in M. Fried, J.-O. Östman and J. Verschueren (eds.) Variation and Change. Pragmatic Perspectives (Handbook of Pragmatics Highlights 6). Amsterdam/Philadelphia: John Benjamins. 239–267. Schneider, K. P. and Barron, A. (eds.) (2008) Variational Pragmatics. A Focus on Regional Varieties in Pluricentric Languages (Pragmatics & Beyond New Series 178). Amsterdam/Philadelphia: John Benjamins. Sell, R. D. (1985) ‘Politeness in Chaucer: Suggestions towards a methodology for pragmatic stylistics’, Studia Neophilologica, 57: 175–185. Shimonomoto, K. (2000) The Language of Politeness in Chaucer. Tokyo: Waseda University.

89

Andreas H. Jucker and Daniela Landert Taavitsainen, I. (1993) ‘Genre/subgenre styles in Late Middle English’, in M. Rissanen, M. Kytö and M. Palander-Collin (eds.) Early English in the Computer Age. Explorations through the Helsinki Corpus (Topics in English Linguistics 11). Berlin: Mouton de Gruyter. 171–200. Taavitsainen, I. and Jucker, A. H. (2008) ‘ “Methinks you seem more beautiful than ever”: Compliments and gender in the history of English’, in A. H. Jucker and I. Taavitsainen (eds.) Speech Acts in the History of English (Pragmatics & Beyond New Series 176). Amsterdam/Philadelphia: John Benjamins. 195–228. Taavitsainen, I. and Jucker, A. H. (2010) ‘Trends and developments in historical pragmatics’, in A. H. Jucker and I. Taavitsainen (eds.) Historical Pragmatics (Handbooks of Pragmatics 8). Berlin/New York: De Gruyter Mouton. 3–30. Taavitsainen, I. and Jucker, A. H. (2015) ‘Twenty years of historical pragmatics: Origins, developments and changing thought styles’, Journal of Historical Pragmatics, 16(1): 1–25. Traugott, E. C. (2004) ‘Historical pragmatics’, in L. R. Horn and G. Ward (eds.) The Handbook of Pragmatics. Oxford: Blackwell. 538–561.

90

10 VARIATIONAL PRAGMATICS1 Anne Barron

Introduction Variational pragmatics is a field of study that aims to systematically describe synchronic variation in the patterns of human interaction within one language due to such factors as region, gender, socioeconomic status, ethnic identity and age. It is situated at the interface of pragmatics and modern dialectology. Variational pragmatics, like cross-cultural pragmatics, historical pragmatics, intercultural pragmatics, interlanguage pragmatics and postcolonial pragmatics, is a branch of pragmatics (cf. Barron and Schneider 2009: 425; Schneider 2010a, for an overview of the relationships between these fields). As a research area, it is a relatively recent arrival, having been proposed as a systematic approach to synchronic intralingual pragmatic variation in the early years of this century (cf. Barron 2005a, 2014, 2015; Schneider and Barron 2008; Barron and Schneider 2009; Schneider 2010a, 2014; cf. Barron and Schneider 2009: 432–434; Schneider 2010a: 254–256 on the development of variational pragmatics). This is not to say that studies on intralingual pragmatic variation according to macro-social factors did not exist before this time. Overall, however, prior to the emergence of this field, systematic research into synchronic intralingual pragmatic variation according to macro-social factors was a research desideratum both in the areas of pragmatics and modern dialectology. In pragmatics research, languages had been implicitly viewed as homogeneous wholes with macro-social variation largely abstracted away. Hence, the norms of interaction in British English, for example, had been implicitly assumed to be the same as those in, for instance, Irish English or Australian English. A similar dearth of research on intralingual pragmatic variation existed in modern dialectology. There, research had focused on synchronic phonetic, phonological, morphological, syntactic and lexical variation. Intralingual pragmatic variation had, however, remained largely undescribed. Schlieben-Lange and Weydt (1978) noted this in the 1970s, and in the early years of this century, the situation had not changed significantly (cf. Clyne et al. 2003: 96; Wolfram and Schilling-Estes 2006; Pichler 2010: 582; Jucker and Taavitsainen 2012; cf. also Barron 2005a: 522–523; Schneider and Barron 2008: 2–3 for an overview). It was in this context that variational pragmatics emerged. Its appearance has put intralingual pragmatic variation on the research agenda. Indeed, Terkourafi in a recent article on variational pragmatics (and on variationist sociolinguistic approaches to pragmatic variation), 91

Anne Barron

writes: ‘Variation in the pragmatic plane has finally made it to the forefront’ (2012: 315). In the following, the approach taken in this ‘budding field’ of variational pragmatics is outlined (Aijmer and Andersen 2011: 4).

Macro-social factors In variational pragmatics, as in modern dialectology, five macro-social factors are distinguished as having a systematic influence on the conventions of language use. These are region, social class, ethnicity, gender and age, although further factors, such as education and religion, may also represent extensions to the list (cf. Coupland 1983; Wolfram and Schilling-Estes 2006; cf. also Schneider and Barron 2008; Barron and Schneider 2009; Schneider 2010a; Placencia 2011). Variational pragmatics aims in the first instance at determining the influence of each of the five macro-social factors on language use in interaction. So, a variational pragmatic study might investigate how requests are realised in Irish English and British English or how twenty- to thirty-year-olds and forty- to fifty-year-olds use a particular discourse marker. Additionally, the question arises as to the nature of the interplay of these factors and also as to the nature of the interaction between macro-social and micro-social factors (i.e. social distance, social dominance, degree of imposition). Both quantitative and qualitative studies have been conducted on the relationship between the macro-social factors and conventions of language use.2 To date, quantitative studies outnumber qualitative studies. As a result, macro-social factors are generally operationalised via geographical, biological and social facts in line with Cheshire’s workable approach to gender. She writes: Speaker sex is intended to be a purposely broad, unrefined social variable that can be easily taken into account at the data-collection stage of research. If all researchers categorize speakers in the same, albeit simplistic way, we can ensure replicability and can draw useful comparisons between studies carried out in a range of communities. (Cheshire 2002: 424–425) In other words, to date the focus in variational pragmatic research is generally on sex rather than on gender as a social category and on geographical domicile rather than on regional identity (cf. also Barron 2005a; Barron and Schneider 2009; Schneider 2010a on identity construction).

Levels of analysis Variational pragmatics distinguishes five levels of analysis, namely the formal level, the actional level, the interactional level, the topic level and the organisational level (cf. Schneider and Barron 2008; Barron and Schneider 2009; Schneider 2010a, 2012a, 2012b, 2014; Barron 2014), without wanting to exclude alternative levels of analysis (cf. Jucker 2008; Schneider 2010a; Placencia 2011; Jucker and Taavitsainen 2012). Empirical analyses may combine a number of levels. Analyses on the formal level focus on the communicative function of individual forms, such as mitigators or discourse markers. This level may be equated with Jucker and Taavitsainen’s (2012) expression level and represents the smallest unit of analysis. Pichler’s (2009) investigation of localised and non-localised full and reduced variants of the discourse markers I DON’T KNOW and I DON’T THINK is an example. Foolen (2011: 221–225) presents an overview of studies of pragmatic markers from a variational perspective which shows clearly that a single form may realise different functions across varieties and, vice versa, that a single function may be realised using different forms across varieties. 92

Variational pragmatics

The actional level deals with speech act analyses. It is equivalent to Jucker and Taavitsainen’s (2012: 299–300) utterance level. Here, the question is posed as to how particular speech acts, for example requests, offers, refusals, are realised in different intralingual varieties. Specifically, analyses centre on pragmalinguistic questions relating to the strategies (conventions of means) and linguistic realisations of strategies realising the individual speech acts (conventions of form) (cf. Clark 1979; cf. also Barron 2005a: 526–529). Analyses are combined with sociopragmatic questions, such as when and where which speech act and speech act strategy is used. In offering, for instance, one might choose between a ‘question desire’ strategy, a ‘state willingness’ strategy and a range of other offer strategies in a particular situation (cf. Barron 2005b). The choice of strategy and its linguistic realisation may vary across intralingual varieties in a particular context according to the macro-social factors. In other words, there may be particular strategies or realisations of offer strategies preferred in or indeed exclusive to one variety relative to another variety, although exclusivity is frequently restricted to the conventions of form. On the level of the conventions of form, for instance, the form (what) do you fancy realises a ‘question desire’ hospitable offer strategy in British English, as for example, in what do you fancy for your lunch? in the initial line of (1) from the British National Corpus (BNC) web. (1) SP:PS0JJ: 

Oh yes, what do you fancy for your lunch? (pause) What do you fancy for your lunch? SP:PS0JL:  Food. (pause) Bit of bread and butter. SP:PS0JJ:  Bread and butter? (pause) Drop of bread and dripping. (BNC: KD3 S_conv)

Indeed, a search for do you fancy in the spoken component of the BNC, encompassing approximately 10 million words, yielded nine instances of hospitable offers of food or drink. The same form is not common in American English. A search for this form in the spoken component of the Corpus of Contemporary American English (COCA), encompassing some 90 million words, yielded only one hit (‘Do you fancy yourself a good mother? Ms-ANDREWS: I hope I’m a good enough mom’), a form which did not realise an offer but rather a request of judgement (cf. Variation and the pragmatic variable on the identification of function). Similarly, I’m good may be used as a refusal of an offer strategy in American English, as in the following example taken from the COCA: (2) # The waitress turned to Charlotte again. ‘Are you sure I can’t get you anything? Maybe an appetizer or a salad?’ # ‘No, I’m good. Really.’ (COCA: FIC Bk: LoveHonorBetray) In contrast, the form I’m good does not realise a refusal of offer in the BNC. Apart from such head act realisations, modification is also a focus of analysis at the actional level (cf. Barron 2005a: 529). So, for instance, alerters, such as dear might be used, as in ‘What do you fancy, dear?’ (BNC: KBW S_conv was) or indeed grounders (i.e. reasons/explanations/ justifications), as in the You must be hungry in the sample utterance ‘You must be hungry. What do you fancy?’ (cf. also Blum-Kulka et al. 1989). The preferred choice of modification in a particular situation may differ across varieties. The focus at the interactional level extends beyond the individual speech act to deal with sequential patterns. This level is partly equivalent to Jucker and Taavitsainen’s (2012: 301–302) conversation level, a level of analysis which includes both the interactional and organisational 93

Anne Barron

level in the present scheme. Questions posed relate to how speech acts combine into larger units of discourse, such as adjacency pairs, interchanges, interactional exchanges or phases. Haugh and Carbaugh (2015), for example, contrast self-disclosure patterns across initial interactions among speakers of American English and among speakers of Australian English. They find American participants to use unprompted self-disclosures, particularly sequence-medial unprompted self-disclosures, more often than Australian participants. In addition, they show both cultures to reveal an orientation to reciprocity insofar as a high level of one particular type of self-disclosing prompts the other participant to self-disclose in the same way more frequently, and vice versa. The topic level is concerned with discourse content, that is, with the propositions of individual utterances as well as with macro-propositions. It addresses, in particular, issues of topic selection and topic management. Schneider (2008) examines the topics for which information is requested in party talk in England, Ireland and the United States. Further questions in this context on the level of topic management which might be addressed include how much small talk is necessary in different varieties before getting to the heart of the interaction. However, variational research is still limited on this level. Finally, the organisational level combines ethnomethodological analysis and conversation analysis. The focus is on turn-taking. Analyses include comparisons of interrupting behaviour across varieties, analyses of overlap, of minimal responses, of back-channels and of inter-turn silence across varieties. Tottie (1991) compares backchannels in British and American English, while McCarthy (2002) examines non-minimal response tokens in British and American spoken conversations.

Variation and the pragmatic variable Variational pragmatics is influenced by – although by no means limited to – the Labovian approach to variation (cf. also Schneider 2010a: 251, 2014: 361–362 for a more in-depth discussion). Traditionally, the concepts of variable and variant are important in variationist sociolinguistics, and it is not surprising, therefore, that the advent of variational pragmatics has recently sparked debate about the applicability of these concepts to pragmatic analyses (cf. Terkourafi 2011: 344, 358). In particular the criterion of semantic, or truth-conditional, equivalence, one of the defining criterion of Labov’s (1966) variable, and its application to the pragmatic context has been a subject of debate not least due to, as Jucker and Taavitsainen put it, ‘the difficulty of arguing that two different pragmatic units constitute different ways of saying the same thing’ (2012: 296) (cf. Pichler 2010: 587–591; Jucker and Taavitsainen 2012: 295–297; Terkourafi 2011, 2012). Pichler (2010: 588) highlights this problem in the case of discourse markers, arguing that semantic sameness as a defining criterion is inappropriate as discourse markers by their very nature are semantically bleached. Similarly, Jucker and Taavitsainen argue that different request strategies, such as a direct performatives (I ask you to. . . .), impersonal constructions, the use of let’s + infinitive or the use of conventionally indirect requests, such as could you + infinitive. . . ., ‘cannot be said to be different ways of saying the same thing’ (2012: 301). Such difficulties suggest the necessity of modifying the original concept of the variable and, indeed, current research recognises the necessity of some adaptation. A number of solutions have been put forward to facilitate the application of the concept to discoursepragmatic features. Dines (1980) and Lavandera (1978) propose functional comparability between the variants of a pragmatic variable as a defining criterion to replace semantic sameness (cf. also Cameron and Schwenter 2013). Terkourafi (2011) takes a relevance theoretical 94

Variational pragmatics

approach to defining how functional equivalence might be interpreted in variational pragmatics. She suggests a procedural definition of the pragmatic variable, conceiving the process of inference as that which remains constant while the forms associated with it vary. As she writes, ‘a single procedural meaning [may be] encoded by different forms. . . . and a single form encod[ed] by different procedural meanings’ (Terkourafi 2011: 366). On the other hand, Pichler, writing on discourse markers (formal level), proposes the criterion of underlying structural similarity as an option while at the same time admitting that ‘some discourse variables might be better conceptualised based on functional comparability between variants (e.g., intensifiers), others based on structural commonality (e.g., general extenders)’ (2010: 590 and 591). On the interactional level, Schneider (2014) proposes a specific selection of initial speech acts (e.g., a greeting, a question identity or a disclose identity) as an example of a pragmatic variable. Jucker and Taavitsainen, also addressing the problem that ‘pragmatic variables rarely – if ever – provide clear cases of saying the same thing in several ways,’ put forward the idea of pragmatic variables (dependent variables) and pragmatic context variables (independent variables) (2012: 296). Pragmatic variables are those that can be realised and include, for instance, discourse ­markers on the formal level (level of expression in Jucker and Taavitsainen 2012) or individual speech acts on the actional level (their utterance level). Pragmatic context variables, on the other hand, influence the realisation of the linguistic items. They relate to the context and include factors such as the presence or absence of particular individuals in an interaction, that is, aspects which influence the realisation of the pragmatic variables. Jucker and Taavitsainen suggest that on the actional level, ‘strategies share a core of functional equivalence even if they cannot be said to be different ways of saying the same thing’ (2012: 301). Indeed, this definition is reminiscent of Schneider’s (2010a) suggestion that the variable on the actional level of analysis is the illocution. In other words, the focus is on individual illocutions and their variants – more precisely on the conventions of means and conventions of form (cf. Clark 1979) and how these vary according to contextual and macro-social factors (cf. also Schneider 2014: 370). Terkourafi, however, criticises this definition of illocutionary identity, claiming that the approach adopted to date in variational pragmatics: relies on the assumption that the analysts can define what constitutes, for instance, a request outside of particular contexts, so that they may then use that definition as a ‘yardstick’ to identify actual occurrences of requests in the data (original emphasis). (Terkourafi 2012: 302 and 314) Terkourafi goes on to say that ‘illocutionary sameness is not available to the analysts before they begin analysis of the data but must be empirically established each time within particular discourses and communities of practice’ (emphasis in original: 302). She supports this claim with reference to the fact that no one-to-one relationship exists between a particular form and illocution. However, in this criticism of variational pragmatic analyses,Terkourafi overlooks the care taken in variational pragmatics to overcome the challenge of identifying function. Firstly, variational pragmatics makes extensive use of elicited data, which provides researchers with contextualised data via, for example, hearer uptake, explicit situational descriptions and explicit communication of illocutionary force. Secondly, variational pragmatic analyses of naturally occurring data do not take form as an indicator of illocutionary force. Rather, the recommendation is to systematically employ speech act identification criteria, such as those of uptake, propositional content and further context, in the identification of illocutionary sameness (cf. Barron 2011, 95

Anne Barron

forthcoming; Flöck 2015).3 Both approaches are discussed in Methodological issues, the section to which we now turn.

Methodological issues Given its interest in language use rather than in the language system per se, analyses in variational pragmatics are not based on intuitive, fabricated data but are rather empirical analyses involving both naturally occurring and elicited data. A basic division may be made in this regard, with naturally occurring data yielding information on the actual use of language in context and experimental data yielding information of prototypical, canonical modes of behavior, also termed ‘cultural models’ (Schneider 2012b) (cf. later in this chapter). Variational pragmatic analyses follow the contrastivity principle according to which linguistic features can be considered variety-specific or variety-preferred only if the variety under study is contrasted with at least one other variety of the same type and language. In other words, the fact that a feature or pattern is used in one variety does not allow for generalised crossvarietal statements (cf. Barron and Schneider 2009: 429–430; Schneider 2012b: 363–364; cf. also McCarthy 2002: 69). Related to the principle of contrastivity is the principle of comparability which necessitates that comparable data sets are employed in variational analyses (cf. Barron and Schneider 2009; Schneider 2010a: 252, 2012b: 364).That is, an analysis of requests among female middle-class speakers of British English should be contrasted with requests by female middleclass speakers in American English of the same age group. Thus, a focus on the influence of one particular independent variable (e.g. region) on the dependent variable (e.g. request) means that all other variables (e.g. social class, ethnicity, gender, age) need, as far as is possible, to be kept constant. The empirical data analysed in variational pragmatics ranges from naturally occurring data gathered using field methods through corpus data to experimental data (cf. also Schneider 2010a: 253–254). As with all research, it is recognised that there is no ‘best method, per se’ in variational pragmatics. Instead, the best method is that which best fits the particular research question at hand (cf. Jucker 2009; Schneider 2012a: 1034–1035). In the following, the strengths and weaknesses of the possible data sources are highlighted. Within variational pragmatics, naturally occurring data have been collected using a range of methods. Bieswanger (2015), in a study of responses to thanks across several regional varieties of English, used field notes in an innovative Labovian-style methodology involving asking strangers for directions (cf. also Rüegg 2014). This data type is ideal for examining what interactants actually do in reactive speech acts, such as response to thanks, and any problems of determining functional sameness would seem trivial. Control of the macro-social factors is somewhat more difficult, much being left to the judgement of the researchers themselves. It is also a time-consuming method and one which may be constrained by the limits of short-term memory. A further method of recording naturally occurring data is non-participative observation. Placencia (2008) employs this method in an analysis of requests in the context of service encounters. Such recordings provide an ideal source of contextualised data when comparability of contexts (e.g. size of shop, size of town in this example), often a challenge, is guaranteed. Besides comparability issues, a further problem of naturally occurring data is the potential difficulty of eliciting sufficient pragmatic features for quantitative analyses, particularly if the focus is on initiative speech acts, such as offers, or dispreferred speech acts, such as refusals. Electronic corpora are a further source of naturally occurring data which are employed, not exclusively, but particularly in analyses at the formal level given the ease of electronic searches. Comparative corpora are particularly suitable for variational pragmatic research. The 96

Variational pragmatics

International Corpus of English (ICE), encompassing several regional varieties of English, most of them freely available, is ideal, for instance, for analyses on the regional level (cf. e.g. Barron et al. 2015). In addition, corpora, such as the demographic component of the British National Corpus (BNC) and the Michigan Corpus of Academic Spoken English (MiCASE), allow analyses focusing on a range of further macro-social factors. Also, the larger corpora include a variety of discourse contexts and so provide detailed insights into the occurrences and contexts of the pragmatic feature under consideration (cf. Schauer and Adolphs 2006: 130–131). Apart from such and other options (cf. David Lee’s overview of corpora at www.uow.edu.au/~dlee/ CBLLinks.htm), a further possibility is the construction of small-scale corpora, as Clancy (2011) advocated. In general, small-scale corpora have the advantage of being an easily accessible source of data which is representative of a particular genre or variety. One current difficulty of corpus research is that electronic searches of corpora remain largely restricted to form due to a lack of pragmatic annotation (cf. Weisser, Chapter 5, this volume). Hence, analysts face the problems of precision and recall (cf. e.g. Jucker 2009). Difficulties of recall relate to the fact that relevant data may go unrecorded. In speech act analyses, for instance, non-conventionally indirect speech acts (e.g. hints, implicit compliments) are not retrievable via form-based searches. Also, less formulaic speech acts, such as realisations of disagreements, are difficult to search for. Difficulties of precision, on the other hand, relate to the fact that searches generate many more concordances than those of interest to the researcher. As such, close attention is demanded in identifying illocutionary force, particularly given the close relationships between speech acts, the lack of prosodic annotation in corpora4 and the fact that the background context, the identity of the speakers in corpora and their relationship towards each other are frequently unknown. Uptake, propositional content and background context (cf. also Terkourafi 2012) are the criteria recommended in establishing illocutionary sameness in such analyses of naturally occurring data (cf. also Variation and the pragmatic variable). In the following, we illustrate each of these criteria in turn using the case of offers and with reference to an analysis of offers in the British and Irish components of the ICE corpus (ICE-GB and ICE-IRE, respectively) (cf. Barron forthcoming). Uptake Hearer uptake is taken as evidence that the speaker’s (S) communicative intention is recognised by the hearer (H) (essential condition) (cf. Copestake and Terkourafi 2010). So, for instance, in (3), A’s response to B’s utterance Do you want a cup reveals that A recognises B’s utterance as an offer by B to get A a cup (commissive aspect of offers) and also as an utterance demanding a response (conditional aspect of offers). (3) B> Look I’m sorry about my political ignorance Do you want a cup cos it’s bigger A> No that’s fine A glass is fine (ICE-IRE: s1a-084) Uptake may be verbal or nonverbal. The latter can be recognised in corpus data if there is some evidence of nonverbal uptake in the data. Uptake is sometimes taken as an obligatory criterion, as Terkourafi (2012) recommended. In such cases, focus is on the pragmatic effect of a particular utterance on the dialogue rather than on a speaker’s intention exclusively. Under this procedure, infelicitous offers in which the illocutionary force is not recognised by H are not included in the analysis. In addition, it 97

Anne Barron

is necessary to be aware that such a convention has an influence on the speech act realisations analysed. In the case of offers, for instance, using uptake as an obligatory condition may mean that less forceful offers are not analysed. Davidson (1984: 103–104) notes that the presence of silence directly after an offer may be a signal of a(n) (upcoming) rejection. Given this potential/actual rejection, S may examine the initial formulation for any inadequacies that may be adversely affecting its acceptability, and then repeat or reformulate the offer in such a way as to deal with the possibility of rejection. Hence, offers which do not yield a verbal uptake may possibly be less forceful and less explicit than such subsequently issued initiative offers. Indeed, this would seem to be the case in (4). If uptake is taken as an obligatory criterion, then A’s utterance Tea in line 1 will not be identified as an offer realisation as there is no uptake following it. Later in the interaction, towards the end of (4), A makes a further attempt to issue the offer of tea – this time using the form Granny do you want tea. Uptake yeah yeah please follows. Thus, if uptake is taken as an obligatory criterion only this latter offer, Granny do you want tea, is coded. (4) A> Tea D> But sure if Daddy and Grandad could put it up It can’t be too hard to do C> several sylls A> What Oh when was that put up B> Last week A> You’ve two teabags in this cup on purpose D> What was put up there B> 4 sylls A> Granny do you want tea C> Yeah yeah please D> I’ll have a cup too thanks (ICE-IRE: s1a-067) Propositional content Propositional content is a further important criterion in identifying illocutionary force (cf. also Copestake and Terkourafi 2010). Offers, for instance, concern a future action A to be carried out by S which requires some effort on the part of S and which is assumed to be beneficial to H. It is clear in the utterance Granny do you want tea in (4) that the tea is potentially beneficial for H and that S will be responsible for providing the tea should H wish it. Further context Clues available in the co-text also aid in identifying illocutionary force. In the case of offers, references to addressee trouble, pre-offers or offer negotiations can signal the presence of an offer. The utterance Shall I lock him up in ICE-GB (s1a-052) is preceded by H’s utterance of the form Sorry I’m not a great lover of dogs, an utterance which informs the offerer that the future act A, to lock up the dog, would be welcome (cf. also Curl 2006; Sidnell 2009: 218; cf. also e.g. Jucker et al. 2008: 282–283 on the presence of references to addressee trouble). These examples show the pattern of identifying illocutionary identity in analyses of naturally occurring data. Here, uptake, propositional content and the context (co-text and background knowledge) of the utterance aid the analyst in establishing illocutionary force and in differentiating speech acts from each other. However, on occasion, as mentioned earlier, a lack of 98

Variational pragmatics

contextual data may make it impossible to establish the illocution at hand. An utterance, such as What do you fancy doing today then? (BNC) in line 5 of (5), may, for instance, be a request for information or an offer as, unlike (4), it is unclear from the context in (5) whether the act at hand is beneficial to the speaker or hearer and consequently whether the felicity conditions for a request or offer avail. Such unclear cases must then be omitted from the analysis. (5) SP:PS0E8: Go up and get the uhum Index catalogue and see if that’s got any more in it. SP:PS0EA: Any more? SP:PS0E8: (unclear)2. SP:PS0EA:  What they (unclear)2. SP:PS0E8: (unclear)2. (unclear). What do you fancy doing today then? SP:PS6ST: Uhum, I don’t know. Just a break (unclear)2. SP:PS0E8: I wonder whether we go (unclear)2. (unclear)2. SP:PS6ST: No. (unclear)2. SP:PS0E8: No, that’s what they want. Just get that in Tunbridge Wells. SP:PS6ST: (unclear)2. (BNC: KCD S_conv)

Apart from naturally occurring data, variational pragmatic research (particularly in the investigation of the actional level) frequently employs elicited data using production questionnaires or role plays which ensure comparable data and functional identity. Discourse completion tasks – also termed production questionnaires (Kasper 2000, 2008; cf. also Félix-Brasdefer and Hasler-Barker, Chapter 4, this volume) – demand informants put themselves in a particular situation and complete the ensuing dialogue. The situational description provides information concerning the micro-social factors of relevance and the choice of informants determines the macro-social factors controlled and/ or investigated. The dialogue itself may already be initiated. In the classic DCT form (cf. Kasper 2000 for variations), this is followed by a gap which informants are to complete and the dialogue generally closes with a hearer response, signaling uptake. However, even in the absence of a hearer response, situations are designed to elicit a single speech act whether via uptake or via implicit or explicit clues given in the situational description. In this way, identity of illocution is guaranteed. Similar clues – some more, some less explicit – are given to informants in a role-play situation to the same effect (cf. Barron 2003). The identification of the elicited speech act – and thus the establishment of illocutionary sameness – is, thus, unproblematic (cf. also Grainger and Harris 2007: 2–3). At the same time, it is possible to allow informants to opt out of a particular speech act (cf. Kasper 2000, 2008). Other advantages of such elicited data relate to the fact that a carefully designed and tested questionnaire or role play (ideally developed using ethnographic data) can give access to large quantities of the speech act or pragmatic feature of interest. Such elicited data is particularly valuable for analyses of speech acts, such as apologies, which are difficult to gather ethnographically (cf. Grainger and Harris 2007: 2–3). A further advantage is that the situational description given in both production questionnaires and role plays enables the researcher to control for or to vary individual factors as required (cf. Schauer and Adolphs 2006: 120; Jucker 2009: 1618). This facilitates focus on the effect of individual factors and combinations of factors. Furthermore, although the development of the instruments themselves is time-consuming, data elicitation is fast, unproblematic and the data are comparable across cultures – providing that assessment tests have been carried out in advance to ensure the comparability of the situational constellations (cf. Blum-Kulka and House 1989; Barron 2003). 99

Anne Barron

Compared to naturally occurring spoken data, the data elicited using instruments of elicitation has proven a valid representation of spoken data (cf. Barron 2003; Kasper 2008; Jucker 2009). However, differences in occurrences have also been recorded in the uses of particular strategies (cf. Kasper 2008: 293–294 for an overview). Also, turns are shorter and many features of interactional discourse, such as laughter, pauses, hesitations and repetitions, are lacking (cf. Schauer and Adolphs 2006: 130). In addition, the contexts of occurrence of the pragmatic feature under investigation are fewer as the number of situations investigated is limited. Having said this, it should be noted that although such elicited data approach naturally occurring spoken data, the data elicited are clearly off-line data (cf. Kasper 2008). As such, it is important to be aware that these instruments elicit stereotypical interactions in the mind of the respondents and so portray the socially accepted use of language in a particular culture (cf. Barron 2005b). As Schneider, writing in the context of variational pragmatics, claims, such data are ideal for identifying ‘what counts as appropriate verbal behavior in a given situation’. He points out that: If the question is not to examine what interactants actually do which may be appropriate or inappropriate, but what they believe one should do, which is considered appropriate, then the best method is an experimental method. . . . .Written production questionnaires. . . . are particularly suitable as they provide informants with time to analyse the situation described in the instructions and thus reduce the accidentialities of spontaneous speech and interaction. Production questionnaires elicit ‘the prototype of the variants occurring in the individual’s actual speech’ (Hill et al., 1986: 353). (Schneider 2012a: 1034) Schneider (2012a) goes on to claim that these behavioural prototypes elicited using production questionnaires are similar to Watts’ (2003: 256–257) concept of ‘first order politeness’ or ‘politic behavior’, that is, appropriate expected behaviour in interaction (cf. also Schneider 2012b and Schneider 2010b: 85 on behavioural scripts). In other words, experimental data allow us to concentrate on what interactants perceive as polite norms. In addition, as Schneider states, with elicited data ‘it is possible to empirically establish different kinds of pragmatic variables and their respective variants’ (2014: 370). In sum then, as mentioned earlier, each method has its strengths and weaknesses. The choice of data thus depends on which data type suits the research question best.

Conclusion Previous research has suggested a number of areas in which further study would be particularly interesting (cf. Barron and Schneider 2009; Schneider 2010a; Barron 2014). Such writings highlight a need for research on a regional level to focus in particular on languages other than English and Spanish. Also, pragmatic variation according to ethnic identity, age and socioeconomic class continue to remain largely desiderata, as also is research relating to the interplay of these and other factors. As regards the levels of analysis, the formal and actional levels of pragmatic analysis have enjoyed most research interest. The remaining levels remain ripe for further research. Further research desiderata relates to the fact that the majority of studies in variational pragmatics to date have focused on everyday conversation.Written genres and variation of these according to macro-social factors has been largely disregarded (cf. also Barron 2012: 287–288) – possibly also due to the absence of this level from the variational pragmatic framework. However, the realisation of these purposes may vary – and often does vary – across cultures (cf. Günthner 2007; Barron 2012). Yajun and Chenggang (2006) is an interesting article in this regard. They address the 100

Variational pragmatics

question as to how contrastive genre analysis can be successfully combined with the study of World Englishes (regional variation in the present context) given that ‘the involvement of discoursal and rhetorical analyses are exceptions rather than regular practice’ in the study of World Englishes. This question is, thus, also a question for variational pragmatics particularly since such discoursal analyses on the level of genre are the exception not only for regional variation, but also for ethnic, gender, age and socioeconomic variation. The variational pragmatic framework needs thus to be extended to include a level of analysis akin to Jucker and Taavitsainen’s discourse domain level, defined as ‘the entire repertoire of texts and genres at the disposal of a discourse community’ (2012: 302). The tertium comparationis, that is, the ‘common platform of reference’ (Krzeszowski 1989: 60, 1990: 15), in such analyses would be the genre itself (cf. Lüger 2005: 170–171), operationalised in practice as similarity of communicative function, supplemented with a comparable context of use (cf. Barron 2012: 10–16, 26–28 for a synthesis on the process of establishing [identity of] communicative purpose). With identity of communicative function as the variable, variational pragmatic research might investigate whether and, if so, to what extent the macro-social factors influence intralingual genre conventions. Investigations could focus on the level of move structure, that is, on the global organisational patterns through which a particular overriding communicative purpose is realised, and also on move register, that is, on how these moves, that is, organisational units, are realised in language. We look forward to further developments in this area.

Suggestions for further reading Barron, A. (ed.) (2015) ‘A variational pragmatic approach to regional variation in language: Celebrating the work of Klaus P. Schneider’ [Special issue], Multilingua, 34(4). This special issue is a celebratory volume to the work of one of the founders of variational pragmatics, Klaus P. Schneider.The volume further develops variational pragmatics as a field of research and focuses on the influence of region on language use. The levels of analysis investigated are broad, including the formal and actional levels, but also the less frequently analysed interactional and topic levels. Barron, A. and Schneider, K.P. (eds.) (2009) ‘Variational pragmatics’ [Special issue], Intercultural Pragmatics, 4(6). This special issue on variational pragmatics provides an introduction to the field of variational pragmatics and illustrates the types of questions and issues addressed in the area. The empirical analyses in the volume address pragmatic variation across region in the first instance. The levels of pragmatic analysis included encompass the formal, actional levels and interactional levels. Jautz, S. (2013) Thanking Formulae in English. Explorations across Varieties and Genres (Pragmatics & Beyond New Series 230). Amsterdam/Philadelphia: Benjamins. This monograph contrasts the forms and functions of thanking formulae in New Zealand English and British English. The empirical analysis focuses on the 1 million-word Wellington Spoken Corpus of New Zealand English and a selection of comparable texts taken from the British National Corpus. Márquez-Reiter, R. and Placencia, M. E. (2005) Spanish Pragmatics. Basingstoke: Palgrave Macmillan. This is a textbook on pragmatics focusing on Spanish. It is innovative in placing particular emphasis on the pluricentrality of the Spanish language, with many examples included from across the varieties of Spanish. It also features a valuable chapter on sociopragmatic variation. Schneider, K. P. and Barron, A. (eds.) (2008) Variational Pragmatics: A Focus on Regional Varieties in Pluricentric Languages (Pragmatics & Beyond New Series 178). Amsterdam/Philadelphia: Benjamins. This seminal volume established variational pragmatics as a field of research. The introductory chapter establishes the rationale for studying variational pragmatics as a separate field of inquiry, systematically sketches the broader approach and presents a framework for further analysis. The papers which follow present empirical variational pragmatic research focusing on regional varieties of pluricentric languages.

Notes 1 I would like to thank Irina Pandarova for helpful comments on an earlier draft. Thanks also to Kerstin Single for support with the preparation of the manuscript. The usual disclaimers apply.

101

Anne Barron 2 The bibliography on variational pragmatics hosted by the University of Bonn (www.linguistics.­unibonn.de/ research/variational-pragmatics/variational-pragmatics-bibliography/) and the bibliography on regional pragmatic variation hosted at the University of London (www.bbk.ac.uk/spanish/our-staff/mariaelena placencia/bibliography-on-regional-pragmatic-variation) provide an overview. 3 Cf. also Félix-Brasdefer (2014) on the role of the felicity conditions and on the role of the hearer’s interpretation of a particular utterance (securing uptake, inviting response) in the pragmatic interpretation of a particular utterance. 4 Note that the general availability of audio files appears to be increasing in recent years.The British component of the International Corpus of English (ICE-GB) is now available aligned with 300 audio recordings. Similarly, sound files are also available for the Santa Barbara Corpus of Spoken American English.

References Aijmer, K. and Andersen, G. (2011) ‘Introducing the pragmatics of society’, in G. Andersen and K. Aijmer (eds.) The Pragmatics of Society. Berlin/New York: Mouton. 1–28. Barron, A. (2003) Acquisition in Interlanguage Pragmatics: Learning How to Do Things with Words in a Study Abroad Context. Amsterdam/Philadelphia: Benjamins. Barron, A. (2005a) ‘Variational pragmatics in the foreign language classroom’, System, 33(3): 519–536. Barron, A. (2005b) ‘Offering in Ireland and England’, in A. Barron and K. P. Schneider (eds.) The Pragmatics of Irish English. Berlin: Mouton. 141–177. Barron, A. (2011) ‘Variation revisited: A corpus analysis of offers in Irish English and British English’, in J. Frenk and L. Steveker (eds.) Anglistentag 2010 Saarbrücken: Proceedings: [Proceedings XXXII]. Trier: Wissenschaftlicher Verlag Trier. 407–419. Barron, A. (2012) Public Information Messages: A Contrastive Genre Analysis of State-Citizen Communication. Amsterdam/Philadelphia: Benjamins. Barron, A. (2014) ‘Variational pragmatics’, in C. A. Chapelle (ed.) The Encyclopedia of Applied Linguistics (EAL: electronic version). Oxford: Wiley-Blackwell. Available from: www.encyclopediaofappliedlinguistics.com (accessed 15 September 2015). Barron, A. (2015) ‘Explorations in regional variation: A variational pragmatic perspective’, Multilingua, 34(4): 449–459. Barron, A. (forthcoming) ‘The speech act of ‘offers’ in Irish English’, World Englishes. Barron, A., Pandarova, I. and K. Muderack (2015) ‘Tag questions across Irish English and British English: A corpus analysis of form and function’, Multilingua, 34(4): 495–525. Barron, A. and Schneider, K. P. (2009) ‘Variational pragmatics: Studying the impact of social factors on language use in interaction’, Intercultural Pragmatics, 6(4): 425–442. Bibliography on regional pragmatic variation at the University of London. Available from: www.bbk.ac.uk/ languages/our-staff/maria-elena-placencia/bibliography-on-regional-pragmatic-variation (accessed 15 September 2015). Bibliography on variational pragmatics at the University of Bonn. Available from: www.linguistics.unibonn.de/research/variational-pragmatics/variational-pragmatics-bibliography/ (accessed 15 September 2015). Bieswanger, M. (2015) ‘Variational pragmatics and responding to thanks – revisited’, Multilingua, 34(4): 527–546. Blum-Kulka, S. and House, J. (1989) ‘Cross-cultural and situational variation in requesting behavior’, in S. Blum-Kulka, J. House and G. Kasper (eds.) Cross-cultural Pragmatics: Requests and Apologies (Advances in Discourse Processes, vol. 31). Norwood, NJ: Ablex. 123‑154. Blum-Kulka, S., House, J. and Kasper, G. (1989) ‘The CCSARP coding manual’, in S. Blum-Kulka, J. House and G. Kasper (eds.) Cross-cultural Pragmatics: Requests and Apologies (Advances in Discourse Processes, vol. 31). Norwood, NJ: Ablex. 273–294. British National Corpus (BNC) (Lancaster BNCweb Server). Available from: http://bncweb.lancs.ac.uk/ bncwebSignup/user/login.php (accessed 15 September 2015). Cameron, R. and Schwenter, S. (2013).‘Pragmatics and variationist sociolinguistics’, in R. Bayley, R. Cameron and C. Lucas (eds.) The Oxford Handbook of Sociolinguistics. Oxford: Oxford University Press. 464–483. Cheshire, J. (2002) ‘Sex and gender in variationist research’, in J. K. Chambers, P. Trudgill and N. SchillingEstes (eds.) Handbook of Language Variation and Change. Oxford: Blackwell. 423–443. Clancy, B. (2011) ‘Complementary perspectives on hedging behaviour in family discourse: The analytical survey of variational pragmatics and corpus linguistics’, International Journal of Corpus Linguistics, 16(3): 371–390.

102

Variational pragmatics Clark, H. H. (1979) ‘Responding to indirect speech acts’, Cognitive Psychology, 11: 430–477. Clyne, M., Fernandez, S. and Muhr, R. (2003) ‘Communicative styles in a contact situation: Two German national varieties in a third country’, Journal of Germanic Linguistics, 15(2): 95–154. Copestake, A. and Terkourafi, M. (2010) ‘Conventionalized speech act formulae: From corpus findings to formalization’, in P. Kühnlein, A. Benz and C. L. Sidner (eds.) Constraints in Discourse (vol. 2). Amsterdam/Philadelphia: John Benjamins. 125–140. Corpus of Contemporary American English (COCA). Available from: http://corpus.byu.edu/coca/ (accessed 15 September 2015). Coupland, N. (1983) ‘Patterns of encounter management: Further arguments for discourse variables’, Language in Society, 12: 459–476. Curl, T. S. (2006) ‘Offers of assistance: Constraints on syntactic design’, Journal of Pragmatics, 38: 1257–1280. Davidson, J. A. (1984) ‘Subsequent versions of invitations, offers, requests, and proposals dealing with potential or actual rejections’, in J. M. Atkinson and J. Heritage (eds.) Structures of Social Action: Studies in Conversational Analysis. Cambridge: Cambridge University Press. 102–128. Dines, E. (1980) ‘Variation in discourse – “and stuff like that” ’, Language in Society, 9: 13–31. Félix-Brasdefer, C. (2014) ‘Speech act sequences’, in K. P. Schneider and A. Barron (eds.) Pragmatics of Discourse. Berlin/New York: Mouton. 323–353. Flöck, I. (2015) ‘Requests in American and British English: A contrastive multi-method analysis’, unpublished doctoral thesis, University of Oldenburg. Foolen, A. (2011) ‘Pragmatic markers in a sociopragmatic perspective’, in G. Andersen and K. Aijmer (eds.) Pragmatics of Society. Berlin/New York: Mouton. 217–242. Grainger, K. and Harris, S. (2007) ‘Special issue: Apologies: Introduction’, Journal of Politeness Research, 3(1): 1–9. Günthner, S. (2007) ‘Intercultural communication and the relevance of cultural specific repertoires of communicative genres’, in H. Kotthoff and H. Spencer-Oatey (eds.) Handbook of Intercultural Communication. Berlin/New York: Mouton. 127–151. Haugh, M. and D. Carbaugh (2015) ‘Self-disclosure in initial interactions amongst speakers of American and Australian English’, Multilingua, 34(4): 461–493. Hill, B., Ide, S., Ikuta, S., Kawasaki, A. and Ogino, T. (1986) ‘Universals of linguistic politeness’, Journal of Pragmatics, 10: 347–371. International Corpus of English (ICE). Available from: http://ice-corpora.net/ice/avail.htm (accessed 15 September 2015). International Corpus of English. British component (ICE-GB). Available from: www.ucl.ac.uk/englishusage/projects/ice-gb/ (accessed 15 September 2015). Jucker, A. H. (2008) ‘Historical pragmatics’, Language and Linguistics Compass, 2(5): 894–906. Jucker, A. H. (2009) ‘Speech act research between armchair, field and laboratory: The case of compliments’, Journal of Pragmatics, 41(8): 1611–1635. Jucker, A. H., Schneider, G.,Taavitsainen, I. and Breustedt, B. (2008) ‘Fishing for compliments: Precision and recall in corpus-linguistic compliment research’, in A. Jucker and I. Taavitsainen (eds.) Speech Acts in the History of English. Amsterdam/Philadelphia: John Benjamins. 273–294. Jucker, A. H. and Taavitsainen, I. (2012) ‘Pragmatic variables’, in J. M. Hernández-Campoy and J. C. CondeSilvestre (eds.) The Handbook of Historical Sociolinguistics. Oxford: Wiley-Blackwell. 293–306. Kasper, G. (2000) ‘Data collection in pragmatics research’, in H. Spencer-Oatey (ed.) Culturally Speaking. Managing Rapport through Talk across Cultures. London/New York: Continuum. 317–341. Kasper, G. (2008) ‘Data collection in pragmatics research’, in H. Spencer-Oatey (ed.) Culturally Speaking. Culture, Communication and Politeness Theory. London: Continuum. 279–303. Krzeszowski,T. P. (1989) ‘Towards a typology of contrastive studies’, in O.Wieslaw (ed.) Contrastive Pragmatics. Amsterdam/Philadelphia: Benjamins. 55–72. Krzeszowski,T. P. (1990) Contrasting Languages.The Scope of Contrastive Linguistics. Berlin/New York: Mouton de Gruyter. Labov, W. (1966) ‘The linguistic variable as a structural unit’, Washington Linguistics Review, 3: 4–22. Lavandera, B. R. (1978) ‘Where does the sociolinguistic variable stop?’, Language in Society, 7: 171–182. Lee, D. (2001) Bookmarks for Corpus-Based Linguists. Available from: www.uow.edu.au/~dlee/CBLLinks. htm (accessed 15 September 2015). Lüger, H.-H. (2005) ‘Kontrastive Text(sorten)analyse: Voraussetzungen, Probleme, Perspektiven‘, in M. Orosz and T. Albrecht (eds) Jahrbuch der ungarischen Germanistik 2004. Budapest/Bonn: DAAD. 169–184.

103

Anne Barron McCarthy, M. (2002) ‘Good listenership made plain: Non-minimal response token in British and American spoken English’, in D. Biber, S. M. Fitzmaurice and R. Reppen (eds.) Using Corpora to Explore Linguistic Variation. Amsterdam: Benjamins. 49–72. Michigan Corpus of Academic Spoken English. Available from: http://quod.lib.umich.edu/m/micase/ (accessed 15 September 2015). Pichler, H. (2009) ‘The functional and social reality of discourse variants in a northern English dialect: I DON’T KNOW and I DON’T THINK compared’, Intercultural Pragmatics, 6(4): 561–596. Pichler, H. (2010) ‘Methods in discourse variation analysis: Reflections on the way forward’, Journal of Sociolinguistics, 14(5): 581–608. Placencia, M. E. (2008) ‘Requests in corner shop transactions in Ecuadorian Andean and Coastal Spanish’, in K. P. Schneider and A. Barron (eds.) Variational Pragmatics: A Focus on Regional Varieties in Pluricentric Languages. Amsterdam/Philadelphia: Benjamins. 307–332. Placencia, M. E. (2011) ‘Regional pragmatic variation’, in G. Andersen and K. Aijmer (eds.) Pragmatics of Society (Handbooks of Pragmatics 5). Berlin: Mouton. 79–113. Rüegg, L. (2014) ‘Thanks responses in three socio-economic settings: A variational pragmatics approach’, Journal of Pragmatics, 71: 17–30. Santa Barbara Corpus of Spoken American English. Available from: www.linguistics.ucsb.edu/research/ santa-barbara-corpus (accessed 15 September 2015). Schauer, G. A. and Adolphs, S. (2006) ‘Expressions of gratitude in corpus and DCT data:Vocabulary, formulaic sequences, and pedagogy’, System, 34: 119–134. Schlieben-Lange, B. and Weydt H. (1978) ‘Für eine Pragmatisierung der Dialektologie’, Zeitschrift für Germanistische Linguistik, 6(3): 257–282 [pp. 257–260 are given twice in this volume, each time with different content]. Schneider, K. P. (2008) ‘Small talk in England, Ireland and the U.S.A.’, in K. P. Schneider and A. Barron (eds.) Variational Pragmatics. A Focus on Regional Varieties in Pluricentric Languages. Amsterdam: Benjamins. 99–139. Schneider, K. P. (2010a) ‘Variational pragmatics’, in M. Fried, J.-O. Östman and J.Verschueren (eds.) Variation and Change: Pragmatic Perspectives (Handbook of Pragmatics Highlights 6). Amsterdam/Philadelphia: Benjamins. 239–267. Schneider, K. P. (2010b) ‘Small talk: Units, sequencing, realizations’, in J. Helbig and R. Schallegger (eds.) Anglistentag 2009 Klagenfurt: Proceedings. Trier: Wissenschaftlicher Verlag Trier. 79–90. Schneider, K. P. (2012a) ‘Appropriate behavior across varieties of English’, Journal of Pragmatics, 44(9): 1022–1037. Schneider, K. P. (2012b) ‘Pragmatic variation and cultural models’, Review of Cognitive Linguistics, 10(2): 346–372. Schneider, K. P. (2014) ‘Comparability and sameness in variational pragmatics’, in S. Mergenthal and R. M. Nischik (eds.) Anglistentag 2013 Konstanz: Proceedings. Trier: Wissenschaftlicher Verlag Trier. 361–372. Schneider, K. P. and Barron, A. (2008) ‘Where pragmatics and dialectology meet: Introducing variational pragmatics’, in K. P. Schneider and A. Barron (eds.) Variational Pragmatics: A Focus on Regional Varieties in Pluricentric Languages. Amsterdam: Benjamins. 1–32. Sidnell, J. (2009) ‘Sequences’, in S. D’hondt, J.-O. Östman and J.Verschueren (eds.) The Pragmatics of Interaction. Amsterdam/Philadelphia: John Benjamins. 215–239. Terkourafi, M. (2011) ‘The pragmatic variable: Toward a procedural interpretation’, Language in Society, 40: 343–372. Terkourafi, M. (2012) ‘Between pragmatics and sociolinguistics: Where does pragmatic variation fit in?’, in J. C. Félix-Brasdefer and K. Dale (eds.) Pragmatic Variation in First and Second Language Contexts: Methodological Issues. Amsterdam/Philadelphia: Benjamins. 295–318. Tottie, G. (1991) ‘Conversational style in British and American English: The case of backchannels’, in K. Aijmer and B. Altenberg (eds.) English Corpus Linguistics. London: Longman. 254–271. Watts, R. (2003) Politeness. Cambridge: Cambridge University Press. Wolfram, W. and Schilling-Estes, N. (2006) American English. Dialects and Variation, 2nd ed., Malden, MA: Blackwell. Yajun, J. and Chenggang, Z. (2006) ‘World Englishes and contrastive rhetoric’, English Today, 2(2): 11–22.

104

11 POSTCOLONIAL PRAGMATICS Eric A. Anchimbe and Richard W. Janney

Introduction I talk country talk with my mother. I talk Pidgin and country talk with my sister and brothers. I talk French when I play with my friends. I talk English and Pidgin in school. (Tanyi 1978: 10)

This statement by an Anglophone child in Yaounde, Cameroon, thirty-eight years ago typifies a situation familiar to language users in many of today’s multiethnic, multicultural, and multilingual postcolonial societies. Each sentence describes the young speaker using a different language with a different partner or partners, in a different situation, in a different role relationship, and with a slightly different identity. The italicised words – ‘language’, ‘partner’, ‘situation’, ‘role relationship’, and ‘identity’ – are baseline dimensions in any type of interaction. In postcolonial contexts, however, each of these dimensions is itself rendered potentially multidimensional by differences in partners’ ethnic roots, regional affiliations, tribal ties, group alignments, religious attachments, and linguistic backgrounds and abilities. Postcolonial societies are characterised by extensive multilingualism and multiethnicity. It is said that on a daily basis, a typical postcolonial urbanite speaks one or more indigenous tribal languages (home languages), a mixed lingua franca (pidgin, creole), and one or more varieties of a former colonial European language (e.g. English, French, Portuguese, Spanish) (Simo Bobda and Mbouya 2005: 2122). Postcolonial speakers acquire proficiency in many languages before reaching adulthood: the oral languages spoken in their villages and surrounding region, the official national European languages taught in school, the pidgins spoken everywhere on the street.Through interaction with partners from different ethnic, social, and occupational backgrounds, they often have a working knowledge of additional languages as well, and a keen awareness of the kinds of people who speak them, the situations in which speaking them is appropriate, and the different purposes for which they can be used in social interaction.The routine use of multiple languages from competing cultures – what Anchimbe (2007) terms linguabridity (see later in this chapter) – is a common feature of communication in many of the world’s modern postcolonial nation-states. While multilingualism on this scale is unusual in most Western contexts, it is a normal fact of life and a natural consequence of mobility and migration in the complex multiethnic societies 105

Eric A. Anchimbe and Richard W. Janney

studied in postcolonial pragmatics (cf. also Cenoz, Chapter 13, this volume). Postcolonial pragmatics is a relatively recent arrival on the pragmatic scene. It takes intermixed languages and communicative practices as its point of departure and studies different forms, functions, and effects of multilingualism and interlingual interaction in non-Western postcolonial contexts. Its goal is to approach postcolonial discourse as a process rooted in the lives of speakers whose communication strategies have been shaped by the heterogeneous postcolonial environments in which they interact on a daily basis. Postcolonial speakers’ identities, attitudes, social affiliations, communicative needs, and interactional strategies are all influenced by the prevailing hybridities of postcolonial life. Postcolonial pragmatics uses discourse as a key to exploring complex forms of identity construction, role projection, group maintenance, gate-keeping, and social empowerment and marginalisation in postcolonial societies. Such features and functions of postcolonial interaction are routinely regulated by choices of languages and interlingual discourse strategies. Postcolonial pragmatics provides a framework for investigating these strategies and gaining linguistic insight into the impact of ethnic, cultural, and lingual diversity on postcolonial interaction. It addresses all manifestations of postcolonial discourse, including face-to-face conversation, print journalism, public broadcasting, audiovisual media, and Internet discourse. Moreover, it approaches these different types of postcolonial discourses on their own terms, prioritising the need to describe their embeddedness in the cultures and ecologies that produce them, free of the constraints of monolingual Western pragmatic frameworks. Postcolonial pragmatics aligns itself with the interests of scholars who have become concerned in recent years about the limitations of individualistic Western pragmatic theories in accounting for collectivistic non-Western pragmatic practices (see later in this chapter). As calls for culturally unbiased pragmatic frameworks have become louder, two movements have emerged dedicated to emancipating research on non-Western pragmatics from the ethnocentric constraints of Western pragmatic theory. Both encourage the development of research perspectives capable of providing emically adequate explanations of non-Western pragmatic practices. Emancipatory pragmatics, called into being a few years ago by Hanks and colleagues (2009), focuses mainly on interaction in relatively homogeneous, historically stable, Asian languages and cultures. Postcolonial pragmatics, introduced at about the same time by Janney (2006, 2009a, 2009b), Anchimbe (2006, 2007, 2011a), and Anchimbe and Janney (2011a, 2011b), investigates interaction in the heterogeneous new postcolonial nation-states that came into existence with the dismantlement of British, French, Portuguese, Dutch, and Belgian colonial empires after World War II. This chapter is structured as follows: after the introduction, Interaction in postcolonial contexts offers an overview of three important linguistic features of interaction in postcolonial societies: multilingualism, code-switching, and language identities. It describes the central roles played by sociocultural expectations (ethnic, ingroup, community), identity projections, and relations between language, society, and identity in postcolonial interaction. Western pragmatics and non-Western pragmatic practices discusses the theoretical challenge of explaining complex non-Western discourse strategies, given the unsuitability of traditional monolingual mainstream Western pragmatic models. Difficulties of applying individualistic, monolinguistic notions of speech acts and principles of politeness to the study of collectivistic multilinguistic discourse are discussed. The section concludes that postcolonial pragmatics requires approaches suited to analysing Non-Western discourse on its own terms. Current issues in postcolonial pragmatics presents a survey of current areas of research in postcolonial pragmatics, illustrating the scope of present studies of inter- and intralingual role relationships (professional, occupational, kinship, gender, age, religion) and interlingual identity constructions (multiple identities, identity opportunism, linguistic victimisation, social affiliation and exclusion, ingroup youth vernaculars). The conclusion calls for more emic descriptions of postcolonial discourse. 106

Postcolonial pragmatics

Interaction in postcolonial contexts Colonisation profoundly affected nearly all indigenous cultures that it touched. Formerly autonomous regional ethnic groups were split apart by arbitrary boundaries, or artificially forced together into large, heterogeneous dependencies with highly multiethnic, multicultural, multilingual populations. Exploitation of resources and the emergence of colonial urban centres triggered accelerated migration and ethnic mixing. Control over these mixed populations was maintained in part linguistically by monolingual rule. The introduction of foreign official written languages and the teaching of these to members of privileged groups disrupted traditional balances of ethnic, cultural, and linguistic influence. Complex forms of ethnic and linguistic rivalry arose, as well as complex patterns of interlingual communication and competition. New commercial, political, civil, and educational social classes emerged in competition with older tribal classes. European religions competed with indigenous religions for the hearts and minds of the people. The outcome of these developments was a complex, hybrid social, political, and linguistic system whose members were forced to acquire equally complex, hybrid adaptive skills in order to successfully cope with its hybrid composition. Languages, language identities, and interlingual group affiliations figured importantly in all of these developments and continue to play an important role in many of today’s modern postcolonial societies. For these to be properly described, a framework that takes into account the complex hybrid structure of these societies is necessary – postcolonial pragmatics is designed to play this role.

Multilingualism Unlike the relatively homogeneous speech communities usually studied in pragmatics, postcolonial societies are heterogeneous mixtures of indigenous languages, hybrid lingua francas, and foreign official languages. Some populations are so linguistically varied that the concept of the speech community itself may be difficult to apply to them. Moreover, migration constantly fuels and remixes their linguistic diversity. Especially in postcolonial urban centres, speakers often use different languages for different interactional purposes. In such an urban scenario, the oral indigenous languages are used as languages of family and tribal solidarity, the mixed languages (pidgins, creoles) serve as pan-ethnic lingua francas, and the written colonial languages function as the official languages of education, the judiciary, commerce, civil control, and national identity. Coping with the different functions and statuses of these languages in postcolonial interaction requires complex linguistic skills. Speakers must interact flexibly with partners from vastly different cultural and linguistic backgrounds, with different levels of interlingual and multilingual competence and different degrees of motivation to use the languages they can speak.There may be languages in the community that they understand but cannot speak, languages they can speak but prefer not to, and languages they do not understand at all. Discourse strategies in these different languages may differ as well, as may their native speakers’ cultural assumptions, attitudes, and characteristic modes of thought. Moreover, strategies from one language may be carried over into the discourse of others.

Code-switching and code-mixing Situational choices of home languages, pidgins, creoles, and official languages have complex pragmatic functions in postcolonial life. In order to be a fully competent member of a modern postcolonial urban society, an individual must be fluent in all of the linguistic domains discussed earlier and must be adept at switching, mixing, and alternating languages strategically in different situations. In many cases, code-switches (switches between languages) operate almost like 107

Eric A. Anchimbe and Richard W. Janney

indirect speech acts and are used to define or redefine speech situations and regulate role relationships. Among Africans living abroad, for example, switching from the national official language (e.g. French, English) to a pidgin can be a means of checking a fellow African’s nationality or regional origin. In a postcolonial urban setting, using a local indigenous language can induce a context of trust with someone from the same tribe or region. In a multiethnic neighbourhood, interjecting a few words of a neighbour’s tribal language into one’s utterances (i.e. code-mixing) can signal friendship or solidarity. On the street, switching to a pidgin can be used to include everyone in the conversation, signal informality, show affiliation, ease tension, shift interactional roles, or make jokes. The overt prestige of the official languages can be used to mark one’s education, superiority, official speaking role, national identity, and so forth. Thus, code-switching and code-mixing are important conversational tools enabling individuals in postcolonial societies to regulate their interpersonal affairs. They are instrumental in marking group alignments, performing social roles, and projecting different situational identities.

Language identities The linguistic negotiation of identities and alignments is a key issue in postcolonial life, where decades of migration have created large populations of individuals who, to different extents, have become uprooted from their homelands, native traditions, and ethnic, tribal, and family identities. To compete successfully in modern urban life, postcolonial speakers are forced to forge flexible new identities as members of mixed social and occupational groups that are often less stable than those of traditional tribal life. Here, choices of languages figure importantly. Choosing to speak a home language, for example, evokes all of the roles, rights, and responsibilities associated with it. These are typically constrained by cultural and tribal values, matrixes of relations to people in one’s family, rules governing respect and proper behaviour, and responsibilities for others’ well-being. A home language identity is one’s most personal language identity. Being regionally restricted, however, it discourages contact with those beyond one’s ethnic inner circle. Speaking an official language, on the other hand, projects an identity associated with established authority and stakes a claim to rights and esteem befitting that authority. An official language identity is a broad, pan-regional identity that may appeal to individuals beyond a person’s ethnic or regional sphere but also excludes the less fluent, uneducated members of the population. Speaking a pidgin or creole, finally, projects the most inclusive and neutral of all possible language identities. While pidgin invites interaction across ethnic and social boundaries, it can also be used defensively to avoid identification with speakers of stigmatised indigenous or official languages (cf.Van De Mieroop, Chapter 16, this volume). Projections of language identities are thus also tools of empowerment or marginalisation that can be used to acknowledge or challenge others’ identities and to create, protect, or conceal one’s own identity. This opens the way for complex forms of situational identity opportunism and linguistic victimisation (Anchimbe 2013: 149–168). Identity opportunism, or speaking a language associated with a group identity felt by the speaker to be of potential immediate advantage, is often done to benefit from the group identified with, to avoid being excluded by it, or to dispel distrust among its members. Linguistic victimisation, a form of social discrimination related to linguism, can take many forms: among these, refusing to use others’ languages although these may be the only ones they understand, obliging others to use one’s own language although they may not be able to, claiming not to understand others’ languages, stigmatising others’ languages, penalising partners for using others’ languages, or banning others’ languages. 108

Postcolonial pragmatics

Western pragmatics and non-Western pragmatic practices Postcolonial interaction has historically been a neglected area of pragmatics study. This was perhaps initially due to an unconscious assumption among pragmaticians that this area of research somehow fits better into the research agendas of sociolinguistics or of linguistic anthropology. An even more basic problem in pragmatic theory itself, however, was the lack of adequate frameworks for analysing multilingual discourse in non-Western cultures. Prior to the first pioneering studies of Asian concepts of politeness in the late 1980s (Matsumoto 1988; Ide 1989; Gu 1990; Mao 1994), little pragmatic work had appeared on discourse in collectivistic non-Western cultures. Cross-cultural pragmatics was still in its infancy (Blum-Kulka et al. 1989; Cogo and House, Chapter 15, this volume), and modern notions of variational pragmatics (Schneider and Barron 2008; Barron, Chapter 10, this volume) and interlanguage pragmatics (Barron 2012; Taguchi, Chapter 14, this volume) were not yet even on the horizon. Research at this time was guided by assumptions about speakers and interaction adopted mainly from monolingual pragmatic models. Theories posited idealised native speakers living in homogeneous Western speech communities, with individualistic Western needs and rational strategies for satisfying these in everyday discourse (Brown and Levinson 1987: 58). For all their plausibility in Western contexts, however, standard pragmatic approaches actually tended to discourage studies of postcolonial interaction more than to encourage them, and to some extent, these same approaches continue to block research today. Austin’s (1962) and Searle’s (1969) work on speech acts rested strictly on intuitions about English utterances. Grice’s (1975) theory of implicature ignored conversational inferences rooted in collectivistic cooperative principles or non-Western conversational maxims. Brown and Levinson’s (1987) account of linguistic politeness provided no convincing means of explaining concepts of face and politeness in many collectivistic non-Western cultures. Goffman’s (1967, 1969) notions of floor holding, floor taking, and floor giving provided little help in explaining interaction in cultures where ownership of the floor is based on tradition, age, or respect rather than negotiation. Nevertheless, in time, the steady output of Western scholarship culminated in a widely accepted, standard body of mainstream pragmatic theory and research driven nearly entirely by Western assumptions that their proponents believed constituted ‘universals of language usage’ (cf. Janney and Arndt 1993). The ad hoc transformation of the West’s emic research perspectives into the prescribed etic standards for the rest of the world made findings about non-Western interaction that were inconsistent with Western views of communication difficult to integrate into standard theories. A situation developed in which findings about postcolonial discourse appeared to be interpretable only as exceptions to or deviations from the mainstream’s monolingual, monocultural models. For many scholars studying postcolonial interaction, it seemed that investigating the consequences of multilingualism and interlinguality – central features of communication in many non-Western societies and key influences on postcolonial interaction – was nearly beyond the theoretical scope of the discipline merely because it contradicted the dominant Western models. To appreciate the dilemma, one need only imagine the situation of a Western researcher forced by mainstream practice to analyse interaction in an American family within a theoretical framework developed to account for Japanese interaction. For scholars working in or from postcolonial contexts, the echoes of Western cultural hegemony were unmistakable. In spite of this, even today, work on postcolonial discourse continues to be influenced to a certain extent by models and assumptions stemming from research on Western interaction. Two examples of this are studies of postcolonial speech acts and politeness strategies. 109

Eric A. Anchimbe and Richard W. Janney

Speech acts Much research on postcolonial speech acts thus far has been produced in the spirit of BlumKulka and her colleagues’ Cross-cultural Speech Act Realisation Project (CCSARP), which began as a comparative study of realisations of ‘requesting’ and ‘apologising’ across Western languages (Blum-Kulka and Olshtain 1984; Blum-Kulka et al. 1989). In the original project, the assumption that contexts of speech act realisation and functions of speech acts remain constant across languages and cultures served as a basis for contrastive studies of their different realisations in German, French, Danish, Hebrew, Russian, and national varieties of English. Using discourse completion tests, researchers collected data on utterances used to realise apologies and requests in predefined speaking contexts (the study’s ‘givens’), by native speakers of different languages (the ‘variables’).While this enabled them to categorise and correlate speech act realisations across the languages studied, it yielded no useful findings about the functions of speech act realisations in real-life contexts. Research on postcolonial speech act realisation has long followed a similar pattern. Token speech act realisations in indigenous languages or in indigenised varieties of English have been compared with native English speech acts. But here, researchers have often had difficulty integrating non-Western speech act realisations and their presumed Western correlates together into unified models. Representative of this work are Kadt’s (1995) study of English and Zulu requests and directives, Lwanga-Lumu’s (2000) study of English and Luganda requests and apologies, Kasanga and Lwanga-Lumu’s (2007) study of English and Setswana apologies, Gough’s (1995) study of Xhosa requests, Dlali’s (2003, 2004) studies of Xhosa complaints and apologies, Egner’s (2006) study of Western and West African promises, and Tinkham’s (1993) study of mitigating devices in Westernised and non-Westernised Indian English directives. The main finding of this research has been that neither the contexts nor the functions of speech act realisations (the CCSARP’s original ‘givens’) can be assumed to remain constant across Western and non-Western cultures. In some cases, the functions of speech acts in postcolonial conversation differ from those of what seem to be similar speech acts in Western conversation. In others, the conditions under which particular types of speech acts are performed differ. Reasons in postcolonial conversation for offering, thanking, promising, agreeing, apologising, congratulating, hedging, or expressing sympathy, regret, or willingness can vary considerably from motives for performing similar acts in Western conversation (cf. Egner 2006; Kasanga 2006; Kasanga and Lwanga-Lumu 2007). Egner (2006), for example, points out that promises in West African cultures are often made in situations where both interlocutors know they will not be kept. They are not commissive speech acts as in Western contexts. The collectivistic logic of such situations dictates that false promises can be acceptable speech acts if performed to avoid disappointing one’s partner. In India, a similar collectivistic logic applies to false assurances like ‘No problem, sir’ in situations where problems are expectable (Janney 2009a). In both cases, lying, as it would be regarded in Western speech act theory (violating Austin’s Sincerity Condition or Grice’s Maxim of Quality), is simply behaving in keeping with a cultural code that requires individuals in certain situations to show respect and concern for others’ feelings, even at the cost of not telling the truth (cf. Heidemann 2006: 218 ff.). Showing empathy and respect in these situations simply has a higher cultural priority than being truthful. Accounting for such variations within Western frameworks, however, is often hindered by a lack of concepts relevant to explaining their underlying cultural logic. Thus, as a result, comparisons of Western/non-Western speech acts often only reveal divergences from Western patterns that they cannot explain. 110

Postcolonial pragmatics

Politeness and face A rather similar predicament has also occurred in studies of face and politeness in postcolonial cultures, where research has been widely influenced by Brown and Levinson’s Politeness: Some Universals in Language Usage (1987). Here, researchers have generally begun by collecting data in postcolonial target languages and then tried to find fits between the data and different categories of positive and negative politeness strategies in Brown and Levinson’s model (Nwoye 1992; Kadt 1998; Mashiri 2009). Alternatively, utterances in native and nonnative English have been compared (Tinkham 1993; Grainger et al. 2010), or utterances in indigenous languages have been compared with English (Kadt 1995; Kasanga and Lwanga-Lumu 2007). In much of this research, Brown and Levinson’s model has served as the grid for identifying and categorising non-Western politeness strategies and as the theoretical framework for explaining them. Research based on this model has revealed many differences between Western and nonWestern concepts of face and politeness that cannot be adequately accounted for in terms of the theory itself. A point of criticism has been the notion of ‘negative face’, or the idea that all humans have a need for, and indeed a natural right to, a certain degree of personal autonomy.To this, non-Western scholars reply that needs for personal autonomy, however natural in individualistic Western cultures, often tend to play only a relatively minor role in collectivistic Asian and African politeness, where the interests of the group often override the needs of the individual (Ide 1989; Gu 1990; Nwoye 1992; Mao 1994). Western negative face needs and negative politeness strategies, they say, have few direct equivalents in these cultures. Many forms of non-Western politeness in need of explanation and integration into pragmatic models have been noted. Ide (1989) points out the need for concepts related to forms of Japanese politeness rooted in notions of wakimae (behaviour in keeping with one’s sense of one’s conventional ‘place’ in a particular situation). Mao (1994) notes a lack of concepts related to Chinese ideas of miànzi (dignity, reputation) and liän (moral character). Nwoye (1992) suggests the need for attention to notions of what the Igbo call ezigbo omume (good behaviour); and several scholars have described forms of Zulu politeness rooted in concepts of hlonipha (respectfulness) and ubuntu (considerateness) (cf. Zungu 1997; Kadt 1998; Kamwangamalu 1999; Finlayson 2002). Research on non-Western face and politeness has shown that Western theories lack ways of explaining the meanings and functions of these concepts in their own collectivistic cultural contexts. Collectivistic politeness requires individuals to observe rules of the different roles they play that are based on complex hierarchical structures transferred into and recreated by their social relationships. Individualistic Western psychological models have historically factored most of this complexity out of consideration, leaving research gaps that can be filled only by culturally informed ethnographic research.

Current issues in postcolonial pragmatics Postcolonial pragmatics has taken shape as a response to this situation. It subscribes to the belief that postcolonial pragmatic practices cannot be adequately explained without reference to the complex, hybrid cultural contexts in which they occur. In line with this belief, it encourages ethnographically informed research oriented towards observing situated communication, hearing postcolonial voices, and attending to the singularities of speech situations and speakers’ pragmatic choices in different instances (Anchimbe 2011a: 10–12). It favours empirical observation and fieldwork over formal logic and supports research based on recordings of real interaction, interviews with local informants, and analyses of public communication of all kinds. Research in postcolonial contexts requires situated awareness and knowledge of the postcolonial speech communities themselves: knowledge of their histories, languages, ethnic demography, traditions, 111

Eric A. Anchimbe and Richard W. Janney

cultural beliefs, tribal and family bonds, and social conventions, as well as knowledge of their collectivistic perceptions of individuals’ roles, rights, and responsibilities. Understanding postcolonial cultural and social experience is essential for analysing postcolonial pragmatic practices. For this reason, postcolonial pragmatics emphasises the importance of developing communitybased explanatory frameworks placing local communication spaces at the centre of theory formation and analysis (Anchimbe and Mforteh 2011: ix). In recent years, research in this emic spirit has uncovered many previously unstudied features of postcolonial pragmatics. With increasing numbers of scholars analysing interaction in postcolonial contexts, the boundaries of pragmatics are expanding and the types of research questions dealt with are changing. By and large, current research focuses on patterns of postcolonial social interaction and the enormous impact of cultural and linguistic diversity on postcolonial pragmatic practices. Due to the constantly evolving social structures of most modern postcolonial communities, the list of issues receiving pragmatic attention grows steadily. The following summary represents only the present state of research in the discipline. It summarises work in the following areas: inter- and intralingual role relationships based on professions, kinship, gender, age, and religion, interlingual identity construction, linguistic opportunism and victimisation, and linguistic in- and outgroup marking.

Interlingual and intralingual role relationships The introduction of official written European languages and the emergence of mixed lingua francas (pidgins, creoles) profoundly influenced traditional patterns of social organisation and communication in the colonised regions. New forms of interaction and social stratification arose based on knowledge of the new languages. While the written European languages (e.g. English, French, Portuguese, Spanish) provided access to prestigious white-collar employment and upward social and professional mobility, the pidgins and creoles brought speakers from different regions and indigenous language communities together and, over time, attained covert prestige as languages of social solidarity.Tasks of postcolonial pragmatics have been to investigate functional relationships (including competition and clashes) between indigenous, mixed, and European languages in modern postcolonial contexts, to study different social roles ascribed to speakers of these languages, and to analyse language choices (including language switches and mixes) as markers of particular situations and role relationships.

Professional and occupational social role relationships Speaking a language unknown to one’s partner or answering a partner in a language diverging from his or her momentary situational expectations is sometimes a pragmatic avoidance strategy used in multilingual societies. Example (1) is from a roadside encounter between a taxi driver and a policeman in Bamenda, an English-speaking town in Cameroon.The policeman addresses the taxi driver in French, the default language of the Cameroonian armed forces, but the taxi driver answers in the local indigenous language. By doing this, the driver recontextualises the situation from a police traffic check into a tribal encounter where tribal norms require a closer role relationship. In calling the policeman by the ethnic social address form ‘Tanyi’ (‘father of twins’), the taxi driver reminds the policeman of his role within their collectivistic system, and the policeman has little choice but to ask the driver to drive on without controlling his car papers (cf. Anchimbe 2011b: 1472–1473). (1) Policeman: (Stops the taxi) Bonsoir. Taxi driver: Tanyi. Policeman: Tu as déjà acheté tes papiers? 112

Postcolonial pragmatics

Taxi driver: Tanyi. (Then silence, stares at him) Policeman: (Waves him to drive on. He does not control the car papers.) Language-based role relationships like this are found in many interactional contexts in postcolonial societies. For example, on university campuses in Cameroon, Nigeria, Ghana, and elsewhere, it is common to hear students speak English or French during their courses and then switch to pidgin in the cafeteria. Anchimbe and Janney (2011a: 1455–1456) describe how an Anglophone radio moderator in Cameroon uses English and pidgin to signal different role relationships to her listeners and an interviewee (2). In introducing the programme, the moderator assumes a professional broadcasting role vis-à-vis the listeners by addressing them in English. With the interviewee, a young schoolgirl who has just given birth, the moderator speaks pidgin, assuming the role of a mother figure reprimanding the girl and advising her about how to deal with her early motherhood in the future. Other contexts involving constructions of language-based professional and occupational roles include courtroom proceedings, political campaigns, and rural and urban health sensitisation campaigns.

Kinship role relationships Notions of family membership and role relationships have been shown to differ widely across postcolonial societies, often in ways contradicting Western concepts of the nuclear family. Notions of father and mother, brother and sister, aunt and uncle tend to be far more broadly conceived in collectivistic cultures than in individualistic Western ones. Wong’s (2006) analysis of uses of aunty in Singapore English and Anchimbe’s (2008) study of uses of uncle in Cameroon English suggest that this is true even of postcolonial societies that are geographically quite far apart. In Wong’s and Anchimbe’s studies, aunty and uncle are shown to have many meanings and contexts of usage in addition to their literal, nuclear family meanings. These usages reflect sociocultural attitudes of respect and deference, and illustrate how choices of kinship terms recreate and relocate relationships with strangers within kinship space, prescribing proper conduct towards others required by the natural order of age. In the example of the Anglophone radio moderator, see (2), the young schoolgirl calls the moderator aunty when the latter takes on the role of a mother figure. She had initially called her Madame – a formal respect address form. (2) Presenter: (to audience) Our first mother on today’s edition is on bed 23. She is so young and she is the mother of a baby boy. Who is this young lady? Good morning. Lady: Good morning, Madame. I am called N.E. . . . Presenter: Weti happen? You know say e no easy for start born pikin wey you di still go for school, no be so? (What happened? You know it is not easy to start having children when you are still in school, don’t you?) Lady: Yes Aunty. Presenter: How, you glad witham? Na mistake you be make? (Well, are you happy about this? Was it a mistake?) Lady: Yes. Na mistake.We just takam as it come. (Yes. It was a mistake.We just have to accept it.) (cf. Anchimbe and Janney 2011a: 1455 for full version) Here, the choice of aunty relocates the interaction which is taking place on radio, in the public domain, to a private family domain where advice can be given and taken without threat to face. 113

Eric A. Anchimbe and Richard W. Janney

Gender role relationships Issues of gender in postcolonial societies are often enmeshed in the fuzzy intersection between traditional indigenous beliefs and Western assumptions about men’s and women’s social roles. Current gender discourses tend to label traditional indigenous beliefs as conservative (bad) and Western gender assumptions as progressive (good). Ellece (2011), for instance, explores traditional vs. progressive rhetorical constructions of men and women in a Botswana television debate about gender roles. The study reveals deep conceptual contrasts between standpoints rooted in tribal tradition and standpoints echoing the ‘progressive’ global media. Similarly, Oduol (2007) discusses the difficulty of reconciling concepts of gender roles in Western gender discourse with indigenous concepts of gender roles in Kenya, explaining how conceptual gaps between the two forms of gender discourse make it difficult for Kenyan women to participate in Western gender debates. In spite of this, postcolonial urbanites’ linguistic choices make clear how gender is understood and woven into conversation about men and women. They navigate between indigenous and Western concepts according to contextual and linguistic expectations, making use of the hybrid structures available to them. While these studies (i.e. Oduol 2007 and Ellece 2011) are predominantly sociolinguistic in nature, gender-related forms of politeness have also been studied in postcolonial pragmatics. Finlayson (2002) discusses the custom of ukuhlonipha (‘respect’) language among Xhosa-speaking women in South Africa. According to Ntsimane (2007: 116), this custom is ‘a way to regulate family and clan relationships and to achieve harmonious co-existence’. It requires ‘especially. . . . women to show respect to men in their behaviour, speech and general conduct, particularly in performing certain rituals of avoidance’. On the political level, Atanga (2010) explains how progressive gender discourse in the Cameroonian parliament is often implicitly countered by female parliamentarians’ desire to be seen as ‘model traditional Cameroonian women’ (Atanga 2010: 93) who are respectful to their husbands. This position is clearly discursively constructed through their participation in parliamentary debates and the roles they occupy in the parliamentary system. More research in this direction will undoubtedly reveal further underlying gender structures and contribute to a better understanding of gender-related postcolonial interaction patterns.

Age role relationships Age is an all-pervasive factor in postcolonial social relationships. Traditional patterns of respect, influence, upward mobility, and social status are all influenced by age-related differentiations between members of the society that are extensively replicated in language. In Nigeria, Adegbija (1989: 169) says, ‘an elder is usually not referred to by name by a younger person; doing so would be considered rude and patently uncouth’. This also applies in Cameroonian cultures where, as Kouega (1998: 46) explains, ‘it is somewhat irreverent to call an elderly person’s name without affixing a honorific title to it’. Many codes of conduct beyond naming regulate younger people’s behaviour towards older individuals. In the following dialogue between a thirty-five- to forty-five-year-old male Cameroonian (YM) and an old man (OM),YM tries to fulfil the societal norm which requires younger people to surrender their seats to older people. (3) YM:  OM:  YM:  OM:  YM: 

Please, Sir, here’s a seat for you (standing up from his chair). Thank you, young man. I will rather stand. Please, please – How can I sit while you stand? Let me stretch my old bones by standing. I understand, but what impression should onlookers have of me? 114

Postcolonial pragmatics

OM:  Never mind. God has seen your heart. YM:  I wish the others could see my heart as well. OM:  Take it easy. YM: Ok. Here, when the old man refuses the offer, the younger man is troubled especially by what Anchimbe (2011c) calls the bystander effect – what other members of the group will think of him if they see him sitting and the old man standing. Age role relationships, therefore, moderate both linguistic and physical behaviour, and to understand the younger person’s statement ‘what impression should onlookers have of me?’ properly, we have to factor in the hybrid patterns of the society: that is, its traditional indigenous base to which new patterns were added during colonisation. One of these new patterns is religion, represented in (3) by the elder man’s statement ‘God has seen your heart,’ intended to compensate for the bystander effect.

Religious role relationships A heritage of colonialism that has had a lasting impact on postcolonial life and interaction is religion. From the first translations of Western religious sacred books into indigenous local languages up to today’s religious practices, Western religion has remained controversial in many postcolonial countries. Its teachings have been represented as unchallengeable and its God has been presented as the solution to all problems. Interesting findings have been reported about translations of Western religious concepts into indigenous languages. Israel (2011), for example, describes how sacred registers of the Tamil language were used in missionary translations of the Bible into Tamil in South India as a cultural tool to ascribe the Bible a superior truth status to rival scriptural traditions. Odidiomo (2011) shows how Nigerian Christians attending church services in English and Yoruba acquire systematically different conceptualisations of the notion of ‘prayer’. Christians of the Yoruba services represent ‘prayer’ as a traditional staff of power used in war in conformance with indigenous beliefs about people’s relations to their gods. For Christians who attend the English services, ‘prayer’ is simply a means of communicating with God. Although both sets of churchgoers in Odidiomo’s study were members of the same tribe (Yoruba) and worshipped in the same church, their religious inclinations were shaped differently by the worldviews communicated through the church services in the two languages. Studying interaction in such contexts requires weaving such religious hybridisms into the analysis.

Multiple interlingual identity constructions Constructing multiple linguistic identities did not begin with colonisation. Most postcolonial societies were multilingual long before they were colonised. Rather, colonisation simply introduced additional languages (e.g. English, French, Spanish) and social icons (professional and occupational roles) into the existing cultural mix. From these, new identities and language roles could be constructed. The original multilingualism of the colonised regions actually facilitated the acceptance and integration of new languages and identities. Nevertheless, pidgins and officially empowered European languages further complicated issues of identity and affiliation and partly redrew the existing linguistic boundaries between precolonial ethnic, regional and social groups. As identities rooted in colonial languages began competing with identities traditionally associated with indigenous oral languages, individuals’ reasons for projecting identities and their notions of relevant ethnic and social linguistic identity models changed. Today, orientations to indigenous and non-indigenous identities and attitudes about ethnic, age, and gender 115

Eric A. Anchimbe and Richard W. Janney

role relationships have changed as well. Modern postcolonial linguistic identities have become hybridised.

Linguabridity, identity opportunism, and linguistic victimisation As indicated briefly earlier, speakers in today’s multilingual, multiethnic, postcolonial contexts tend to have hybrid language identities, or identities that shift or become intermixed from situation to situation with their choices of languages. Hybrid language identities often result from alternations between or mixtures of different languages in conversation. They are outcomes of what Anchimbe (2007) calls ‘linguabridity’, the ability of many multilingual speakers to use and identify with multiple competing languages at once. Linguabridity is noticeable in conversation through speakers’ seemingly spontaneous adaptations to changes in the conversational situation necessitating shifts between languages, speech roles, or language identities. Such shifts can be unconscious or intentional. This opens the way for different forms of strategic language switching or mixing that enable speakers to take advantage of situational opportunities or avoid threats by projecting different identities or playing different roles (Anchimbe 2006 calls this ‘identity opportunism’). Speakers unable to successfully play this adaptive game may be subject to linguistic victimisation, which takes many forms (discussed earlier). Sala (2011), for example, analyses features of English pronunciation among native speakers of Lamnso, an indigenous language in northwest Cameroon, who are frequently victimised because of their accent when they speak English. To overcome such victimisation, some speakers adopt other accents that are more acceptable, especially in contexts where victimisation is probable. Postcolonial speakers must thus be adept at constructing believable language identities and switching fluently between identities in order to remain credible for their partners and avoid victimisation.

Ingroup youth vernaculars Postcolonial youth identities often find expression in youth group vernaculars. Meierkord (2011) shows how youth in rich Nairobi neighbourhoods have adopted ‘Engsh’, a self-invented English vernacular combining (American) English and Swahili on an English syntactic base structure, as a status dialect for their social identity. It enables these speakers to distinguish themselves from members of the identity group formed around ‘Sheng’, a Kenyan urban slum vernacular combining English and Swahili on a Swahili base structure. Engsh and Sheng thus serve in Nairobi both as social class markers and as codes for ingroup communication.

The ‘ethnification’ of colonial official languages The label ‘ethnic’ in most postcolonial societies is synonymous with the innermost layer of multiple identities, often typified by belonging to a given geographical region and using a particular indigenous or regional language. In recent times, however, issues of ethnicity and language identity have been made unnecessarily complex in some countries by politically motivated state language policies. Wolf (2001) and Anchimbe (2011a) draw attention to the increasing use of the terms ‘Anglophone’ and ‘Francophone’ as near-ethnic markers in Cameroon. According to Wolf (2001: 223), ‘the feeling of unity [has become] so strong that “being Anglophone” denotes a new ethnicity, transcending older ethnic ties’. In this new Anglophone ethnicity, the English language is a central component along with certain Anglo-Saxon character traits often claimed to have been taken over by Cameroonians in the parts of the country (northwest and southwest regions) formerly colonised by Britain. 116

Postcolonial pragmatics

Ingroup and outgroup memberships and relationships In collectivistic African cultures, ingroup safety, prestige, and esteem vis-à-vis other groups are traditionally assured through members’ readiness to fight for their group, to behave in ways that portray their group as the ideal, and to close ingroup boundaries to non-group members. Exclusion occurs through the use of ingroup language to shut out outgroup members, the performance of ingroup rituals, and the victimisation of outgroup members who are unable or unwilling to conform with ingroup rules and values. Such group defence strategies figure tragically in Ola Rotimi’s drama The gods are not to blame (1971) in (4), where Odewale, in trying to defend the honour of the tribe he thinks is his, mistakenly kills his own father. (4) Aderopo:  It is nothing, your highness. . . . It is the way the gods meant it to happen. Odewale: Do not blame the gods. Let no one blame the powers. My people, learn from my fall. The powers would have failed if I did not let them use me. They knew my weakness: the weakness of a man easily moved to the defence of his tribe. I once slew a man in Ede. I could have spared him. But he spat on the tribe I thought was my tribe. The man laughed, and laughing he called me a ‘man from the bush tribe of Ijekun’. And I lost my reason. The excerpt shows how strong tribal group allegiances can be and how each group member has a duty to defend them. Intergroup rivalries influence postcolonial interaction on many levels. Besides tribal rivalries, there are rivalries between ethnic groups, religious groups, social classes, gender groups, or language communities.

Conclusion Given the complexity of modern postcolonial societies – their extreme multilingualism, multiethnicity, colonial prehistories, and competing traditional and modern cultural value systems – the list of topics inviting further postcolonial pragmatic research could be expanded almost indefinitely. Ethnic and linguistic diversity have had enormous impacts on social interaction and communication in postcolonial societies. Postcolonial pragmatics encourages the development of emic approaches to investigating this diversity and its effects and resists attempts to explain principles of postcolonial pragmatics by appealing uncritically to theories of monolingual Western discourse. This is because, as Thomason and Kaufman (1988: 4) explain, ‘[T]he history of a language is a function of the history of its speakers, and not an independent phenomenon that can be thoroughly studied without reference to the social context in which it is embedded’. A commitment to emically adequate, community-based pragmatic approaches is at the same time a call for a re-ordering of research priorities.There remains a need in postcolonial pragmatics for a shift from rationalistic Western research paradigms to new, collectivistic non-Western paradigms. Research on non-Western pragmatics has traditionally suffered from a lack of focus on the kinds of common cultural knowledge that matter pragmatically in postcolonial contexts. All too often, researchers have ignored obvious lacks of fit between Western models and their own knowledge of pragmatic practices in their cultures. Some have felt it necessary to tailor their data, findings, and discussions to the available Western research on their subjects, even at the cost of producing results with little practical explanatory value in their own cultures. Postcolonial pragmatics resists this practice in the belief that it can deal adequately with its subject only by establishing a healthy scholarly independence from ethnocentric Western theory. In the 117

Eric A. Anchimbe and Richard W. Janney

spirit of Franz Boas’s Handbook of American Indian Languages (1911), it maintains that the facts of pragmatics in postcolonial contexts must be allowed to emerge on their own, without the distorting imposition of European templates on them. Scholars worldwide are invited to begin participating in this dynamic new field of study and contributing to broadening its growing research agenda.

Suggestions for further reading Anchimbe, E. A. (2014) (ed.) Structural and Sociolinguistic Perspectives on Indigenisation. Dordrecht: Springer. Focuses on pragmatic indigenisation features in English, French, and Pidgin in Cameroon, applying variational and postcolonial pragmatic frameworks. Anchimbe, E. A. and Janney, R.W. (eds.) (2011b) ‘Postcolonial pragmatics’ [Special issue], Journal of Pragmatics, 43(6). Collection of articles on various pragmatic practices in postcolonial contexts including the Caribbean (Trinidad and Tobago and French Guiana), Cameroon, China, and Haiti. Atanga, L. L., Ellece, S. E., Litosseliti, L., and Sunderland, J. (eds.) (2013) Gender and Language in Sub-Saharan Africa:Tradition, Struggle and Change. Amsterdam: Benjamins. Engages with gender discourses in African indigenous and urban societies calling for more investigations into these rich and diversified societies which have either remained unstudied or have been studied through Western gender lenses. Mulo Farenkia, B. (ed.) (2008) Linguistic Politeness in Cameroon. Frankfurt am Main: Peter Lang. Describes politeness strategies in English, French, and some indigenous Cameroonian languages signalling the insufficiency of Western theories in accounting for them. Mulo Farenkia, B. (2014) Speech Acts and Politeness in French as a Pluricentric Language. Berlin: LIT Verlag. Offers a description of the realisation of compliment and compliment response speech acts in two postcolonial contexts, Cameroon and Canada, from a postcolonial pragmatic and a variational pragmatic perspective. Sommer, G. and Vierke, C. (eds.) (2011) Speech Acts and Speech Events in African Languages. Köln: Köppe Verlag. Illustrates the realisation of speech events, rather than simply speech acts, in various African languages. Focus is on complaints, requests, apologies, politeness, and face.

References Adegbija, E. (1989) ‘Lexico-semantic variation in Nigerian English’, World Englishes, 8(2): 165–177. Anchimbe, E. A. (2006) ‘Hybrid linguistic identities in postcolonial Africa:The intricacy of identity opportunism in multilingual Cameroon’, in F. Heidemann and A. De Toro (eds.) New Hybridities: Societies and Cultures in Transition. Leipzig: Olms. 237–261. Anchimbe, E. A. (2007) ‘Linguabridity: Redefining linguistic identities among children in urban areas’, in E. A. Anchimbe (ed.) Linguistic Identity in Postcolonial Multilingual Spaces. Newcastle: Cambridge Scholars Publishing. 66–86. Anchimbe, E. A. (2008) ‘ “Come greet Uncle Eric” – Politeness through kinship terms’, in Farenkia, B. M. (ed.) Linguistic Politeness in Cameroon. Frankfurt am Main: Peter Lang. 109–120. Anchimbe, E. A. (2011a) ‘Postcolonial linguistic voices: Stitching together identity choices and their representations’, in E. A. Anchimbe and S. A. Mforteh (eds.) Postcolonial Linguistic Voices: Identity Construction and Representations. Berlin: De Gruyter Mouton. 3–21. Anchimbe, E. A. (2011b) ‘On not calling people by their names: Pragmatic undertones of sociocultural relationships in a postcolony’, Journal of Pragmatics, 43(6): 1472–1483. Anchimbe, E. A. (2011c) ‘ “Take a beer” – “Thanks. Sorry, I prefer another day”: A postcolonial pragmatic perspective on offers and offer refusals’, in J. Frenk and L. Steveker (eds.) Anglistentag 2010 Proceedings. Trier: Wissenschaftlicher Verlag Trier. 421–431. Anchimbe, E. A. (2013) Language Policy and Identity Construction:The Dynamics of Cameroon’s Multilingualism. Amsterdam: Benjamins. Anchimbe, E. A. and Janney, R. W. (2011a) ‘Postcolonial pragmatics: An introduction’, Journal of Pragmatics, 43(6): 1451–1459.

118

Postcolonial pragmatics Anchimbe, E. A. and Janney, R.W. (eds.) (2011b) ‘Postcolonial pragmatics’ [Special issue], Journal of Pragmatics, 43(6). Anchimbe, E. A. and Mforteh, S. A. (eds.) (2011) Postcolonial Linguistic Voices: Identity Construction and Representations. Berlin: De Gruyter Mouton. Atanga, L. L. (2010) Gender, Discourse and Power in the Cameroonian Parliament. Bamenda: Langaa Publishers. Austin, J. L. (1962) How to Do Things with Words. New York: Oxford University Press. Barron, A. (2012) ‘Interlanguage pragmatics: From use to acquisition to second language pedagogy’, Language Teaching, 45(1): 44–63. Blum-Kulka, S., House, J., and Kasper, G. (eds.) (1989) Cross-cultural Pragmatics: Requests and Apologies. Norwood: Ablex. Blum-Kulka, S. and Olshtain, E. (1984) ‘Requests and apologies: A cross-cultural study of speech act realisation patterns (CCSARP)’, Applied Linguistics, 5(3): 196–213. Boas, F. (1911) Handbook of American Indian Languages, vol. 1. Bureau of American Ethnology, Bulletin 40. Washington, DC: Government Printing Office. Brown, P. and Levinson, S. C. (1987) Politeness: Some Universals in Language Usage. Cambridge: Cambridge University Press. Dlali, M. (2003) ‘The speech act of complaint in isiXhosa’, South African Journal of African Languages, 23(3): 131–143. Dlali, M. (2004) ‘The speech act of apology in isiXhosa educational contexts’, South African Journal of African Languages, 24(2): 118–131. Egner, I. (2006) ‘Intercultural aspects of the speech act of promising: Western and African practices’, Intercultural Pragmatics, 3(4): 443–464. Ellece, S. (2011) ‘Gender and cultural identity in a television show in Botswana’, in E. A. Anchimbe and S. A. Mforteh (eds.) Postcolonial Linguistic Voices: Identity Construction and Representations. Berlin: De Gruyter Mouton. 51–76. Finlayson, R. (2002) ‘Women’s language of respect: “Isihlonipho sabafazi” ’, in R. Mesthrie (ed.) Language in South Africa. Cambridge: Cambridge University Press. 279–296. Goffman, E. (1967) Interaction Ritual: Essays on Face-to-Face Behaviour. New York: Pantheon Books. Goffman, E. (1969) Strategic Interaction. Philadelphia: University of Pennsylvania Press. Gough, D. (1995) ‘Some problems for politeness theory: Deference and directness in Xhosa performative requests’, South African Journal of African Languages, 15(3): 123–125. Grainger, K., Mills, S., and Sibanda, M. (2010) ‘ “Just tell us what to do”: Southern African face and its relevance to intercultural communication’, Journal of Pragmatics, 42(8): 2158–2171. Grice, H. P. (1975) ‘Logic and conversation’, in P. Cole and J. L. Morgan (eds.) Speech Acts (Syntax and Semantics, vol. 3). New York: Academic Press. 41–58. Gu,Y. (1990) ‘Politeness phenomena in modern Chinese’, Journal of Pragmatics, 14(2): 237–257. Hanks, W. F., Ide, S., and Katagiri, Y. (2009) ‘Towards an emancipatory pragmatics’, Journal of Pragmatics, 41(1): 1–9. Heidemann, F. (2006) Akka Baakka. Berlin: LIT Verlag. Ide, S. (1989) ‘Formal forms and discernment: Two neglected aspects of universals of linguistic politeness’, Multilingua, 8(2–3): 223–238. Israel, H. (2011) ‘Contesting the sacred in Tamil: Missionary translations and Protestant scriptures in colonial South India’, in E. A. Anchimbe and S. A. Mforteh (eds.) Postcolonial Linguistic Voices: Identity Construction and Representations. Berlin: De Gruyter Mouton. 167–182. Janney, R. W. (2006) ‘Postcolonial pragmatics’, Plenary paper presented at the International Conference on Pragmatics, Pune, India, 15–17 December 2006. Janney, R. W. (2009a) ‘Toward a postcolonial pragmatics’, in B. Fraser and K. Turner (eds.) Language in Life, and a Life in Language: Jacob Mey – A Festschrift. Bingley: Emerald. 203–212. Janney, R. W. (2009b) ‘Pragmatics in postcolonial contexts: For Paul Mbangwana’, Annals of the Faculty of Arts and Social Sciences. [Special edition in honour of Professor Paul N. Mbangwana]. Yaoundé, Cameroon: University of Yaoundé. 101–108. Janney, R. W. and Arndt, H. (1993) ‘Universality and relativity in cross-cultural politeness research: A historical perspective’, Multilingua, 12(1): 7–34. Kadt, E. de (1995) ‘The cross-cultural study of directives: Zulu as a non-typical language’, South African Journal of African Linguistics, 27: 45–72. Kadt, E. de (1998) ‘The concept of face and its applicability to the Zulu language’, Journal of Pragmatics, 29(2): 173–191.

119

Eric A. Anchimbe and Richard W. Janney Kamwangamalu, N. (1999) ‘Ubuntu in South Africa: A sociolinguistic perspective to a pan-African concept’, Critical Arts, 13(2): 24–41. Kasanga, L. A. (2006) ‘Requests in a South African variety of English’, World Englishes, 25(1): 65–89. Kasanga, L. A. and Lwanga-Lumu, J. C. (2007) ‘Cross-cultural linguistic realisation of politeness: A study of apologies in English and Setswana’, Journal of Politeness Research, 3(1): 65–92. Kouega, J. P. (1998) ‘Loans from some indigenous languages in Cameroon’, Alizés 16. Available from: http://laboratoires.univ-reunion.fr/oracle/documents/266.html (accessed 12 June 2008). Lwanga-Lumu, J. C. (2000) Requests and Apologies in Luganda and English. Johannesburg: University of Witwatersrand. Mao, L. M. (1994) ‘Beyond politeness theory: “Face” revisited and renewed’, Journal of Pragmatics, 21(5): 451–486. Mashiri, P. (2009) ‘Saying “no” without saying “no”: Indirectness and politeness in Shona refusals’, Journal of Language and Communication, 3(2): 121–146. Matsumoto,Y. (1988) ‘Reexamination of the universality of face: Politeness phenomena in Japanese’, Journal of Pragmatics, 12(4): 403–426. Meierkord, C. (2011) ‘U r ma treasure bila measure: Identity construction in Kenya’s multilingual spaces’, in E. A. Anchimbe and S. A. Mforteh (eds.) Postcolonial Linguistic Voices: Identity Construction and Representations. Berlin: De Gruyter Mouton. 25–50. Ntsimane, R. (2007) ‘The ukuhlonipha code of respect: Gender and cultural tensions around the Zulu nurses. The case of the Emmaus Mission Hospital’, Studia Historiae Ecclesiasticae, XXXIII(2): 115–133. Nwoye, G. O. (1992) ‘Linguistic politeness and socio-cultural variations of the notion of face’, Journal of Pragmatics, 18(4): 207–243. Odidiomo, F. (2011) ‘What mental images reveal about religious lexemes in Yoruba and English in presentday Nigerian churches’, in E. A. Anchimbe and S. A. Mforteh (eds.) Postcolonial Linguistic Voices: Identity Construction and Representations. Berlin: De Gruyter Mouton. 183–200. Oduol, J. A. (2007) ‘English as lingua franca: A threat to multilingualism in gender discourse. The Case of Dholuo of Kenya’, in E. A. Anchimbe (ed.) Linguistic Identity in Postcolonial Multilingual Spaces. Newcastle: Cambridge Scholars Publishing. 92–118. Rotimi, O. (1971) The Gods Are Not to Blame. Oxford: Oxford University Press. Sala, B. M. (2011) ‘ “Lamnso” English: A study in ethnic variation in Cameroon English’, in E. A. Anchimbe and S. A. Mforteh (eds.) Postcolonial Linguistic Voices: Identity Construction and Representations. Berlin: De Gruyter Mouton. 241–261. Schneider, K. P. and Barron, A. (eds.) (2008) Variational Pragmatics: A Focus on Regional Varieties in Pluricentric Languages. Amsterdam: Benjamins. Searle, J. R. (l969) Speech Acts. Cambridge: Cambridge University Press. Simo Bobda, A. and Mbouya, I. F. (2005) ‘Revisiting some linguistic concepts and beliefs in the light of the sociolinguistic situation of Cameroon’, in J. Cohen, K. T. McAlister, K. Rolstad, and J. McSwan (eds.) Proceedings of the 4th International Symposium on Bilingualism. Somerville: Cascadilla Press. 2122–2132. Tanyi, R. (1978) ‘Bilingualism and the Young Cameroonian School Child’, unpublished DES dissertation in English studies, University of Yaounde. Thomason, S. and Kaufman, T. (1988) Language Contact, Creolisation and Genetic Linguistics. Berkeley: University of California Press. Tinkham, T. (1993) ‘Sociocultural variation in Indian English speech acts’, World Englishes, 12(2): 239–247. Wolf, H.-G. (2001) English in Cameroon. Berlin: Mouton de Gruyter. Wong, J. (2006) ‘Contextualising aunty in Singaporean English’, World Englishes 25(3–4): 451–466. Zungu, P. (1997) ‘Some aspects of hlonipha in Zulu society’, Language Matters, 28(1): 171–181.

120

12 GENDER AND SOCIOPRAGMATICS1 Janet Holmes and Brian W. King

Introduction This chapter explores some of the ways in which sociopragmatic research has made an increasingly valuable contribution to our understanding of the relationship between language and gender over the past few decades.We use the term sociopragmatics to cover research which examines the relationship between social context and discourse, and we focus in particular on the pragmatics of power and politeness in language and gender research.We begin with a discussion of the terms sex and gender. Researchers working in the areas of pragmatics and sociolinguistics generally treat the term sex as a label for the relevant biological category and gender as a label for the sociocultural category. For both categories, despite the widespread popular belief that just two subcategories are relevant, there is interesting evidence of complexity. As Bing and Bergvall comment: If the boundaries are not problematic, it is curious that so much energy is expended to reinforce them and to render invisible large numbers of people, including homosexuals, bisexuals, eunuchs, hermaphrodites, transvestites, transsexuals, transgendered and intersexed individuals, and others who assume social and sexual roles different from those that their cultures legitimize. (1996: 6) Similarly, the distinction between apparently feminine and masculine behaviours is not at all clear-cut. Stereotypical or normative expectations about appropriately feminine and masculine ways of behaving, including talking, differ across societies and speech communities. One widely cited example is the contrasting norms of Malagasy vs. European societies. Malagasy has two broadly distinguishable styles of talking: resaka, an informal style women and men alike use in everyday interaction, and kabary, a more formal style that only men use in specific contexts such as religious rituals, and for indirect criticism in competitive displays where success leads to greater social prestige (Keenan 1989). In general, the women’s style of speaking is considered direct and confrontational; women criticise directly and bargain aggressively in the marketplace. The men’s kabary style is subtle and non-confrontational, associated with skill and learning, and generally regarded as more polite. As Bucholtz (2003: 47) notes, “Keenan does not explicitly contrast this pattern with the scholarly and popular view, common at the time she did her research, of Western women’s speech as indirect and men’s as direct”, though many other 121

Janet Holmes and Brian W. King

scholars have done so. The example makes clear that what is regarded as “feminine” or “masculine” is culture and context-dependent. It is also evident that dimensions such as (in)directness and (im)politeness are matters of degree rather than absolutes, and sociopragmatic research has made a major contribution by exploring these complexities in the past decade in particular. In what follows, we first describe the ways in which theoretical approaches to research on the relationship between language and gender have developed since the 1970s; we then explore the insights gained from sociopragmatic research, focussing on the themes of politeness and power and also queer-oriented approaches to sociopragmatics, illustrating with examples from our own research. Next we discuss the directions in which we see this field developing, and we conclude with a summary section and suggestions for further reading.

Historical perspectives on language and gender research Language and gender research which focussed on pragmatic issues was kick-started by Robin Lakoff, an outstanding pragmaticist, who profoundly influenced the direction of language and gender research worldwide with her ground-breaking article (1973) and book (1975) Language and Woman’s Place. She addressed two fundamental dimensions of the interaction of language and gender, namely the language used by and about women, claiming that both contributed to perceptions of women as weak and indecisive, and suggesting that women should adopt male ways of speaking if they wanted to be taken seriously. Lakoff ’s provocative and quite specific hypotheses about what she considered deficiencies in the ways in which American women spoke compared to men generated a huge amount of quantitative research in the late 1970s and early 1980s. Women, she suggested, used more of a number of specific usages than men did. Consequently, sociologists and psychologists counted the adverbs, tag questions, colour words, and specific adjectives (e.g. cure, divine) that Lakoff had identified as gender specific to test out her claims. Findings were very disparate and often conflicting (e.g. Crawford 2003; Dubois and Crouch 1975). Sociolinguists, however, noted that Lakoff ’s hypotheses about a range of superficially distinct linguistic forms (such as stress, tag questions, and modal adverbs) were unified by an underlying analysis of two basic pragmatic functions, namely, hedging and boosting (e.g. Holmes 1984, 1995; Talbot 1998). This meant that each form needed to be considered in context to ascertain its social meaning and pragmatic force. By contrast, the disparate linguistic forms that became the focus of quantification (typically lexical items easily identified by a computer program which ignored their different meanings) had no such claims to coherence. Moreover, sociolinguists also noted Lakoff ’s comments on the relevance of social context, as well as her fundamental political argument about power and politeness. Subsequent research supported her claims that the frequency with which particular forms occurred typically depended on their sociopragmatic meaning in a specific interactional context, and also on power relations rather than on the sex of the speaker (e.g. O’Barr and Atkins 1980; Holmes 1986). Specifically, researchers argued that differences between women’s and men’s use of language were best accounted for by attending to societal power relations. Hence, in contexts where men were more powerful and influential, they often dominated the discourse: they talked more than women, took more speaking turns, and disruptively interrupted other speakers more often (e.g. Eakins and Eakins 1979). Women, on the other hand, were more likely in such situations to use encouraging feedback and facilitative questions and to contribute talk when it was helpful to the addressee (e.g. Holmes 1992). From here it was a short step to an argument that such differences were the result of the different socialisation experiences of boys and girls. Deborah Tannen is perhaps the best-known advocate of the “cultural difference” explanation for men’s and women’s ways of talking. Her book You Just Don’t Understand, published in 122

Gender and sociopragmatics

1990, drew on a wide range of sociolinguistic research to support the argument, first proposed by Maltz and Borker (1982), that boys and girls are socialised differently and consequently adopt different (sub)cultural frameworks. Males, she argues, develop a status-oriented frame, while females are connection-oriented. Hence their ways of talking are different, and this can lead to cross-cultural misunderstandings: for example in couple’s talk, men try to solve women’s problems, while the connection-oriented women just want a sympathetic ear; men won’t ask directions because that would put them in a one-down position; men like to hold the floor in contexts where talk may enhance status; women prefer to talk in the comfort of intimate settings; men focus on “report” talk while women prefer “rapport” talk.2 It is easy to see why such “men do this; women do that” generalisations became a target for critical comment, though in fact Tannen was careful in basing her claims on published research, and she also pointed out exceptions where there was evidence of them. However, some took issue not just with the generalisations, but also with the theoretical “two cultures” explanation, which they regarded as inadequate because it ignored the discursive effects of unequal power distribution and differential social status (e.g. Troemel-Ploetz 1991; Freed 1992). Inevitably the next wave of gender and language research gave detailed attention to contextual sources of variation in accounting for differently gendered talk. Although any analysis which treats gender as distinct from biological sex (as with Tannen) can be classified as “social constructionist”, this term has come to be associated with “postmodernist” approaches to social construction (see Cameron 2005). As part of a postmodernist turn, there was a shift away from treating female/male and woman/man as rigid cultural categories in favour of viewing the “doing of gender” as an ongoing performance as opposed to one that is stabilised in childhood (Cameron 2005). In other words, gender is performed (i.e. it is something we do) rather than an expression of what we are; it is not a fact but an act (Nelson 1999).Therefore, gender is never truly finished, and these gender performances are viewed during analysis as changeable and dependent on context. This postmodernist leaning coincided with the turn (mentioned earlier) toward contextbased variation, and also matched a shift to the pluralisation of gender into masculinities and femininities. That is, attention turned to the ways in which gender intersects with other social constructs such as class, race, and sexual identity to produce various types of masculinity and femininity (e.g. middle class vs. working class, black vs. white, or gay vs. straight). Another way to put this is that attention turned to the gendered politics that exist within femininity and within masculinity (Johnson 1997) rather than always focussing on differences between masculinity and femininity. The tendency towards pluralisation of gender has led to an interest in how nuanced versions of gender and sexuality might influence the assumptions we bring to sense-making during conversation. Thus, sociopragmatic analysis explores how masculinities, femininities, and sexualities play a role in such sense-making. Pragmatics holds that meaning is grounded at least partly in inference rather than being taken just from the words spoken. That is, a pragmatic approach to analysis asks how conversational participants work out what is meant on the basis of abstract discourse strategies instead of simply attending to the words and their grammatical combinations (Cameron 1998). As part of this process, certain assumptions are brought to bear by participants (based on their past experience and knowledge of the world) in order to decide what is being implied or indirectly suggested, requested, and refused. As examples will demonstrate, gender can make a difference to these implied meanings and their interpretation. Cameron (2005) points out that a pragmatic approach to analysis is particularly useful in moving beyond a further debate which rages within all social constructionist circles (and has for some time). The debate is about whether gender is (a) globally relevant to interaction even in the light of diversity and local variability (a stance termed gender realism); or whether it (b) must be explicitly 123

Janet Holmes and Brian W. King

mentioned in talk to be relevant (a strong form of gender relativism). Cameron (2005) suggests that this grounding of meaning in inference can actually be identified in both gender realist and gender relativist analyses. Even in cases where gender is explicitly referred to, the strictest “relativist” analyst also has to “go beyond the data” to explain why it is made relevant. ‘To answer the question “why that now?” it is often necessary to presuppose an answer to the prior question “why that ever?” ’ (Cameron 2005: 329). In other words, background assumptions tend to preexist the conversation being analysed; therefore moving beyond the realist-relativist debate involves everyone investigating how participants make use of information that is not explicitly encoded in the spoken words. As outlined previously, postmodernist versions of social constructionism treat gender and sexuality as ongoing performances rather than as pre-given or stabilised. However, “queer” approaches to inquiry have begun to take very seriously the idea that the “taken-for-granted” status of heterosexuality (i.e. heteronormativity) needs to be critically investigated more completely. That is, there is a need to reveal how heterosexuality is made to look like common sense, while other forms of sexual desire (e.g. same-sex attraction) or sexual identity (e.g. gay identity) are positioned as unusual or “abnormal” in spite of their constant presence (Sauntson 2008: 282).When performances of straight identity or heterosexual desire are treated as common sense while other sexual identities and sexual desires are not treated that way, it becomes clear that those commonsense constructs are being placed in a very powerful position. By looking more closely at the ongoing social construction of heterosexuality in conversation we can gain more insight into the sociopragmatics of power and gender. In other words, by looking at the assumptions people make in conversation and how they work out various implied meanings (i.e. inferences) we can gain a lot of insight into how power, gender, and sexuality come together.This type of analysis is exemplified later in Queer-oriented approaches to sociopragmatics. Finally in this section, it is worth considering the influence of ideology on what is valued in a society. Keenan (1989) notes, for instance, that Malagasy language ideologies privilege the indirect style men use, especially in oratory, evaluating it as skilled and artful, while denigrating direct language as unsophisticated. In other communities where men’s discourse tends to be more direct, it is directness which is highly valued. Clearly the influence of ideology is an important consideration when examining gendered talk in different communities (see Queer-oriented approaches to sociopragmatics).

Critical issues in sociopragmatics and language and gender research Power, politeness, and gender There is a huge amount of research which could be considered in this section. We have chosen to focus on examples of how issues of power and politeness emerge in sociopragmatic research which examines the relationship between gender and language. Power can take many forms: for example, physical, economic, ideological (see Kiesling 1997: 68); here we are concerned with the ways in which sociopragmatic strategies support or contest particular power relations in interaction.Tannen (1987: 5) points out that in any particular interaction, different participants may have different kinds of power which they exercise in different ways. In other words, she suggests that it is impossible to identify the power in a situation. Rather, power is dynamically constructed and exercised, both implicitly and explicitly, in different aspects of a specific interaction; different participants manifest power in diverse ways as they construct their own identities and roles in response to the behaviour of others in particular sociocultural contexts (Keating 2009). Current analyses of power in interaction tend, then, to take a dynamic approach, recognising that power is constantly being constructed, negotiated, maintained, and re-asserted, as people interact. Another useful insight, especially in the context of language and gender research, emerges from critical discourse analytical (CDA) research which highlights the “systemic” nature of power 124

Gender and sociopragmatics

(Fletcher 1999, Wodak 1999), that is, the way the norms of the most powerful group are assumed and taken for granted in most situations. This approach strives to identify evidence of the covert exercise of power in, for example, the unobtrusive, “naturalised” conversational strategies through which power and gender relations are constructed and reinforced in everyday, unremarkable, takenfor-granted interactions (cf. Fairclough 1992; see also Holmes 2005). Consider how sexist jokes often go unchallenged, or the frequent assumption that the most senior male has the right to lead the discussion, or the fact that boys frequently get most attention in mixed-sex classrooms. Analysing politeness in interaction has also occupied the attention of myriad researchers, resulting in many different definitions and approaches (cf. Archer, Chapter 29, this volume). Again, it is fruitful to treat politeness as a complex concept, dynamically constructed and exercised, both implicitly and explicitly, in specific interactions in specific sociocultural contexts. Politeness is often equated with attention to face needs (Brown and Levinson 1987), to “relational work” (Locher and Watts 2005), or to establishing and maintaining solidarity and “rapport” (Spencer-Oatey 2008).We illustrate these points by considering how power and politeness manifest in workplace contexts, and especially in research on gender and leadership.

Power and masculinity Power is a central component of leadership, since leaders are people with “legitimate power”, that is, power due to position (French and Raven 1959). They may also have “expert power”, that is, power derived from expertise and skills (French and Raven 1959), and in some contexts, gendered sociopragmatic behaviour may also contribute to the construction of leadership identity. Holmes (2006a), Mullany (2007), and Baxter (2010), for example, all demonstrate that effective communicators, both female and male, make use of a wide range of sociopragmatic strategies to enact power in workplace interaction.3 Some of these are considered normatively masculine or “power-oriented” strategies such as directing meeting talk, ratifying decisions, issuing relatively “bald” directives, using controlling questions, and contesting the statements of others, while other strategies are regarded as more normatively feminine or “politeness-oriented”, such as expressing approval, attenuating disagreement, hedging negatively affective speech acts, using humour as a “sweetener”, and providing a good deal of supportive feedback. The limitations of normatively gendered views of talk are well illustrated by these workplace studies, which demonstrate the flexibility of effective leaders in selecting the most effective sociopragmatic strategy in specific contexts. On the other hand, it is also important to recognise that interactional norms may be gendered. Particular contexts and occupations, such as the police force or the army vs. primary school teaching and the nursing profession, may be heavily dominated by men vs. women, and consequently relatively masculine vs. feminine interactional norms may obtain (e.g. McElhinny 1992; see also Holmes 2014). Baxter (2011) provides further support for this analysis in her examination of the discourse of female leaders in British male-dominated organisations. She uses the term “double-voiced discourse” (Bahktin 1994) to describe the ways in which these women used language consciously and strategically to compensate for their marginalised status in large, male-dominated organisations, while Baxter and Wallace (2009) demonstrate how, through their on-the-job talk, UK builders construct building sites as masculine workplaces. The Language in the Workplace Project (LWP) recordings in factories and constructions sites illustrate the complex ways in which power is enacted and gender identity constructed in these typically hierarchical and “masculinist” settings (Holmes 2006a). Directives are predominantly unmitigated (just take these staples; grab another bloody ladder; drill those; hold it like right there); swear words and jocular abuse are much more frequent than in white-collar professional settings, and social talk often revolves around sport rather than family, for example (Holmes and Stubbe 125

Janet Holmes and Brian W. King

2015). These findings are not surprising; they confirm societal understandings of normative sociopragmatic gender patterns. It is interesting to consider, then, how power is enacted in a less typical setting, where pre-allocated roles and responsibilities are not relevant. The following example illustrates how power is abrogated in a situation where all participants are, in principle, equal in status. Discussions held at different tables were recorded in a series of public meetings organised to discuss the preservation of a particular city suburb as a historic area. A table host (TH) was appointed for each group to lead the discussion, and our analysis clearly demonstrated interesting differences in the way different THs exercised authoritative vs. facilitative sociopragmatic strategies in enacting these roles (Lazzaro-Salazar, Holmes, Marra, and Vine, 2016). The most authoritative of the THs controlled the discussion by first asking questions, which the other members of the group answered, and then commenting on their responses. (1) illustrates these strategies. (1)4 (Transcription conventions are provided at the end of this chapter.) Context: talk at one discussion table in a public meeting forum. TH is the table host. F represents unidentifiable female contributors.   1. TH:  2. F:   3. TH:   4. F:   5. TH:   6.   7.   8.   9. 10. 11. F: 12. TH: 13. 14. FX: 15. TH:

but you’re talking about the tools you’re not talking about overall yeah are are you willing to lose the history of SUBURB /no no\ /but\ you’re you’re talking about the tools and the rules \ I’m talking about the big picture it’s (. . . . . .. . ..) but what you’re all looking at when it says retain you’re looking at retaining specific rules and things in the district plans or whatever what I’m trying to suggest is a much bigger picture /mm\ /of of \\ of what your area looks like what it feels like what it is to live in your community (.°.°.) /it’s a real (mix.°.°.)\ yes that’s right it’s //your vision your vision (Lazzaro-Salazar et al., 2016)

TH states his view of what the group is saying, for example, you’re talking about the tools (lines 1, 5) evaluates responses, yes that’s right (line 15), poses questions, for example, are you willing to lose the history of SUBURB (line 3), and asserts repeatedly what he thinks the group should be focussing on, the big picture (line 6), a much bigger picture (line 10). Nobody at the table seems willing to contest or question the TH’s statements; rather their responses endorse his view (lines 2, 11) and indicate their understanding of what it feels like to live in this community (line 14), and someone even responds cooperatively to his rhetorical question (lines 3–4). This is a nice illustration of how power and control can be subtly exerted through the use of strategic questions and challenges. Summarising is another such strategy, also illustrated here; the TH critically summarises the group’s position as focussed on the tools and the rules (line 5). Summary is a powerful sociopragmatic device since, typically, the summariser’s definition of issues is then on record. Allocating turns to speakers who agree with your viewpoint is another powerful strategy that we identified as a means of subtly influencing the outcome of a discussion (LazzaroSalazar et al., 2016). 126

Gender and sociopragmatics

Such controlling discursive behaviour, enacting power both explicitly and more covertly, especially in relatively public or formal contexts, is generally regarded as normatively masculine in Western societies; and there is evidence that this perception is also relevant in some Asian contexts (Schnurr and Chan 2005; Chan 2007, 2008; Saito 2011). These strategies are effective means of constructing authority and power, and there is good evidence that they are effectively used by women leaders when judged appropriate to achieving their transactional goals (Holmes 2006a; Mullany 2007; Baxter 2010).

Politeness, relational talk, and femininity To illustrate more relationally oriented sociopragmatic behaviour, we focus on humour. Humour is a particularly productive and sophisticated sociopragmatic strategy to consider in order to better understand the complexities of the relationship between gender and discourse in different sociocultural contexts, as our research has demonstrated (Holmes and Marra 2002a, 2002b). Humour is undoubtedly more frequent in relaxed interactions between friends than in business meetings (Jenkins 1985; Hay 1995, 2002; Holmes and Marra 2002b), but an examination of its distribution in workplace interaction is remarkably revealing. Stereotypically, men are regarded as having a better sense of humour than women, while women are stereotypically allocated the role of responders (Crawford 1995; Evans Davies 2006; Kotthoff 2006). But the reality is rather different, as many researchers have demonstrated (e.g. Crawford 2003). Our analysis of the amount of humour contributed by women and men in twenty-two workplace meetings, for instance, clearly challenged the negative stereotype of women as humourless creatures: the average ratio for women was twenty-five instances per 100 minutes compared to the men’s ratio of fourteen instances per 100 minutes. Moreover, the very presence of women tended to be associated with higher levels of humour: as the proportion of female participants in a meeting increased, so did the amount of humour (Holmes et al. 2001). Turning to different functions and styles of humour, gender stereotypes are again easily identified. Aggressive, confrontational humour, teasing, and jocular abuse are stereotypically regarded as “masculine” (e.g. Tannen 1990; Coates 2003; Kotthoff 2006), or “masculinist” discourse (Baxter 2003), while supportive and collaborative humour is stereotypically considered more “feminine” (e.g. Tannen 1990; Eder 1993; Coates 1996; Kotthoff 2006). Again the reality is more complex. There is certainly some evidence that certain workplaces tend to be characterised by relatively contestive and challenging styles of interaction. Plester and Sayers (2007), for example, studying IT companies, identified jocular abuse, aggressive barbs, and potentially hurtful teasing, which they described as banter or “taking the piss”. Our LWP research identifies similar patterns in IT companies, contexts in which males tend to dominate numerically (Holmes 2006b). On the other hand, there is also evidence that women use jocular abuse both in relaxed contexts (Hay 1995, 2002) and in workplace contexts, as (2) illustrates. (2) Context: Three women of equal status in a government department discussing proposals they are working on. 1. Val: and Celia’s finished her proposals I’m sure [laughs] 2. Cel: on the last one 3. Val: ah you sod All: [laugh] (Holmes and Stubbe 2015) 127

Janet Holmes and Brian W. King

This example also illustrates that swear words do occur in white-collar workplaces, and that despite societal norms, women use them as well as men; however, relevant, this is a noticeably rare and relatively mild instance. While humour may be used strategically or subversively in workplace interaction (Holmes 2000, 2006b), there are also many examples in our data where both women and men contribute to a more supportive style of relationally oriented humour that is often characterised as normatively feminine (Tannen 1990; Coates 1996). Such humour often serves to develop and maintain solidarity and collegiality. (3) illustrates some of the sociopragmatic complexities of this style of humour in a white-collar professional workplace.This team does not meet together face-to-face very often and thus constructing solidarity is an important function of such meetings. (3) Context: Planning meeting of a group of four male and four female regional managers of a national organisation. There is a lot of laughter and overlapping concurrent unintelligible talk throughout this excerpt. 1. Pen: the fact that we don’t go to Malt [name of a town in New Zealand] 2. How: mm 3. Pen: doesn’t mean that people from Malt can’t 4. Sco: yeah 5. Pen: go somewhere to get help mm cos they were interested enough t6. Ral: if you live in Malt you need to go somewhere /(to get help)\ /[general laughter]\ 7. 8. Sco: there is actually quite a big consultancy in Malt 9. How: is there? 10. Sco: yeah 11. Hen: I was told many years ago that Malt /was the\ 12. Mal: /Malt\ 13. Hen: /heart of the\wife swapping area for [name of province] 14. Mal: /(Malt)\ [pronounced with local pronunciation] 15. Sco: /isn’t\ it Malt that had the highest rate of um 16. Pen: /ex ex nuptial\ birth- births /ex\ 17. Sco: /s t oh no that’s Hopeville\ /highest\ the highest s t d rate per capita 18. Pen: /Malt had th-\the highest 19. Sco: /Malt or Hopeville [laughs]\ 20. Kir: /did they?\ 21. Pen: rates of ex ex nuptial births at one point. . . . 22. Mal: it’s the alcohol that does it 23. How: [laughs] it’s the alcohol 24. [General laughter and overlapping talk continues throughout next section] 25. Pen: poor old Malt 26. Kir: we should be there 27. Sco: we should be there 28. Pen: we should be there [general laughter] 29. (Holmes 2006a) This is a section from a much longer humorous sequence. Its main function seems to be to provide light relief during a two-day strategic planning meeting; everyone is involved and it is 128

Gender and sociopragmatics

good-humoured and entertaining. The humour revolves around the disadvantages of living in a particular small, isolated, rural town, here pseudonymed Malt. Participants, male and female, work together to develop the topic at length, making more and more outrageous claims about the horrors of life in the town. Each utterance supports and further develops the proposition of the previous contributor. There is a great deal of fast and frequently overlapping speech, as well as laughter throughout. The contributions are closely integrated stylistically with one person filling in another’s gaps and answering their queries. There are requests for confirmation (e.g. lines 9, 15, 20), and a great deal of repetition extending from single words and phrases, for example, highest (lines 15, 17, 18) and outpost (lines 31, 32, 34), to whole clauses, for example, the refrain we should be there is repeated by several different voices at different pitches and volumes in a way which is strongly reminiscent of the different parts in a motet or madrigal (lines 26, 27, 28). More typically in our workplace interactions, humour is used to soften a directive or a criticism, as in (4) where Sam comments good-humouredly on the fact that the chair, Jill, tends to run long meetings. (4) Context: Board meeting at a small IT company. 1. Sam: 2. Jill:

ke- keep going until there’s only one person standing [laughs] oh you’ve been to our board meetings before [laughs](LWP corpus)

Jill’s laughing response to Sam’s sarcastic comment makes clear that the implied criticism is made in a context where good collegial relations prevail. As a final example of relational humour, it is, interestingly, often found in medical settings (Pizzini 1991; Ragan 2000). In (5), humour is used to distract the patient during a painful procedure. (5) Context: The nurse is in the patient’s room taking out a cannula (a thin tube inserted into a vein or body cavity to administer medicine, drain off fluid, or insert a surgical instrument.   1. Sophie: so I’m staying with mum and dad   2. Tara: wicked yeah  3. Sophie: yeah   4. Tara: a few home comforts don’t /hu- don’t hurt\  5. Sophie: /oh mum would\ insist on it even if I don’t want to /she’d insist yeah [laughs]\   6. Tara: /yeah [laughs]\ mothers have that right /don’t they\   7. Sophie: /they do\   8. Tara: yeah /okay\   9. Sophie: /well I don’t mind\ cos I get treated like a queen /so [laughs]\ /exactly \ lap it up 10. Tara: (Holmes and Major 2002: 8) During all this talk, Tara is taking out the cannula. The colloquial, friendly, relationally oriented chat, full of humour, helps distract Sophie from the discomfort involved. The overlapping talk indicates the informality and good rapport between the two young women. The talk focusses on Sophie, and on the comfort and care she can look forward to from her mother when she leaves the hospital. The two women co-construct a collusive bond, expressing a shared wry, yet appreciative attitude to the way mothers fuss over their daughters. Establishing friendly rapport 129

Janet Holmes and Brian W. King

like this obviously helps the patient feel more relaxed in this clinical setting, and this clearly assists the nurse to achieve her goals as well. This is normatively gendered talk both in topic and style (Holmes 2006a). This section has examined some ways in which power and politeness are sociopragmatically negotiated in interaction using data on meetings to illustrate how gender stereotypes and discursive patterns play out in different contexts.

Queer-oriented approaches to sociopragmatics As outlined in Historical perspectives on language and gender research, the queer critique directly addresses the exclusionary nature of binary systems, or the way in which binary categorisation has the effect of erasing certain categories which fail to fit neatly into such polar opposites. Even with good intentions, the analyst can be constrained by “heteronormatively realised” social categories (e.g. “woman and man”, “female and male”, “gay and straight”), and thus erase subjectivities which fall outside the narrow field of view such categories afford. This is because although such dyads are inadequate classification systems for describing the world around us (see Introduction), they are widely believed to constitute the extent and range of genders, bodies, and sexual identities available. (Socio)pragmatics has proved a useful tool in queer-oriented approaches to linguistic identity construction, and thus has taken a prominent place in theorisation about speech acts and performativity (Motschenbacher 2010: 36–37). For example, Morrish and Sauntson (2007) identify two pragmatic functions of what is termed the closet. These two functions, concealment and disguisement, envelop a range of linguistic features which serve to “police” or “gate-keep” the potential audience of sexual identity performances. With concealment, an ingroup code is used which is not intelligible to outsiders at all, while disguisement deploys features which the wider public finds intelligible as language, but which they decode differently. A hypothetical example is the question ‘Are you a friend of Dorothy?’ to mean ‘Are you gay?’ (or something similar). Only the ingroup will interpret it in a non-heteronormative way, and so risk of exposure is minimised if remaining in the closet is one’s desire. This use of inference to communicate in code can be seen as a variation on “strategic moves” of intentional misunderstanding (Cameron 1998). Cameron identifies these as moves in which a person exercises power through the purposeful “incorrect” interpretation of another speaker’s utterance (at the level of inference) in order to satisfy their own agenda. In the case of disguisement, the speaker is strategically making use of the outgroup’s misunderstanding of inference in order to disempower them. Inference and its interpretation also play an important role in resistance to heteronormativity as well as its propagation. This is because power lies in the heteronormative assumptions that speakers bring to reasoning as part of making and interpreting utterances. Whether intentional or not, the normalisation of heterosexuality (i.e. its stabilisation as the only truly “normal” sexuality) is often achieved in the pragmatics of silence or omission (see Kulick 2005; Motschenbacher 2010), and/or in the positioning of/by others in conversation and possible alignments or resistance to such positioning. Data from sexuality education classrooms provide an interesting illustration of the effects of inference on heteronormativity’s reinforcement or avoidance. (6)–(9) were recorded in a lesson with senior secondary school students in New Zealand (King 2011). Their teacher, Mr J, has asked them to complete an activity in which they must ask each other questions about heterosexuality; questions typically asked only about homosexuality. As the lesson is set up by their 130

Gender and sociopragmatics

teacher and the students engage with it, positionings are inferred. In the following two examples, pronoun choice positions those present in very different ways. (6) Context: Students in a sexuality education lesson have just been completing the activity. 1. Ruby: [reading] do you think straights flaunt their sexuality (1) if so (.) why 2. Callum: um 3. Ruby: I think SOME straights do 4. Callum: yeah like (.) it would be easier for them to like (1) 5. y’know flaunt their sexuality because more people out there are straight? but (.) if like (.) someone who was homosexual 6. 7. was around homosexual people they’d be the same? 8. Ruby: true (King 2011) (7) Context: Following the activity in (6), the teacher asks them how they felt about the questions. 1. Mr J 2. 3. Caitlin 4. 5. 6. 7. Mr. J 8. 9.

unless anyone else wants to i’m gonna ask Caitlin (5) are you ready to answer ↑ i was just talking to Sarah about how it was quite easy because like (2) it’s about heterosexuality and like we ARE↑ and so like it makes it more comfortable cool (1) so the assumption there is that (3) WE means the class↑ (King 2011)

In (6), Callum uses the third-person plural pronouns them and they to refer to both heterosexual people and homosexual people (lines 4, 5, and 7). As a result, he positions neither Ruby nor himself (nor anyone else in the room) as either. This sits in contrast with (7) in which Caitlin says the questions were easy to respond to because we are heterosexual (line 5). Mr J then conducts his own pragmatic analysis of this, pointing out that there is an assumption buried in her use of the pronoun we (line 9). She has positioned the whole class as heterosexual and, by doing so, foreclosed on other possibilities. This tendency to position everyone in the room as heterosexual can be very powerful and insidious regardless of a speaker’s intentions or level of awareness. In addition to the pronominal effect exemplified earlier, (8) demonstrates that nonheteronormative assumptions can subtly keep the sexualities of people in the room open via other linguistic means. This example also demonstrates that this “opening up” is rather unstable and can quickly be undermined, yet again. In this excerpt, students are about to start the discussion activity. Mr J is outlining the purpose of the task. Italicised, underlined sections comprise a small group conversation taking place at the same time as the whole class discussion, which appears in larger type. (8) 1. Mr J  2. 3. 4.

now (1) these are questions about sexual orientation now (2) for answering these questions we’re asking that you (.) um think aBOUT the question and think about the BEST way you can answer that question 131

Janet Holmes and Brian W. King

  5. all right ((papers rustling)) the questions are all about heterosexuality   6. does everyone know what that means  7. Matt yeah=  8. Luana =no  9. Matt different 10. Liam normal 11. Codey can you /define it\ 12. Liam /as in\ 13. Callum boy likes girl 14. Liam yeah 15. Mr J define it↑ 16. it’s /people who are interested in the opposite sex\ 17. Liam /that’s what WE conside\/ normal 18. Codey EH ((heightened pitch)) 19. Mr J ok↑ 20. Luana 21. Luana yes 22. Mr. J do you understand what a heterosexual is 23. Luana yeah he’s telling me ((indicating Liam)) 24. Mr J /people interested in the opposite sex\ 25. Liam /people in the opposite sex\ 26. into the opposite sex 27. Mr J now 28. Callum so not gay 29. Luana you mean interested in men 30. Liam that’s the SAME sex 31. Callum just straight not gay 32. Liam /yeah\ 33. Mr J /we’re looking\ at ((picks up paper)) 34. assumptions societal assumptions and myths about homosexuals 35. 36. but the WAY we’re doing that is by answering a lot of questions about heterosexuals 37. 38. (1) does everyone understand what we’re doing↑ 39. Many yeah/yes/mhm [but murmured only] 40. Mr J the outcome is we’re looking at the societal myths of (.) about homosexuals 41. 42. myths and assumptions (King 2011: 123–124) Mr J’s language is carefully worded in these instructions, presumably so as to avoid ascribing to the students any particular sexual orientation. Pragmatically, this avoidance eventuates from the absence of the word “normal” from his definition of heterosexuals (i.e. people interested in the opposite sex – lines 16 and 24). In addition, Mr J repeatedly says heterosexuals and homosexuals, avoiding (consciously or unconsciously) any reliance on the pronouns “us” and “them”. For example throughout the passage he avoids saying “about them”, but rather says about homosexuals (lines 35 and 41), and such an avoidance of pronouns is

132

Gender and sociopragmatics

marked in English (Wales 1996: 30–31). However, Liam does suggest (only to his group) that we consider heterosexuality normal (lines 10 and 17). This statement works against Mr J’s lack of normalisation of heterosexuality, and (similarly to [7]) the inclusive we to some extent ascribes heterosexuality to all present via its assumption that we are all people who possess a heteronormative gaze. This slippage provides insight into the stability of heterosexuality. Another theme which pragmatic analysis can investigate is the power inherent in heterosexuality’s tendency to seem “invisible” or at least “transparent” enough not to be noticed. Example (9) demonstrates that naming heterosexuality actually confuses some students. This is because they do not know the word and so infer that its referent cannot be normal because it is being named; it must be something “queer.” We would like to propose that these pragmatic functions can be called gazing and averting, processes in which minority sexualities are “gazed at” whilst that gaze is “averted” from heterosexuality. (9) Context: Students are completing the discussion activity described in (6).  1. Luana: [reading] how can you tell if someone is hesterohex- het-ter-ro-sexual straight   2. Lito: what’s (1) hetero (.) sexual   3. Luana: (6) like you know how Amiria is a girl but she likes being a boy?   4. Lito: oh a TOMboy   5. Callum: no that’s not=   6. Luana: =OH no that’s NOT it   7. Callum: um (.) well you can’t really tell by looking at someone   8. but (.) if you see them with like (.) a member of the opposite sex   9. in like a relationship well I guess you can tell that they’re heterosexual 10. Luana: AH ok 11. Callum: heterosexual’s just the opposite of HOMosexual 12. Luana: (5) like so if a gay person is (1) gay 13. Lito: 14. Callum: they’re homosexual if a boy likes a boy they’re homosexual 15. 16. if a boy likes a girl they’re heterosexual? 17. Luana: AWW /DUHuhhUHH\ 18. Lito: /so that’s\ normal 19. Luana: so a girl likes a boy that’s heter- (reading) heter-ro-sex-ual 20. Callum: yeah 21. Luana: NOW I get it (King 2011) Thus we see that Luana presumes heterosexuality must mean the same thing as tomboy (line 3), and begins to apply her gaze to it as a marked positioning. But Callum corrects her and explains the term’s meaning (lines 7–9). It is interesting to note that he avoids using the term “normal” and thus avoids speaking heteronormatively (just as he does with pronoun use in [6]). However, heteronormativity raises its head again in Lito’s final turn, as she replaces the term heterosexual with normal (line 18). Once again our gaze is averted from heterosexuality.

133

Janet Holmes and Brian W. King

Future directions Fruitful research into gender and sociopragmatics is needed in numerous areas, but a topical avenue would be to pursue multimodal analysis. Although research into multimodal sociopragmatics is beginning to proliferate (see Norris 2011), at present there is a dearth of scholarship focusing on this topic in relation to gender. In the area of media advertising some have begun to do so (see e.g. Velasco-Sacristán and Fuertes-Olivera 2006; Del Saz-Rubio and PennockSpeck 2009). Additionally, analyses of interaction between social actors have also benefitted from a multimodal pragmatic approach (e.g. Ferré 2011). In fact, Goodwin and colleagues (2002) incorporated gender into their investigation of body language, movement, and inference a decade ago, but there has been little (if any) similar published research since. These studies have all opened vistas for research which will hopefully soon be explored in more detail. Sociopragmatic research in the area of language and gender has moved through several stages in the past four decades. An early focus on ways in which women’s ways of talking could be distinguished from men’s speech styles evolved into more contextualised and nuanced approaches which recognised that pragmatic strategies and devices constituted resources available for constructing complex social identities, including gender identities. Treating gender as a performance led to an appreciation of the gendered politics that exist within femininity and masculinity.“Queer” approaches fruitfully questioned the “taken-for-granted” status of heterosexuality (i.e. heteronormativity), providing convincing evidence of the importance of exploring the pragmatic inferences we all constantly make in everyday interaction.

Conclusion Exploring the themes of power and politeness, this chapter has illustrated the richness of sociopragmatic research which focuses on the relationship between gender and language. Using data from workplace contexts, the analysis has suggested how power and authority may be constructed through subtle discourse choices, such as strategic questions and challenges and skilfully positioned summarising, and masculinity performed through the use of contestive and aggressive discourse, including, in some contexts, jocular abuse and swearing. Similarly, good collegial relations are created and maintained through the skilful use of pragmatic strategies such as complimenting, expressing approval, and, more subtly, through hedging disagreement and criticism, and the strategic use of humour and small talk. Data from sexuality education classrooms has highlighted the usefulness of sociopragmatic research in exploring the stabilisation of heterosexuality. This takes place regardless of speaker intention in some cases, through nuanced usage of pronominal reference. Additionally, the avoidance of heteronormative talk can be achieved through the avoidance of pronouns. In addition, we have seen the power of “averting our gaze” from heterosexuality, as students secure their understanding of the term through its “normal”, unremarkable status. In sum, this chapter has surveyed and illustrated a range of approaches with the aim of demonstrating the insights that a sociopragmatic analysis of the relationship between language and gender can provide.

Suggestions for further reading Eckert, P. and McConnell-Ginet, S. (2013) Language and Gender, 2nd ed. Oxford: Blackwell. A definitive social constructionist approach to the analysis of gendered discourse, including discussion of “the gender order”.

134

Gender and sociopragmatics Ehrlich, S., Meyerhoff, M., and Holmes, J. (eds.) (2014) The Handbook of Language, Gender and Sexuality. Malden, Oxford: Wiley-Blackwell. This is a comprehensive reference text with chapters by leading scholars on a wide variety of different theoretical approaches to the relationship between language, gender, and sexuality. Mills, S. and Mullany L. (2011) Language, Gender and Feminism: Theory, Methodology and Practice. London: Routledge. This book provides a thorough and accessible introduction to contemporary language and gender studies and addresses in detail the differences and similarities between second-wave and third-wave feminist analysis. Motschenbacher, H. (2010) Language, Gender and Sexual Identity: Poststructuralist Perspectives. Amsterdam: John Benjamins. This is an accessible introduction to queer linguistics which provides a clear outline of language, gender, and sexual identity scholarship. It is a useful resource for gaining understanding of critical approaches to heteronormativity and gender binarism. Talbot, M. M. (1998) Language and Gender: An Introduction, 2nd ed. Oxford: Polity Press. This is a very readable and comprehensive introduction to language and gender research with especial attention to critical discourse analysis approaches and evidence of gendered discourse in the media.

Notes 1 We would like to thank Sharon Marsden for her assistance with references and a literature survey which we have made use of in this chapter. 2 See also Coates (1996, 2003). 3 Saito (2011) makes the same point about Japanese leaders. 4 This example is slightly edited for ease of reading.

References Bahktin, M. M. (1994) The Bakhtin Reader: Selected Writings of Bakhtin, Medvedev and Voloshinov. London: Edward Arnold. Baxter, J. (2003) Positioning Gender in Discourse: A Feminist Methodology. Basingstoke: Palgrave Macmillan. Baxter, J. (2010) The Language of Female Leadership. Basingstoke: Palgrave Macmillan. Baxter, J. (2011) ‘Survival or success? A critical exploration of the use of “double-voiced discourse” by women business leaders in the UK’, Discourse & Communication, 5(3): 231–246. Baxter, J. and Wallace, K. (2009) ‘Outside in-group and out-group identities? Constructing male solidarity and female exclusion in UK builders’ talk’, Discourse & Society, 20(4): 411–429. Bing, J. M. and Bergvall,V. L. (1996) ‘The question of questions: Beyond binary thinking’, in V. L. Bergvall, J. M. Bing, and A. F. Freed (eds.) Rethinking Language and Gender Research: Theory and Practice. London: Longman. 1–30. Brown, P. and Levinson, S. (1987) Politeness. Cambridge: Cambridge University Press. Bucholtz, M. (2003) ‘Theories of discourse as theories of gender’, in J. Holmes and M. Meyerhoff (eds.) The Handbook of Language and Gender. Oxford: Blackwell. 43–68. Cameron, D. (1998) ‘ “Is there any ketchup,Vera?”: Gender, power and pragmatics’, Discourse & Society, 9(4): 437–455. Cameron, D. (2005) ‘Relativity and its discontents: Language, gender, and pragmatics’, Intercultural Pragmatics, 2(3): 321–334. Chan, A. (2007) ‘Same context, different strategies: A company director’s discourse in business meetings’, Journal of Asian Pacific Communication, 17(1): 61–81. Chan, A. (2008) ‘Meeting openings and closings in a Hong Kong company’, in H. Sun and D. Z. Kádár (eds.) It’s the Dragon’s Turn: Chinese Institutional Discourse(s). Bern: Peter Lang. 181–229. Crawford, M. (1995) Talking Difference: On Gender and Language. London: SAGE. Crawford, M. (2003) ‘Gender and humor in social context’, Journal of Pragmatics, 35(9): 1413–1430. Coates, J. (1996) Women Talk: Conversation between Women Friends. Cambridge, MA: Blackwell Publishers. Coates, J. (2003) Men Talk: Stories in the Making of Masculinities. Oxford: Blackwell. Dubois, B. L. and Crouch, I. (1975) ‘The question of tag questions in women’s speech:They don’t really use more of them, do they’, Language in Society, 4(3): 289–294.

135

Janet Holmes and Brian W. King Eakins, B. and Eakins, G. (1979) ‘Verbal turn-taking and exchanges in faculty dialogue’, in B.-L. Dubois and I. Crouch (eds.) The Sociology of the Languages of American Women. San Antonio, TX: Trinity University. 53–62. Eder, D. (1993) ‘ “Go get ya a French!” Romantic and sexual teasing among adolescent girls’, in D. Tannen (ed.) Gender and Conversational Interaction. Oxford: Oxford University Press. 17–31. Evans Davies, C. (2006) ‘Gendered sense of humor as expressed through aesthetic typifications’, Journal of Pragmatics, 38(1): 96–113. Fairclough, N. (ed.) (1992) Critical Language Awareness. London: Longman. Ferré, G. (2011) ‘Functions of three open-palm hand gestures’, Multimodal Communication, 1(1): 5–20. Fletcher, J. K. (1999) Disappearing Acts: Gender, Power and Relational Practice at Work. Cambridge, MA: MIT Press. Freed, A. F. (1992) ‘We understand perfectly: A critique of Tannen’s view of cross-sex communication’, in K. Hall, M. Bucholtz, and B. Moonwomon (eds.) Locating Power: Proceedings of the Second Berkeley Women and Language Conference, (vol. 1). Berkeley, CA: Berkeley Women and Language Group. 144–152. French, J. R. P. and Raven, B. (1959) ‘The bases of social power’, in D. Cartwright and A. Zander (eds.) Group Dynamics. New York: Harper & Row. 150–167. Goodwin, M. H., Goodwin, C., and Yaeger-Dror, M. (2002) ‘Multi-modality in girls’ game disputes’, Journal of Pragmatics, 34(10–11): 1621–1649. Hay, J. (1995) ‘Gender and humor: Beyond a joke’, unpublished master’s thesis,Victoria University of Wellington, New Zealand. Hay, J. (2002) ‘Male cheerleaders and wanton women: Humour among New Zealand friends’, Te Reo, 45: 3–36. Holmes, J. (1984) ‘ “Women’s language”: A functional approach’, General Linguistics, 24(3): 149–178. Holmes, J. (1986) ‘Functions of you know in women and men’s speech’, Language in Society, 15(1): 1–21. Holmes, J. (1992) ‘Women’s talk in public contexts’, Discourse & Society, 3(2): 131–150. Holmes, J. (1995) Women, Men and Politeness. London: Longman. Holmes, J. (2000) ‘Politeness, power and provocation: How humour functions in the workplace’, Discourse Studies, 2(2): 159–185. Holmes, J. (2005) ‘Power and discourse at work: Is gender relevant?’, in M. Lazar (ed.) Feminist Critical Discourse Analysis. London: Palgrave. 31–60. Holmes, J. (2006a) Gendered Talk at Work. Oxford: Blackwell. Holmes, J. (2006b) ‘Sharing a laugh: Pragmatic aspects of humor and gender in the workplace’, Journal of Pragmatics, 38(1): 26–50. Holmes, J. (2014) ‘Language and gender in the workplace’, in S. Ehrlich, M. Meyerhoff, and J. Holmes (eds.) The Handbook of Language, Gender and Sexuality, 2nd ed. Hoboken: Wiley & Sons. 433–451. Holmes, J. and Major, G. (2002) ‘ “It’s getting a bit desperate isn’t it.” Communication on the ward’ (Staff Newsletter of Capital & Coast District Health Boards) Staff Matters, (41): 8. Holmes, J. and Marra, M. (2002a) ‘Humour as a discursive boundary marker in social interaction’, in A. Duszak (ed.) Us and Others: Social Identities across Languages, Discourses and Cultures, Amsterdam: John Benjamins. 377–400. Holmes, J. and Marra, M. (2002b) ‘Having a laugh at work: How humour contributes to workplace culture’, Journal of Pragmatics, 34(12): 1683–1710. Holmes, J., Marra, M., and Burns, L. (2001) ‘Women’s humour in the workplace: A quantitative analysis’, Australian Journal of Communication, 28(1): 83–108. Holmes, J. and Stubbe, M. (2015) Power and Politeness in the Workplace, 2nd ed. London/NewYork: Routledge. Jenkins, M. M. (1985) ‘What’s so funny? Joking among women’, in N. Caskey, S. Bremner, and B. Moonwomon (eds.) Proceedings of the First Berkeley Women and Language Conference. Berkeley, CA: Berkeley Women and Language Group, University of California. 131–151. Johnson, S. (1997) ‘Theorising language and masculinity: A feminist perspective’, in S. Johnson and U. H. Meinhof (eds.) Language and Masculinity. Oxford: Blackwell. 8–26. Keating, E. (2009) ‘Power and pragmatics’, Language & Linguistics Compass, 3(4): 996–1009. Keenan, E. (1989) ‘Norm-makers, norm-breakers: Uses of speech by men and women in a Malagasy community’, in R. Bauman and J. Sherzer (eds.) Explorations in the Ethnography of Speaking, 2nd ed. Cambridge: Cambridge University Press. 125–143. Kiesling, S. F. (1997) ‘Power and the language of men’, in S. Johnson and U. H. Meinhof (eds.) Language and Masculinity. Oxford: Blackwell. 65–85.

136

Gender and sociopragmatics King, B. W. (2011) Linguistic Negotiations of Sexual Agency in Sexuality Education. Wellington, NZ: Victoria University of Wellington dissertation. Available from: http://researcharchive.vuw.ac.nz/handle/10063/2273 (accessed 16 September 2015). Kotthoff, H. (2006) ‘Gender and humor: The state of the art’, Journal of Pragmatics, 38(1): 4–25. Kulick, D. (2005) ‘The importance of what gets left out’, Discourse Studies, 7(4–5): 615–624. Lakoff, R. (1973) ‘Language and woman’s place’, Language in Society, 2(1): 45–80. Lakoff, R. (1975) Language and Woman’s Place. New York: Harper & Row. Lazzaro-Salazar, M., Holmes, J., Marra, M., and Vine, B. (2016) ‘Doing power and negotiating through disagreement in public meetings’. Pragmatics & Society, 6(3): 444–464. Locher, M. A. and Watts, R. J. (2005) ‘Politeness theory and relational work’, Journal of Politeness Research, 1(1): 9–33. Maltz, D. H. and Borker, R. A. (1982) ‘A cultural approach to male-female miscommunication’, in J. J. Gumperz (ed.) Language and Social Identity. Cambridge: Cambridge University Press. 196–216. McElhinny, B. (1992) ‘ “I don’t smile any more”: Affect, gender and the discourse of Pittsburgh police officers’, in J. Coates (ed.) Language and Gender: A Reader. Oxford: Blackwell. 309–327. Morrish, L. and Sauntson, H. (2007) New Perspectives on Language and Sexual Identity, Basingstoke: Palgrave Macmillan. Motschenbacher, H. (2010) Language, Gender and Sexual Identity: Poststructuralist Perspectives. Amsterdam: John Benjamins. Mullany, L. J. (2007) Gendered Discourse in the Professional Workplace. Basingstoke: Palgrave Macmillan. Nelson, C. (1999) ‘Sexual identities in ESL: Queer theory and classroom inquiry’, TESOL Quarterly, 33(3): 371–392. Norris, S. (2011) ‘Three hierarchical positions of deictic gesture in relation to spoken language: A multimodal interaction analysis’, Visual Communication, 10(2): 129–147. O’Barr, W. and Atkins, B. (1980) ‘ “Women’s language” or “powerless language?” ’, in S. McConnell-Ginet, R. Borker, and N. Furman (eds.) Women and Language in Literature and Society. New York: Praeger. 93–110. Pizzini, F. (1991) ‘Communication hierarchies in humour: Gender differences in the obstetrical/gynaecological setting’, Discourse & Society, 2(4): 477–488. Plester, B. A. and Sayers, J. G. (2007) ‘Taking the piss: Functions of banter in the IT industry’, Humor, 20(2): 157–187. Ragan, S. L. (2000) ‘Sociable talk in women’s health contexts: Two forms of non-medical talk’, in J. Coupland (ed.) Small Talk. London: Longman. 269–287. Saito, J. (2011) ‘Managing confrontational situations: Japanese male superiors’ interactional styles in directive discourse in the workplace’, Journal of Pragmatics, 43(6): 1689–1706. Sauntson, H. (2008) ‘The contributions of queer theory to gender and language research’, in K. Harrington, L. Litosseliti, H. Sauntson, and J. Sunderland (eds.) Gender and Language Research Methodologies. New York: Palgrave Macmillan. 271–282. Del Saz-Rubio, M. M. and Pennock-Speck, B. (2009) ‘Constructing female identities through feminine hygiene TV commercials’, Journal of Pragmatics, 41(12): 2535–2556. Schnurr, S. and Chan, A. (2005) ‘ “Are you the project manager?” How a leader ensures the compliance of his subordinates’, in S. Schnurr, A. Terraschke, and A. C. K. Chan (eds.) Proceedings of the 1st International Linguistics and Literary Studies Postgraduate Conference. Wellington, NZ:Victoria University of Wellington. 27–37. Spencer-Oatey, H. (2008) ‘Face, impoliteness and rapport’, in H. Spencer-Oatey (ed.) Culturally Speaking: Culture, Communication and Politeness Theory. London: Continuum. 11–47. Talbot, M. (1998) Language and Gender: An Introduction. Oxford: Polity Press. Tannen, D. (1987) ‘Remarks on discourse and power’, in D.Tannen (ed.) Power through Discourse. Norwood, NJ: Ablex. 3–18. Tannen, D. (1990) You Just Don’t Understand:Women and Men in Conversation. New York: Morrow. Troemel-Ploetz, S. (1991) ‘Selling the apolitical’, Discourse & Society, 2(4): 489–502. Velasco-Sacristán, M. and Fuertes-Olivera, P. A. (2006) ‘Towards a critical cognitive-pragmatic approach to gender metaphors in advertising English’, Journal of Pragmatics, 38(11): 1982–2002. Wales, K. (1996) Personal Pronouns in Present-Day English. Cambridge: Cambridge University Press. Wodak, R. (1999) ‘Critical discourse analysis at the end of the 20th century’, Research on Language and Social Interaction, 32(1–2): 185–193.

137

Janet Holmes and Brian W. King

Transcription conventions [laughs] : : SOME . . . . . .. \ . . . . . . /. . . . . .. \\ . . . ( ) (hello) .°.°. = (1) (.) ↑  

Paralinguistic features and editorial information in square brackets, colons indicate start/finish Strong stress Simultaneous speech Unclear utterance Transcriber’s best guess at an unclear utterance Section of transcript omitted Turn continuation Elapsed time of silence in seconds Tiny gap, less than one second Rising intonation in the preceding syllable Falling and then rising intonation in the preceding syllable Rising and then falling intonation in the preceding syllable

138

13 BILINGUALISM AND MULTILINGUALISM1 Jasone Cenoz

Introduction The increasing mobility of the population, Internet communication and the spread of English as the language of wider communication have contributed to the development of bilingualism and multilingualism all over the world.The term bilingualism, usually referring to ‘two languages’ (Greek prefix ‘bi’ = two), has been the term most widely used because most research has focused on two rather than three or more languages. However, the term multilingualism (Latin prefix ‘mult ’= much, many) has gained currency in recent years and usually refers to two or more languages, including bilingualism as a type of multilingualism. The terms bilingualism and multilingualism are used in different ways and can refer to individual competence, but also to language use in society. Indeed, the European Commission defines multilingualism as ‘the ability of societies, institutions, groups and individuals to engage, on a regular basis, with more than one language in their day-to-day lives’ (European Commission 2007: 6). This definition has two characteristics. It covers both the individual and social levels and is based on language use. We deal with each in turn in the following.

Multilingualism: an individual and a social phenomenon Individual and social multilingualism are linked to each other insofar as it is more likely that individuals living in a multilingual community speak more than one language in their day-today lives than those in monolingual contexts. The Council of Europe (2007: 8) uses the term plurilingualism (Latin prefix ‘pluri’ = several, many) to refer to the ‘repertoire of varieties of language which many individuals use’. According to this view, a person as a social agent: has proficiency of varying degrees, in several languages, and experience of several cultures. This is not seen as the superposition or juxtaposition of distinct competences, but rather as the existence of a complex or even composite competence on which the user may draw. (Council of Europe 2001: 168)

139

Jasone Cenoz

The term plurilingualism (fr. plurilingualisme; esp. plurilingüismo) is often used with the same meaning as individual multilingualism, but is also widely used in some countries, such as France or Spain, instead of the term multilingualism. However, plurilingualism is also frequently seen to highlight the interaction between the languages and individual agency to a larger extent than multilingualism (Moore and Gajo 2009; Canagarajah and Linayage 2012). The scope of multilingualism is broad, not only because it is a world phenomenon, but also because its study can be approached from a range of different perspectives, including linguistics, psycholinguistics, sociolinguistics, anthropology or education. One dimension of multilingualism that links language proficiency to language use is receptive multilingualism, which ‘refers to language constellations in which interlocutors use their own language while speaking to each other’ (Bahtina and Ten Thije 2012: 1). Receptive multilingualism has a long tradition in some areas of Europe such as Scandinavia. Another distinction that has a long tradition is that between additive and subtractive multilingualism. Baker (2011: 72) relates these dimensions of bi/multilingualism to the enrichment or loss of the minority language when learning a second language. In the case of additive bi/ multilingualism, the second language does not prevent the development of the first language, while in subtractive bilingualism the second language replaces the first language. This distinction is often related to élite and folk bi/multilingualism, and is closely linked to the status of the languages. Élite bi/multilingualism takes place when speakers of high-status languages, such as English, French or German, learn additional high-status languages either as a personal choice or as part of their education. In the case of folk bi/multilingualism, speakers of minority languages, such as Welsh or Quechua, have the obligation to learn the national language.

Multilingualism: focus on language use As we have seen, the European Commission’s definition of multilingualism is based on the use of two or more languages rather than on proficiency. This is a common perspective currently. Lüdi and Py (2009: 158), for instance, consider multilinguals to be those who use two languages and who can switch from one language to the other. Similarly, Li defines a multilingual individual as ‘anyone who can communicate in more than one language’ (2008: 4).These definitions based on use are in clear contrast with maximalist definitions requiring ‘native control of two languages’ such as the one proposed by Bloomfield (1933: 55). Communicating in more than one language implies competence in several areas as it is defined in different models of communicative competence (Canale and Swain 1980; Bachman 1990; Celce-Murcia et al. 1995). One of these areas is pragmatic competence, defined as ‘the competence in conveying and understanding communicative intent’ (Celce-Murcia et al. 1995: 17). The focus of bilingualism/multilingualism research on language use links the field to that of interlanguage pragmatics, ‘the branch of second language research which studies how nonnative speakers. . . . understand and carry out linguistic action in a target language and how they acquire L2 pragmatic knowledge’ (Kasper 1992: 203). Research on interlanguage pragmatics has looked at how learners produce, understand and acquire communicative action in a second language. It is a research area which has undergone considerable development since the 1990s (Barron 2012). Research studies have looked at the linguistic resources used to express speech acts and social actions and to analyse interactional competence (Ishida 2013; Taguchi, Chapter 14, this volume). Similarly, the focus on language use in research on bilingualism/multilingualism means that the field also intersects with that of intercultural pragmatics, intercultural pragmatics having, according to Kecskes (2014: 18), its focus on ‘intercultural communication that involves interactions among people from different culture using a common language’ (see also Cogo 140

Bilingualism and multilingualism

and House, Chapter 15, this volume). In the remainder of this chapter, research on the main areas of bi/multilingualism will be discussed. Special attention will be paid to the contribution of this research to pragmatics.

Current work Current work on bilingualism and multilingualism covers many areas and uses different research methodologies. In this section, we summarise the main lines of research in the study of bi/multilingualism at both the individual and social levels. First we look at the effect of bilingualism on metalinguistic awareness and the acquisition of additional languages, then we focus on the multilingual mind, the process of becoming bi/multilingual, multilingualism as a social construct and finally on minority languages.

The effect of bilingualism on metalinguistic awareness and the acquisition of additional languages One of the main areas of study in individual bi/multilingualism takes a psycholinguistic approach and looks at differences between monolinguals and bilinguals in some aspects of language processing. Bilinguals have been reported, for instance, to develop a higher level of metalinguistic awareness than monolinguals, with metalinguistic awareness referring to the ability to reflect about language and focus objectively on its structure and functions. Advantages associated with bilingualism have been reported in selective attention and inhibitory control (Bialystok et al. 2008). Craik and colleagues (2010) also report that bilingualism can delay the onset of Alzheimer’s disease. Enhanced metalinguistic awareness has been associated with advantages for the acquisition of additional languages.The fact that bi/multilinguals already have at least two linguistic systems at their disposal and also the fact that they have had experience acquiring a second language can explain bilinguals’ faster rates of acquisition in third language acquisition as compared to monolinguals learning a second language (Cenoz 2009). The differences between monolinguals and bilinguals have also been observed in the study of speech acts. Safont Jordà (2005) and Safont Jordà and Alcón Soler (2012) confirmed that bilinguals outperformed monolinguals in the use of requests and request modifiers. Safont Jordà (2005) compared Catalan-Spanish bilinguals to Spanish monolinguals learning English and showed that bilinguals had developed a higher level of pragmalinguistic awareness when formulating requests. Also, Safont Jordà and Alcón Soler (2012) reported that bilinguals made more progress than monolinguals when learning internal and external modifiers in English as a third language relative to monolinguals learning a second language. Alcón Soler (2012) analysed metapragmatic awareness by comparing two types of bilinguals, productive and receptive, learning an additional language. The study by Alcón Soler includes a pedagogical intervention to develop metapragmatic awareness in the speech act of refusal. Productive bilinguals developed a higher level of pragmalinguistic awareness than receptive bilinguals as a result of the pedagogical intervention. These findings associate a higher level of bilingualism with advantages in the acquisition of pragmatic competence. Alcón Soler (2012) explains that this can be due to bilinguals’ enhanced communicative sensitivity, higher capacity to store information, better control over intended meaning or better control over the variables affecting language use. The comparisons between monolinguals and bilinguals in the acquisition of pragmatic competence in a third language have shown that it is necessary to take previous language learning experience(s) into account as an additional factor in language learning. Having said that, 141

Jasone Cenoz

however, it is also necessary to conduct such studies in other contexts and with other language combinations and for other speech acts to see if these findings can be supported.

The multilingual mind Neurolinguistics is an area of bilingualism and multilingualism that has undergone important developments recently. Technological developments such as the use of event-related-potentials (ERP), functional magnetic resonance imaging (fMRI) or positron emission tomography (PET) allow for the analysis of some brain functions that were previously inaccessible.These techniques open up new possibilities in the study of bilingual and multilingual processing. However, results are in many cases still exploratory. Psycholinguistic research in bilingualism and multilingualism has looked at language comprehension and production in the multilingual mind. The area that has received most attention in recent years is the multilingual lexicon, with the mechanisms of lexical access being explored with the aim of investigating if all languages of a multilingual are activated at the same time and also of exploring the factors influencing the activation (see e.g. Dijkstra 2009). Some studies have focused on pragmatics. For example, Siegal and colleagues (2009) analysed the differences in pragmatic competence in Italian monolingual and Slovenian-Italian bilingual children. They used a Conversational Violations Test consisting of twenty-five short conversational exchanges. The children, who were four to six years old, were asked to choose between two utterances, one that violated Gricean conversational maxims and one that did not. Bilingual children outperformed monolingual children. In a second experiment reported in the same article, the performance of Italian monolingual, Slovenian monolingual and SlovenianItalian bilingual children was compared. The results of the second experiment confirmed the advantages of bilingual children in detecting violations of conversational maxims. According to Siegal and colleagues (2009), this finding may be due to a cognitive compensatory mechanism by means of which the development of pragmatic competence somehow compensates for the delay in vocabulary development in the weaker language (a typical feature of early bilinguals). Research on the multilingual mind has also looked at conceptual representation. The idea that differences between monolinguals and multilinguals in the conceptual base are not only quantitative but also qualitative in nature has gained currency recently (Kesckes 2010). Pavlenko (2011) explores these differences further and shows how the acquisition of additional languages implies cognitive restructuring as seen in visual perception, inner speech or gesturing (see also Jarvis and Pavlenko 2008; Cook and Bassetti 2011). Cognitive restructuring takes place at different levels. In the case of the lexicon, bilinguals do not just have two labels for one object. In many cases, cognitive restructuring occurs because lexical categories are not the same in different languages and some concepts are culture-specific. Pavlenko and Malt (2011) reported on how Russian and English speakers differed in the way they name drinking containers such as cup, mug and glass.

Becoming bi/multilingual One of the areas that has attracted increased attention in recent years is early bilingualism and multilingualism. Most studies look at the simultaneous acquisition of two languages and compare bilingual development to first language acquisition in monolinguals (De Houwer 2009). These studies are usually longitudinal and focus on specific aspects of phonetics, grammar, vocabulary or pragmatics. An important issue in the area of pragmatics is the development of individuals’ ability to use each language in the appropriate context. Several factors have been related to the 142

Bilingualism and multilingualism

development of appropriate language choice, including the one parent-one language policy, parental discourse or relative language proficiency (Mishina-Mori 2011). There are also some studies on early trilingualism. Barnes (2006) reports the development of questions and their functions in English by a trilingual child in Basque, Spanish and English. This study focuses on language development in a situation in which exposure to English is almost limited to the home context. The child’s questioning behaviour in English shares some of the characteristics of the development of questions reported in studies conducted in English-speaking contexts. At the same time, the child’s development of pragmatic competence reveals some particular characteristics because of the interaction with the other two languages. In another study, Barnes (2011) compared the speech acts used by the mother, a native speaker of English, and the same child.The results of the study reveal that the child is able to formulate a mitigated request and a non-imperative request, which are less common in Basque and Spanish than in English. Barnes also reports that the child’s production is not just a reflection of the mother’s request productions. Becoming bi/multilingual is very common in school contexts. Schools can have an important influence on language learning, but also on the status and values associated with different languages.There are important decisions about which languages are studied as subjects or which languages are used as languages of instruction. In many cases, the language that schoolchildren use at home is not taught at school. Nowadays English is taught as a second or foreign language in most European schools outside English-speaking countries, and the use of English as a language of instruction is becoming popular in CLIL and immersion programs (Cenoz 2015). In some cases, English is taught along with other regional and state languages. Research on language learning in school contexts is at the intersection of bi/multilingualism and second/ third language acquisition studies. Some students may already be early bilinguals and others can acquire a second language at school and become bilingual while they are also learning an additional language. The goal of an educational program can be bilingualism and multilingualism, and this is achieved through the process of acquiring second and third languages (Gorter 2013). Cross-linguistic influence is a central focus of research when more than two languages are involved. In this area, multidirectionality (Jarvis and Pavlenko 2008) has proven an important concept. Speakers can, for instance, use resources from their first language in their second language and from their first and/or second language in their third language. However, the second language can also influence the first language and the third language can influence the second and/or the first language. This was shown in a study of pragmatic competence reported by Safont Jordà (2013). She conducted a longitudinal study dealing with requestive behaviour in three languages (i.e. Catalan, Spanish and English) over a period of 2.6 to 3.6 years. The introduction of a third language (i.e. English) in the subject’s linguistic repertoire promoted the use of conventionally indirect forms in the three languages, with the child’s requestive behaviour in Catalan and Spanish changing as a result of learning how to make requests in English. Portolés (2015) conducted a study with children in the same contexts and reported that pragmatic awareness increased with multilingual proficiency. Fouser (2001), on the other hand, is a further study that has looked at cross-linguistic interaction in the acquisition of a third language. He examined the influence of L2 Japanese into L3 Korean in the production of speech acts by two speakers of English as a first language. Fouser used a Discourse Completion Task (DCT), a language choice questionnaire and a short writing task. In contrast to these studies, results revealed very limited transfer from Japanese (L2) into Korean (L3) in the formulation of speech acts. In another study on the acquisition of pragmatic competence in the third language, Koike and Palmiere (2010) did not find clear patterns of transfer from the first language in the acquisition of Portuguese as a third language by English L1 speakers proficient in Spanish, Spanish-English bilinguals and Spanish L1 speakers. 143

Jasone Cenoz

Multilingual practices, multilingualism as a social construct and multilingual identities Research on bilingualism and multilingualism has also focused on the way bi/multilinguals communicate with specific attention paid to code-mixing, code-switching and translanguaging. Code-mixing generally refers to switches between languages at the word level, while codeswitching describes any switch at both the word and sentence levels (Baker 2011: 107). Li and Wu (2009: 193) point out that code-switching is ‘the most distinctive behaviour of the bilingual speaker’. The study of code-mixing and code-switching has a long tradition, and has been carried out from a grammatical, psycholinguistic and sociolinguistic perspective (GardnerChoros 2009). The term translanguaging is a translation of the Welsh trawsieithu, which Williams (2002) first used to refer to the educational practice of alternating the use of Welsh and English in the classroom. Nowadays its scope is broader and it refers to multiple discursive practices. García and Sylvan (2011: 389) define translanguaging as ‘the process by which bilingual students and teachers engage in complex discursive practices in order to “make sense” of, and communicate in, multilingual classrooms.’ Li considers that translanguaging includes the full range of linguistic performances of multilingual language users for purposes that transcend the combination of structures, the alternation between systems, the transmission of information and the representation of values, identities and relationships. (2011: 1223) The concept of translanguaging goes further than an alternation between linguistic codes and challenges the idea of codes as fixed entities. The idea is that speakers use linguistic features at their disposal to communicate. The concept of translanguaging is different from code-switching because it considers that the different languages that form the multilingual speaker repertoire are integrated (García and Li 2014; Cenoz and Gorter 2015). Code-switching and translanguaging are seen as resources in bilingual and multilingual communication that speakers also use to establish social roles and identities (Zhu 2008). As Romaine (1995: 143) pointed out some years ago, bilingual speakers can mix and borrow when they do not know a word, but very often they just have a wider choice of linguistic means to express their communicative intent than monolingual speakers. Baker (2011: 108–109) presents thirteen functions for code-switching. Many of these are clearly relevant to pragmatics. Examples include the use of mixed utterances to reinforce a request, to clarify a point or to change the tone of the conversation. For example, a teacher in a second language class can switch to the students’ first language to reinforce a command or a request. Some trends in the study of bilingualism and multilingualism adopt a more sociological and anthropological perspective and have been influenced by the work of the French sociologist Pierre Bourdieu (1991).This research often adopts a poststructuralist critical approach and looks at the way the use of linguistic resources can be constrained by institutional ideologies and how linguistic practices represent power relations (Heller 2007). Multilingual identities have also been described as dynamic and subject to negotiation (Blackledge and Creese 2010; cf. also Van De Mieroop, Chapter 16, this volume). Language choice and translanguaging are not only dependent on the availability of the linguistic resources that the multilingual individual has at his/her disposal, but also on choices individuals make in the construction of identity. Indeed, translanguaging can also serve as an identity marker for 144

Bilingualism and multilingualism

some groups. Studies of multilingual communication show how speakers may cross ethnic and social borders and use languages and varieties that belong to other groups (Jørgensen 2008; Rampton and Charalambous 2012).

Minority languages Research on the sociolinguistic and educational aspects of bilingualism and multilingualism has undergone important development in areas where minority languages are spoken. Speakers of minority languages are bilingual or multilingual because they need to learn at least one majority language, and very often an additional international language. In contrast, speakers of majority languages can often ‘decide’ if they wish to become bilingual or multilingual. The use of minority languages in education is usually the result of language planning as an effort to protect and develop their acquisition and use, and, indeed, lesser used languages can provide a sense of identity – apart from being used for communication they can carry symbolic value. However, as Gorter (2013) points out, schools alone cannot safeguard the future of a minority language. Language choice is one of the most interesting aspects of pragmatic research when a minority language is used at school. Bilingual children are influenced by the status of the majority language and often use the majority rather than the minority language even when they are expected to use the minority language (see e.g. Thomas et al. 2012). Gorter (2013) also points out that minority language revitalisation processes have become multilingual because they do not involve only the minority and majority language, but also international and immigrant languages.

Critical issues and topics Research on bilingualism and multilingualism has traditionally taken monolingualism as a reference. Language development in early bilinguals is systematically compared to monolingual children of the same age so as to see if their development is satisfactory according to the monolingual norm (Baker 2011: 9–10). Similarly, older learners and users of several languages are expected to have the same linguistic abilities as the monolingual speakers of each of the languages in their linguistic repertoire (Cenoz and Gorter 2011). It is quite common to refer to delays on some aspects of early bilingual development, such as vocabulary, or to consider that bilinguals and multilinguals are deficient as compared to monolingual speakers. Cook (2012), who proposed the concept of multicompetence some years ago, considers that a bilingual person is different from a monolingual and cannot be seen as a deficient monolingual. He gives two arguments to justify the use of multicompetence as a different type of competence from that of monolingual speakers.The first is the spread of multilingualism particularly in Asian and African communities where multiple languages are used. The second reason is related to the rights of individuals because using the native speaker as a reference is judging a group by the standards of another (see also Firth and Wagner 1997; Seidlhofer in press). In fact, multilingual speakers use languages in a different way to monolinguals. Monolingual speakers use one single language in all situations and for all domains, but multilingual speakers do not use each of their languages for the same purposes in all communicative situations, in the same domains or with the same people (Moore and Gajo 2009). Furthermore, multilingual speakers can navigate between languages and use elements of different languages when communicating with other multilinguals. As we have already seen, multilinguals have particular characteristics at the cognitive level (Kesckes 2010; Cook and Bassetti 2011; Pavlenko 2011). It is also important to take into account that multilingual competence is not fixed but dynamic and linked to exposure and language use (Jessner 2009). 145

Jasone Cenoz

An important related point is the fixed vs. permeable boundaries between languages. Traditional approaches to the study of bilingualism and multilingualism have analysed languages separately as if they were completely independent of each other (Cenoz and Gorter 2011). However, the study of the outcomes of bilingualism and the study of cross-linguistic interaction and multilingual language practices show that multilingual speakers use languages as a resource without establishing rigid boundaries between them. Research that analyses the written production of multilingual children in two and three languages has shown that general writing strategies transfer cross-linguistically (Cenoz and Gorter 2011; Soltero-González et al. 2012).This indicates that there is interaction between the languages and that multilingual speakers use their linguistic repertoire as a resource even when languages are kept separate in school contexts. The results of these studies can have important implications for language teaching because in multilingual educational programs teachers could use their student’s linguistic resources when teaching languages. Cenoz and Gorter (2011) advocate a ‘focus on multilingualism’ that takes into account the characteristics of the multilingual speaker, the interaction between languages and the multilingual practices in the social context for research on bilingualism and multilingualism. A multilingual perspective that rejects the traditional monolingual approach to the study of bilingualism, multilingualism and second language acquisition is a critical issue that is considered revolutionary because it goes against the deep-rooted tradition of the native speaker as a reference in language learning (Kramsch 2012). Taking into account that this multilingual approach to bilingualism and multilingualism challenges the whole research tradition in applied linguistics, it is not surprising that the studies that have explored pragmatics using this approach are very few and exploratory. Already some years ago, Kecskes and Papp (2000) found that multilingual speakers transfer concepts and skills not only from the first to the second language, but also from the second language to the first. They attributed this interaction to the existence of a common underlying conceptual base. Almost a decade before, Blum-Kulka (1991) had proposed the Intercultural Style Hypothesis to refer to an intercultural speech style involving elements of two languages but distinct from either. This hypothesis is based on research studies showing that requests issued in English (L1) and Hebrew (L2) by American immigrants to Israel who were fully competent in the two languages differed significantly from both Israeli and American patterns. Similarly, Cenoz (2003) carried out a study on the formulation of requests in English by university students in the Basque Country. She reported that advanced learners of English developed an intercultural style common to Spanish and English with some similarities and differences as compared to native speakers of Spanish and English. The fact that second language learners differ from native speakers of the target language when using the target language is not remarkable if we take into account that most second language learners do not achieve native L2 proficiency. The main finding of these studies is the influence of the second language on the first and the development of an intercultural style which is common in the two languages of these emergent multilinguals. It is necessary to ‘focus on multilingualism’, as Cenoz and Gorter (2011, 2015) propose, and look at the multilingual speaker’s whole linguistic repertoire to understand this intercultural style.

Future directions Research on bilingualism and multilingualism has undergone many developments in recent years. Bilinguals and multilinguals have some particular characteristics when compared with monolinguals. Research on cognitive conceptualisation and metalinguistic awareness show that there are differences between monolinguals and bi/multilinguals at the cognitive level because one of the implications of becoming bi/multilingual is that the content of existing concepts is modified (Kecskes 2010; Pavlenko 2011). At the same time, monolinguals and multilinguals have 146

Bilingualism and multilingualism

been shown to differ in some aspects of metalinguistic awareness (Bialystok et al. 2008). Research on language practices shows how bilinguals and multilinguals use their linguistic resources in their repertoires in different ways from monolinguals by code-switching and translanguaging. As we have already seen, these findings indicate that bilinguals and multilinguals are different from monolinguals. This calls for research approaches that focus on multilingualism as an alternative to the strong tradition of looking at languages separately (Cenoz and Gorter 2011, 2015). Research on bilingualism and multilingualism opens new directions in pragmatics as well. One direction is to continue exploring the acquisition of pragmatic competence by comparing second and third language learners. This line of research has looked at the acquisition of a limited number of speech acts and has reported advantages in metapragmatic awareness on part of bilinguals (Safont Jordà 2005; Alcón Soler 2012). These advantages need to be confirmed by extending the research to other language combinations and to other areas of pragmatics. It is also necessary to conduct further research so to investigate whether bilinguals and multilinguals adopt the same intercultural style for all their languages as exploratory studies have reported (Cenoz 2003). As we have seen, research on multilingual practices – that is, on code-switching, code-mixing and translanguaging – has undergone important developments in recent years, but more work needs to be done on the pragmatics of these practices. Finally, a related aspect that deserves attention is the dynamic character of multilingualism. As Jessner (2009: 272) pointed out, multilingualism can no longer be seen as static and fixed because the constant interaction between a speaker’s multiple languages creates new structures and emergent properties that are not found in monolingual systems. Longitudinal studies of the pragmatic aspects of multilingualism could contribute to the analysis of these dynamics.

Conclusion This chapter looks at central issues in research on bilingualism and multilingualism. This is a research area characterised by its diversity. The diversity of multilingualism applies to the phenomenon itself because it takes place in any part of the world, at any age, with different combinations of languages and affecting different domains in life. It is not surprising that multilingualism is studied by scholars in anthropology, sociology, linguistics, neurolinguistics, sociolinguistics, psycholinguistics or education, among others. These disciplines focus on different aspects of multilingualism and also use different theoretical approaches and research methodologies. The study of pragmatics as related to bilingualism and multilingualism can be a very productive area that needs more development. Studies in different areas of bilingualism and multilingualism could benefit by focusing more on pragmatic competence and analysing pragmatic aspects in bilingual and multilingual interaction. At the same time the study of interlanguage pragmatics could benefit from adopting a more multilingual perspective and considering future multilingualism as a resource and multilingual speakers as different from monolingual speakers. Globalisation and its implications, including an international economy and individual mobility, are providing more opportunities to carry out research on bi/multilingualism and pragmatics. This research will not only increase our knowledge of these fields, but also contribute to better international communication.

Suggestions for further reading Bathia, T. K. and Ritchie, W. C. (eds.) (2012) The Handbook of Bilingualism and Multilingualism, 2nd ed. London: Blackwell. This volume looks at bilingualism and multilingualism from both an individual and societal perspective.

147

Jasone Cenoz Cenoz, J. and Gorter, D. (eds.) (2015) Multilingual Education. Between Language Learning and Translanguaging. Cambridge: Cambridge University Press. This edited volume looks at multilingualism from a holistic perspective, taking the learners’ whole multilingual repertoire into account. Groot, A. M. B., de (2011) Language and Cognition in Bilinguals and Multilinguals: An Introduction. New York: Psychology Press. This monograph looks at individual multilingualism from a psycholinguistic approach. Martin-Jones, M., Blackledge, A. and Creese, A. (eds.) (2012) The Routledge Handbook of Multilingualism. London: Routledge. This edited volume looks at discourses about multilingualism in social, cultural and political contexts. Safont Jordà, M. P. (2005) Third Language Learners: Pragmatic Production and Awareness. Clevedon: Multilingual Matters. This monograph brings together the study of the effects of bilingualism on third language acquisition and interlanguage pragmatics

Note 1 This chapter was written with the assistance of the MINECO/FEDER Grant EDU2015-63967-R and Basque government funding for the research group Donostia Research on Education and Multilingualism (DREAM), UFI 11/54.

References Alcón Soler, E. (2012) ‘Teachability and bilingualism effects on third language learners’ pragmatic knowledge’, Intercultural Pragmatics, 9(4): 511–541. Bachman, L. (1990) Fundamental Considerations in Language Testing. Oxford: Oxford University Press. Bahtina, D. and Ten Thije, J. D. (2012) ‘Receptive multilingualism’, in C. A. Chapelle (ed.) The Encyclopedia of Applied Linguistics. Malden, MA: Blackwell. Baker, C. (2011) Foundations of Bilingual Education and Bilingualism, 5th ed. Bristol: Multilingual Matters. Barnes, J. (2006) Early Trilingualism: A Focus on Questions. Clevedon: Multilingual Matters. Barnes, J. (2011) ‘The influence of child-directed speech in early trilingualism’, International Journal of Multilingualism, 8: 42–62. Barron, A. (2012) ‘Interlanguage pragmatics: From use to acquisition to second language pedagogy’, Language Teaching, 45: 44–63. Bialystok, E., Craik, F. I. M. and Luk, G. (2008) ‘Cognitive control and lexical access in younger and older bilinguals’, Journal of Experimental Psychology: Learning, Memory, and Cognition, 34(4): 859–873. Blackledge, A. and Creese, A. (2010) Multilingualism: A Critical Perspective. London: Continuum. Bloomfield, L. (1933) Language. New York: Holt. Blum-Kulka, S. (1991) ‘Interlanguage pragmatics: The case of requests’, in R. Phillipson, E. Kellerman, L. Selinker, M. Sharwood Smith and M. Swain (eds.) Foreign/Second Language Pedagogy Research. Clevedon: Multilingual Matters. 255–272. Bourdieu, P. (ed. by J. B. Thompson) (1991) Language and Symbolic Power, translated by G. Raymond and M. Adamson. Cambridge: Polity Press. Canagarajah, A. S. and Liyanage, I. (2012) ‘Colonial and post-colonial contexts of multilingualism in Sri Lanka’, in M. Martin-Jones, A. Blackledge and A. Creese (eds.) Routledge Handbook of Multilingualism, London: Routledge. 49–65. Canale, M. and Swain, M. (1980) ‘Theoretical bases of communicative approaches to second language teaching and testing’, Applied Linguistics, 1(1): 1–47. Celce-Murcia, M., Dörnyei, Z. and Thurrell, S. (1995) ‘A pedagogical framework for communicative competence. A pedagogically-motivated model with content specifications’, Issues in Applied Linguistics, 6: 5–35. Cenoz, J. (2003) ‘The intercultural style hypothesis: L1 and L2 interaction in requesting behaviour’, in V. Cook (ed.) Effects of the Second Language on the First. Clevedon: Multilingual Matters. 62–80. Cenoz, J. (2009) Towards Multilingual Education: Basque Educational Research from an International Perspective. Bristol: Multilingual Matters.

148

Bilingualism and multilingualism Cenoz, J. (2015) ‘Content-based instruction and content and language integrated learning: The same or different?’, Language, Culture and Curriculum, 28: 8–24. Cenoz, J. and Gorter, D. (2011) ‘Focus on multilingualism: A study of trilingual writing’, The Modern Language Journal, 95: 356–369. Cenoz, J. and Gorter, D. (eds.) (2015) Multilingual Education. Between Language Learning and Translanguaging. Cambridge: Cambridge University Press. Cook, V. (2012) ‘Multicompetence’, in C. A. Chapelle (ed.) The Encyclopedia of Applied Linguistics. Malden, MA: Blackwell. Cook,V. and Bassetti, B. (eds.) (2011) Language and Bilingual Cognition. New York: Psychology Press. Council of Europe (2001) Common European Framework of Reference for Languages: Learning, Teaching, Assessment. Cambridge, UK: Cambridge University Press. Available from: www.coe.int/t/dg4/Linguistic/ Source/Framework_EN.pdf (accessed 28 April 2015). Council of Europe (2007) Guide for the Development of Language Education Policies in Europe. From Linguistic Diversity to Plurilingual Education. Strasbourg, France: Council of Europe, Language Policy Division. Available from: www.coe.int/t/dg4/linguistic/Guide_niveau3_EN.asp#TopOfPage (accessed 28 April 2015). Craik, F. I. M., Bialystok, E. and Freedman, M. (2010) ‘Delaying the onset of Alzheimer’s disease: Bilingualism as a form of cognitive reserve’, Neurology, 75(19): 1726–1729. De Houwer, A. (2009) An Introduction to Bilingual Development. Bristol: Multilingual Matters. Dijkstra, A. (2009) ‘The multilingual lexicon’, in D. Sandra, J.-O. Östman and J.Verschueren (eds.) Cognition and Language Use (Handbook of Pragmatics Highlights, vol. 3). Amsterdam: John Benjamins. 369–388. European Commission (2007) Final Report. High Level Group on Multilingualism. Available from: www. google.es/search?q=Final+Report.+High+Level+Group+on+Multilingualism&oq=Final+Report .+High+Level+Group+on+Multilingualism&aqs=chrome..69i57.529j0j7&sourceid=chrome&es_ sm=93&ie=UTF-8 (accessed 18 September 2015). Firth, A. and Wagner, J. (1997) ‘On discourse, communication, and (some) fundamental concepts in SLA research’, The Modern Language Journal 81 (3): 285‑300. Fouser, R. (2001) ‘Too close for comfort? Sociolinguistic transfer from Japanese into Korean as an L3’, in J. Cenoz, B. Hufeisen and U. Jessner (eds.) Crosslinguistic Influence in Third Language Acquisition: Psycholinguistic Perspectives. Clevedon: Multilingual Matters. 149–169. García, O. and Sylvan, C. E. (2011) ‘Pedagogies and practices in multilingual classrooms: Singularities in pluralities’, The Modern Language Journal, 95(3): 385–400. García, O. and Li, W. (2014) Translanguaging: Language, Bilingualism and Education. Hampshire: Palgrave Macmillan. Gardner-Chloros, P. (2009) Code-Switching. Cambridge: Cambridge University Press. Gorter, D. (2013) ‘Multilingual interaction and minority languages: Proficiency and language practices in education and society’, Language Teaching, 48(1): 82–98. Heller, M. (ed.) (2007) Bilingualism: A Social Approach. Basingstoke: Palgrave. Ishida, M. (2013) ‘Second language pragmatic development’, in C. A. Chapelle (ed.) The Encyclopedia of Applied Linguistics. Malden, MA: Blackwell. Jarvis, S. and Pavlenko, A. (2008) Crosslinguistic Influence in Language and Cognition. New York: Routledge. Jessner, U. (2009) ‘The role of metalinguistic knowledge in L2 and L3 development: A dynamic-systemstheory perspective’, Modern Language Journal, 92(2): 270–283. Jørgensen, J. N. (2008) ‘Polylingual languaging around and among children and adolescents’, International Journal of Multilingualism, 5(3): 161–176. Kasper, G. (1992) ‘Pragmatic transfer’, Second Language Research, 8(3): 203–231. Kecskes, I. (2010) ‘Dual and multilanguage systems’. International Journal of Multilingualism, 7(2): 91–109. Kecskes, I. (2014) Intercultural Pragmatics. Oxford: Oxford University Press. Kecskes, I. and Papp, T. (2000) Foreign Language and Mother Tongue. Mahwah, NJ: Lawrence Erlbaum. Koike, D. and Palmiere, D. (2010) ‘First and second language pragmatics in third language oral and written modalities’, Foreign Language Annals, 44(1): 80–104. Kramsch, C. (2012) ‘Authenticity and legitimacy in multilingual SLA’, Critical Multilingualism Studies, 1: 107–128. Li, W. (2008) ‘Research perspectives on bilingualism and multilingualism’, in W. Li and M. Moyer (eds.) Blackwell Guide to Research Methods in Bilingualism and Multilingualism. Hoboken:Wiley-Blackwell. 1–17. Li,W. (2011) ‘Moment analysis and translanguaging space: Discursive construction of identities by multilingual Chinese youth in Britain’, Journal of Pragmatics, 43(5): 1222–1235.

149

Jasone Cenoz Li, W. and Wu, C. (2009) ‘Polite Chinese revisited: Creativity and the use of codeswitching in the Chinese complementary school classroom’, International Journal of Bilingual Education and Bilingualism, 12(2): 193–211. Lüdi, G. and Py, B. (2009) ‘To be or not to be. . . . a plurilingual speaker’, International Journal of Multilingualism, 6(2): 154–167. Mishina-Mori, S. (2011) ‘A longitudinal analysis of language choice in bilingual children: The role of parental input and interaction’, Journal of Pragmatics, 43(13): 3122–3138. Moore, D. and Gajo, L. (2009) ‘Introduction: French voices on plurilingualism and pluriculturalism:Theory, significance, and perspectives’, International Journal of Multilingualism, 6(2): 137–153. Pavlenko, A. (ed.) (2011) Thinking and Speaking in Two Languages. Bristol: Multilingual Matters. Pavlenko, A. and Malt, B. C. (2011) Kitchen Russian: Cross-linguistic differences and first-language object naming by Russian-English bilinguals. Bilingualism: Language and Cognition, 14: 19–45. Portolés, L. (2015) Multilingualism and Very Young Learners: An Analysis of Pragmatic Awareness and Language Attitudes. Boston: Mouton de Gruyter. Rampton, B. and Charalambous, C. (2012) ‘Crossing’, in M. Martin-Jones, A. Creese and A. Blackledge (eds.) Routledge Handbook of Multilingualism. London: Routledge. 482–498. Romaine, S. (1995) Bilingualism, 2nd ed. Oxford: Blackwell. Safont Jordà, M. P. (2005) Third Language Learners: Pragmatic Production and Awareness. Clevedon: Multilingual Matters. Safont Jordà M. P. (2013) ‘Early stages of trilingual pragmatic development. A longitudinal study of requests in Catalan, Spanish and English’, International Journal of Pragmatics 59: 68–80. Safont Jordà, M. P. and Alcón Soler, E. (2012) ‘Teachability of request act peripherical modification devices in third language learning contexts’, in M. Economidou-Kogetsidis and H.Woodfield (eds.) Interlanguage Request Modification (Pragmatics & Beyond New Series 217). Amsterdam: John Benjamins. 275–313. Seidlhofer, B. (in press) ‘English as a lingua franca and multilingualism’. Encyclopedia of Language and Education. Berlin: Springer. Siegal, M., Iozzi, L. and Surian, L. (2009) ‘Bilingualism and conversational understanding in young children’, Cognition, 110(1): 115–122. Soltero-González, L., Escamilla, K. and Hopewell, S. (2012) ‘Changing teachers’ perceptions about the writing abilities of emerging bilingual students: Toward a holistic bilingual perspective on writing assessment’, International Journal of Bilingual Education and Bilingualism, 15(1): 71–94. Thomas, E. M., Lewis, W. G. and Apolloni, D. (2012) ‘Variation in language choice in extended speech in primary schools in Wales: Implications for teacher education’, Language and Education, 26(3): 245–261. Williams, C. (2002) Ennill iaith: Astudiaeth o sefyllfa drochi yn 11–16 oed [A Language Gained: A Study of Language Immersion at 11–16 Years of Age]. Bangor, UK: School of Education. Available from: www.bangor. ac.uk/addysg/publications/Ennill_Iaith.pdf (accessed 18 September 2015). Zhu, H. (2008) ‘Duelling languages, duelling values: Codeswitching in bilingual intergenerational conflict talk in diasporic families’, Journal of Pragmatics, 40(10): 1799–1816.

150

Pragmatics and culture

14 INTERLANGUAGE PRAGMATICS A historical sketch and future directions Naoko Taguchi

Introduction Interlanguage pragmatics (ILP), a branch of second language acquisition (SLA), examines second language (L2) learners’ knowledge, use and development in performing sociocultural functions. L2 learners need linguistic forms and skills to perform everyday social functions in the target language. At the same time, because our way of speaking is determined by context – to whom we are talking and under what circumstances – learners need to know which forms are appropriate to use in what situations. Hence, linguistic knowledge and sociocultural knowledge of social conventions, customs, and norms of interaction are two layers of pragmatic competence. The process of learning these knowledge bases, individual variation between learners in the process and factors affecting the process are the focal objects of inquiry in ILP research. The original definition of ILP goes back to Kasper and Dahl (1991: 216), who stated that ‘interlanguage pragmatics will be defined in a narrow sense, referring to nonnative speakers’ (NNSs’) comprehension and production of speech acts, and how their L2-related speech act knowledge is acquired’. This definition has since evolved to reflect a more holistic concept of language use in social interaction. Kasper and Rose (2003), for instance, claim that ILP examines how nonnative speakers comprehend and produce actions in a target language, and how L2 learners develop the ability to understand and perform actions in a target language. More recently, Bardovi-Harlig (2010: 219) underlines that pragmatics and pragmatic acquisition in ILP encompasses both form and use. ‘[Pragmatics] bridges the gap between the system side of language and the use side, and relates both of them at the same time. Interlanguage pragmatics brings the study of acquisition to this mix of structure and use’. Corresponding to the increasing body of definitions of ILP, the scope and number of empirical studies in the field has expanded over time. This chapter surveys major findings from the 1980s to 2015. It first describes changes in the theoretical construct of pragmatic competence and illustrates the shift in the conceptualisation of pragmatic abilities from an individual- to an interaction-oriented view. This section concludes with a new conceptualisation of pragmatic competence, which reflects the notion of the intercultural speaker. Following this, a historical sketch of ILP research since the 1980s is presented by surveying common research topics across time periods. I will explain in this section how cross-linguistic studies dominated the field in the 1980s and 1990s, and how the research focus in the 1990s shifted to the instruction 153

Naoko Taguchi

and assessment of pragmatic competence. In the same period, the body of longitudinal studies directly addressing acquisitional pragmatics expanded and moved away from the dominant practice of cross-sectional investigation prominent up to that time.The first decade and a half of this century saw a further growth of instructional, assessment and acquisitional research, characterised as a more explicit application of mainstream SLA theories to ILP studies. These changes will be illustrated by summarising key studies in each period, and this chapter will conclude with directions for future research.

Definitions of pragmatic competence This section describes how the conceptualisation of pragmatic competence has evolved from the individualistic view to more interactional and intercultural understanding of that competence.

Pragmatic competence in models of communicative competence The concept of pragmatic competence originated in Dell Hymes’ theoretical model of communicative competence. Hymes (1972) claimed that language knowledge entails both grammatical and sociocultural knowledge that determine the appropriateness of language use in context. Drawing on Hymes’ insight, several models of L2 communicative competence emerged in the field in the 1980s and 1990s (Canale and Swain 1980; Bachman 1990; Bachman and Palmer 1996, 2010). These models emphasised the multidimensionality of language ability, and situated pragmatic competence as a requisite component within the model. Canale and Swain’s (1980) model was a forerunner in this trend, maintaining that successful communication entails an efficient integration of grammatical, sociolinguistic, discourse, and strategic competencies. Canale and Swain’s model, however, did not sufficiently distinguish between sociolinguistic and pragmatic competence, nor did it explicitly articulate pragmatic competence within the model. Pragmatic competence was part of sociolinguistic competence, which involves the ability to interpret and produce an utterance in context. Bachman (1990) and Bachman and Palmer (1996, 2010) more fully developed pragmatic competence as a competence in its own right. In Bachman and Palmer’s (1996, 2010) framework, language knowledge consists of organisational and pragmatic knowledge. Organisational knowledge in this framework dealt with formal aspects of language (grammar and textual aspects), whereas pragmatic knowledge concerned language use in relation to language users and language use settings. Two types of pragmatic knowledge were distinguished, namely functional knowledge, which enables us to interpret relationships between utterances and the communicative goals of language users (e.g., knowledge of how to perform the speech act of request), and sociolinguistic knowledge, which enables us to interpret or create utterances that are appropriate to specific language use settings (e.g., which forms to use to make a request in situation).

Pragmatic competence in interaction A characteristic of these early models of communicative competence is that they treated pragmatic competence as a psycholinguistic ability that exists within individuals as a stable trait, independent from context. In later research, pragmatic competence has been incorporated into a broader conceptual framework that focuses on the dynamic and dialogic aspects of communication. Most notable in this trend is the emergence of interactional competence as an alternative to the traditional models of communicative competence (Young 2011). Interactional competence 154

Interlanguage pragmatics

views language knowledge and ability as locally situated and jointly constructed by participants in discourse. To this end, ability and context are connected. Learners’ resources are not set in advance but are dependent on the specifics of the dynamic social context. Traditionally, research in pragmatics assumed a one-to-one correspondence between an utterance and force. For instance, the speech act of request is often associated with conventional forms, such as could you and may I. Traditional ILP research has been to identify those linguistic forms and semantic moves that convey illocutionary force in a particular language, and compare these with learners’ forms in order to determine a learner’s level of pragmatic competence. An example of this trend is seen in the popularity of the discourse completion test (DCT) used to assess pragmatic competence (Bardovi-Harlig 2010, 2013). In the DCT, participants read a situational scenario and produce a speech act (e.g., request) given in the scenario. Hence, the DCT can elicit participants’ knowledge of normative speech act expressions, but it does not tell us much about their speech act performance. In contrast to what is assumed in the DCT task, in reality, a speech act is not a pre-planned action, nor does it occur in isolation from discursive context. It arises in the course of conversation through participants’ mutual understandings of the topic and through reactions to each other’s contribution to the ongoing discourse. Indeed, the conventional forms of request may or may not appear in the speech act, depending on how a conversation unfolds. Instead, nonlinguistic forms (e.g., facial expressions, intonation, and pause) may convey the request intention. Or the intention may be negotiated among participants over multiple turns and jointly constructed in interaction. Traditionally, ILP research has disregarded the interactive and dynamic nature of a speech act that emerges from the interplay of context, action and linguistic resources. In general, learners’ pragmatic competence has been examined only in predictable contexts by using an instrument such as DCT, with the focus only on the linguistic forms used to convey a particular illocution. Learners’ ability to adapt and reciprocate in a changing context has been simply discounted. The notion of a changing context is, however, fundamental to the understanding of interactional competence. Interactional competence involves a variety of resources that learners bring to the joint construction of discourse and meaning making. These resources include knowledge of rhetorical scripts, lexis, and syntax specific to the practice, the turn-taking system, topic management, repair, and recognition and production of boundaries between speech activities (e.g., Kasper 2006;Young 2011). The contribution of interactional competence to ILP is evident in the recent use of conversational analysis (CA) to study action, meaning and context (Taguchi and Roever 2017). CA utilises the emic approach to analyse talk-in-interaction to reveal how participants co-­construct an action sequentially in turns, and design their turns to jointly accomplish the activity at hand (Kasper 2006). CA has started to make inroads into ILP, as seen in a growing number of studies that have analysed L2 pragmatics behaviour from a CA perspective (e.g., Al-Gahtani and Roever 2014; Flores-Ferrán and Lovejoy 2015; also see Ross and Kasper 2013, on assessment of L2 pragmatics in interaction). Indeed, the concept of interactional competence closely aligns with contemporary definitions of pragmatics, the other parent discipline of ILP besides SLA. For instance, LoCastro (2003: 15) defines pragmatics as ‘the study of speaker and hearer meaning created in their joint actions that include both linguistic and non-linguistic signals in the context of socioculturally organised activities’. This definition points to the primary object of study in pragmatics – speaker-hearer interaction in a socioculturally bounded act. Interactional competence complements this practice by providing concrete frameworks and empirical means to analysing a pragmatic act in a situated interaction. 155

Naoko Taguchi

Pragmatic competence in intercultural communication Up to this, I have discussed pragmatic competence within SLA, focusing on the origin of the theoretical construct and its evolution from pragmatics-within-individuals to pragmatics-in-interaction-in-context. However, the more recent view of pragmatic competence in social interaction also has a synergy with the field of intercultural communication, which studies intercultural interaction as a cultural practice (Kecskes 2014) (for a review, see Taguchi and Roever, 2017). In intercultural interaction, communication is always a dynamic process wherein collaboration and negotiation constantly take place with the goal of mutual understanding among speakers from different cultures. Interactants’ interactional competence is at stake, and conversation skills including knowledge of turn-taking and adjacency pairs, topic management, repairs, and paralinguistic activities are directly related to the goal of mutual understanding. The complexity of intercultural interaction lies in the fact that these skills are often culturally specific, and speakers bring their own norms to communication. Recent research on lingua franca communication has revealed that participants constantly negotiate interactional norms, standards of politeness and directness, communication styles, and cultural conventions as interactions unfold (e.g., Kecskes 2014; Cogo and House, Chapter 14, this volume). Participants either interpret others based on their own L1 conventions or create a whole new standard of communication. Intercultural competence is broadly defined as ‘a complex of abilities needed to perform effectively and appropriately when interacting with others who are linguistically and culturally different from oneself ’ (Fantini and Tirmizi 2006: 12). Some models of intercultural competence focus on stages of development by defining intercultural competence as the ability to move from an ethnocentric to an ethnorelative worldview (Bennett and Bennett 2004). Others emphasise a specific set of elements that form the basis of one’s potential to succeed in intercultural encounters. For example, Byram (1997) proposes five aspects of intercultural competence: attitudes, knowledge, skills of interpreting and relating, skills of discovery and interaction, and critical cultural awareness. Language competence is recognised as the core of intercultural competence, as shown in its explicit mention in many of the models. Fantini (2012), for instance, stresses the importance of language proficiency in intercultural competence, arguing that developing intercultural competence with language competence promotes full access to a new culture. Similarly, under the term intercultural speaker, Byram (2012: 89) emphasises that being and acting interculturally involves ‘both intercultural competence and linguist/communicative competence, in any task of mediation where two distinct linguacultures are present’. Despite this recognition, curiously, the fields of intercultural and linguistic studies have developed separately. None of the models of intercultural competence provides detailed descriptions or linguistic analyses of intercultural interaction to point out what linguistic abilities are needed for successful intercultural communication. Similarly, despite the extensive literature on the models of communicative competence, the concept of intercultural competence is largely absent from linguistic research. Recent literature has noted this separatism and called for more explicit integration of linguistic and intercultural competences in research and teaching (Spencer-Oatey 2010; Byram 2012; Fantini 2012). As Spencer-Oatey (2010) argues, the study of intercultural competence is clearly an area to which pragmatics research can contribute. Pragmatics research into intercultural interaction and lingua franca communication has revealed characteristic communication patterns in intercultural exchanges. Pragmatic studies can provide authentic interaction data to illustrate a successful cross-cultural interaction. At the same time, the framework of intercultural competence can be useful for ILP in advancing our conceptualisation of pragmatic competence.The characteristic of pragmatic competence 156

Interlanguage pragmatics

(e.g., ability to interact and perform language functions in context) can be situated within some of the core constructs of intercultural competence such as communicative awareness and intercultural empathy. Such a conceptualisation would go beyond the traditional scope of pragmatic competence focused on how learners perform a pragmatic act in the L2 and extend the concept to an understanding of how learners successfully participate in intercultural interaction. Situating ILP in a broader scope of intercultural studies is timely, because in today’s multilingual society, the goal of language learning is not to become a native speaker, but to become an intercultural speaker who is linguistically and interculturally competent – a person who is sensitive to other cultures and aware of his/her own cultural position to mediate across linguistic and cultural boundaries (Byram 2012; Wilkinson 2012). Pragmatic competence can serve as a resource that assists in this process of mediation. Reconceptualising pragmatic competence to reflect this notion of the intercultural speaker will elevate the practice of ILP research from SLA matters alone to the area of global citizenship. At the same time, pragmatic insights into intercultural interaction will help move beyond the current practice of description of intercultural competence to the analysis of acquisition of that competence. We look forward to future interdisciplinary research in this area.

Historical sketch of ILP research This section presents a walkthrough of ILP studies from the 1980s to 2015. By reviewing the field chronologically, I illustrate changes in primary research topics over time. My focus is on cross-linguistic analyses of pragmatic behaviours, longitudinal studies, and instructional studies in ILP.

Cross-linguistic and cross-sectional comparisons of pragmatic behaviours Without question, cross-linguistic studies of pragmatics dominated in the early years of ILP as seen in a bulk of studies produced in the 1980s and 1990s that analysed pragmatic behaviours across languages. Such studies were based on a premise that different cultures have different ways of encoding pragmatic notions of politeness or directness into linguistic behaviours. Although in principle all aspects of pragmatics are subject to cross-cultural comparisons, scholarly interest in this area has concentrated on two areas, namely speech acts and politeness. Studies compared linguistic expressions used in speech acts, and variation in the use of these expressions corresponding to the contextual parameters of interlocutor relationships, power, and degree of imposition. A milestone project that set a trend in this practice was the Cross-Cultural Speech Act Realisation Project (CCSARP) by Blum-Kulka and colleagues (1989). The CCSARP analysed requests and apologies in seven languages using the discourse completion test (DCT). By categorising speech act expressions across languages using a single coding framework, the researchers revealed culturally specific features of speech acts by investigating contrastively how many types of expressions exist in a language, which expressions are considered direct or indirect, and how they vary in different situations. The coding framework and DCT instrument facilitated many replication studies, which provided empirical descriptions of speech acts across cultures. This trend continues today (cf. e.g. Chen 2010, Netz & Lefstein, 2016). These cross-linguistic findings were informative for interlanguage pragmatics analysis because they identified areas in which L1 (first language) pragmatic behaviours differ from L2 behaviours. In fact, a major contribution of the CCSARP was that it collected data from multiple L2 groups and analysed similarities and differences between L2 patterns and those of native 157

Naoko Taguchi

speakers. Results revealed that learners were more verbose and direct in making a request, and used fewer syntactic downgraders than native speakers. The findings highlighted potential areas of pragmatic failure stemming from L1-L2 differences, but also revealed a nuanced picture of L1 transfer, showing learners also avoid transferring language-specific patterns to the L2 (BlumKulka et al. 1989). Following this practice, subsequent studies in this period described differences between native speakers and L2 learners in speech acts, many of which appeared in seminal books and review articles in the 1990s (e.g., Kasper and Rose 1999). In the same period, this comparativedescriptive practice expanded to include L2 groups of different levels of proficiency, length of formal study, and duration of residence in a target country. These in turn contributed to the popularity of cross-sectional inquires into pragmatic development, which contrast data from two or more groups based on differences in proficiency level or length of residence. Any differences between groups were attributed to changes that the learners exhibit at different stages of their L2 learning and thus provide developmental insights. Findings generally suggested a positive effect of proficiency and length of residence on increased pragmatic competence (see Kasper and Rose 1999 for a review of these studies). The scope of cross-sectional investigation has expanded over time with an addition of pragmatic features other than speech acts, for example, formulaic expressions (Bardovi-Harlig 2009), interactional discourse markers (Wei 2011) and the sequential organisation of argumentative discourse (Dippold 2011). Dippold (2011) compared the sequential organisation of argumentative discourse between three L2 German groups of different lengths of study. Learners of lower proficiency levels used a short, two- or three-turn discourse structure including only one adjacency pair of an opinion followed by an agreement or disagreement. In contrast, higherproficiency learners could engage in an extended discourse by relating each turn to the preceding turn.

Longitudinal investigations in ILP While cross-sectional studies have some interesting insights to offer to researchers interested in the development of pragmatic competence, true development can be observed only through a longitudinal design that traces the same participant(s) over an extended period of time. Despite its importance, longitudinal investigation remains relatively unexplored in the field. In fact, the first longitudinal study in ILP was Schmidt (1983), and it was almost a decade later before the second study appeared in the field (Ellis 1992). Schmidt (1983) conducted a case study of a Japanese artist, Wes, who was naturalised in Hawaii. He analysed Wes’s development on four sub-competencies of Canale and Swain’s communicative competence model over a period of three years and found a marginal development in Wes’s sociolinguistic competence, as seen in an increased use of formulaic expressions and lexical items to convey intentions (e.g., please for a request and maybe for a suggestion). Ellis (1992), on the other hand, examined two ESL learners’ developments in performing requests in a classroom setting. He observed change in three stages: (1) a pre-basic stage where learners conveyed a request intention in a context-dependent, minimalistic manner; (2) a formulaic stage where learners performed requests with unanalysed formulas; and (3) an unpacking stage where formulas were incorporated into productive language use with the use of conventional request forms. Although in documenting changes in L2 pragmatic acts, these early studies established a model for longitudinal investigation, only a small number of studies followed this lead, resulting in a general consensus that ILP research has primarily focused on pragmatic use, rather than on acquisition (Bardovi-Harlig 2000; Kasper and Rose 2003; Taguchi 2010). Indeed, Kasper and 158

Interlanguage pragmatics

Rose’s (2003) review of development in ILP listed only nine longitudinal studies. A subsequent overview of development by Bardovi-Harlig (2000) listed fewer than a dozen. The most recent review by Taguchi (2012) found twenty-three unique studies over five languages (eleven English, four Japanese, three French, three German, and two Indonesian) in three areas: comprehension of implicature (four studies), perception/recognition of pragmatic features (five studies), and production of pragmatic functions (fourteen studies). In the area of implicature comprehension, findings largely support the notion that learners progress from the stage where implied meaning is marked via strong signals (e.g., universal or shared conventionality between L1 and L2) to the stage where meaning does not involve those signals and thus requires more extensive inferencing (e.g., Bouton 1994; Taguchi 2012). Taguchi (2012), for instance, examined the comprehension of indirect refusals (conventional implicature) and indirect opinions (non-conventional implicature) among Japanese learners of English. Learners’ comprehension was tracked over two semesters using a multiple-choice listening test. Comprehension was faster and more accurate, and development more profound for conventional indirect refusals than for non-conventional indirect opinions. In the area of pragmatic perception, all studies were conducted in a study abroad context (e.g., Matsumura 2001; Kinginger 2008). They reveal learners’ gains over time, but also find individual variation due to different qualities and quantities of experience. The area of pragmatic production reveals a slow development of L2 pragmalinguistic forms. Research shows that learners usually begin with overgeneralisation of a few forms over a range of functions. They gradually expand their pragmalinguistic repertoire by adopting new formfunction mappings into their systems. These findings are illustrated in studies in two major categories: form-to-function studies and function-to-form studies. The former studies examine how a particular form becomes target-like in function, by tracking changing (or expanding) functions of one form over time (e.g., Hellermann 2009; Ishida 2011), while the latter examine how a particular function becomes target-like in form over time, by analysing changing forms in performing the same function (e.g., Barron 2003; Nguyen 2011). An example of a form-to-function analysis is Hellermann’s (2009) study of one L2 English learner’s use of the word no over fifty weeks.Video recordings of the learner’s classroom interaction showed that the learner first used the unmitigated no in direct correction, but later she expanded the function of no, using it in repair and humour. She also started to mitigate no with hesitation and hedging (e.g., well) to demonstrate her orientation towards dispreferred response. In a function-to-form analysis, on the other hand, Nguyen (2011) examined an L2 English learner’s change in the function of response over five weeks. Videotaped student-professor meetings showed that as the student’s role shifted from a passive meaning receiver to an active meaning contributor, her response action changed from using minimal responses in a delayed manner to using expanded responses in an immediate manner, and eventually to initiating a topic herself. While these longitudinal studies have primarily analysed changing patterns in the pragmatic system, a recent trend in the longitudinal practice is to present a more context-oriented account of learners’ changes in conjunction with individual and contextual factors. This practice corresponds to the current epistemological shift in SLA to focus on the dynamicity and complexity of language development in a social context. Dynamic systems theory (Verspoor et al. 2011) and chaos/complexity theory (Larsen-Freeman and Cameron 2009) are major proponents of this new epistemological trend. They share the following views: (1) Language development is a non-linear, non-static process shaped by socially co-regulated interactions of multiple influences. 159

Naoko Taguchi

(2) Language development is inseparable from context. Language emerges via interactions between agents (individuals) and with their environments. (3) Language development entails intra- and inter-variability. Variability data provide idiosyncratic details of individual learners’ developmental trajectories that are otherwise masked out in the analysis of group-level means. In my view, only a few ILP studies conform to some of the methodological principles of dynamic, complexity systems research. Those studies are Ohta (2001), Barron (2003), Kinginger (2008), and Taguchi (2012). Using naturalistic recordings of classroom interactions, Ohta investigated two L2 Japanese learners’ development of acknowledgement and alignment expressions. She identified developmental patterns of these expressions and revealed classroom-specific experiences (e.g., teacher input, peer-to-peer talk) that affected the patterns. Barron (2003) examined the development of the speech act of requests, offers, and refusals of offers in L2 German over a fourteen-month study-abroad program. DCT data revealed only modest progress in learners’ ability, which was closely related to individual experiences. Kinginger (2008) examined the development of the awareness of sociolinguistic forms (e.g., address terms, colloquial expressions) in L2 French during a semester abroad in France. Pre- and posttest comparisons revealed considerable individual variation in learners’ changes. Qualitative data from interviews, journals, and diaries showed the variation to come from differences in the amount and intensity of the sociocultural contact learners experienced.These three studies are all book-length longitudinal studies that used a mixed-methods approach – the combination of qualitative and quantitative methods. By collecting the individual- and contextuallevel data, these studies illustrated how the developmental paths converge or diverge across individuals. Taguchi (2012) attempted a more explicit application of the complex, dynamic systems perspective to pragmatic development. The study showed that Japanese ESL students’ production of high-imposition speech acts (making a high-stake request of a teacher or expressing an opinion to a teacher on a serious matter) showed little progress after a year. This was displayed in the students’ interaction styles with native speaker instructors. In the real-life pragmatics of expressing disagreement with their teachers about course assignments, students often used strong modals (e.g., should) and used an explicit expression of dislike (I don’t like X.). The interview data revealed that teachers were often so keen on getting students’ feedback that they did not care much about the direct manner of speech, either neglecting to correct students’ misuse of pragmalinguistic forms or feeling no need to correct them. These findings are in line with the dynamic, complexity systems’ views of language development. The restricted improvement with high-imposition speech acts was a product of the intricate interaction among the subsystems, elements, agents, and processes. Agents (teachers, students) and elements (learning expectations, bilingual context) in the environment co-adapted to each other, giving rise to students’ slow progress towards sociolinguistic norms of interaction in high-imposition speech acts. These findings emphasise the strength of longitudinal design in understanding pragmatic development. Longitudinal studies that combine systematic data collection on quantitative change with qualitative analyses of context and individuals can provide an account of the intricate relationship between pragmatic change, individual differences, and context. A mixed-­ methods approach can shed light on the complexity and dynamicity of pragmatic development in which multiple factors – learners’ subjectivity, stance, affect, resources, and interaction in the target language – are interconnected and jointly influence the evolving pathways towards increased pragmatic competence. 160

Interlanguage pragmatics

Instructional studies in ILP In the final subsection of the historical sketch, I outline key findings in instructional studies in ILP (cf. also Basturkmen and Nguyen, Chapter 40, this volume). Similar to cross-sectional and longitudinal studies, instructional studies are concerned with change and factors affecting the change, but more precisely, they focus on direct teaching as the independent variable and measure its impact on learning outcomes. In other words, they concentrate on changes in pragmatic knowledge from pre- to post-test. Studies are largely quasi-experimental, comparing learners who received instruction to those who did not, or examining two or more groups under different treatment conditions. Some studies almost give rise to longitudinal research by examining instructional effects over an extended observation period (Alcón-Soler 2015). Instructed ILP is a growing area of research, supported by mounting empirical studies published since the 1990s, as well as edited volumes focusing on instructed pragmatics (for a review, see Taguchi and Roever 2017). Jeon and Kaya (2006) located thirty-four instructional studies, of which thirteen were subjected to a quantitative meta-analysis. Takahashi (2010) found a total of forty-nine studies. The most recent review, Taguchi (2015), found fifty-eight studies over six target languages (thirty-eight English, four Spanish, nine Japanese, three French, two German, and two Chinese) in a range of pragmatic targets (e.g., speech acts, implicature, routines, reactive tokens, discourse markers, address forms, hedging, and epistemic markers). Studies in the 1990s revealed that most pragmatic features are teachable, meaning that instruction helps boost learners’ pragmatic development. The next decade evolved around a question of efficacy: what instructional methods best promote pragmatics learning? This question was taken up by a line of intervention studies that compared the effects of certain teaching methods over others by measuring the degree of learning from pre- to post-instruction. The comparison between explicit and implicit teaching has generated by far the most empirical findings. The former typically involves explicit metapragmatic explanation followed by focused practice, while the latter withholds explanation but tries to develop learners’ implicit understandings of the targets through consciousness-raising tasks and implicit feedback. These studies were motivated by Schmidt’s (2001) Noticing Hypothesis, which claims that learners’ attention to linguistic forms, their functions and relevant contextual features is necessary for pragmatic input to become intake, leading to acquisition. Explicit/implicit teaching is a way of promoting this awareness of the target form-function-context mappings and subsequent internalisation of them. An example of the explicit-implicit comparison is found in Fordyce’s (2014) study. He compared the effect of explicit and implicit treatment on Japanese EFL learners’ use of epistemic stance markers (modal expressions that convey speakers’ psychological states). This study is unique in two aspects. First, it examined the long-term instruction effect by giving a delayed post-test five months after the treatment. Second, the study compared the effects among three L2 groups of different proficiency to see if they benefit from instruction differently. Explicit treatment in this study involved teacher explanation of epistemic markers, along with exposure to the targets via enhanced input (forms in bold type or in a larger font), consciousness-raising activities, feedback on students’ epistemic forms, and a quiz. In contrast, implicit treatment did not attempt to draw learners’ attention to the forms: rather students were exposed to texts that contained epistemic forms. Both groups improved after the treatment, but the explicit group outperformed the implicit group on the frequency and range of the epistemic forms they used in the post-test essays, supporting the Noticing Hypothesis. There was no effect of proficiency: all groups improved. Previous studies have generally confirmed the superiority of the explicit over implicit method, although the length of instruction, types of pragmatic targets and outcome measures 161

Naoko Taguchi

moderated the observed learning benefits (Jeon and Kaya 2006; Takahashi 2010; Taguchi 2015). However, the problem with the explicit-implicit comparison is that studies have operationalised the explicit/implicit dichotomy differently. While availability of metapragmatic explanation was one of the criteria in many explicit treatments, operationalisation of implicit treatment differed across studies: some studies used only input flood like Fordyce’s study, while others used tasks to draw learners’ attention to the pragmatic targets via consciousness-raising tasks (Derakhshan and Eslami 2015). As a result, the implicit condition itself formed a continuum from absolute implicit to more explicit end of implicit treatment. In addition, explicit treatment typically involved more than just metapragmatic explanation, often combined with implicit activities such as consciousness-raising tasks (Félix-Brasdefer 2008). As a result, we do not know whether the observed benefits of explicit over implicit treatment are solely due to the explicit method. Glaser (2013) recently reinforced this observation. She claimed that the explicit-implicit opposition simplifies the approach to teaching pragmatics because the explicit teaching is automatically equated with the provision of pragmatic explanation and the implicit teaching with a lack of explanation. However, a further approach that combines the explicit explanation with inductive rule discovery is also possible. Glaser argued for the benefit of the explicit-inductive approach in which the instructor provides language examples first, encouraging learners to discover regularities among the examples, and later addresses underlying rules explicitly. Another shortcoming of the explicit-implicit opposition is that findings are inconclusive depending on the type of outcome measures. In Takahashi’s (2010) review of twenty-one studies on explicit/implicit treatment, eleven studies reached mixed findings, of which eight used multiple outcome measures. For instance, Fukuya and Martínez-Flor (2008) revealed a contrast between online and offline tasks in moderating the instructional effects of the speech act of suggestion.The explicit group received metapragmatic information, whereas the implicit group did role play with recasts. Results revealed an advantage for the explicit instruction. It outperformed the implicit counterpart on the phone message task (a suggestion recorded on the answering machine), but both groups gained in the email task (a suggestion sent via email). While the Noticing Hypothesis and explicit-implicit comparison have dominated the field for a long time, we have witnessed a gradual expansion of intervention studies adopting different theoretical frameworks, including: Li’s (2012) study using skill acquisition theory (DeKeyser 2007),Van Compernolle’s study (2014) under the framework of sociocultural theory (Vygotsky 1978), and Taguchi and Kim’s study (2014) within Swain and Lapkin’s (1998) concept of collaborative dialogue, and Kim and Taguchi’s study (2015) using Cognition Hypothesis (Robinson, 2011) (for a review, see Taguchi 2015). It is hoped that this diversity in theoretical backgrounds will continue with more empirical findings.

Conclusion In this chapter, I have presented a summary of the rapid development of ILP literature over the past few decades. I have illustrated the shift in the view of pragmatic competence in ILP from pragmatic competence as an individual trait to pragmatics as a co-construction of a socioculturally bounded act in interaction. I also pointed out that, since such interactions are situated in intercultural settings, gains in pragmatic competence contribute to intercultural competence and to interactants’ ability to function effectively and appropriately in a new culture. Parallel to such changes in the conceptualisation of pragmatic competence in ILP, empirical interests have also shifted over time, namely from cross-linguistic/cross-sectional studies that primarily focus on pragmatic use to longitudinal studies that focus on acquisition. In the same period, the field has fully embraced the instructional practice of ILP. After surveying the literature I see five main directions for future research. 162

Interlanguage pragmatics

Future direction (1): expanding the scope of pragmatic competence Traditional models of communicative competence define pragmatic competence as the ability to perform language functions appropriately in a social situation, contending that learners need to possess a range of linguistic forms and semiotic resources in their repertoire, choose appropriate forms according to the situation, and use them effectively to achieve communicative goals. Although the problem of the treatment of static contexts is noteworthy in these models (see earlier in this chapter on pragmatic competence in the models of communicative competence), this definition collaborates with some of the often-cited dimensions of multicultural competence and intercultural competence, namely adaptability, flexibility, and variability. For instance, Canagarajah (2007: 932) states that, ‘[M]ultilingual competence is open to unpredictability. It refers to the ability to find a fit or an alignment between the linguistic resources they bring and the context of communication’. The flexibility and adaptability aspect, however, has not been fully incorporated into ILP research, because most studies assign fixed variables to context (power, distance, and degree of imposition), and draw a one-to-one correspondence between context and forms, neglecting multiplicity and dynamicity of context. Future research can be more creative in designing tasks that reveal learners’ ability to move between multiple contexts and to adapt, align and reciprocate their pragmatic behaviors in a dynamic, changing context.

Future direction (2): longitudinal studies in a language other than English In the current landscape, the majority of longitudinal and instructional studies concentrate on L2 English (see Taguchi 2010, 2015).There is no doubt that the weighty influence of L2 English studies in the field needs to be remedied in the future by looking at pragmatic development and instructional treatment in languages other than English.To this end, the growing scope of crosslinguistic analysis into rather under-studied languages (e.g., Arabic, Chinese, Vietnamese) will be useful in the future. Correspondingly, future research might pay more attention to features of pragmatics-specific-to-languages. Many previous studies have applied a top-down procedure and employed previously existing speech act categories and implicature coding schemes to investigate patterns of development or to design instructional materials. However, future studies would benefit from a bottom-up analysis, identifying pragmalinguistic and sociopragmatic features that are unique to a specific language and crucial for learning that language. Individual languages have different pragmatic devices in their structure and discourse. They use different linguistic means to convey appropriate levels of politeness or to communicate meaning indirectly. Future research should explore central characteristics of the construct of pragmatic competence specific to individual languages, and link these to empirical methods through which pragmatic development can be observed.

Future direction (3): individual differences and instructional outcomes Following the call in recent reviews (Takahashi 2010), a dynamic interaction of learnability, pragmatic targets, instructional methods, and learning outcomes is needed in future instructed ILP research. So far, only Fordyce (2014) compared learning outcomes among learners of different proficiency levels. Further research should address what kinds of target features should be taught to what kinds of learners (in terms of proficiency or individual difference factors) 163

Naoko Taguchi

using what kinds of methods (e.g., implicit or explicit; input-based or output-based). This line of research will advance our understanding of the relationship between pragmatic learnability and instructional intervention.

Future direction (4): advanced pragmatic competence A relatively underexplored area of research is the investigation into advanced pragmatic competence. The majority of the previous ILP studies dealt with L2 learners at low to intermediate levels enrolled in language courses for one to four years. As such, the level of pragmatic competence targeted in these studies is limited to the low- to mid-level pragmatic behaviours, such as production of speech acts in a few utterances, comprehension of implicature in short, artificial dialogues, use of basic turn-taking structure and minimum interaction devices and knowledge of one or two forms of sociolinguistic variation. As a result, we do not know the types of pragmatic tasks that very advanced speakers can handle or the types of challenges still left for them in their development of pragmatic competence. Advanced pragmatic competence is an important concept in the current era of global mobility, because language learning goes beyond the classroom and directly impacts career success in international assignments and participation in global virtual teams. In these high-stake situations, advanced proficiency, characterised by the ability to handle a variety of communicative tasks in formal and informal exchanges, or the ability to cope with linguistic challenges stemming from unexpected turns of events, is crucial.The ILP field can make a contribution to the understanding of advanced proficiency and cross-cultural adaptability if it can find ways to operationalise advanced pragmatic competence, and design methods to elicit and examine the construct.There are several examples of advanced L2 pragmatics, such as Ikeda’s (2009) study on Japanese honorifics and Dippold’s (2011) study on argumentative discourse in German. Louw and colleagues (2010) use a pedagogical intervention on complex pragmatic skills in job interviews via simulation. Future expansion is to be seen in this line of study.

Future direction (5): native speaker variation in pragmatic behaviours On the same topic of advanced pragmatics, we have to be aware that native speaker variation typically exists in any pragmatic behaviours, and the degree of variation might be large in advanced pragmatic competence. Although to my knowledge no studies have empirically compared the extent of native speaker variation over different pragmatic functions and tasks, it is plausible that the more complex the pragmatic act is, the greater the variation becomes among people who perform the act. An example is Japanese honorifics. Honorifics systems are complex linguistically but more so socioculturally because they reflect social actions that speakers accomplish with polite language. Speakers use honorifics not just to conform to socially agreed norms in a situation, but they also use them strategically in order to construct their social identity (e.g., Geyer 2013). In other words, the degree of honorifics speakers use is a reflection of the social selves they want to project – how polite or casual they want to sound in a certain situation, which in turn leads to considerable situational and individual variation in their use (cf.Van De Mieroop, Chapter 16, this volume). In today’s society, where multiple ethnic groups and languages constitute the demographic make-up of one place, uniform native speaker standards do not exist, nor are they relevant to the evaluation of pragmatic competence (cf. Barron, Chapter 10, this volume).Yet using monolingual norms remains the mainstream practice in traditional ILP research. A fruitful direction would seem to be to investigate the degree of variation in pragmatic behaviours among local 164

Interlanguage pragmatics

speakers, and cultivate ways to incorporate variation into the analysis of interlanguage pragmatics. For example, researchers might establish an assortment of acceptable target pragmatic behaviours in a given community, rather than just single norms, and interpret learners’ pragmatic behaviours in consideration of these variable behaviours. These future directions will move ILP research forward and contribute to the accumulated knowledge about L2 learning and development in the broader field of SLA.

Suggestions for further reading Bardovi-Harlig, K. (2013) ‘Developing L2 pragmatics’, Language Learning, 63: 68–86. This paper links L2 pragmatics research and the larger field of SLA research by surveying literature in five areas: (1) the design and evaluation of pragmatic tasks; (2) the task design for implicit and explicit pragmatic knowledge; (3) the measurement of pragmatic development; (4) the interface between grammar and pragmatics; and (5) the effect of learning context on pragmatic development. Barron, A. (2012) ‘Interlanguage pragmatics: From use to acquisition to second language pedagogy’, Language Teaching, 45: 44–63. This paper presents a historical timeline of L2 pragmatics research by surveying the key publications that have advanced the field. References and summaries of fifty-seven studies are presented in four categories: (1) use and acquisition; (2) pedagogy; (3) social-affective factors; and (4) appropriateness of an L2 pragmatic norm for learners. Taguchi, N. (2015) ‘ “Contextually” speaking: A survey of pragmatic learning abroad, in class, and online’, System, 48: 3–20. This paper synthesises key findings in these three contexts, and compares and contrasts the opportunities and challenges involved in each context with the overall aim of revealing how each context supports pragmatic learning and development.

References Alcón-Soler, E. (2015) ‘Instruction and pragmatic change during study abroad email communication’, Innovation in Language Learning and Teaching, 9(1): 34–45. Al-Gahtani, S. and Roever, C. (2014) ‘Preference structure in L2 Arabic requests’, Intercultural Pragmatics, 11(4): 619–643. Bachman, L. F. (1990) Fundamental Considerations in Language Testing. New York: Oxford University Press. Bachman, L. F. and Palmer, A. S. (1996) Language Testing in Practice: Designing and Developing Useful Language Tests. Oxford: Oxford University Press. Bachman, L. F. and Palmer, A. S. (2010) Language Assessment in Practice. Oxford: Oxford University Press. Bardovi-Harlig, K. (2000) ‘Pragmatics and second language acquisition’, in R. B. Kaplan (ed.) Oxford Handbook of Applied Linguistics. Oxford: Oxford University Press. 182–192. Bardovi-Harlig, K. (2009) ‘Conventional expressions as a pragmalinguistic resource: Recognition and production of conventional expressions in L2 pragmatics’, Language Learning, 59: 755–795. Bardovi-Harlig, K. (2010) ‘Exploring the pragmatics of interlanguage pragmatics: Definition by design’, in A. Trosborg (ed.) Handbook of Pragmatics: Pragmatics across Languages and Cultures. Berlin: Mouton de Gruyter. 219–260. Bardovi-Harlig, K. (2013) ‘Developing L2 pragmatics’, Language Learning, 63: 68–86. Barron, A. (2003) Acquisition in Interlanguage Pragmatics: Learning How to Do Things with Words in a Study Abroad Context. Amsterdam/Philadelphia: John Benjamins. Bennett, J. M. and Bennett, M. J. (2004) ‘Developing intercultural sensitivity: An integrative approach to global and domestic diversity’, in D. Landis, J. M. Bennett and M. J. Bennett (eds.) Handbook of Intercultural Training. Thousand Oaks: Sage. 147–165. Blum-Kulka, S., House, J. and Kasper, G. (1989) Cross-Cultural Pragmatics: Requests and Apologies. Norwood, NJ: Ablex. Bouton, L. (1994) ‘Can NNS skill in interpreting implicature in American English be improved through explicit instruction? A pilot study’, Pragmatics and Language Learning Monograph Series, 5: 89–109. Byram, M. (1997) Teaching and Assessing Intercultural Communicative Competence. Philadelphia: Multilingual Matters.

165

Naoko Taguchi Byram, M. (2012) ‘Conceptualizing intercultural (communicative) competence and intercultural citizenship’, in J. Jackson (ed.) The Routledge Handbook of Language and Intercultural Communication. London: Routledge. 85–97. Canagarajah, S. (2007) ‘Lingua franca English, multilingual communities, and language acquisition’, Modern Language Journal, 91: 923–939. Canale, M. and Swain, M. (1980) ‘Theoretical aspects of communicative approaches to second language teaching and testing’, Applied Linguistics, 1: 1–47. Chen, R. (2010) ‘Compliment and compliment response research: A cross-cultural survey’, in A. Trosborg (ed.) Handbook of Pragmatics: Pragmatics across Languages and Cultures. New York/Berlin: Mouton de Gruyter. 79–101. DeKeyser, R. (2007) Practice in a Second Language: Perspectives from Applied Linguistics and Cognitive Psychology. Cambridge: Cambridge University Press. Derakhshan, A. and Eslami, Z. (2015) ‘The effect of consciousness-raising instruction on the pragmatic development of apology and request’, TESL-EJ, 18(4): 1–24. Dippold, D. (2011) ‘Argumentative discourse in L2 German: A sociocognitive perspective on the development of facework strategies’, Modern Language Journal, 95: 171–187. Ellis, R. (1992) ‘Learning to communicate in the classroom: A study of two language learners’ requests’, Studies in Second Language Acquisition, 14(1): 1–23. Fantini, A. E. (2012) ‘Language: An essential component of intercultural competence’, in J. Jackson (ed.) The Routledge Handbook of Language and Intercultural Communication. London: Routledge. 263–278. Fantini, A. E. and Tirmizi, A. (2006) ‘Exploring and assessing intercultural competence’, World Learning Publications, paper 1. Available from: http://digitalcollections.sit.edu/cgi/viewcontent.cgi?article=1001 &context=worldlearning_publications (accessed 2 July 2015). Félix-Brasdefer, J. C. (2008) ‘Pedagogical intervention and the development of pragmatic competence in learning Spanish as a foreign language’, Issues in Applied Linguistics, 16(1): 49–84. Flores-Ferrán, N. and Lovejoy, K. (2015) ‘An examination of mitigating devices in the argument interactions of L2 Spanish learners’, Journal of Pragmatics, 76: 67–86. Fordyce, K. (2014) ‘The differential effects of explicit and implicit instruction on EFL learners’ use of epistemic stance’, Applied Linguistics, 35(1): 6–28. Fukuya,Y. and Martínez-Flor, A. (2008) ‘The interactive effects of pragmatic-eliciting tasks and pragmatics instruction’, Foreign Language Annuals, 41(3): 478–500. Geyer, N. (2013) ‘Discernment and variation: The action-oriented use of Japanese addressee honorifics’, Multilingua, 32(2): 155–176. Glaser, K. (2013) ‘The neglected combination: A case for explicit-inductive instruction in teaching pragmatics in ESL’, TESL Canada, 30(7): 150–163. Hellermann, J. (2009) ‘Practices for dispreferred responses using no by a learner of English’, International Review of Applied Linguistics, 47(1): 95–126. Hymes, H. D. (1972) ‘On communicative competence’, in J. B. Pride and J. Holmes (eds.) Sociolinguistics: Selected Readings. Harmondsworth: Penguin. 269–293. Ikeda, K. (2009) ‘Advanced learners’ honorific styles in emails and telephone calls’, in N. Taguchi (ed.) Pragmatic Competence. Berlin: Mouton de Gruyter. 69–100. Ishida, M. (2011) ‘Engaging in another person’s telling as a recipient in L2 Japanese: Development of interactional competence during one-year study-abroad’, in G. Pallotti and J. Wagner (eds.) L2 Learning as Social Practice: Conversation-Analytic Perspectives. Honolulu: University of Hawai’i, National Foreign Language Resource Center. 45–85. Jeon, E.-H. and Kaya, T. (2006) ‘Effects of L2 instruction on interlanguage pragmatic development’, in N. John and L. Ortega (eds.) Synthesizing Research on Language Learning and Teaching. Philadelphia: John Benjamins. 165–211. Kasper, G. (2006) ‘Beyond repair: Conversation analysis as an approach to SLA’, AILA Review, 19: 83–99. Kasper, G. and Dahl, M. (1991) ‘Research methods in interlanguage pragmatics’, Studies in Second Language Acquisition, 13(2): 215–247. Kasper, G. and Rose, K. (1999) ‘Pragmatics and SLA’, Annual Review of Applied Linguistics, 19: 81–104. Kasper, G. and Rose, K. (2003) Pragmatic Development in a Second Language, (Language Learning Monograph Series vol. 3). Malden, MA/Oxford, UK: Wiley-Blackwell. Kecskes, I. (2014) Intercultural Pragmatics. New York: Oxford University Press. Kim, Y. and Taguchi, N. (2015) ‘Promoting task-based pragmatics instruction in EFL classroom context: The role of task complexity’, Modern Language Journal, 99: 656–677.

166

Interlanguage pragmatics Kinginger, C. (2008) ‘Language learning in study abroad: Case studies of Americans in France’, The Modern Language Journal, 92: 1–124. Larsen-Freeman, D. and Cameron, L. (2009) Complex Systems and Applied Linguistics. Oxford: Oxford University Press. Li, S. (2012) ‘The effects of input-based practice on pragmatic development of requests in L2 Chinese’, Language Learning, 62(2): 403–438. LoCastro,V. (2003) An Introduction to Pragmatics: Social Action for Language Teachers. Ann Arbor: University of Michigan Press. Louw, K. J., Derwing,T. M. and Abbott, M. L. (2010) ‘Teaching pragmatics to L2 learners for the workplace: The job interview’, The Canadian Modern Language Review, 66(5): 739–758. Matsumura, S. (2001) ‘Learning the rules for offering advice: A quantitative approach to second language socialization’, Language Learning, 51(4): 635–679. Netz, H. and Lefstein, A. (2016) ‘A cross-cultural analysis of disagreements in classroom discourse: Comparative case studies from England, the United States, and Israel’, Intercultural Pragmatics, 13: 211–255. Nguyen, H. T. (2011) ‘A longitudinal microanalysis of a second language learner’s participation’, in G. Pallotti and J. Wagner (eds.) L2 Learning as Social Practice: Conversation-Analytic Perspectives. Honolulu: University of Hawai’i, National Foreign Language Resource Center. 17–44. Ohta, A. (2001) Second Language Acquisition Processes in the Classroom: Learning Japanese. Mahwah, NJ: Lawrence Erlbaum. Robinson, P. (2001) ‘Individual differences, cognitive abilities, aptitude complexes and learning conditions in second language acquisition’, Second Language Research, 17: 368–392. Ross, S. J. and Kasper, G. (2013) Assessing Second Language Pragmatics. Basingstoke: Palgrave MacMillan. Schmidt, R. (1983) ‘Interaction, acculturation, and the acquisition of communicative competence’, in N. Wolfson and E. Judd (eds.) Sociolinguistics and Language Acquisition. Rowley, MA: Newbury House. 137–174. Schmidt, R. (2001) ‘Attention’, in P. Robinson (ed.) Cognition and Second Language Instruction. Cambridge: Cambridge University Press. 3–32. Spencer-Oatey, H. (2010) ‘Intercultural competence and pragmatics research: Examining the interface through studies of intercultural business discourse’, in A. Trosborg (ed.) Handbook of Pragmatics: Pragmatics across Languages and Cultures. Berlin: Mouton de Gruyter. 189–218. Swain, M. and Lapkin, S. (1998) ‘Interaction and second language learning: Two adolescent French immersion students working together’, Modern Language Journal, 82(3): 320–337. Taguchi, N. (2010) ‘Longitudinal studies in interlanguage pragmatics’, in A. Trosborg (ed.) Handbook of Pragmatics: Pragmatics across Languages and Cultures. Berlin: Mouton de Gruyter. 333–361. Taguchi, N. (2012) Context, Individual Differences, and Pragmatic Competence. New York/Bristol: Multilingual Matters. Taguchi. N. (2015) ‘Instructed pragmatics at a glance: Where instructional studies were, are, and should be going’, Language Teaching, 48(1): 1–48. Taguchi, N. and Kim,Y. (2014) ‘Collaborative dialogue in learning pragmatics: Pragmatic-related episodes as an opportunity for learning request-making’, Applied Linguistics, 2014: 1–23. Available from: http://applij. oxfordjournals.org/content/early/2014/07/29/applin.amu039.full.pdf+html (accessed 10 July 2015). Taguchi, N. and Roever, C. (2017) Second Language Pragmatics. New York: Oxford University Press. Takahashi, S. (2010) ‘Assessing learnability in second language pragmatics’, in A. Trosborg (ed.) Handbook of Pragmatics: Pragmatics across Languages and Cultures. Berlin: Mouton de Gruyter. 391–421. Van Compernolle, R. A. (2014) Sociocultural Theory and Instructed L2 Pragmatics. Bristol/New York: Multilingual Matters. Verspoor, M., De Bot, K. and Lowie,W. (2011) A Dynamic Approach to Second Language Development: Methods and Techniques. Amsterdam/Philadelphia: John Benjamins. Vygotsky, L. S. (1978) Mind in Society: The Development of Higher Psychological Processes. Harvard: Harvard University Press. Wei, M. (2011) ‘Investigating the oral proficiency of English learners in China: A comparative study of the use of pragmatic markers’, Journal of Pragmatics, 43(14): 3455–3472. Wilkinson, J. (2012) ‘The intercultural speaker and the acquisition of intercultural/global competence’, in J. Jackson (ed.) The Routledge Handbook of Language and Intercultural Communication. London: Routledge. 296–309. Young, R. (2011) ‘Interactional competence in language learning, teaching, and testing’, in H. Hinkel (ed.) Handbook of Research in Language Learning and Teaching. New York: Routledge. 426–443.

167

15 INTERCULTURAL PRAGMATICS Alessia Cogo and Juliane House

Introduction Intercultural pragmatics is a field of research characterised by a concern with language use in (oral or written) discourse between members of different mother tongues and cultures. To understand the nature of intercultural pragmatics as a distinctive field of inquiry, it is first necessary to set it off from two other well-known, adjacent research paradigms: cross-cultural pragmatics and interlanguage pragmatics. According to Kecskes (2012: 68), both cross-cultural and interlanguage pragmatics are based on three theoretical constructs: Gricean pragmatics (1975), Brown and Levinson’s (1987) politeness theory and the so-called interlanguage hypothesis (cf. Selinker 1972). The most important and long-standing common research foci of cross-cultural and interlanguage pragmatics have been politeness and pragmatic transfer, reflecting an underlying monolingual and cross-cultural stance that derives from a belief in the independence of linguistic and cultural systems as well as an assumption of the universality of politeness and the cooperative principle. Studies in cross-cultural pragmatics, a popular research paradigm since the early 1980s, have often dealt with the realisation of speech acts and speech act sequences (Blum-Kulka et al. 1989; Kachru 1998; Cheng 2012), but also with contrastive pragmatic discourse studies (House 1982; Moder and Martinovic-Zic 2004) and cross-cultural misunderstanding (House 2000; House et al. 2003), as well as cross-cultural pragmatic failure (Thomas 1983). A common focus in all of these studies has been on comparing different lingua-cultural norms as these become evident in how members of different lingua-cultures behave in certain situational contexts. In interlanguage pragmatics, a term Blum-Kulka and colleagues (1989) first used to characterise part of the classic study on cross-cultural pragmatics, the focus has moved towards looking at second language learners and their acquisition and use of pragmatic norms in L2 in the course of their development of pragmatic competence inside and outside instructional environments (Kasper and Blum-Kulka 1993; House 1996; Bardovi-Harlig and Hartford 2005; cf. Taguchi, Chapter 14, this volume). While interlanguage pragmatics can be described as part of second language acquisition research (Kasper 1996), which focuses on comparing learners’ pragmatic behaviour with the

168

Intercultural pragmatics

more often than not unattainable native speaker pragmatic competence, intercultural pragmatics reflects a development away from such comparisons and away from the essentialist assumption of preexistent, static pragmatic norms. In intercultural pragmatics, the focus is thus on ‘language use rather than language development and acquisition and on the sociopragmatic and sociocultural focus of language choice’ (House 2007: 14). In the following chapter, we introduce the areas that have shaped and continue to shape intercultural pragmatic research and then focus on a perspective of current work in this field, that is on lingua franca pragmatics. The necessarily selective nature of this chapter meant we had to prioritise themes and topics and discard areas of investigation which, though relevant for intercultural pragmatics, cannot find suitable space if treated briefly here. These are the use of intercultural pragmatic strategies in different domains, such as business meetings, computermediated communication and higher education, among others, and the relation between pragmatic aspects and issues of identity and culture. However, in the parts that follow, we have tried to make reference to studies that cover these areas, where possible.

Areas of intercultural pragmatic research As outlined in Blum-Kulka and colleagues (2008), there are at least four major areas of intercultural pragmatics: (1) historical pragmatics devoted to the study of written texts from earlier periods with an emphasis on the culturally unique and particular, for example, the use of certain discourse markers or norms of politeness in selected texts and historical contexts; (2) intercultural workplace communication, where universal principles of interaction in multilingual communities of practice in the workplace may arguably be revealed; (3) interculturality as emergent discursive construction involving ongoing negotiation between participants in intercultural events that may foreground or suppress cultural distinctiveness; and (4) globalisation of discourse exploring inter alia the features of English as a lingua franca in an increasingly globalised world. In what follows, these themes will be briefly described.

Historical intercultural pragmatics This field of intercultural pragmatic inquiry aims to extend our knowledge and understanding of texts emanating from periods other than the current one, focusing on links between language, society and culture. While linguistic anthropology stretches our knowledge and imagination in space, historical pragmatics does so in time, opening up distant discursive worlds of the past (cf. Jacobs and Jucker 1995; Jucker et al. 1999). This extension of the familiar present to the past permits us to evaluate current models on a broader database, facilitating more valid, potentially universal claims. Another important contribution of historical pragmatics to the field of intercultural pragmatics are the deeper insights it may afford into processes of linguistic variation and change. Further, historical pragmatic studies may re-construct shifts in social relations, linguistic norms and conventions and communicative preferences. Historical pragmatic studies of fictional and nonfictional historical texts typically look at the realisation of specific linguistic forms, such as discourse markers and connectives, or at phenomena like politeness, speech acts, speech act sequences, terms of address, meta-communication and reflections on dialogues. In all of these research strands, insights may be gained into communicative norms holding in the distant past (Jacobs and Jucker 1995: 10). Taken together, historical pragmatics is important and useful as a basis for providing new ideas and data for analysis, and for historically validating current linguistic and cultural hypotheses and theorems.

169

Alessia Cogo and Juliane House

Intercultural workplace communication This research area comprises at least the following three major fields: (1) organisational communication with its long tradition of research on organisation or workplace communication, including the study of different organisational cultures or subcultures (cf. e.g., Martin 1992; Boden 1994); (2) an equally established tradition exists in the field of social behaviour in institutional settings, where interactions in institutions have been subjected to intensive qualitative micro-analysis. Of the genres analysed in this research strand, medical discourse has been investigated particularly frequently (cf. e.g., Heritage and Maynard 2006); (3) much work on intercultural workplace communication has been done in the context of language for specific purposes and foreign or second languages (cf. e.g., Ehlich and Wagner 1995; Firth 1995; Bargiela-Chiappini et al. 2013). Among the major issues discussed in the field of intercultural workplace communication are cultural differences between the interactants (cf. Bargiela-Chiappini and Harris 1997), the dominant role of English as a lingua franca (ELF) in the international workplace (cf. Firth 2009 and see Current work on ELF pragmatics), and the ‘normality’ of international workplace interactions (cf. e.g. Firth 1996; Rasmussen and Wagner 2002).The work by Firth in this latter area is most influential: he describes the successful ‘normalising’ behaviour of interactants in ELF workplace interactions achieved jointly by ELF users’ interactional work for immediately relevant local purposes. Another important research strand can be located at the intersection of professionally motivated social encounters and relations (cf. e.g., Day et al. 2008). On the basis of detailed qualitative micro-analytical studies designed to uncover the complex connection between professional work communication and social networking, it is claimed that a major characteristic of workplace interactions is what one may call their inherent relational instrumentality.The establishment of social connections is revealed as being subservient to the goal of the business conversation, social relationships being part and parcel of reaching this overriding and preset goal. Much further research remains to be carried out on the interplay between professional communication and participants’ drawing on their multilingual and multicultural resources, if one is to better understand the meaning of multilingual interactional competence in professional intercultural contexts.

Interculturality as discursive construction One of the major concerns of intercultural pragmatics research has always been an examination of communication between members of different lingua-cultural communities. An important assumption underlying such research has been that interactants’ membership in different lingua-cultural communities is liable to render any interaction they are involved in per definitionem ‘intercultural’. This membership was thus thought to determine how interactants perceive and perform in these intercultural contexts, and how it may be a direct cause of so-called intercultural misunderstandings (cf. e.g., House et al. 2003). Recently, an alternative discursive-­ constructionist view of what happens in such ‘intercultural’ encounters has been gaining ground. Here, research focuses on how participants themselves perceive and handle such interactions, how relevant the ‘interculturality’ is for them, and how it influences the nature of the interaction. Conversation analysis and membership categorisation analysis are used here to find out how interactants co-construct themselves as members of a particular lingua-cultural group, how they invoke, reaffirm or, indeed, resist being considered a member of a particular group and ideology, and how they go about claiming expertise in terms of cultural and linguistic knowledge (cf. e.g., Nishizaka 1995 for Japanese and Zhu 2013 for Chinese interactants). Such an approach is designed to reveal how interactants’ shifting identities, understanding and behaviour emanate from the interaction and become publicly known in and through it. 170

Intercultural pragmatics

Studies in the discursive constructionist framework have focused on the resources and practices on which interactants rely when they make claims to membership in different linguacultural communities. The studies have shown that L1 speakers do indeed orient sometimes to pre-fixed ethno-cultural borders, claiming sole membership to cultural histories and artefacts effectively excluding non-L1 speakers as cultural outsiders (cf. e.g., Yoshino 1992). In other contexts, however, interactants in ‘intercultural’ encounters are found to emphasise shared identities, orienting towards the inter or the trans in intercultural encounters without relying on an all-decisive and omnipresent membership in an L1 cultural group (cf. e.g., Kim 2008 who notes shifts between foregrounding and backgrounding of cultural matters often depending on topic, expertise and emergence of social relations in the interactions). Of particular interest in the discursive-constructionist approach has been the investigation of the role of language in multilingual settings. While in other approaches, a division of individuals into native speakers and nonnative speakers has been commonly considered to correlate strongly with linguistic knowledge, behaviour and group affiliation, the discursive-constructionist paradigm regards native/non-nativeness as never pre-fixed but potentially emerging in and out of the interaction. Of particular concern in the context of English as a lingua franca (see Current work on ELF pragmatics) is the idea of replacing the notion of ‘nativeness’ with a concept of ‘expertise’ (Rampton 1990) in the use of a certain language and in the practices of a certain domain. This expertise is considered open to variation and change over time and in different communities of practice. It is regarded as a fluid resource available for achieving affiliation and identification.

Globalisation of discourse In the field of globalised discourse, computer-mediated linguistic aspects play an important role. In the following they will be briefly characterised. Linguistic aspects of globalised discourse can be located at various linguistic levels, for example, at the lexical, semantic, pragmatic-discourse and socio-semiotic levels. At the lexical level, globalised discourse has often been characterised as featuring a large amount of so-called internationalisms (cf. e.g., Braun et al. 2003), and, here especially, ‘Anglicisms’. Such borrowings have either been categorically condemned for damaging local languages in their expressive and functional potential, or they have been looked upon more positively as facilitating intercultural communication processes by creating common lexical reservoirs (see Multilingual resources). At the semantic level, globalised trends have been identified in the semantic development of routine formulae and illocutionary force indicating devices (IFIDS) such as please, sorry or thank you. The semantic flexibility of such seemingly fixed items has often been underrated in intercultural contexts. As Terkourafi (2011) has shown, such borrowings tend to serve functions that already exist in the receiving languages, while at the same time contributing to the development of new, additional functions from their original functions in English (see also Heine and Kuteva 2005, who make a similar point with regard to language change through language contact). At the pragmatic, discourse level, globalised norms of written discourse in various genres seem to ‘drift’ towards English-based rhetorical structures. English-based forms of rhetorical patterns have been observed to filter into academic, scientific and economic discourse in many other languages. Indeed, recent studies on intercultural rhetoric (cf. e.g., Canagarajah 2007; Connor 2012), point out that since cultures are in principle hybrid and dynamic, negotiation and accommodation processes tend to be set in motion in any text production. At the socio-semiotic level, intercultural globalised discourse has been described as an assemblage of ‘globalised linguistic signs’ that lead to the creation of new globalised multilingual 171

Alessia Cogo and Juliane House

landscapes (cf. Gorter 2006). Linguistic landscapes are an important new research strand in intercultural pragmatics which may reveal how written language is made visible in public, often urban spaces in hitherto unexplored ways. Much research has been done on East Asian megacities like Seoul, Shanghai, Hong Kong, Singapore or Tokyo (cf. e.g., Backhaus 2006; Kim 2008), illustrating the increased usage of multilingual and multicultural signs in a globalised urban world. An important area in studies of globalised intercultural discourse is concerned with the use of modern technology. Computer-mediated communication, including many Internet domains revealing influential new communicates of practice, have thus become increasingly popular research foci. Many studies in this paradigm look at the influx of English words into blogs or television commercials in other languages (cf. e.g., Lee 2006). Such imports are remarkable, because they do not fill a lexical gap in the receiving language; rather, perfectly simple words are more often than not easily available in the receiving language, but they are strategically replaced by English words in order to achieve certain effects. For instance, the use of the English word car in a blog is chosen to add a certain pragmatic function like advertising one’s ‘global identity’, modernity or rebellion in computer-mediated communication. Clearly, the English language is here instrumentalised as a resource for interculturalising a native language. It remains an open question, however, whether the Internet is on its way to becoming an ‘equalising’ force, an all-embracing ‘global language’ in a ‘virtual universe’ able to create an egalitarian ubiquitous society without political, social or linguistic borders – a type of universal intercultural communication – or whether it is an elitist tool for promoting more inequality between the haves and the have-nots.

Current work on ELF pragmatics Current work in intercultural pragmatics has more recently delved into the contexts where English is used in its global role. Findings in this area have shown the intercultural nature of lingua franca communication, where more attention is dedicated to working towards pragmatic understanding, focusing on how interactants construct and negotiate understanding and how they solve miscommunication problems. An overview of current work in English as a lingua franca (ELF) pragmatics can be divided into four main areas: negotiation of meaning (including resolution of non-understanding), use of interactional elements (including discourse markers), idiomatic expressions and multilingual resources.

Negotiation of meaning Negotiation of meaning is possibly the area that has been most developed in ELF pragmatics. Extensive work has been done on the strategies used to construct and negotiate meaning and resolve issues of non-understanding. Researchers have focused especially on two types of strategies: those moves that are performed after the trouble source and those that occur before any signal of non-understanding has taken place in conversation (cf. Cogo and Dewey 2012: 120–135). Pre-emption signals draw attention to a specific conversational item before any obvious nonunderstanding has occurred. Examples are repetition and paraphrasing, both proactive measures employed to avert problems of understanding.The following extract comes from Kaur’s study of ELF pragmatic strategies among Malaysian postgraduate students. Here the participants discuss a course-related issue.   1. K: no I- I mean it’s erm:. . . . (3.4) this is to my understanding I   2. think that erm er: it-it’s like erm:. . . . (1.2) erm it’s like erm why   3. SME in each regions need to go into e-trade I think this my 172

Intercultural pragmatics

  4.   5. V:   6.   7.   8.   9. 10. 11. 12. 13. K:

understanding of something like this. no not- not need to go into e-trade. that’s not what we’re considering. this global response has to::. . . . (0.7) has to deal with. . . . (0.7) either. . . . (0.7) positive response or negative response. now if we say this is why they have to go to e-trade that means only positive. . . . (2.7) do you understand me? this global response we must think. . . . (0.6) if it is positive. . . . (0.6) or negative in a particular country, we must state it and then. . . . (0.7) give the instances °mm°(Kaur 2009: 112)

With this example, Kaur emphasises how both paraphrasing and comprehension checks are used to enhance ‘understanding and to possibly even secure it in the event that shared understanding has not yet been achieved’ (2009: 113). In line 9, after disagreeing with K and explaining that this global response has to deal with either positive response or negative response (lines 6–9), V makes a pre-emptive move designed to clarify any possible misunderstanding by producing the comprehension check do you understand me? This check is uttered just after a prolonged silence of 2.7 seconds, a comparatively long pause in conversational analysis (Jefferson 1989). K’s silence during the long pause is also a failure to take up the turn and may constitute a signal of nonunderstanding on the part of K. To pre-empt any of K’s doubts, V follows the comprehension check in lines 9–11 with a further paraphrase of lines 6–9. Pre-empting strategies, such as pre-emptive comprehension checks and paraphrases, can prove crucial as they require speakers to engage in a constant joint effort to monitor understanding at every turn and notice, even when explicit signals of non-understanding are absent. This constant monitoring and attention underlines how mutual understanding in intercultural exchanges is not taken for granted, but co-constructed and checked on a turn-by-turn basis. A number of studies (e.g., Pitzl 2005 Lichtkoppler 2007; Cogo 2009; House 2010; Mauranen 2012; among others) have shown that strategies such as repetition, both self-repetition and other-repetition, and rephrasing are commonly used for this kind of interactional monitoring in intercultural communication. Furthermore, different kinds of repetitions, ranging from wordby-word repetition to more rephrasing, are used for various functions. Repetition, for instance, can be used both to draw attention to a possible non-understanding and to solve a communication problem afterwards. House has pointed out in various ELF studies (e.g., 2010, 2011, 2012) that so-called Represents (cf. Edmondson 1981) are often used in ELF interactions to pre-empt potential communication problems. Represents – in the literature also known as echo-, mirror- or shadow-elements – are multi-functional gambits which serve to repeat or ‘represent’ (part of) previous speakers’ moves. Represents fulfill different functions: (1) as a strategy with which speakers’ working memory in comprehension and production is deliberately supported; (2) a coherence-creating strategy with which lexical-paradigmatic clusters are systematically built up for speaker and addressee; (3) a signal of receipt and confirmation of comprehension of one’s interactant; (4) a meta-­ communicative procedure that strengthens interactants’ awareness of their own and others’ talk. Here are two examples of the strategic use of Represents taken from the corpus of the Hamburg ELF project ‘Multilingualism and Multiculturalism in German Universities’: A: and you mean that English (2) is really getting important or taken for the education because the grammar is syntactic erm the grammar is very easy B: is easy is very easy 173

Alessia Cogo and Juliane House

C: if you start speaking English in France they will answer you in French D: answer in French that’s true E: that’s true The repetition of answer in French by D is not just a simple echoing of the interlocutor’s words, but it provides a more prominent agreement and therefore affiliation to the other speaker (C). Mauranen (2012) illustrates a similar use of other-repetition for showing a more explicit agreement in the following conversation from a seminar on internal medicine. S6: and the ASAT is all in the cytoplasm and in mitochondria so [that’s why] S5: [yeah] S6: the ALAT rises more easily S3: and is more specific S6: yeah more specific that’s right, S5: and then there are the serum (Mauranen 2012: 222–223; transcript adapted for readability) The repetition of more specific by S6 has the effect of emphasising the agreement, which is already expressed by ‘yeah’, thus making the message more explicit. The ‘added value’ (Mauranen 2012: 222) of the repetition should not be underestimated. Simply using other agreement tokens, such as yes or yeah, would not have had the same effect. Here S6 is not only agreeing, but siding with the interlocutor and showing that they align with their assessment. Overall, the repetition emphasises the stance of the speaker and contributes to making the message clearer and more explicit. Other studies (Mauranen 2006; Lichtkoppler 2007; Cogo 2009; Kaur 2012; Björkman 2013; Gotti 2014) have shown that the various kinds of repetition, represents and rephrase/paraphrase are often used as explicitness strategies. Research has highlighted how increasing explicitness is an important process behind other strategies emerging from lingua franca communication, such as metadiscourse and utterance completion, which make the purpose and understanding of an utterance explicit. Moreover, the various functions of repetition can be interpreted as a strategic answer to ELF users’ particular needs. Repetitions typically occur in genres, such as psychotherapeutic interviews, instructional discourse and aircraft control discourse, where information is deliberately and routinely restated to ensure understanding. The fact that ELF users make regular use of this convention to ensure proper intake shows that their strategic communicative competence is well developed. Overall, the main aim of intercultural pragmatic strategies used in ELF settings is getting the message across rather than focusing on form, or how the message is formulated. An interesting observation to make here is that recent findings in ELF empirical naturally occurring data (as opposed to set-up, experimentally collected data) have shown that the let-it-pass strategy (also called topic abandonment) is not frequent in ELF communication (cf. Firth 1996 and Mauranen 2006). This is very likely due to the purpose-oriented nature of the communication events researched (business context in Cogo 2012 and Pitzl 2005; higher education context in Björkman 2013), which unlike small talk or phatic communication, is meant to pay attention to the specific aims and its implications if these are not achieved. Such findings also support earlier observations (House 2003; Mauranen 2006) that communication problems are not very frequent in ELF discourse, largely because speakers resort precisely to these pre-emptive and negotiating strategies to ensure possible misunderstandings are either avoided or carefully negotiated. 174

Intercultural pragmatics

Interactional elements Making meaning in intercultural encounters is not only a matter of negotiating and solving communication problems, but certain discourse skills related to the management of the interaction are also needed. Intercultural speakers seem to use various interactional elements such as discourse markers and backchannelling, which aim at managing and co-constructing successful discourse. Discourse markers, like other interaction managing items, have the function of expressing meanings of information management, and they also serve the purpose of marking the interpersonal relations between the interlocutors. They range from very short, fixed expressions to longer units of more or less variable sequence. House (2009a) and Baumgarten and House (2010b) suggest that intercultural speakers may not only use different discourse markers but also attribute different functions to them. For instance, in the case of the discourse marker you know, pragmatic research on NS discourse emphasises the interpersonal function, which is often used in the construction of intersubjectivity, or appealing to the addressee’s sharing of relevant knowledge (cf. e.g., Östman 1981). However, in intercultural exchanges, the relational function does not seem to be the most salient. Here, the discourse marker occurs frequently in considered, task-oriented talk (as opposed to small talk and informal exchanges). It is used to reinforce the speaker’s position, especially in close occurrence with conjunctions, as exemplified in the following extract: JYPY:  but you know like they are doing some RIDICULOUS things there was bff (1.5) (House 2009a: 182) The use of you know in these cases also adds focus to the topic at hand – the speaker uses this discourse marker to draw the interlocutor’s attention to what follows. In other words, House shows how this discourse marker is not necessarily related to intersubjectivity, politeness or relational purposes; rather it acquires different functions in intercultural exchanges and seems to contribute mainly to the discourse intensity and organisation of communication. Another frequent interactional element is the minimal response or backchannel item (see, e.g., Cogo 2009; Kalocsai 2011; Wolfartsberger 2011), that is, short verbal and nonverbal signals given to the interlocutors to indicate that s/he can continue speaking, such as yeah, ok, mhm mhm. Initial research in intercultural pragmatics (cf. House 1996) showed that L2 English speakers make limited use of discourse markers of the backchannelling kind, but more recent investigation into naturally occurring interactions shows that backchannels are common and used for various functions. Baumgarten and House (2010a) and House (2013), for instance, draw a distinction between the agreement marker yes and the use of yeah as a presentation marker at the beginning or at the final stage of a turn. At the beginning of an utterance, the marker is often used as a face-keeping device, to display attentiveness and gain time (similarly to well), while at the end of the utterance, it is used for positive final emphasis. The following example, taken from Baumgarten and House (2010a), shows the marker in a turn-initial position in academic discourse at an oral examination. E: ehm (2) why is age relevant/ (1.5) or how is age relevant in the process of second language acquisition? S: yeah, in the. . . . (1.5) if you just start with the process of learning a language In this instance, S is clearly not using yeah as an agreement token, but as a way of introducing the answer and, probably, of displaying attention, while gaining time to think about the answer. 175

Alessia Cogo and Juliane House

Mauranen explores how various kinds of interactive elements can be used in talk when managing the topic of conversation, when organising the succession of turns and when displaying stance. One example she provides from her English as a Lingua Franca in Academic settings (ELFA) corpus is the phraseological unit in my point of view (a conflation of in my view and from my point of view). As she puts it, this uncommonly long phraseological unit ‘is at the interface of linguistic convention and creativity’ (2009: 231), in the sense that it uses known lexical material which is displayed in an unconventional sequence. As well as being used in turn-taking, this expression signals the opinion of the speaker, but also functions as a marker of a divergent view from that of the interlocutors. Backchannels, phraseological units and discourse markers of these kinds all contribute to supporting meaning-making in intercultural communication. Other phenomena that facilitate communication in this sense are idioms and idiomatic expressions.

Idioms and idiomatic expressions In any natural language, idiomatic expressions are used for effective communication and economy of effort, and intercultural exchanges are no exception. In these exchanges, however, in order to avoid problems with what has been termed unilateral idiomaticity (the use of idioms which interlocutors do not share, see Seidlhofer 2011), speakers do not assume general knowledge of the formulaic pattern and take care in explaining or negotiating the meaning of idioms used. While previous studies seem to point towards avoidance of idiomaticity, recent data has shown that ELF speakers do not avoid using idiomatic expressions, but they avail themselves of the open-choice principle, creating new idioms or adapting the ones available to them. From naturally occurring data taken from the Vienna Oxford International Corpus of English (VOICE) corpus, Seidlhofer (2009; 2011) provides evidence of ELF speakers dynamically co-constructing idiomatic expressions in discourse. S3: we don’t have to think ONLY to European master multilingual problems [and] things like that but we also have to (.) S2: [mhm] S3: to think about (.) ENdangered (.) [er] S1: [mhm] S3: [fiel-] S2: [species] S3: species @ yes @ @ {later on in the conversation} S1: you you’re talking about (.) er: er (.) I saw (.) er two factors the enDANgered (.) factor (.) so to say. you have low numbers and you have to have a (.) hh to have critical masses. (Seidlhofer 2009: 206–207) The example shows how speakers cooperate to modify a conventional phrase and create a new collocation. The conventional collocation of endangered species is known, and the speakers seem to be aware that this is the most common collocation, which is indicated by their laughter in the final line of the excerpt (transcription convention: @), but S3 still uses endangered with field (of study). This new collocation is later exploited various times in the conversation, in collocations such as endangered factor, endangered program and endangered study (Seidlhofer 2011: 139). Here, speakers go beyond just accepting the non-conventional phrasing; they also contribute to building a shared ELF repertoire in interaction. This is also an example of the kind of speakers’ engagement in communication and alignment with co-speakers in the effort to 176

Intercultural pragmatics

create a successful exchange. Similarly, Pitzl’s (2009) research has shown how speakers rephrase idiomatic expressions, building on their metaphoricity and creatively changing them to the extent that certain innovative expressions may come to show signs of re-metaphorisation (Pitzl 2012). This process, apart from contributing to the meaning making of the exchange, may also serve to establish a sense of playfulness and ingroup belonging. Idiomatic expressions are also found to serve discourse strategies such as introducing a topic and drawing attention to an item of discussion, often accompanied by paraphrase or listener-oriented strategies, which facilitate understanding of an idiom (Francheschi 2013). Idiomatic expressions are also found to occur in the original language, that is, in the L1 or other language of the repertoires of participants who make use of their multilingual resources to enhance meaning in the intercultural exchange.

Multilingual resources A common element of intercultural communication in certain ELF contexts is the exploitation of multilingual resources, involving the strategic use of language alternation or the reliance on speakers’ multilingual background for cognitive transfer phenomena. Research has shown that multilingual strategies are used to various ends, which are often overlapping and inter-related and addressed: (1) the sharing of a sense of non-nativeness (Cogo 2009, 2012; Hülmbauer 2009); (2) the collaborative construction of meaning (Cogo 2010, 2012; Vettorel 2014; and (3) the creation of a sense of intercultural community membership or identity, among others. Across most of these studies is an understanding that using multilingual resources involves the development of accommodation practices, which ensure sensitivity to speakers’ cultural backgrounds and linguistic repertoires, while adapting their resources for communicative effectiveness. One way of prioritising communicative effectiveness involves exploiting speakers’ shared non-nativeness. For instance, Hülmbauer (2009) shows how ELF speakers rely on their shared sense of non-nativeness by making use of the pool of plurilingual resources at their disposal. This may take place in the form of overt code-switching moves, but also in more covert transfer phenomena and the use of cognates (Hülmbauer 2011). House (2012) discusses cases where ELF interactants code-switch into their L1 where this L1 is shared by at least one of the speakers. Such code-switching tends to take place at certain parts of the talk, that is, in phatic and routinised opening and closing phases and at topic transition points. The following is an example of code-switching into German in the closing phase of an ELF academic advising session taken from the corpus of the Hamburg project on English as a lingua franca: ‘Multilingualism and Multiculturalism in German Universities’: P: [. . .] mister (name) has made some publication in Germany about this how he comes to this number this is for example the big discussion the be the debate about this number (.) it is something which must be in your thesis S: {fast} ja P: okay? for this YES and erm haben sie noch was? nee des is der erste teil war fertig S: ja Other studies explore how speakers may engage in a collaborative co-construction of meaning, while signalling their community membership and making code-switching an intrinsic part of their ELF interactions. In the following example, a negotiation and expansion of meaning by using multilingual resources becomes an opportunity to display intercultural identities. A group of colleagues engages in small talk and comments on a set of pictures posted by an engaged 177

Alessia Cogo and Juliane House

couple on a personal website. S1 says he has to go to the couple’s wedding in France and starts describing the couple’s pictures. 18 S1 no it’s nice .°.°. they have picture of them 19 S2 =eh? 20 S1 =pictures of them you know. . . . in Australia [in Kathmandu in Tibet like 21 22 S2 [@@@@ 23 S3 they sent pictures. . . . [on the Internet? 24 S1            [it’s nice but it’s a bit 25 S3 =too much eh? 26 S1 =cheesy 27 S1 [YE::AH 28 S3 [YE::AH 29 S2 yeah a bit too much I think @@ 30 S1 so. . . . blue flower we say. . . . fleur bleue 31 S3 why?. . . . [to say that it’s cheesy? 32 S1        [fleur-yeah. . . . fleur bleue means. . . . you know when you have these pictures with little 33 34 angels of 35 S2 a:::h  [yeah 36 S3       [yeah 37 S1 fleur bleu 38 S2 kitsch- [kitschig 39 S1        [kitschig yeah @@@ (Cogo 2010: 301) In this extract we can observe numerous aspects addressed in this chapter. For a start, the cheesy issue is an opportunity to engage in negotiation of meaning, where the adjective cheesy (line 26) is taken up by the participants and negotiated by means of code-switching into French first and German later, thus searching for common understanding and nuance of meaning. While S1 attempts to provide a different version of the cheesy meaning with fleur bleue, he pre-empts a possible non-understanding with a paraphrase of the meaning, that is, the pictures of little angels carrying blue flowers, which are considered (in S1’s sociocultural community) cheesy. Then, S2 offers her own interpretation with kitschig in line 38, the German version of the adjective, and they confirm understanding with various repetitions and confirmatory discourse markers (yeah and laughter). This extract also offers an example of the sort of interactional supportive work explored in Interactional elements, in their use of backchannelling, discourse markers and utterance completions, which is characteristic of ELF intercultural exchanges. Moreover, this conversation seems to go beyond the search for meaning and negotiation of understanding. Participants bring their own sociocultural background and build on other’s contributions, while at the same time creating a sense of comity, solidarity and ingroup belonging. Overall, ELF research has developed exponentially in recent years, and the initial emphasis on systematic and recurrent features has been replaced by a focus on the flexibility and fluidity of ELF, which, in pragmatic terms, has translated into more research on pragmatic processes and accommodation (cf. Jenkins et al. 2011). Lately, more work has also been done on emergent multilingual practices, which highlight the complexity and fluidity of ELF intercultural pragmatics. These include translanguaging practices (cf. Cogo 2012 and 2015), which involve 178

Intercultural pragmatics

mobilising a repertoire of resources in a flexible and integrated way, and go beyond the stable and fixed separation between languages.

Conclusion From the discussion of current work offered in this chapter, it seems clear that ELF is ‘an opensource phenomenon’ (House 2014: 364), which is constantly adapted by its users and varies according to the context where it is used.The fluidity of this kind of communication, the ‘inherent variability’ of the lingua franca factor (Firth 2009) and the super-diversity of its contexts (Cogo 2012) imply that language in an ELF context is always and constantly under negotiation and reliant on adaptations which make use of the users’ linguacultural repertoires. In current research on ELF, this characteristic has been reflected in the dilemma most scholars face when dealing with natural conversational data in the field – that is, how to explore ELF’s contingent fluidity and hybridity while not denying the ‘observed regularities’. As highlighted in every aspect of what we discussed in this chapter, ELF discourse displays more flexibility and adaptability than generally observed in other kinds of communication. Current research in this field maintains that intercultural talk, maybe more than in any other natural communication, depends on the interlocutors’ cooperation, which involves both coconstructing a shared repertoire and collaborating in accommodating unconventional expressions. As Jenkins maintains, accommodation is ‘possibly the single most important pragmatic skill in ELF communication’ (Jenkins 2011: 928). Accommodation is a key element for successful communication, and this is shown in current research (see Jenkins et al. 2011) as well as in the examples analysed here. Constant intercultural adaptations, together with fluidity and hybridity, which are characteristic of ELF intercultural encounters, raise the issue of how to deal with pedagogical applications, specifically in relation to the topic of this chapter, the issue of how to apply lingua franca pragmatics to the language teaching classroom. ELF research so far has pointed to features and processes of successful communication, which can usefully find classroom adaptations, and work on applications of these for language teaching is well under way (Bayyurt and Akcan 2015; Bowles and Cogo 2015; Vettorel 2015). Some of these features and processes have been mentioned in the strategies explored in this chapter, but more research in this direction is needed, especially in collaboration with language teaching practitioners.

Suggestions for further reading Cogo, A. and Dewey, M. (2012) Analysing English as a Lingua Franca: A Corpus-Driven Investigation. London: Continuum. This monograph focuses on ELF spoken communication and provides empirical analysis of lexicogrammatical and pragmatic aspects of ELF, emphasising the interconnection between the two and the role of accommodation for processes of adaptability, fluidity and innovation. The last chapter also provides implications for applied linguistics and ELT. House, J. (ed.) (2009a) ‘The pragmatics of English as a Lingua Franca’ [Special issue], Intercultural Pragmatics, 6(2). This special issue on the pragmatics of ELF is a collection of five papers exploring different aspects of intercultural pragmatics. Firth’s paper explores one of the characteristics of ELF, namely its inherent variability. House’s contribution is an empirical investigation of the pragmatic marker you know, Seidlhofer’s is on accommodation and the idiom principle, Mauranen’s is on chunking and emphasises regularity and variation in ELF phraseological units, and finally Pickering focuses on intonation as a pragmatic resource in negotiation sequences. Mauranen, A. (2012) Exploring ELF. Academic English Shaped by Non-native Speakers. Cambridge: Cambridge University Press.

179

Alessia Cogo and Juliane House This monograph explores the use of ELF in academic contexts and is based on analysis of empirical data from the ELFA corpus.The pragmatic chapters on metadiscourse, negotiation, repetition and rephrasing provide a wealth of examples of communication in a high-stakes context like academia. Seidlhofer, B. (2011) Understanding English as a Lingua Franca. Oxford: Oxford University Press. This monograph is a conceptual and empirical exploration of ELF based on the VOICE corpus. It challenges some widespread assumptions about ‘English’ and shows how ELF should be understood as a dynamic, adaptable and creative use of English.

References Backhaus, P. (2006) ‘Multilingualism in Tokyo: A look into the linguistic landscape’, International Journal of Multilingualism, 3: 52–66. Bardovi-Harlig, K. and Hartford, B. (eds.) (2005) Interlanguage Pragmatics. Exploring Institutional Talk. Mahwah, NJ: Lawrence Erlbaum. Bargiela-Chiappini, F. and Harris, S. (eds.) (1997) The Language of Business: An International Perspective. Edinburgh: Edinburgh University Press. Bargiela-Chiappini, F., Nickerson, C. and Planken, B. (eds.) (2013) Business Discourse, 2nd ed. London: Palgrave Macmillan. Baumgarten, N. and House, J. (2010a) ‘Discourse markers in high stakes ELF academic interactions: Oral exams’, paper presented at the Vienna ELF conference,Vienna. Baumgarten, N. and House, J. (2010b) ‘I think and I don’t know in English as lingua franca and native English discourse’, Journal of Pragmatics, 42(5): 1184–1200. Bayyurt,Y. and Akcan, S. (eds.) (2015) Current Perspectives on Pedagogy for English as a Lingua Franca. Berlin: De Gruyter Mouton. Björkman, B. (2013) English as an Academic Lingua Franca. Berlin: De Gruyter. Blum-Kulka, S., Blondheim, M., House, J., Kasper, G. and Wagner, J. (2008) ‘Intercultural pragmatics, language and society’, in P. van Sterkenburg (ed.) Unity and Diversity of Languages. Amsterdam: Benjamins. 155–173. Blum-Kulka, S., House, J. and Kasper, G. (eds.) (1989) Cross-cultural Pragmatics: Requests and Apologies. Norwood, NJ: Ablex. Boden, D. (1994) The Business of Talk: Organizations in Action. Cambridge: Polity Press. Bowles, H. and Cogo, A. (eds.) (2015) International Perspectives on English as a Lingua Franca: Pedagogical Insights. Basingstoke: Palgrave. Braun, P., Schaeder, B. and Volmert, J. (2003) Internationalismen. Tübingen: Niemeyer. Brown, P. and Levinson, S. (1987) Politeness. Some Universals in Language Usage. Cambridge: Cambridge University Press. Canagarajah, S. (2007) ‘Lingua franca English, multilingual communities and language acquisition’, The Modern Language Journal, 91: 923–939. Cheng, W. (2012) ‘Speech acts, facework, and relationship building’, in J. Jackson (ed.) The Routledge Handbook of Language and Intercultural Communication. London: Routledge. 148–163. Cogo, A. (2009) ‘Accommodating difference in ELF conversations: A study of pragmatic strategies’, in A. Mauranen and E. Ranta (eds.) English as Lingua Franca. Studies and Findings. Newcastle on Tyne: Cambridge Scholars Publishing. 254–273. Cogo, A. (2010) ‘Strategic use and perceptions of English as a lingua franca’, Poznań Studies in Contemporary Linguistics, 46(3): 295–312. Available from: http://core.ac.uk/download/pdf/22845.pdf (accessed 1 June 2015). Cogo, A. (2012) ‘ELF and super-diversity: A case study of ELF multilingual practices from a business context’, Journal of English as a Lingua Franca, 1(2): 287–313. Cogo, A. (2015) ‘Complexity, negotiability, and ideologies: A response to Zhu, Pitzl, and Kankaanranta et al.’, Journal of English as a Lingua Franca, 4(1): 149–155. Cogo, A. and Dewey, M. (2012) Analysing English as a Lingua Franca: A Corpus-Driven Investigation. London: Continuum. Connor, U. (2012) Intercultural Rhetoric in the Writing Classroom. Ann Arbor: University of Michigan Press. Day, D., Jensen, A. and Millar, S. (2008) ‘Making communication work: Instrumental and relational aspects of intercultural communication in the globalized workplace’, paper presented at the 18th International Congress of Linguists, Seoul.

180

Intercultural pragmatics Edmondson, W. (1981) Spoken Discourse: A Model for Analysis, (Longman Linguistics Library, vol. 27). London: Longman. Ehlich, K. and Wagner, J. (eds.) (1995) The Discourse of Business Negotiations. Berlin: Mouton de Gruyter. Firth, A. (ed.) (1995) The Discourse of Negotiation. Oxford: Pergamon. Firth, A. (1996) ‘The discursive accomplishment of normality: On ‘lingua franca’ English and conversation analysis’, Journal of Pragmatics, 26: 237–258. Firth, A. (2009) ‘The lingua franca factor’, Intercultural Pragmatics, 6(2): 147–170. Franceschi,V. (2013) ‘Figurative language and ELF in university settings’, Journal of English as a Lingua Franca, 2(1): 75–99. Gorter, D. (2006) ‘The study of the linguistic landscape as a new approach to multilingualism’, International Journal of Multilingualism, 3: 1–6. Gotti, M. (2014) ‘Explanatory strategies in university courses taught in ELF’, Journal of English as a Lingua Franca, 3(2): 337–361. Grice, H. P. (1975) ‘Logic and conversation’, in D. Davidson and G. Harmon (eds.) The Logic of Grammar. Encino, CA: Dickerson. 64–75. Heine, B. and Kuteva, T. (2005) Language Contact and Grammatical Change. Cambridge: Cambridge University Press. Heritage, J. and Maynard, D. (eds.) (2006) Communication in Medical Care: Interaction between Primary Care Physicians and Patients. Cambridge: Cambridge University Press. House, J. (1982) ‘Opening and closing phases in German and English dialogues’, Grazer Linguistische Studien, 16: 52–82. House, J. (1996) ‘Developing pragmatic fluency in English as a foreign language: Routines and metapragmatic awareness’, Studies in Second Language Acquisition, 18(2): 225–252. House, J. (2000) ‘Understanding misunderstanding: A pragmatic-discourse approach to analysing mismanaged rapport in talk across cultures’, in H. Spencer-Oatey (ed.) Culturally Speaking. Managing Rapport through Talk across Cultures. London: Continuum. 145–164. House, J. (2007) ‘What is an “intercultural speaker”?’, in E. Alcón Soler and M. P. Safont Jordà (eds.) Intercultural Language Use and Language Learning. Berlin: Springer. 7–22. House, J. (2009a) ‘Subjectivity in English as lingua franca discourse: The case of you know’, Intercultural Pragmatics, 6(2): 171–193. House, J. (ed.) (2009b) ‘The pragmatics of English as a lingua franca’ [Special issue], Intercultural Pragmatics, 6(2). House, J. (2010) ‘The pragmatics of English as a lingua franca’, in A. Trosborg (ed.) Pragmatics across Languages and Cultures (Handbook of Pragmatics, vol. 7). Berlin: de Gruyter Mouton. 363–390. House, J. (2011) ‘Global and intercultural communication’, in G. Andersen and K. Aijmer (eds.) Pragmatics of Society (Handbook of Pragmatics, vol. 5). Berlin: de Gruyter Mouton. 607–626. House, J. (2012) ‘Subjectivity in English lingua franca interactions’, in N. Baumgarten, I. Du Bois and J. House (eds.) Subjectivity in Language and in Discourse. Bingley, UK: Emerald. 139–156. House, J. (2013) ‘Developing pragmatic competence in English as a lingua franca: Using discourse markers to express (inter)subjectivity and connectivity’, Journal of Pragmatics, 59: 57–67. House, J. (2014) ‘English as a global lingua franca: A threat to multilingual communication and translation?’, Language Teaching, 47(3): 363–376. House, J., Kasper, G. and Ross, S. (eds.) (2003) Misunderstanding in Social Life. Discourse Approaches to Problematic Talk. London: Longman. Hülmbauer, C. (2009) ‘We don’t take the right way.We just take the way that we think you will understand’ – The shifting relationship of correctness and effectiveness in ELF communication’, in A. Mauranen and E. Ranta (eds.) English as Lingua Franca. Studies and Findings. Newcastle on Tyne: Cambridge Scholars Publishing. 323–347. Hülmbauer, C. (2011) ‘Old friends? – Cognates in ELF communication’, in A. Archibald, A. Cogo and J. Jenkins (eds.) Latest Trends in ELF Research. Newcastle on Tyne: Cambridge Scholars Publishing. 139–161. Jacobs, A. and Jucker, A. (1995) ‘The historical perspective in pragmatics’, in A. Jucker (ed.) Historical Pragmatics. Pragmatic Development in the History of English. Amsterdam: Benjamins. 3–33. Jefferson, G. (1989) ‘Preliminary notes on a possible metric which provides for a “standard maximum” silence of approximately one second in conversation’, in D. Roger and P. Bull (eds.) Conversation: An Interdisciplinary Perspective. Cleveland, Philadelphia: Multilingual Matters Ltd. 166–196.

181

Alessia Cogo and Juliane House Jenkins, J. (2011) ‘Accommodating (to) ELF in the international university’, Journal of Pragmatics, 43(4): 926–936. Jenkins, J., Cogo, A. and Dewey, M. (2011) ‘Review of developments in research into English as a lingua franca’, Language Teaching, 44(3): 281–315. Jucker, A., Fritz, G. and Lebsanft, F. (eds.) (1999) Historical Dialogue Analysis. Amsterdam: Benjamins. Kachru, Y. (1998) ‘Culture and speech acts: Evidence from Indian and Singapore English’, Studies in the Linguistic Sciences, 28(1): 79–98. Kalocsai, K. (2011) ‘The show of interpersonal involvement and the building of rapport in an ELF community of practice’, in A. Archibald, A. Cogo and J. Jenkins (eds.) Latest Trends in ELF Research. Newcastle on Tyne: Cambridge Scholars Publishing. 113–137. Kasper, G. (ed.) (1996) ‘The development of pragmatic competence’ [Special issue], Studies in Second Language Acquisition, 18(2). Kasper, G. and Blum-Kulka, S. (eds.) (1993) Interlanguage Pragmatics. Oxford: Oxford University Press. Kaur, J. (2009) ‘Pre-empting problems of understanding in English as lingua franca’, in A. Mauranen and E. Ranta (eds.) English as Lingua Franca. Studies and Findings. Newcastle upon Tyne: Cambridge Scholars Publishing. 107–123. Kaur, J. (2012) ‘Saying it again: Enhancing clarity in English as a lingua franca (ELF) talk through selfrepetition’, Text  & Talk, 32(5): 593–613. Kecskes, I. (2012) ‘Interculturality and intercultural pragmatics’, in J. Jackson (ed.) The Routledge Handbook of Language and Intercultural Communication. London: Routledge. 67–84. Kim, S.-D. (2008) ‘The linguistic landscape of East Asian cities from the perspective of intercultural pragmatics’, paper presented at the 18th International Congress of Linguists, Seoul. Lee, J. S. (2006) ‘Linguistic constructions of modernity: English mixing in Korean television commercials’, Language in Society, 35: 95–91. Lichtkoppler, J. (2007) ‘ “Male. Male.” – “Male?” – “The sex is male.” The role of repetition in English as a lingua franca conversations’, Vienna English Working Papers, 16(1): 39–65. Martin, J. (1992) Cultures in Organizations:Three Perspectives. Oxford: Oxford University Press. Mauranen, A. (2006) ‘Signaling and preventing misunderstanding in English as lingua franca communication’, International Journal of the Sociology of Language, 177: 123–150. Mauranen, A. (2009) ‘Chunking in ELF: Expressions for managing interaction’, Journal of Intercultural Pragmatics, 6(2): 217–233. Mauranen, A. (2012) Exploring ELF. Academic English Shaped by Non-native Speakers. Cambridge: Cambridge University Press. Moder, C. L. and Martinovic-Zic, A. (eds.) (2004) Discourse across Languages and Cultures. Amsterdam: Benjamins. Nishizaka, A. (1995) ‘The interactive construction of interculturality: How to be a Japanese with words’, Human Studies, 18: 301–326. Östman, J.-O. (1981) You Know: A Discourse-Functional Approach. Amsterdam: Benjamins. Pitzl, M.-L. (2005) ‘Non-understanding in English as a lingua franca: Examples from a business context’, Vienna English Working Papers, 14(2): 50–71. Available from: http://anglistik.univie.ac.at/fileadmin/ user_upload/dep_anglist/weitere_Uploads/Views/Views0502ALL_new.pdf (accessed 26 June 2015). Pitzl, M.-L. (2009) ‘We should not wake up any dogs’: Idiom and metaphor in ELF’, in A. Mauranen and E. Ranta (eds.) English as Lingua Franca. Studies and Findings. Newcastle upon Tyne: Cambridge Scholars Publishing. 298–322. Pitzl, M.-L. (2012) ‘Creativity meets convention: Idiom variation and re-metaphorization in ELF’, Journal of English as a Lingua Franca, 1(1): 27–55. Rampton, B. (1990) ‘Displacing the “native speaker”: Expertise, affiliation, and inheritance’, ELT Journal, 44(2): 97–101. Rasmussen, G. and Wagner, J. (2002) ‘Language choice in international telephone conversations’, in L. Kang Kwon and T. Pavlidou (eds.) Telephone Calls: Unity and Diversity in Conversational Structure across Languages and Cultures. Amsterdam: Benjamins. 111–131. Seidlhofer, B. (2009) ‘Accommodation and the idiom principle in English as a lingua franca’, Intercultural Pragmatics, 6(2): 195–215. Seidlhofer, B. (2011) Understanding English as a Lingua Franca. Oxford: Oxford University Press. Selinker, L. (1972) ‘Interlanguage’, International Review of Applied Linguistics, 10(3): 31–54. Terkourafi, M. (2011) ‘Thank you, sorry and please in Cypriot Greek. What happens to politeness markers when they are borrowed across languages’, Journal of Pragmatics, 43: 218–235.

182

Intercultural pragmatics Thomas, J. (1983) ‘Cross-cultural pragmatic failure’, Applied Linguistics, 4(2): 91–112. Vettorel, P. (2014) English as a Lingua Franca in Wider Networking. Berlin: De Gruyter Mouton. Vettorel, P. (2015) New Frontiers in Teaching and Learning English. Newcastle upon Tyne: Cambridge Scholars Publishing. Wolfartsberger, A. (2011) ‘ELF business/business ELF: Form and function in simultaneous speech’, in Archibald, A., Cogo, A. and Jenkins, J. (eds.) Latest Trends in ELF Research. Newcastle on Tyne: Cambridge Scholars Publishing. 164–183. Yoshino, K. (1992) Cultural Nationalism in Contemporary Japan. London: Routledge. Zhu H. (2013) Exploring Intercultural Communication: Language in Action. London: Routledge.

Transcript conventions = latching (i.e. speech following the previous turn without a pause) short pause (unmeasured) . . . (0.5) measured pause [abc] overlapping speech XX unintelligible passage [. . .] omitted transcript °soft° speech which is softer than the surrounding talk @ laughing

183

16 IDENTITY AND MEMBERSHIP Dorien Van De Mieroop

Introduction When talking about a person’s identity, people tend to refer to some kind of core of a human being, and, in essence, attempt to answer the compelling question ‘who this person is’. On second thought, though, one soon realises that grasping the complexity of a person by answering a single question is quite an absurd endeavour. There are many reasons for this, the most obvious being that people change in the course of their lives, enacting different identities. Some of these identities have a certain chronological sequencing (e.g. child, teenager, young adult), while others may remain constant our entire lives (e.g. woman, Caucasian). Others may be present simultaneously, but are typically drawn upon in specific contexts (e.g. mother, wife, friend, employee), or they may gain a more or less prevalent status, occurring throughout a number of different contexts (e.g. American). But even interpreting this variety of potential identities as complementarily defining who a person is – or rather may be – means barely scratching the surface of the great complexity of the identity of a single person. After all, we think of so much more when we try to voice who we are: we may be fans of a particular soccer club or music genre and (sometimes) dress, speak and act accordingly; we may surprise ourselves sometimes because we arduously defend feminist viewpoints in a specific conversation even though we did not perceive ourselves as feminists at all; or we may speak with a pronounced Welsh accent and this may sometimes highlight our Welsh identity, while at other times it may not. In light of this vastness and fluidity of potential identities that may all be relevant to a single human being, it becomes quite clear what identity is not, namely: . . . identity does not signal that stable core of the self, unfolding from beginning to end through all the vicissitudes of history without change; the bit of the self which remains always – already ‘the same’, identical to itself across time. (Hall 1996: 17) In other words, instead of looking at the identity a person has, we view identities – in the plural – as the result of a constantly negotiated process, meaning that: . . . identities are never unified and. . . . increasingly fragmented and fractured; never singular but multiply constructed across different, often intersecting and antagonistic, 184

Identity and membership

discourses, practices and positions. They are subject to a radical historicization, and are constantly in the process of change and transformation. (Hall 1996: 17) Identities are thus viewed as multiple and fragmentary constructions closely related to the specific way they are situated in their local interactional contexts. This view will be discussed in more detail in the following section.

Identity through a social constructionist lens Researchers studying identity belong to a wide diversity of disciplines, ranging from anthropology, psychology, sociology, history and literature through to linguistics. Within the latter field, identity is studied by (critical) discourse analysts, interactional sociolinguists, conversation analysts and discursive psychologists. On the one hand, this diversity of disciplines and perspectives bestows the cross-disciplinary field of identity analysis with a vast methodological and theoretical diversity which may hinder its accessibility to newcomers to this research field. On the other hand, this diversity adds to the richness of the field, since the differing viewpoints continuously challenge researchers to critically scrutinise their own analytical approach by comparing it with others, and indeed, such interaction is especially important when studying such an intangible concept as identity. Important for the success of communication between researchers within the linguistic field is the common influence of one fundamental, underlying approach, namely ‘social constructionism’ (Berger and Luckmann 1967). According to this view, reality is a social construct, constituted by the interaction between individuals and their world. In other words, researchers share a common view on reality not as some kind of entity existing objectively and independently which people talk about, but rather as reality created through language in the course of human interaction. Naturally, this social constructionist view has important implications for the interpretation of the concept of identity, which, as sketched earlier, rejects essentialist, conventional views of identity as ‘singular or unitary’ and people as ‘self-contained, separate, independent and consistent across situations’ (Wetherell and Maybin 1996: 223–229). Instead, social constructionism views identities as ‘multifaceted’, emerging ‘in fields of meaning and practices which are socially and culturally organized’, and the person as ‘social through and through’ (Wetherell and Maybin 1996: 223–229). According to this approach, people may construct identities in communication which may disappear again as quickly as they have emerged, they may draw on identities that have emerged previously (e.g. being a woman) to differing degrees or they may construct identities from scratch on the spot, and these identities then only derive their meaning from the local interaction and negotiation with the other interlocutors. Such identities may be contradictory at times, and they have, as Zimmerman (1998: 90) terms it, different ‘home territories’, meaning that there are different types of identities, namely discourse, situated and transportable identities (1998: 90). Discourse identities are identities ‘integral to the moment-by-moment organization of the interaction’, such as the identity of current speaker, listener or storyteller. Situated identities are related to particular types of situations (e.g. business meetings) in which people ‘display an orientation to. . . . particular identity sets’ (Zimmerman 1998: 90), as for example enacting the role of chairperson of a meeting. Finally, transportable identities are those identities which individuals seem to carry with them. They are, for example, related to a person’s gender, ethnicity or age. These are ‘potentially relevant in and for any situation and in and for any spate of interaction’ (Zimmerman 1998: 90). However, given the social constructionist lens through which identities are studied, we need to be careful with assuming that there 185

Dorien Van De Mieroop

are some transportable identities which are always latently present, as if they are temporarily stored, but always ready to be taken from the shelf when needed in an interaction. According to social constructionist approaches, this is not the case; rather these identities are seen as being enacted each time they emerge, with mild to significant differences in identity performances depending on what people are doing at that specific point in time and with whom. For example, a woman may enact her female identity quite differently when being among other women as compared to when being among men, and when being among younger women than among older women, and depending on the specific situations she is in. The relationality principle as identified by Bucholtz and Hall is relevant in this context, as this principle emphasises that identities ‘acquire social meaning in relation to other available identity positions and other social actors’ (Bucholtz and Hall 2005: 598). Of similar relevance is Bucholtz and Hall’s (2005: 606) partialness principle which highlights that identity constructions are often partially deliberate, habitual and interactional. Consequently, even though a woman’s identity performances may differ significantly, one might be tempted to group them under the label of ‘female identity’. Such a grouping, however, would – in social constructionist terms – wrongly suggest that transportable identities are – to some degree – monolithic. It would as such conceal their intrinsic diversity.

Identity and membership: social identity theory Identity is often also explicitly related to membership, as the following definition by Kroskrity illustrates (2000: 111). He writes: ‘Identity is defined as the linguistic construction of membership in one or more social groups or categories.’ When talking about such group memberships, researchers usually draw on insights from social psychology, and more specifically on Tajfel’s (1981) social identity theory, which focuses on the crucial link between identity and group memberships. Tajfel argues that individuals create and can define their place in society through these memberships. As such, they are essential for the individual’s definition of his or her selfconcept (Tajfel 1981: 255). According to social identity theory, each person is a member of a number of different social groups with membership defined on the basis of variables such as similarity, proximity and common fate. In other words, an individual is more likely to become a member of a social group if s/he shares physical features with the ingroup members, if s/he is spatially close to them or if s/he encounters events on his/her life path that are similar to those of ingroup members. Since these variables may change over time, group memberships are also dynamic processes which individuals may shift in and out of. Such changes in group membership may be intentional and based on an individual’s comparison of different social groups and on his or her evaluation of the different group qualities. Members of a social group tend to regard the ingroup in terms of qualities that are considered in a more positive light than those of the outgroup. Furthermore, members of the outgroup are often perceived in a reductive way, not as individuals anymore, but as stereotypical representations of the average outgroup member. Social identity theory also recognises that though each individual is a member of a number of different social groups, these may differ in salience. For example, patriots may find their group membership on the basis of their nationality more important than their group membership based on racial criteria, while racists may see this the other way around. The salience of these group memberships is fluid since memberships are also closely related to members’ – ever changing – social contexts (Tajfel 1982: 2–3). Individuals who perceive a specific group membership, for example on the basis of racial criteria, as relatively unimportant, are almost forced to act in terms of group membership in some situations (Tajfel 1981: 239) as, for instance, when they find themselves in a situation in which they continuously have to face racist remarks and 186

Identity and membership

racial discrimination. If this situation persists, the initially unimportant racial group membership increases in salience, and this leads to a changing perception of the members of this ingroup. Similarities among these ingroup members are emphasised, while the differences with the outgroup members become more apparent, and encounters with the outgroup tend to be regarded from an intergroup rather than an interpersonal perspective. However, even though social psychologists emphasise the dynamic nature of group memberships, social identity theory is fundamentally irreconcilable with the social constructionist perspective currently shared by linguistic approaches to identity. This is mainly due to the fact that social psychologists tend to look for a priori, essentialist categories that reside in individuals, often on the basis of studies in experimental settings that thus occur in a social vacuum. Furthermore, the ingroup-outgroup opposition, which is so central in social identity theory, does not allow for the extreme fluidity of ephemeral, sometimes contradictory, identities that are locally negotiated on a turn-by-turn basis in interactions. Finally, in contrast to social constructionist approaches, language receives only limited attention from a social psychological perspective. Hence, the concepts from social identity theory introduced here serve mainly as inspiration for social constructionist approaches, even though they can be integrated in a meaningful way into discursive analyses as well when taking an emic, rather than the typically social psychological etic, perspective (Van De Mieroop 2015).

Identity and membership in globalised, multicultural societies Currently, much research on identity and membership takes the multicultural nature of our modern day societies into account. This is not surprising given that globalisation and migration have caused significant societal changes which have led to a considerable increase of intercultural contact between individuals. This has not been inconsequential for human beings’ notions of self, as individuals have been increasingly confronted with differences. In addition, since people tend to define themselves comparatively in terms of ‘what’ they are as well as ‘what’ they are not (e.g. not male, not black) (cf. Baynham 2006: 384), the increase in intercultural contact has led to a further scrutiny of individuals’ self-definitions. Also, it has contributed to the growing interdisciplinary interest in identity in the past decade. The multiplication of the occasions for contact with the other has brought with it a problematization of the concept of identity itself and an effort to understand the relationship between people’s sense of membership in a community, the beliefs and social practices that define that sense of membership, and its expression and manifestation in social behavior. (De Fina 2006: 351) Due to these fundamental changes in the constitution of societies around the world, the link between identity and social group membership, which had long been established by social psychologists within the context of social identity theory (see earlier in this chapter), has also received increased emphasis in studies of identity that follow a social constructionist approach. Furthermore, although the focus of these social constructionist studies remains on the close scrutiny of the role that language plays in the concrete communicative practices under study, the need to take elements from the macro level into account when studying the micro level has also become more important. Indeed, there has been ‘a growing uneasiness’ with approaches that look at local, turn-by-turn linguistic constructions without scrutinizing their ‘embeddedness into larger social and political processes’ (De Fina and Baynham 2005: 4). After all, the 187

Dorien Van De Mieroop

way societies deal with minority groups (e.g. immigrants and incoming refugees, as well as marginalised and underprivileged groups within the society) largely influences the dominant discourses regarding the effect these social processes have on the community. These effects may vary between two poles of a continuum from reflecting an openness towards diversity and framing it as an enrichment of the ‘original’ society on one hand, to frantically protecting the ‘original’, national identity on the other hand. The governing atmosphere in a particular community may promote certain dominant discourses and, conversely, yield so-called silenced stories, namely, ‘those stories that are ‘ignored/silenced/othered by the public discourse on displacement, migration and settlement’ (De Fina and Baynham 2005: 2). Tellers of such stories need to resist the particular society’s dominant discourses and produce counternarratives, in which the construction of social identities, group memberships and ingroup-outgroup dichotomies are often heavily negotiated. Apart from narratives, often considered ‘prime vehicle[s] for expressing identity’ (De Fina 2015: 351) (see later in this chapter), group memberships also play an important role in studies focusing on the construction of identity in day-to-day language use, for example in the workplace. Workplaces are typically sites in which individuals discursively construct their identities as professionals and members of an institutional collective. These professional and institutional identities also interact with the context in which they are constructed, since it is against the backdrop of contextual norms that identities are shaped and reshaped. These norms, surpassing the local, interactional level, can be situated on three levels as defined by Holmes and colleagues (2011). They include (1) the community of practice norms, pointing at the typical ways in which interactions within institutional teams take place; (2) the organisational norms, referring to the institutional expectations regarding appropriate communicative behaviour; and (3) the larger, societal norms, which may entail particular shared ideas about how, for example, men and women or experts and novices should behave vis-à-vis each other. These norms are not static entities, but rather interact with each other and are constantly renegotiated and potentially challenged by interlocutors. Crucial, however, is that minority group norms intersect with these three sets of norms (Holmes et al. 2011: 19) and also potentially clash and compete with these other norms that are present in workplaces. These minority group memberships, related to a number of transportable identities (e.g. ethnic or religious identities), are ‘prone to be in a tense relation with power in the workplace’ (Van De Mieroop and Clifton 2012: 195). This is linked to the fact that in many societies, such nonprofessional group memberships are normatively expected to be irrelevant in institutional contexts, since making relevant that an employee is a Muslim or a Polish immigrant, for example, would in particular situations be (perceived as) a sign of discrimination against this minority group member.

Critical issues and topics: where and how to analyse identity? Having described the issues surrounding the rather elusive concept of identity in our multicultural societies, we now turn to the discussion of the ways in which identity is actually analysed. Given the fact that the study of identity has received a significant amount of interest in many disciplines, it is quite difficult to give hard and fast rules of how such analyses are carried out in a general sense, even within the linguistic field of study. Furthermore, due to the crucial importance of taking a social constructionist perspective, a bottom-up approach to the data is strongly favoured, meaning also that the application of a priori categories is avoided. Before giving more details about this bottom-up, data-driven approach, we first turn to the types of data on which identity analyses are usually conducted. These have been briefly mentioned, but are discussed in more detail in the following. 188

Identity and membership

Data types Approaches to identity, such as (critical) discourse analysis, conversation analysis, discursive psychology and interactional sociolinguistics, tend to focus on naturally occurring oral data that are recorded and then transcribed. Depending on the specific approach favoured, these transcriptions may be highly detailed, using for example the Jeffersonian conversation analytic conventions (Jefferson 1984), or a set of conventions adopted to serve a particular analytical purpose (e.g. the conventions of the Wellington Language in the Workplace Project; see Holmes et al. 2011). In some approaches, such as in conversation analysis, the data under study usually consist of naturally occurring everyday interactions, in which identities are typically highly fluid, often shifting from one brief turn to the next as they are under constant negotiation between the interlocutors. Other approaches have developed a particular interest in narrative because it is considered ‘a privileged locus for the negotiation of identities’ (De Fina et al. 2006: 16). This is due to the fact that – as in all forms of discourse – the self is created through language, but in this particular type of communication, identity construction has a very central position, even to the extent that one could say that ‘the purpose of narrating is precisely the creation of an autonomous, unique self in discourse’ (Johnstone 1996: 56). Such narratives can occur in everyday interactions, in which case they can take many different forms, ranging from ‘highly monologic to highly collaborative tellings; from past to future and hypothetical events, from long and performed to fragmented and elliptical tellings’ (Georgakopoulou 2007: 17). In addition, they can be obtained through research interviews, in which interviewees are usually invited to tell their life story, hence resulting in more prototypical autobiographical narratives about past events.The former everyday narratives, called ‘small stories’, usually contain continuously shifting constellations of heavily negotiated identity claims, while this is often to a much lesser extent the case in the latter types of narratives, called ‘big stories’ (Georgakopoulou 2007). However, having said this, big stories are also subject to negotiation as they are ‘complex communicative encounters, co-constructed between researcher and researched’ (Georgakopoulou 2015: 264). This has long been acknowledged in qualitative studies (see e.g. Cicourel 1964) and it has – although only relatively recently – resulted in an explicitly interaction-oriented approach to the study of identity in these ‘big story’-data by narrative analysts (De Fina 2009).

How to analyse identity: theoretical issues As mentioned earlier, researchers who approach the study of identity from a social constructionist perspective are cautious with a priori assumptions of any kind. Thus, it does not follow that because one has collected data in an intercultural context, that ethnic identities will be made relevant in the interaction. Similarly, in doctor-patient interactions not only the situated identities of doctor and patient may be talked into being in the course of a consultation. Likewise, if analyses have uncovered that an interlocutor makes relevant his ethnic identity in one interaction, this does not mean that the same will happen in another interaction. In other words, the analyst’s aim is to always investigate the manner in which identity is constructed and negotiated in its local, interactional context. The analyses often draw on the principle of indexicality and/ or on the insights from membership categorisation analysis, both discussed successively in the following. The principle of indexicality (cf. e.g., Bucholtz and Hall 2005) is originally a linguistic anthropological principle. It establishes a link between linguistic forms and extralinguistic reality by focusing on the ways in which linguistic forms index social meaning and construct identity positions. As such the principle facilitates linguistically oriented researchers of identity to analyse 189

Dorien Van De Mieroop

the indirect relationship existing between language and identity. These links between linguistic utterances and social meaning are never pre-established because social meaning is always embedded in the specific local interactional context in which a particular linguistic form occurs at a particular point in time. Indeed, a particular form may indexically point to a different social meaning on other occasions or when uttered by other speakers. For example, the use of legal jargon may index a lawyer’s professional identity in an interaction with a client, but, when used by a defendant, it may also signal the client’s identity as an experienced convict. The list of linguistic forms with an indexical force is not only context dependent; it is also very long, consisting of many highly divergent linguistic resources that are situated on the whole length of the continuum between micro and macro. Furthermore, such a list is necessarily incomplete, since new social meanings for specific forms will emerge and others will disappear at some point in time. However, in order to give a more concrete idea of the types of linguistic forms which are typically analysed due to their potential indexical value, two frequently analysed linguistic resources in identity studies are briefly discussed in the present context, namely pronominal usage and linguistic codes. First, we focus on pronouns since these deictic linguistic elements have such an obvious extralinguistic meaning in that they, at the level of discourse identities, signal the relatively basic division of roles between speakers and listeners. However, their indexical force extends well beyond the division of these roles since a switch between singular and plural pronominal forms often points to the invocation of, or, alternatively, dissociation from group membership. For instance, many studies have found that interlocutors who consistently use the first-person plural pronominal form for self-reference highlight their identity as a member of one or another social group. Furthermore, these interlocutors often also refer to members of an opposing social group in collective terms, thus setting up an us-them opposition which not only highlights the interlocutors’ group membership even further, but also polarises the two social groups and frames the talk from an intergroup instead of an interpersonal perspective. Second, the use of particular linguistic codes has also been extensively studied for their indexical value. Importantly, this value should not be regarded as static, since it is not because a speaker uses a specific dialect that this is a deterministic marker of his or her social class or age. Rather, recent studies have drawn attention to the fact that speakers often switch from one language variety to the next, depending on what they are doing and with whom they are interacting. A teenager, for example, may use English as an index of ‘coolness’, or a father may construct his parent identity by using a language his children cannot understand when discussing sensitive topics with his partner. Thus the analysis now typically focuses on the local social meaning of the use of a linguistic code in its particular context. This use may index ingroup identity, but it may just as well make many other potential identity claims. Apart from using such indirect means, interlocutors may also construct their identities and establish their membership of a particular social group explicitly, namely through the use of labels and categories. An interlocutor may, for example, construct her ethnic and gender identity explicitly, saying As a black woman, I think it is important that. . . .. With such an utterance, she makes her membership of the group of women as well as of the group of black people relevant in the ongoing interaction, perhaps also with the aim of setting up an ingroup with some of the interlocutors, while excluding others. The analysis of the use of such labels in interactions has been developed in membership categorisation analysis (MCA), a field of analysis strongly influenced by the seminal work of the ethnomethodologist Harvey Sacks on categorisation. Sacks drew attention to the ‘inference rich’ nature of categories, meaning that: ‘a great deal of the knowledge that members of a society have about the society is stored in terms of these categories’ (Sacks 1992: 40–41). The use of particular categories implies a number of category-bound 190

Identity and membership

features, but this process also works the other way around, namely that by mentioning a number of associated characteristics, a category may also be implied. Importantly, though, these categories are approached in an anti-cognitivist way in CA and MCA, in the sense that researchers avoid applying their own interpretation of the link between certain categories and features. Also, the application of a priori established categories on the basis of, for example, an interlocutors’ physical features is strictly avoided. Instead, an emic perspective is taken, meaning that the analysis strongly focuses on the interlocutors’ orientations to the talk of others, scrutinising how such categories are made relevant in the turn-by-turn setup of an interaction and how the participants’ own invocation of specific category-bound features functions in its particular interactional environment. Such an ethnomethodological approach thus views identities as locally occasioned phenomena of which the interactional consequentiality is of the utmost importance (Antaki and Widdicombe 1998). However, such a strictly anti-cognitivist perspective on categorisation is not shared by all researchers taking a discourse analytical approach to identity, since, as De Fina (2006: 355) argues, interactants manage to interpret categories even though ‘much of what is being conveyed about category membership is a matter of shared understanding and implicit meanings’. She is thus in favour of an approach to categorisation in which analysts can, on the one hand, ‘link local identities to shared ideologies and beliefs’, while, on the other hand, still taking a social constructionist perspective to the data, foregrounding the fact that ‘the construction and presentation of identity is a process in constant development and that one of its crucial sites of negotiation is interaction’ (De Fina 2006: 355). Hence, the insights from membership categorisation analysis are often selectively drawn on in discourse analytical approaches to identity.

How to analyse identity: a concrete example In this section, one short data excerpt is analysed to give a more concrete idea of how these different analytical concepts and methodological perspectives are integrated in order to gain a multifaceted insight into the fluid character and interactional nature of identity construction. For this purpose, I draw on interview data collected in 2009, in which a number of Belgian women of Moroccan descent were interviewed by a ‘native’1 Belgian woman about their lives and career choices (for a more elaborate analysis of interviews from this corpus, see Van De Mieroop 2012 and 2015). For this chapter, I selected one interview with a thirty-three-year-old secondary schoolteacher (IE) who obtained a master’s degree in French and Spanish. I discuss an excerpt from the interview in which the interviewer (IR) probes for situations in which the interviewee was the victim of racism.The interviewee first states that she has never encountered racist reactions, except when she is accompanied by her mother because the latter wears a headscarf, unlike the interviewee herself. She concludes her answer to this question in the first two lines of the fragment, after which the interviewer asks a follow-up question: 1. IE 2. 3. IR 4. 5. IE 6. 7. 8. 9. 10.

Ze zien da ni he. Da ge Marokkaanse ze, dus eu:h. Ze laten u me rust. (.) En wa merkt ge dan als ge me uw moeder over straat wandelt= =Wel, ni zo lang geleden stonden wij aan de kassa in ((shop’s name)). En der zat een-eu:h ouwe, altijd ouwe mensen trouwens, een ouwe mens achter ons aan ‘t aanschuiven. En ik stootte per ongeluk tegen hem °en ons moeder stond naast mij° 191

Dorien Van De Mieroop

11. En euh, ik zeg: ‘Sorry’. 12. En vies dat die keek, me zo’n gezicht. Ik zag da, ge voelt da direct eh. 13. 14. En dan daarna had kik mijn ((shop’s loyalty card)) op de lopende: op de band gelegd. 15. 16. En ons moeder die van haar. En die van hem stond daar, lag er ook nog. 17. 18. En ik ging euh mijn ((shop’s loyalty card)), en die zo: ‘Ge gaat die van mij ni pikken he!’ 19. 20. IR @[@ 21. IE   [@ •hh Ik zeg: ‘Da’s ni de bedoeling.’ 22. Allez, maar ge denkt dan van: ‘Ik ga daar mijn energie ni insteken.’ 23. 24. Zo’n dingen. 1. IE They don’t see that hey. 2. That you are a Moroccan woman, so e:rm. 3. IR They leave you alone. (.) And what do you notice then 4. if you’re walking in the street with your mother= 5. IE =Well, not so long ago 6. we were at the cash register in ((shop’s name)). And there 7. was a-e:rm old, always old people by the way, 8. an old person queuing behind us. 9. And I accidentally bumped into him 10. °and our mother stood next to me° 11. And erm, I say: ‘Sorry’. 12. And a wry look that he gave me, with such a face. 13. I saw that, you feel that directly hey. 14. And then afterwards I had put my ((shop’s loyalty card)) 15. on the conveyer: on the belt. 16. And our mother hers. 17. And his stood there, still lay there as well. 18. And I went erm my ((shop’s loyalty card)), and he like: 19. ‘You are not going to steal mine hey!’ 20. IR @[@ 21. IE  [@ •hh I say: ‘That’s not the intention.’ 22. Well, but you think then: 23. ‘I am not going to put my energy into that.’ 24. Such things.2 Throughout this fragment, a number of identities are constructed and, even though the bulk of the fragment consists of a narrative of personal experience occurring between the interviewee, her mother and an antagonist, these identities are often explicitly related to group memberships. First, the most obvious way in which this happens is through the use of explicit labels, such as the nationality label in line 2. In Dutch, the specific word (Marokkaanse) also indicates female gender (hence the translation Moroccan woman), as such not only making relevant the interviewee’s nationality, but also her gender, thus highlighting these two group memberships. Furthermore, in the course of the story, the interviewee also indicates that the antagonist is an 192

Identity and membership

old person (lines 7–8), and she explicitly remarks that this is not a coincidence (line 7: always old people by the way). As such, she draws on a collectivity of old people and links this group to the category of racists. The link to the category of racists is constructed on two levels. First, the story is explicitly framed as an example of a racist incident, and second, the antagonist is actually constructed as racist in an implicit way. This happens at the end of the excerpt, when the interviewee is making the point of her story (line 19), namely that the antagonist accused the interviewee of wanting to steal his loyalty card.This implicit construction is carried out by voicing the antagonist’s perception of the close link between membership of the group of thieves and that of Moroccans. Such a conflation of two outgroup memberships leads to a prejudiced qualification (namely: ‘Moroccans are thieves’). Having such prejudices is a characteristic closely associated with membership of the category of racists. As such, the antagonist’s membership of the racist category is established through this category-bound feature. Second, the argumentative setup of the story strongly contributes to the projection of this racist category on the antagonist since it is contrasted with the interviewee’s display of polite behaviour in lines 10–11, where the interviewee apologises when accidently bumping into the antagonist. This apology indexes her identity as a well-behaving, polite person, which then leaves no doubt about the falseness of these accusations. As such, the accusations are framed as based only on the physical appearance of the interviewee’s mother, whose presence is explicitly mentioned in line 10. Third, the interviewee’s pronominal usage is also interesting since it on the one hand clearly marks her disaffiliation from the outgroup of racists, who are being referred to as a collectivity through the use of the third-person pronominal form they (line 1). On the other hand, however, the interviewee also refrains from constructing us-them oppositions, for example between Belgians and Moroccans, or racists and non-racists. Instead, she uses the first-person plural pronominal form only to refer to the story-related ingroup of protagonists, consisting of herself and her mother (lines 6 and 8). When referring to more general groups, such as those of Moroccan women (line 2) or victims of racism (lines 13 and 22), the interviewee switches to the use of the second-person pronominal form you, voiced in Dutch by means of the colloquial pronominal form ge. In this case, this second-person pronominal form actually indexes the self, since in all three instances, this form could be replaced by I or we. In particular the use of the latter form would invoke a group membership and would construct a dichotomy between groups. However, the interviewee refrains from using these first-person plural pronominal forms and uses the generic you form instead. This generic form creates a distance between the speaker and herself as one of the protagonists, but it also involves the audience since an appeal is made to the listener’s personal experience with such situations and how to react to these (see O’Connor 1994), as such suggesting general consensus on the interviewee’s perception in line 13 and her thoughts in line 22. This pronominal usage thus indexes the speaker’s de-emphasis of her personal membership of the group of Moroccans. So even though this category is invoked at the start of the excerpt (line 2), the pronominal usage surrounding it downplays this membership and even blurs this, in the sense that the use of the you forms in lines 13 and 22 frame the interviewee’s behaviour as generic human reactions, instead of reactions typical of one or another specific ingroup (in this case e.g. Moroccans, women or victims of racism). As such, the fine-grained analysis of the specific way in which the interviewee discursively negotiates these identities demonstrates that the application of a priori categories based on demographic facts (in this case, of the interviewee as a woman and as of Moroccan descent) would gloss over the way the interviewee oscillates between the invocation and the minimisation of her group membership of Moroccan women. In this brief fragment, the role of the interviewer is rather limited, but in other cases, the way certain questions, for instance, are formulated 193

Dorien Van De Mieroop

implies certain category projections upon the interviewee which can then be either explicitly countered, or accepted. This may lead to more complex discursive negotiations of identities, the discussion of which would extend beyond the scope of this chapter (for an example, see Van De Mieroop and Clifton 2014).

Conclusion The complexity of the analysis of identity has almost been a leitmotiv throughout this chapter.This is due to a number of factors: first and foremost, the inherent elusiveness of the concept and the ephemeral status of its realisations have not made it an easy job for researchers to get a grip on identities. Also the multidisciplinary interest in the study of identity, causing a certain degree of theoretical and methodological fragmentation, as well as the indirect link between language and identity and the contextual embeddedness in increasingly complex multicultural societies, contributes to the difficulty to analytically grasp identities. Furthermore, this complexity can only but increase since scholarly interest in the multimodal dimensions of identity construction, including the analysis of gestures, posture and eye gaze, has been growing steadily and will be an important perspective for future analyses of identity constructions. Yet, the cross-fertilisation between the micro and the macro level of analysis, as well as between related, yet methodologically diverse, disciplines, can only further our understanding of how people interactively construct fleeting constellations of identities in increasingly diverse post-modern societies.This multifaceted and multilevel perspective on the study of identity makes it a challenging, but also a rewarding endeavour.

Suggestions for further reading Antaki, C. and Widdicombe, S. (1998) Identities in Talk. London: Sage. In this collection of research articles, identity is studied at the micro level from an ethnomethodological or conversation analytical perspective. Benwell, B. and Stokoe, E. (2006) Discourse and Identity. Edinburgh: Edinburgh University Press. The authors adopt an interdisciplinary approach in this book and describe and explain how identity can be analysed in a range of different discursive environments throughout a number of theoretical as well as empirical chapters. Bucholtz, M. and Hall, K. (2005) ‘Identity and interaction: A sociocultural linguistic approach’, Discourse Studies, 7(4–5): 585–614. This article provides a framework for the analysis of identity based on five principles; each principle is illustrated by means of examples. De Fina, A. (2003) Identity in Narrative: A Study of Immigrant Discourse. Amsterdam: John Benjamins. In this monograph, De Fina closely scrutinises Mexican undocumented immigrants’ identity constructions in their narratives of personal experience, while also linking these findings to the wider social context. De Fina, A., Schiffrin, D. and Bamberg, M. (eds.) (2006) Discourse and Identity. Cambridge: Cambridge University Press. This book consists of an interesting collection of chapters by a number of top researchers in the field of discourse analysis, all focusing on the discursive analysis of identity construction.

Notes 1 The word ‘native’ is between quotation marks because of the false contrast between the interviewer and the interviewees. Since the latter are all second- or third-generation immigrants, meaning they were born and raised in Belgium, they do not differ from the interviewer’s ‘native’ status in this respect. 2 The translation is as close as possible to the Dutch original, which sometimes causes atypical formulations in English. These are usually based on incomplete sentences or peculiar formulations in Dutch, while at other times they may reflect idiomatic phrasings in Dutch. In this fragment, for example, the interviewee uses the highly frequent colloquial idiomatic phrase ons moeder twice (lines 10 and 16).

194

Identity and membership This is literally translated to our mother, even though this may sound a bit strange and rather marked in English. Needless to say, the analyses are based on the Dutch original. The data were transcribed using the Jeffersonian transcription conventions, as explained at the end of the chapter, but normal sentence punctuation (e.g. capitals, full stops) is used for the reader’s ease.

References Antaki, C. (2002) An Introductory Tutorial in Conversation Analysis. Available from: www-staff.lboro.ac.uk/ ssca1/notation.htm (accessed 22 June 2015). Antaki, C. and Widdicombe, S. (1998) Identities in Talk. London: Sage. Baynham, M. (2006) ‘Performing self, narrative and commuity in Moroccan narratives of migration and settlement’, in A. De Fina, D. Schiffrin and M. Bamberg (eds.) Discourse and Identity. Cambridge: Cambridge University Press. 376–397. Berger, P. L. and Luckmann, T. (1967) The Social Construction of Reality. Harmondsworth: Penguin. Bucholtz, M. and Hall, K. (2005) ‘Identity and interaction: A sociocultural linguistic approach’, Discourse Studies, 7(4–5): 585–614. Cicourel, A.V. (1964) Method and Measurement in Sociology. New York: The Free Press. De Fina, A. (2009) ‘Narratives in interview – The case of accounts: For an interactional approach to narrative genres’, Narrative Inquiry, 19(2): 233–258. De Fina, A. (2015) ‘Narrative and identities’, in A. De Fina and A. Georgakopoulou (eds.) The Handbook of Narrative Analysis. Chichester: Wiley Blackwell. 351–368. De Fina, A. and Baynham, M. (2005) ‘Introduction’, in M. Baynham and A. De Fina (eds.) Dislocations/ Relocations: Narratives of Displacement. Manchester: St Jerome. 1–10. De Fina, A., Schiffrin, D. and Bamberg, M. (2006) ‘Introduction’, in A. De Fina, D. Schiffrin and M. Bamberg (eds.) Discourse and Identity. Cambridge: Cambridge University Press. 1–23. Georgakopoulou, A. (2007) Small Stories, Interaction and Identities. John Benjamins: Amsterdam. Georgakopoulou, A. (2015) ‘Small stories research: Methods – analysis – outreach’, in A. De Fina and A. Georgakopoulou (eds.) The Handbook of Narrative Analysis. Chichester: Wiley Blackwell. 255–271. Hall, S. (1996) ‘Introduction: Who needs ‘identity’?’, in S. Hall and P. du Gay (eds.) Questions of Cultural Identity. London: Sage. 1–17. Holmes, J., Marra, M. and Vine, B. (2011) Leadership, Discourse and Ethnicity. Oxford: Oxford University Press. Jefferson, G. (1984) ‘Transcription notation’, in J. M. Atkinson and J. Heritage (eds.) Structures of Social Interaction. Cambridge: Cambridge University Press. 191–222. Johnstone, B. (1996) The Linguistic Individual. New York/Oxford: Oxford University Press. Kroskrity, P.V. (2000) ‘Identity’, Journal of Linguistic Anthropology, 9 (1–2): 111–114. O’Connor, P. E. (1994) “You could feel it through the skin’: Agency and positioning in prisoners’ stabbing stories’, Text, 14(1): 45–75. Sacks, H. (1992) Lectures on Conversation (edited by Gail Jefferson with introductions by Emanuel A. Schegloff, vols. 1–2). Oxford, UK: Blackwell. Tajfel, H. (1981) Human Groups and Social Categories. Cambridge: Cambridge University Press. Tajfel, H. (1982) ‘Introduction’, in H. Tajfel (ed.) Social Identity and Intergroup Relations. Cambridge: Cambridge University Press. 1–11. Van De Mieroop, D. (2012) ‘The discursive construction of gender, ethnicity and the workplace in second generation immigrants’ narratives: The case of Moroccan women in Belgium’, Pragmatics, 22(2): 301–325. Van De Mieroop, D. (2015) ‘Social identity theory and the discursive analysis of collective identities in narratives’, in A. De Fina and A. Georgakopoulou (eds.) Handbook of Narrative Analysis. Chichester: Wiley-Blackwell. 408–428. Van De Mieroop, D. and Clifton, J. (2012) ‘The interplay between professional identities and age, gender and ethnicity: Introduction’, Pragmatics, 22(2): 193–201. Van De Mieroop, D. and Clifton, J. (2014) ‘The discursive management of identity in interviews with female former colonials of the Belgian Congo: Scrutinizing the role of the interviewer’, Pragmatics, 24(1): 131–155. Wetherell, M. and Maybin, J. (1996) ‘The distributed self: A social constructionist perspective’, in R. Stevens (ed.) Understanding the Self. London: Sage. 219–279. Zimmerman, D. H. (1998) ‘Identity, context and interaction’, in C. Antaki and S. Widdicombe (eds.) Identities in Talk. London: Sage. 87–106.

195

Dorien Van De Mieroop

Transcription conventions (based on Antaki 2002) Example from excerpt

Explanation

20 IR @[@ 21 IE  [@ 4 IR wandelt= 5 IE =Wel eu:h een•hh (.) @ °en ons moeder stond naast mij ° Bedoeling ((shop’s name))

Square brackets aligned across adjacent lines denote the start of overlapping talk The equals sign shows that there is no discernible pause between words (latching) Colons show that the speaker has stretched the preceding sound A dash shows a sharp cut-off Audible in-breath Just noticeable pause Laughter token °-signs indicate a quiet speaking volume Underlined sounds are louder A change of the original words in order to ensure anonymity

196

17 FOLK PRAGMATICS Dennis R. Preston and Nancy Niedzielski

Introduction Folk linguistics (FL) is the study of what non-linguists believe about language, and it has been an area of interest at least since Polle (1889). We consider every approach (psycholinguistic, historical, sociolinguistic, etc.) and every level (phonetics, phonology, morphology, syntax, etc.) open to FL, and, as in all modern studies of ethnographic belief and practice, there is no caricature of the folk as unintelligent or backward; they are simply not professionals, and the data gathered from them may correspond to scientific knowledge, be at odds with it, or even involve areas that scientific approaches have not touched on. There has been a broad focus in FL on both approaches and levels (e.g., Niedzielski and Preston 2003), but folk pragmatics (FP) did not emerge as a specific area of interest until Niedzielski and Preston (2003). This chapter aims to define FP more carefully, review work that has been done, and suggest profitable areas and means of investigation for future research. We take a very broad view of pragmatics (e.g., at least all the main section headings of Levinson 1983: deixis, conversational implicature, presupposition, speech acts, and conversational structure), but are open to the idea that a sociologically facing “sociopragmatics” is as much a part of pragmatics as a grammatically facing “pragmalinguistics” (Leech 1983: 11). That makes a hard boundary between pragmatic and sociolinguistic research interests difficult to define, but we deal here with the areas most traditionally associated with pragmatics.1 Most important to this chapter, however, is the idea that FP, like all FL, is a study of “knowledge about,” not of “knowledge how to.” Competent speakers of any language have incorporated the pragmatic rules of behaviour into their linguistic repertoire as surely as they have those of every other linguistic level, and these are the rules professionals seek to discover, but, just as they are unlikely to be aware of the rules that govern phonetics, phonology, morphosyntax, and semantics, they may be unable to give any overt account of the rules that guide pragmatic behavior, even if one allows for nonprofessional terminology that nevertheless captures the essence of the rule. Silverstein (1981) is an excellent account of those sorts of linguistic facts that ordinary speakers may or cannot have overt knowledge of, many of them pragmatic, and a fuller account of Silverstein’s observations as they apply to FL is available in Preston (1996, 2016). Blum-Kulka and Sheffer (1993: 215–219) is an extended discussion of both availability to and accuracy of respondent comment on pragmatic facts. Some difficulties with the distinction 197

Dennis R. Preston and Nancy Niedzielski

between “know about” and “know how to” will be discussed, both in our examples of FP and in our section on methodologies (cf. How to get FP data and what to do with it). In general, however, we will not be concerned with any “folk theory” that may indeed inform pragmatic behaviour but is not within the awareness or working memory of the folk respondent; for example, we do not deal here with what Kay calls a “folk theory”: What the words we shall be concerned with here are about is language and speech, and the folk theory we shall be looking for is the tacit and mostly unconscious theory of language and speech we invoke when we employ certain parts of the lexicon of English. (Kay 1987: 68) Indeed, any underlying theory of language that can be extracted from numerous examples of language use is the goal of most linguistic research. In Niedzielski and Preston (2003), for example, we conclude that an important language ideology that informs much US FL belief as well as language behaviour is the underlying idea that a standard language is an extra-cognitive reality, a sort of Platonic ideal never quite realised by even ordinary, well-meaning users and befouled completely by speakers of ‘dialects’ (309–314). This is, however, our conclusion, one based on a large number of FL investigations; it is not the explicit comment of the folk themselves. These considerations lead us to propose the following outline of the position of FP within the larger framework of professional (or “scientific”) pragmatic study (SP). Pragmatic rules may be A. shared by FP and SP, or B. not shared by FP and SP, in which case 1. 2. 3. 4.

FP rules are not accurate, or SP rules are inaccessible to FP, or SP rules have not been noticed by FP, or FP accurate rules have not been “discovered” (or perhaps misinterpreted) by SP.

In light of this, we suggest that some established SP discovery procedures have in fact relied on FP. Suppose we want to discover the rules for complimenting.Yuan did the following: The informants were invited to share their opinions or experiences of giving and receiving compliments in the dialect. They were also asked to recall the last two compliments they had given and received, how the compliments were responded to, who the interactants were, and the contexts in which the compliments occurred. (Yuan 2001: 275) Jucker (2009: 1620) comments on this, suggesting that “[It] appears that such an approach can indeed offer insights into native speaker perceptions of this particular speech act.” We are in complete accord; this is a very good way to discover the FP (“speaker perceptions”) of complimenting, and it is illustrative of almost all the work in FL in which respondents have been asked to reflect on, talk about, or carry out fully conscious tasks. We are of a similar opinion about many other valuable approaches, which we find to be examples of FP rather than ones that elicit actual pragmatic behaviour. For example, there is a rich literature on the expression of various speech acts (thanking, offering, and receiving compliments, etc.), particularly in the contrastive tradition, in which speech act routines from 198

Folk pragmatics

different cultures are studied, often in the context of second language teaching and learning (e.g., Kasper and Blum-Kulka 1993). Most studies in this tradition elicit data using oral or written discourse completion tasks (DCT) or role plays in which respondents are presented with a situation and asked to offer appropriate responses. For example, Eisenstein and Bodman (1993) include the following in their list of DCTs devised to elicit acts of giving thanks:   8. You are married. Both you and your spouse work.You come home late from work and find that your spouse has done some work around the house that you had promised to do, but not had a chance to do. 10. You have just gotten a new and better job. A friend at the office tells you she has organized a farewell party for you. (Eisenstein and Bodman 1993: 76) These studies are usually quantitative, with many respondents, but in some cases, respondents themselves or others are invited to comment on the responses, adding a qualitative dimension. Although we do not doubt the value of the tasks set to respondents in nonnaturalistic settings, we find them to involve the sorts of conscious filtering that in our opinion makes their data FP. This is essentially the position taken by Félix-Brasdefer and Hasler-Barker (Chapter 4, this volume) in their chapter on the elicitation of data. They refer there to both “perception” and “production” of pragmatic rules; we label the perception side of this “FP,” although that should not be confused with the uptake of a rule in a setting in which conscious attention is avoided, exactly the sort of thing usually sought in experimental pragmatics. In other words, these investigations are good groundwork for professionals to look further (with observational, experimental, or other pragmatic research techniques) into the clues given by the folk respondents. Will it turn out that professionals will “discover” rules characterised in such investigations, turning B. 4 folk responses into the shared rules of A. 1? Or will they more likely find some accurate hints (B. 4s) as well as some that are not (B. 1s)? We strongly believe that even data from the B. 1 territory will not have been a waste of the professional’s time. First, what the folk believe about pragmatics is an interesting ethnographic area of research in itself. As Hymes (1972: 39) observed: “If the community’s own theory of linguistic repertoire and speech is considered (as it must be in any serious ethnographic account), matters become all the more complex and interesting.”2 Second, we fully expect SP and FP to contrast and that observations of that contrast will be interesting to both folk and scientific accounts. Grundy (1995: 139), for example, provides an interesting discussion of the contrast between the FP and SP characterisations of “politeness” in Britain, the former focusing on negative politeness while the latter is, in his words, “a value-neutral way in which ‘politeness’ is used as a term of pragmatic description available to all communicative instances of language use.” Third, even those “false” FP beliefs of B. 1 may provide guideposts for the professional. After all, they are nonprofessional reflections on just those areas that interest professionals. The Japanese dialectologist Misao Tôjô indicated years ago that the first thing he did when conducting a dialect survey in a new area was to talk to local people about local speech (1954), and then use the folk information as a guide to the scientific work in the next step.3 We will, however, not extensively review earlier work in pragmatics that used what we consider to be means that elicit FL data. Since we are sociolinguists, we fear the “observer’s paradox” (Labov 1972): the more aware one is of their language, the more likely they are to behave unnaturally, in a way that is not consistent with their vernacular behaviour. Although we fear this in the study of pragmatics as well, we nevertheless focus here on direct attempts to find out what non-linguists overtly believe about pragmatic behaviours, taking comfort in both their obvious ethnographic value and their potential as guideposts for professionals. 199

Dennis R. Preston and Nancy Niedzielski

We first provide data that fits our definition of FP from five areas: indirectness, implicature, indexicality, appropriateness, and politeness. These examples are not meant to restrict FP to comments on these areas, but we hope that they expose some of the breadth of such investigation and the considerable richness of the data while at the same time suggest that many other areas could be ripe for investigation.

Folk pragmatic data Indirectness Many examples of respondent reports on “not really meaning what we say” could be collected, and they may give additional insight into the role of indirection (and its relationship to politeness and other matters) in FP accounts. In Niedzielski and Preston (2003), we reported a respondent’s account of an international student’s embarrassing encounter with a native American English-speaking fellow student. The American had said, “Why don’t you come visit me some time?”; the international student had taken that to be a sincere invitation, but when she turned up on her fellow student’s doorstep, she saw in her face and manner that it was in fact not an expected visit; it was our respondent who uttered the FP notion we focus on here. It’s a way you say good-bye. A: Fieldworker: It’s a way (of) say good-bye? Yeah, sometimes what is necessarily SAID is not – what is actually meant. A: (Niedzielski and Preston 2003: 247) Unfortunately, A does not go on to indicate what strategies are used to interpret such nonliteralness, but if our interviewer had been focusing on FP, perhaps such an extended search could have been made. We comment on such methodological opportunities in the next section (cf. How to get FP data and what to do with it). Indirectness is the folk notion also expressed in some respondents’ comments on the passive construction. In a lengthy discussion, one respondent notes that “it takes – (.hhh) uh because the speaker doesn’t have repsonsi- he takes responsibility away from the uh – from the action” (Niedzielski and Preston 2003: 292). Although this focuses on the speaker as agent, it can be extended in folk commentary; in the same interview the speaker’s teenage son notes that “Politicians use that [i.e., the passive] a lot. Because y- they don’t say why” (Niedzielski and Preston 2003: 293). In both these cases, the indirectness involved is the folk notion of hiding the agents and thus shielding them from blame or the accusation of having made an assertion.4

Implicature That faulty interpretations of speech acts may be made is a frequent characteristic of popular culture jokes and media culture humour. One of the authors of this chapter was once subjected to the prescriptivist speech act humour in which a request was intentionally misunderstood as an information-seeking question: A: B: A: B:

Got a dollar (standing in front of a drink machine). Yeah (doing nothing). Well, can I borrow it? Sure; I just thought you wanted to know if I had one. 200

Folk pragmatics

At least in American English, diners are also used to this: A: B: A: B:

Can you pass me the salt? Yes (doing nothing). Well, will you pass it to me or not? Oh, I thought you just wanted to know if I could.

We suspect that in many languages with strong standard language ideologies that numerous examples based on prescriptivist leanings can be found. Media culture also displays the humour of indirectness; the cartoon Zits, which details the interactions of a teenage boy with his parents and others, once featured this conversation: So, what do you think? Boy: Father: I don’t know. Ask Mom. Boy (to Mom): Mom, what does Dad think? (Zits 29 May 2013) Of course “Ask Mom” carries the implicitly recovered “. . . . what she thinks,” but the son’s willful, playful, or even dense interpretation is the source of the humour, and raises further the question of the degree of overt pragmatic knowledge required of the reader to understand the humour.

Indexicality The use of linguistic expressions to index or point to social groups is one of the major interests of sociolinguistics and is the one that perhaps overlaps most with pragmatic concerns, and the extended notion of indexicality that involves the transfer of person characteristics to linguistic expression is frequently touched on in folk accounts. This semiotic process is called iconicity: “the interpretation of linguistic form not just as a dependable index of a social group but as a transparent description of the distinctive qualities of the group” (Woolard 1998: 19). When respondents carry out such tasks with full consciousness of what is required of them, we take that activity to be metapragmatic and belong to FP as well as the other areas with which it has been more closely associated. Since such data are commonly reported in folk linguistic, social psychological, anthropological, and sociolinguistic literature, we will cite them sparingly here, but we believe they need to be reconsidered from an FP point of view. In work in perceptual dialectology, the branch of FL in which respondents are asked to identify regional speech, one task asks them to outline and label on a blank map speech regions of a selected area. In many US maps, non-New Yorkers do not just identify the region as distinct, but also note that the language there is “fast & rude” or “scratch & claw” (Niedzielski and Preston 2003: 60, 62), an assignment of the indexicality of beliefs about regional character to the language of the region. New Yorkers do not suffer alone from such indexical transfer: in many hand-drawn maps of the US South, one may find respondent labels that are just as clearly transfers from beliefs about persons to language: “hicks,” “hillbillies” (2003: 57), “Courteous and Gentlemanly,” and “Also Spoken by Ignorants” (2003: 63), that last revealing both a positive and negative indexicality that suggests sensitivity to context, an obvious pragmatic concern, is important as well in the elicitation of so-called language attitudes (e.g., Preston 2010). When Northern US respondents are confronted with a Likert-scale task that evaluates the speech of 201

Dennis R. Preston and Nancy Niedzielski

various regions, that format clearly highlights both of the indexicals: they evaluate their own region’s speech for notions of standardness (e.g., good English, smart, educated, normal), but the South’s as better than their own for such solidarity notions as polite, friendly, and down to earth (Preston 1999a: 364). Such iconic indexing is by no means limited to US areas; Inoue (1995) outlines studies of perceptual dialectology of Japan in which the dialect of the north (particularly the prefecture of Tohoku) is characterised as having an extremely negative “intellectual” character but a neutral “emotional” one, while the Tokyo (Kanto) area has an extremely positive intellectual but also neutral emotional image. One must go to the Osaka (Kansai) area to find extremely positive emotional regard (coupled with rather negative intellectual characteristics). Preston (1999b) and Long and Preston (2002) within the same perceptual dialectological tradition provide many more examples: Belgium, Canada, Cuba, England, France, Italy, Japan, Korea, Germany, Hungary, Mali, Spain, Switzerland, the Netherlands, Turkey, the United States, and Wales. We believe that many studies of language attitudes, both within the FL and social psychological traditions, might be reinvestigated for their pragmatic significance, so long as one admits to the role of person indexicality which appears to be in many cases the source of language evaluations through the process of iconisation.

Appropriateness Niedzielski and Preston (2003) uncovered considerable folk comment on the inappropriateness of taboo language for women. (1) J: I would never have s- I didn’t say “shit” until I had five kids. All: ((laugh)) J: And then that covered the only word I knew. All: ((laugh)) J: And then I had a difficult time with i- I was- an- it was difficult for me, (.hhh) and even to this day if I say that to= [ Uh huh. S: J: =H, – wh- he says – “That’s not ladylike” or something like that. “Watch your language.” See it’s offends him. (.hhh) He isn’t offended if HE says anything he WANTS, (.hhh) but if I say something like that, why then it’s uh unladylike. (Niedzielski and Preston 2003: 186–187)5 (2) S: How about th- those young people today, if they say words like that. – Well, uh perhaps you- your son and your – daughter-in-law – they – they speak words like that. How do they feel among themselves. [ Well, you don’t find- I don’t find it in any of my daughters or daughter-in-laws. J: S: They – think [ J:  They don’t talk like that. (Niedzielski and Preston 2003: 187) 202

Folk pragmatics

These interpretations are consistent with other folk respondents’ characterisations of the cause for different language practices, at least in US society, for men and women. (3) B: I think men are given more freedom (1.0) in a lot of ways than women. Maybe that’s just one (1.0) example of that- that men have more choices to- (1.0) to talk the way they want to. If men prefer to use slang compared to standard language maybe they have:- It’s more socially acceptable. (1.0) You know, cultures: historically are worried about women not being prim and proper.You know women can’t be promiscuous. It’s okay for men to be. Well the way not to be promiscuous or immoral or not virtuous is to be very limited in what you do. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. B: I think language is a part of that. That men – so therefore women are given less freedom. Because if women are given a freedom they are going to – They are not going to control their sex- their- their sex drive. Do you think that makes sense? Z: Yes, it does. B: I think in general men have- are given more freedom than women, and that’s just an example that they have more freedom to speak what they want to speak. (Niedzielski and Preston 2003: 194) Although many more studies of appropriateness could be cited, it should be clear that it is deeply embedded in contextuality, not only in terms of performer identity, but also in terms of the entire context, which clearly includes interlocutor identity as well.

Politeness Niedzielski and Preston (2003) has a lengthy discussion of politeness that is triggered by a Taiwanese Chinese speaker’s difficulty in knowing the age-relation status of an interlocutor relative to his/her brothers and/or sisters – a necessity for Chinese interaction.6 (4) H: Bu- so here, sometime I’m confused when people say ‘Oh uh:’ – uh – when they say ‘Oh, Angela? Oh she’s my – sister.’ Then I’m- I’m think- I’m thinking – i- sister is younger or is the older G: Don’t be afraid to ask. – Ask. (.hhh) And th-an- and nobody’s offended. And they’ll say, ‘Oh, OK, uh oldest or – uh eldest.’ ( ) Or they’ll say ‘youngest,’ yeah. Don’t feel afraid. Uh – it’s a good way to – start a conversation. Really. – ‘Do you have brother or sister?’ You ask, you know ‘older or younger, elder or younger.’ H: I mean not im- no: im- I mean: this is not impolite, right.= G: =No, no it’s not improper. A: It’s not improper.When I make new friends ( )- when I make new friends at school, I ask if- they- they’ve like got a sister, ask is it- is she older (). (.hhh) I always ask cause I want to know. K: It’s not a problem – to ask about that. H: OK. – And uh – some custom may be difference, uh – u- I mean many – in the United States people I mean sometimes girl: uh they don’t want to – uh: – ask them ‘How old are you’ or something ( ). Or- or- say uh ‘You look uh fami- uh mi- uh – uh:’- no not familiar, I mean: mature or something.Yes, so it is- it’s polite or impolite say- ‘You look mature.’ G: It’s not a good thing to say. Probably it’s – because in America, – women – uh-= H: =How about men. 203

Dennis R. Preston and Nancy Niedzielski



G: Yeah a man could care- a man could care less. He would probably say ‘thank you.’ (.hhh) But to a woman her beauty, whether she’s beautiful or not, is not important. (.hhh) Her beauty is important to her. And- and most women, want to stay young. You could- y- the best compliment you could say to someone, rather than saying ‘mature,’ (.hhh) is say ‘Well, you look very young.’ And right away you’ve made a friend. I mean= [ K:       He said that to me, see no wonder we’re friends.              [ G: =              – (.hhh) that’s Se- see an- and the idea’s if you said to someone and said ‘Oh, that’s you- th- th- that’s you’re younger sister? ‘Thought that was your OLDer sister,’ That would be a compliment, because you’re telling them that theythey are younger – LOOKing than they actually are. (.hhh) and youth is important. No- mi- th- might not be right, may not be correct, (.hhh) but in our society it- it is. Women want to be. (.hhh) That’s why on birthdays, – they really don’t want to tell their birthday. ‘My God, I’m having my thirtieth birthday.’ Or ‘Oh my God, I’m going to die, my fortieth birthday.’ (.hhh) You know the guy says ‘What is it,’ you know ‘Well, I’m going to have my fortieth today, you know, it’s ((claps)) party time.’ No one cares, you know. But to the woman i- it is a sign tha- that she may be losing some of her beauty, some of her attractiveness. A- and we do. Breasts, you know, face, (.hhh) an- and the- and the – culture of ours- our – you know the men’s magazines and everything, the bustier the better, you know. (.hhh) Uh kind of thing. So it’s it’s uh – (.hhh) I won’t say it’s necessarily true once the door shuts and you’re in your family, because then it’s reality. (.hhh) But the idea of the stereotype is that the big busted woman: thin-waisted and all that you know is the ideal. K: [[In the United States. H: [[OK. [ G: In the United States.    [ A:     Yeah, that’s why you see=   [ H:    So is A: =them, like that in commercials.   [ H:   So is that OK, I say uh ‘You look young.’ Is it O= G: =That’s good. H: That’s good? Because, you know what, uh – wh- when I- I thought uh- can I say uh ‘You look young’ because I’m afraid maybe she was – will think – uh ‘You think me childish’ or something. G: No no, she won’t think that. – She’ll think the opposite. [ [ A: No Not unless you’re= [ H: No? A: =speaking to someone like me. Or a young teenager, then-= G: =Yeah yeah, if it’s a teenager, then they want to be an adult. So they want to grow up. And then you could say ‘Oh you look mature,’ and they would be real happy with you. 204

Folk pragmatics

When you’re saying to a woman ‘You look mature,’ you may be giving a backhanded insult, telling her, ‘Well you look nice, but you’re looking old.’ H: So I can only say ‘You’re mature-‘ – to: a y- a: to a teenage here. [ K: Young: teenager. [ G: Teenage- to a teenager girl, that would be, or even a guy, ‘You look mature,’ wouldwould uh because they are trying to be adults. They are try [ K:    Most men would accept that as a – compliment. You know, and YOUNG girls, teenage girls, that’s a compliment. Older women, would consider it in a- uncomfortable statement. H: But- OK what’s the best way to: – de- to say, I mean – for older women like ( )    [ K:     Well – you- you- when you asked me how old I was, when we were (.hhh) together that day (.hhh) uh in- was in the spring. And I told you I was – a- forty, years old, and you said ‘You look very YOUNG for you AGE,’ – That was great. G: Oh that’s a compliment.That’s good.Write- write that down, memorise it, and keep it. H’s fear of being impolite by enquiring about age triggers lengthy explanations from the family, particularly from G. The factors discussed involve both sex and age and deal for the most part with how to avoid impoliteness, although the notion of complimenting arises as well. Early in the conversation, G reassures H that to ask if one is an older or younger brother or sister is not impolite; in fact, he equates “impoliteness” with “impropriety.” (He also makes one of the few comments we have in our data on conversational matters by suggesting that such an enquiry would even be a good conversation starter.) H opens the possibility that women in particular might object to having their age queried, and asks specifically if the word mature might be inappropriate; G immediately assures him it would, noting that men could care less, but that women are especially concerned with beauty, a factor he implicates is tied to youth. In this part of the conversation, he first raises the idea that to indicate how young a woman looks will make a friend and elaborates on it by suggesting that it would be a nice compliment to tell an older sister that she looks like the younger one. He very clearly suggests the conventionality of suggesting a woman looks older than she is by indicating that it might not be right, may not be correct. His performance contrast between female horror at turning thirty (I’m going to die) and the male response to turning forty (party time) solidifies his sex stereotyping. H raises the possibility that telling a woman that she looks young might run the risk of having her interpret such a remark as a reference to childishness, and, although G denies it, A, the young daughter, notes that it might be inappropriate if directed to her, and G quickly agrees that younger people like to be thought of as mature. He adds the interesting classificatory label backhanded insult to the use of mature to an older woman. This part of the interaction concludes with K remembering that, on her first meeting with H (in a graduate course), he remarked that she looked very young for her age. G immediately notes that that was a good compliment and advises H to write that down, memorise it, and keep it. In this short interaction, part of a much longer one, the respondents touch on politeness (and especially the avoidance of impoliteness), conversational structure (i.e., an appropriate topic for a conversation starter), and complimenting. All this is heavily contextualised by the respondents’ views of age and sex, making it a very fruitful site for the identification and interpretation of FP. 205

Dennis R. Preston and Nancy Niedzielski

How to get FP data and what to do with it The kinds of FP data FP data come from non-linguists’ conscious responses to and comments on pragmatic concerns. These data may come from set tasks which directly evoke pragmatic details (e.g., DCTs, role plays) or from tasks that seek other FL details but nevertheless involve pragmatic factors (e.g., language attitude studies of many sorts, perceptual dialectology). We believe our primary examples here show, however, that the richest data come from simply talking to people about language, the primary vehicle in Niedzielski and Preston (2003), the first modern large-scale attempt to elicit and characterise FL in general. The nature of such talk about language varies according to the specific contextual details.Was the talk about language simply overheard conversations in public that a scholar might make note of? Did it derive from explicit talk about pragmatic concerns introduced by the fieldworker? Were pragmatic concerns extracted from broader FL conversations in which respondents were led into the general subject of language or even specific topics, but not pragmatic ones specifically? Were the data “discovered” in fieldworker interviews that did not even introduce language as a topic? Were the respondents interviewed separately or in groups? Were the respondents known to the fieldworker, and was the fieldworker from the same or from a different cultural background? In the previously cited conversation on asking about age and politeness, the fieldworker’s foreign status no doubt elicited detailed cultural descriptions that would have been thought to be obvious to a native.

FP data sources FL and therefore FP data are widely available; sit around and listen to conversations in public, and the chances of language arising as a topic are very good. That the topic will be pragmatic is not guaranteed. Hence, such a method is not a particularly efficient way to acquire FP data. Wolfson and Manes (1980) and Herbert (1991) have used it, however, for the collection of actual pragmatic speech behaviour, although the data were collected by a large team of fieldworkers. A clear alternative is the participant-observer approach, in which fieldworkers insert themselves into the target community and, as members of it, observe natural behaviour; in this case, such behaviour would include conversation about language (FL) with a pragmatic focus (FP).This method has the advantage of authenticity but the disadvantage of long-term fieldworker commitment. Interviews are also a possible data source. It will not take much imagination to turn a sociolinguistic interview (e.g., Labov 1984) into one that will elicit FP data. The topically organised folk linguistic interviews in Niedzielski and Preston (2003) provide a good model.The selection of respondents for interviews or even for the task-oriented elicitations mentioned earlier should reflect, however, the demographic specifications that the researcher has in mind in order to limit the generality of the sorts of claims that might later be made. People are not always needed. Folk comments on pragmatic topics may be gleaned from any media or popular culture source (as we have shown from cartoons and popular culture), and one may search as well the growing electronic corpora now available, although, as with free conversational data, the incidence of folk pragmatic commentary may be sparse. The written record is also important, for if there is a historical pragmatics (cf. Jucker and Landert, Chapter 9, this volume), there is surely a historical FP. Although historical pragmatics has gained recent attention, a leading characterisation of the field (Jacobs and Jucker 1995) does not allow for FP, outlining only the recovery of pragmatic information from texts, which we 206

Folk pragmatics

take to be pragmatic behavior but not commentary on it. There is, however, a rich tradition of comment on language from non-linguists that needs to be exploited. In his Images of English (1991), Bailey collects and interprets interesting and important FL comments from the past. For example, he cites this mid-sixteenth-century remark on the perception of and attitudes towards diversity in English: We see that few of ani nation, doo come to the perfait speche of their mother toung in all their life. . . . . So that yf they heare their neyghbour borne of their next Citie, or duelling not past one or two dais Iourney from theim, speaking some other word then is (in that place) emongst theim used, yt so litell contenteth their eare, that (more than folishli) they seem the stranger were therefore worthei to be derided, and skorned. (31, quoted from Danielsson 1955: 115) If one worked in the pragmatics of sixteenth-century English, such FL data as this would surely be important, for they touch on both the intelligibility of inter-dialectal discourse and the regard held for other varieties. In the pragmatic world, of course, they touch on the indexicality of persons (who are worthy “to be derided and scorned”) and will no doubt form important historical backgrounds for our discovery of the beginnings of the iconisation process in which the characteristics of “scorned people” may be transferred to language details themselves. FP data is all around, but for a realistic approach we may need to elicit it through tasks such as those outlined earlier in the discussion of DCTs, role plays, and assessment tasks or look for it in our own interviews or in searches of existing data bases and the historical record.

FP data analysis Once the data are in hand, we may make use of them in various ways. We will not deal here with the statistical analyses of task-oriented research, but focus instead on conversational data. The first, and simplest, analytic approach is to restate the FP data in ways that will allow them to be compared to the findings of SP (so that the sorts of comparisons and contrasts suggested earlier may be carried out). Since most of the FP commentary we have presented here is embedded in interaction, we also suggest that various methods of interactional analysis may be rewarding.We are not looking for pragmatic behaviour in interaction, but for accounts of pragmatic rules and behaviour; therefore, we must use topic-oriented rather than purely structural approaches to conversation and discourse (e.g., Preston 1994). One such way to investigate FP data is through the employment of pragmatic concepts themselves, and one that investigates the presupposed (as well as asserted) positions taken in folk conversations about language has proved rewarding. We illustrate this potential by pointing out the possibility of extracting pragmatic presuppositions, which are related to lexical and structural triggers (e.g., Levinson 1983: 181–185). In the following exchange, a Taiwanese fieldworker (C) discusses African American English with an African American friend from Michigan (D) in an FL interview. (5) 1 C: We uh – linguistics, in this field, uh – from the book I s- I mean, I saw from the book that – many linguists quite interest in black English. So could you tell me – a little bit about – your dialect? 2 D: Dialects. 3 C: Heh yeah 207

Dennis R. Preston and Nancy Niedzielski

4 All: ((laugh))  [ 5 D:  Well, uh: – well – see the world’s getting smaller.There’s not – even among all the ethnic groups we’re-=      [   [ 6 C:     ((laughs)) I- I mea- do you have7 D: =we’re getting- getting less and less of dialectual in- inFLUence. (.hhh) Uh I’mhappen – not to be – from the South, . . . (Preston 1994: 286–287) Without an account of presuppositions, this discourse is difficult to interpret, particularly the content of 5–7 D. The first key is in the presupposition(s) of So could you tell me a little bit about your dialect (1 C). Your dialect presupposes the existence of “dialect(s)” and that “you” are the speaker of one. D’s perception of C’s presuppositions leads to the otherwise difficult to understand assertions in 5–7 D: The world’s getting smaller. We’re getting less and less of dialectual influence (i.e., there are fewer and fewer dialects) I happen not to be from the South. The world’s getting smaller is an explanation of why there are fewer dialects (education, media, mobility, etc.), but D’s next assertion, that there are fewer dialects, responds to C’s presupposition that there are such things at all. Finally, and more subtly, D confirms C’s presupposition that dialects exist, but, for him (D), they exist only in such places as the South. In other words, if C had been lucky enough to encounter a speaker from the South, he might have had his request for information about “your dialect” fulfilled. How can we make sense of D’s observation that he is not from the South unless it is in some way related to his response to C’s request for information about D’s dialect (and embedded in D’s assertion that there are fewer dialects)? Recall that Michiganders find their own region “correct” but also find the South very salient and that salience is undoubtedly related to its incorrectness; that is, it is “a dialect.” Presuppositional work, however, may also explain why D “happens” not to be from the South. Why does he not simply say, “I am not from the South”? Happen belongs to a group of implicative verbs (Levinson 1983: 181) and presupposes “inadvertence,” “lack of planning,” or “by chance.” D “happens” not to be from the South because it is only a case of bad luck that C picked on a respondent who was not from the South (and could therefore not fulfill his request for “dialect” information). Work on discourse, then, from many perspectives, may reveal not only what speakers have said or asserted (the conscious), but also what they have associated, entailed, and presupposed. Recall that in the earlier conversation on politeness, the primary respondent never directly asserts that age and beauty are connected. This distinction between conscious and unconscious performances and responses raises a difficulty in characterising FP. If investigative techniques uncover unconscious folk characterisations of pragmatic concerns, couldn’t these be quite simply the actual foundations of pragmatic behaviour? Parts of the cognitive underpinnings of ideologies, attitudes, and beliefs about language may be easily and automatically accessed; others may require (or allow) the application of conscious resources. Although the automatic retrieval path may be privileged in some way for its more likely correspondence to the rules of ordinary behaviour, the fully conscious responses might reflect these rules as well. Perhaps this is a question to be left to the emerging field of 208

Folk pragmatics

experimental pragmatics, although that work has focused to date on respondent interpretations of sentence-level meanings that address the semantics-pragmatics border (e.g., Meibauer and Steinbach 2011). We will leave for others to decide whether presupposed and implied material is “fully,” “partly,” or “not at all” under the conscious control of the respondents. We know that this muddies the distinction between FP and pragmatics proper (i.e., SP), but we believe the use of such categories so enhances the interpretive side of the project that we are willing to risk that bit of ambiguity.

Conclusion In summary, although much of FL has not focused on pragmatic concerns directly, opportunities surely await those who find value in the ideologies, attitudes, beliefs, and the like of ordinary people. If FL is valuable for the reasons outlined earlier, then so is FP, and we hope in future reviews that we will be able to add a great deal more information from the folk perspective to the concerns of pragmatics research.

Suggestions for further reading Hoenigswald, H. (1966) ‘A proposal for the study of folk-linguistics’, in W. Bright (ed.) Sociolinguistics. The Hague: Mouton. 16–26. This is the modern call-to-arms for the investigation of FL in general, highlighting the ethnographic and theoretical justifications. Niedzielski, N. and Preston, D. R. (2003) Folk Linguistics, rev. pbk. edn. Berlin/New York: Mouton de Gruyter. This is an extensive study of directed FL conversations between linguistically trained fieldworkers and folk respondents in southeastern MI USA; the respondents come from a variety of status, educational, age, and ethnic backgrounds. It surveys experimental (assessment) and task-oriented (e.g., perceptual dialectological) work, as well as theoretical and general concerns. Preston, D. R. (1996) ‘Whaddayaknow: The modes of folk linguistic awareness’, Language Awareness, 5(1): 40–74. This article surveys the kinds of linguistic knowledge available to folk respondents and offers an application of Silverstein’s (1981) ideas to the field. Preston, D. R. (2004) ‘Folk metalanguage’, in A. Jaworski, N. Coupland, and D. Galasinski (eds.) Metalanguage: Social & Ideological Perspectives. Berlin: Mouton de Gruyter. 75–101. This article surveys three sorts of metalanguage (naturally occurring, non-focused, and topical) and discusses their various applications to FL as well as more general sociolinguistic concerns. Silverstein, M. (1981) ‘The limits of awareness’, Sociolinguistic Working Paper #84. Austin, TX: Southwest Educational. This is the most influential and detailed paper available on what sorts of linguistic knowledge the folk may have; it focuses on pragmatic concerns.

Notes 1 We specifically reject any caricature of sociolinguistics as “mainly concerned with the systematic linguistic correlates of relatively fixed and stable social variables. . . . on [sic] the way an individual speaks” (Thomas 1995: 185; italics in original), and we find odd any definition that assigns to pragmatics but not to sociolinguistics the research goal of “describing the linguistic correlates of relatively changeable features of that same individual (such as relative status, social role) and the way in which the speaker exploits his/ her (socio)linguistic repertoire in order to achieve a particular goal” (Thomas 1995: 185; italics in original). One need mention only the sociolinguistic use of the stylistic continuum (whether operationally or observationally approached) to show that sociolinguistics is not only concerned with dynamic and changeable linguistic features selected from a repertoire to make an utterance appropriate, but has also made such dynamic concerns an integral part of the theoretical foundations of the discipline.

209

Dennis R. Preston and Nancy Niedzielski 2 One might assume that “ethnopragmatics” characterises FP from this ethnographic perspective, but work associated with this label emphasises cross-cultural pragmatics that attempts to deal with speech practices from an “insider” perspective (e.g., Goddard 2011), and the method of analysis proposed does not attend primarily to folk commentary on or belief about language use. 3 We would be remiss, however, not to point out that in spite of Hoenigswald’s (1966) insistence on the importance of folk opinion about language, some have resisted any such consideration. The Bloomfield family, for example, entertained one another around the dinner table in the evenings reporting “stankos” they had heard during the day. They invented the term to describe ignorant comments about language (Hall 1950). 4 This folk notion corresponds to the linking of hegemonic or deceitful intents to the use of specific grammatical structures, an analytic strategy often made use of in critical discourse analysis (e.g., Fairclough 1989). 5 The conversational transcriptions in this chapter follow Gail Jefferson’s suggestions (e.g., Schenkein 1978: xi–xvi) (cf. Appendix). 6 These data are taken from FL interviews between fieldworkers and southeastern Michigan respondents for Niedzielski and Preston (2003). The data given here have not been published previously. The interviewer (H) is a male Taiwanese linguistics graduate student; the interviewees are G, a middle-aged, male, European American junior high school teacher; K, G’s wife, a civil service and telephone company worker; and A, G and K’s daughter, a fourteen-year-old female junior high school student. The taperecorded interview was conducted in their family home. The transcript has been edited considerably to remove back-channel and other matters nonessential for the subject at hand.

References Bailey, R. W. (1991) Images of English. Ann Arbor: University of Michigan Press. Blum-Kulka, S. and Sheffer, H. (1993) ‘The metapragmatic discourse of American-Israeli families at dinner’, in G. Kasper and S. Blum-Kulka (eds.) Interlanguage Pragmatics. New York/Oxford: Oxford University Press. 196–223. Danielsson, B. (1955) John Hart’s Works on English Orthography and Pronunciation. Pt. 1. Stockholm: Almqvist and Wiksell. Eisenstein, M. and Bodman, J. (1993) ‘Expressing gratitude in American English’, in G. Kasper and S. BlumKulka (eds.) Interlanguage Pragmatics. Oxford: Oxford University Press. 64–81. Fairclough, N. (1989) Language and Power. London: Longmans. Goddard, C. (ed.) (2011) Ethnopragmatics: Understanding Discourse in Cultural Context, (Applications of Cognitive Linguistics, vol. 3). Berlin: Mouton De Gruyter. Grundy, P. (1995) Doing Pragmatics. London: Longman. Hall, R. A., Jr. (1950) Obituary of Leonard Bloomfield, Lingua, 2: 117–123 (cited from Charles F. Hockett (ed.) (1970) A Leonard Bloomfield Anthology. Bloomington/London: Indiana University Press. 547–553. Herbert, R. K. (1991) ‘The sociology of compliment work: An ethnocontrastive study of Polish and English compliments’, Multilingua – Journal of Cross-cultural and Interlanguage Communication, 10: 381–402. Hoenigswald, H. (1966) ‘A proposal for the study of folk-linguistics’, in W. Bright (ed.) Sociolinguistics. The Hague: Mouton. 16–26. Hymes, D. (1972) ‘Models of the interaction of language and social life’, in J. J. Gumperz and D. Hymes (eds.) Directions in Sociolinguistics:The Ethnography of Communication. New York: Holt, Rinehart and Winston. 35–71. Inoue, F. (1995) ‘Classification of dialects by image – English and Japanese’, in W. Viereck (ed.) Zeitschrift für Dialektologie und Linguistik; Verhandlungen des Internationalen Dialektologenkongresses, Bamberg, 29.7. — 4.8.1990. (4 vols.). Stuttgart: Steiner. 75–77, 355–368. Jacobs, A. and Jucker, A. H. (1995) ‘Introduction: The historical perspective in pragmatics’, in A. H. Jucker (ed.) Historical Pragmatics. Amsterdam/Philadelphia: John Benjamins. 3–33. Jucker, A. H. (2009) ‘Speech act research between armchair, field and laboratory: The case of compliments’, Journal of Pragmatics, 41: 1611–1635. Kasper, G. and Blum-Kulka, S. (eds.) (1993) Interlanguage Pragmatics. New York/Oxford: Oxford University Press. Kay, P. (1987) ‘Linguistic competence and folk theories of language’, in D. Holland and N. Quinn (eds.) Cultural Models in Language and Thought. Cambridge: Cambridge University Press. 67–77.

210

Folk pragmatics Labov, W. (1972) Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Labov,W. (1984) ‘Field methods of the project on linguistic change and variation’, in J. Baugh and J. Sherzer (eds.) Language in Use: Reading in Sociolinguistics. Englewood Cliffs, NJ: Prentice-Hall. 28–53. Leech, G. N. (1983) Principles of Pragmatics. Longman House, Burnt Mill, Harlow: Longman. Levinson, S. C. (1983) Pragmatics. Cambridge: Cambridge University Press. Long, D. and Preston, D. R. (eds.) (2002) Handbook of Perceptual Dialectology, vol. 2. Amsterdam: Benjamins. Meibauer, J. and Steinbach, M. (2011) Experimental Pragmatics/Semantics. Amsterdam/Philadelphia: John Benjamins. Niedzielski, N. and Preston, D. R. (2003) Folk Linguistics, rev. pbk edn. Berlin/New York: Mouton de Gruyter. Polle, F. (1889) Wie denkt das Volk über die Sprache? Gemeinverständliche Beiträge zur Beanwortung dieser Frage. Leipzig: Teubner. Preston, D. R. (1994) ‘Content-oriented discourse analysis and folk linguistics’, Language Sciences, 16(2): 285–330. Preston, D. R. (1996) ‘Whaddayaknow: The modes of folk linguistic awareness’, Language Awareness, 5(1): 40–74. Preston, D. R. (1999a) ‘A language attitude approach to the perception of regional variety’, in D. R. Preston (ed.) Handbook of Perceptual Dialectology, vol. 1. Amsterdam: Benjamins. 359–373. Preston, D. R. (ed.) (1999b) Handbook of Perceptual Dialectology, vol. 1. Amsterdam: Benjamins. Preston, D. R. (2010) ‘Variation in language regard’, in P. Gilles, J. Scharloth, and E. Zeigler (eds) Variatio Delectat: Empirische Evidenzen und theoretische Passungen sprachlicher Variation (für Klaus J. Mattheier zum 65. Geburtstag). Frankfurt am Main: Peter Lang. 7–27. Preston, D. R. (2016) ‘Whaddayaknow now’, in A. Babel (ed.) Awareness and Control in Sociolinguistics. Cambridge: Cambridge University Press. 177–199. Schenkein, Jim. 1978. Studies in the Organization of Conversational Interaction. New York/San Francisco/ London: Academic Press. Silverstein, M. (1981) The Limits of Awareness (Sociolinguistic Working Paper #84). Austin, TX: Southwest Educational Development Laboratory. Thomas, J. (1995) Meaning in Interaction: An Introduction to Pragmatics. Harlow: Longman House. Tôjô, M. (1954) Nihon Hôgengaku [Outline of Japanese Dialectology]. Tokyo:Yoshikawa Kôbundô. Wolfson, N. and Manes, J. (1980) ‘The compliment as a social strategy’, Papers in Linguistics, 13(3): 391–410. Woolard, K. A. (1998) ‘Introduction: Language ideology as a field of inquiry’, in B. B. Schieffelin, K. A. Woolard, and P. V. Kroskrity (eds.) Language Ideologies: Practice and Theory. Oxford: Oxford University Press. 3–47. Yuan,Y. (2001) ‘An inquiry into empirical pragmatics data-gathering methods: Written DCTs, oral DCTs, field notes, and natural conversations’, Journal of Pragmatics, 33: 271–292. Zits. May 29, 2013. A King Features Syndicated Cartoon by Jerry Scott and Jim Borgman.

Transcript conventions [[ [ ] == CAPS .hhh hhh - - — : (( )) ( )

Overlapping units of talk from two or more utterance beginnings Site where an overlap begins, placed at the point of the overlap between the lines of two speakers Site where an overlap ends, if not obvious from line length Link an unbroken continuation of speech, even though there may be overlapping material Contrastively stressed syllable Audible breath in Audible breath out Clipped or cutoff word or syllable (e.g., “hap-“ for cut-off “happen”) Shot, untimed pause (e.g., “Bill - decided”) Long, untimed pause (e.g., “Bill - decided”) Segment lengthening (e.g., “will:” or even “will::”) Laughter, noises, etcv . . . Unintelligible stretch of speech, in some cases with an inserted guess

211

Linguistic pragmatics

18 INTENTION (INCLUDING SPEECH ACTS)1 Jesús Navarro

Introduction This chapter deals with some of the different senses that the term ‘intention’ has in pragmatics, and the way they are related to each other. I shall begin by distinguishing one employment of the term, which belongs to our folk psychological practices of understanding actions in terms of reasons, from a more technical use, related to the aboutness of language. After a brief historical sketch, I describe the intentional approach to pragmatics as an attempt to account for the intentionality of language (its aboutness) in terms of intentional action. I will do so by explaining the basic tenets of two very influential proposals: the Gricean theory of conversation and speech act theory. This chapter finishes with a review of contemporary debates on the foundations of pragmatics where intentions have a predominant role to play.

‘Intention’: two definitions There are two very different notions, or clusters of notions, commonly associated with ‘intention’ and various cognate terms. Firstly, the word ‘intention’ stems from our ordinary practices of understanding actions in terms of goals and reasons. For instance, there could be different ways of accounting for an agent’s performance in intentional terms, according to Anscombe’s (1957) classical approach to the issue: (1) Mary turned on the light intentionally. (2) While walking the corridor, Mary had the intention to turn on the light. (3) Mary pressed the switch with the intention to turn on the light. Attaining a unified account of these multiple uses has become one of the main tasks of the theory of action (Setiya 2011), but for our topic it may suffice to know that most authors converge today on what is called the standard theory of action. This is the idea that intentions accomplish such an explanatory role because they are psychological states in the minds of agents with causal effects on their behaviour. In this example, the light turned on as a causal effect of Mary’s cognitive state of intention. Therefore, such an account is causal and cognitive. Even if it is well

215

Jesús Navarro

extended, it has been contested on several occasions, and we must bear in mind that it is just one option among others in order to make sense of this folk psychological practice. Secondly, there is a technical use of the same term ‘intention’, usually (but not exclusively) associated with its cognate ‘intentionality’ – a seeming homonymy that is utterly misleading. To have intentionality, be intentional, or be intentionally directed is to be about something else, or being referred to something else, something that may not even exist. Coined by medieval scholastics (the Latin verb intendo means to point at, or aim at something), the term was recovered by nineteenth-century psychologist and philosopher Franz Brentano (1995: 68), who famously claimed that all mental states, and only them, have intentionality.That means that beliefs, desires, and also intentions (in the former cognitive sense) are psychological states directed towards something beyond themselves. For instance: John’s belief that the shop is open is about the shop, which is not in his mind, but out there. Mary’s intention to turn on the light, which is a mental state occurring in her mind, is about the light, which is at the other side of the corridor. Despite Brentano, there seem to be many non-mental candidates for intentionality, at least apparently. Some of them are present in the natural world, as when we claim that smoke on the horizon informs us about fire, footprints on the floor tell us that some animal has been there, or stripes in a bug’s abdomen alert us to its poison; some others belong to human culture, like signs, maps, or language, the latter of which is a prominent case of intentionality since linguistic expressions have meanings. Members of this highly heterogeneous group of phenomena will require very different explanations (Searle 1983; Haugeland 1990; Millikan 2004), and those explanations must somehow fit together. Finding out the way they do is one of the main tasks of the theory of content. The intentional approach to language is an attempt to bring both senses of ‘intention’ together by reaching an explanation of the intentionality of language in terms of the intentional performances of social agents. In a nutshell, the idea is that language has a meaning, and is thus intentional (in the sense of aboutness), because it is employed by people with certain intentions (in the folk psychological sense).This insight was in the origins of pragmatics as a study of language in use, and different champions in this field often differ on the way the intentions of agents may help us explain the intentionality of their language. But, before facing those alternatives, a quick look back at history may help us understand their significance.

Historical perspective Intentional approaches to language appeared quite recently in the analytic philosophy of language. At the beginning of the twentieth century, authors like Gottlob Frege, Bertrand Russell, or the young Ludwig Wittgenstein were focused on the problem of intentionality in its two main varieties: thought and language. One of their main aims was to explain what it is that allows our thoughts and language to be about the world, a problem they approached in a twofold way: by means of semantics on one hand, and syntax on the other. In contrast to this, in the interwar period, philosophers like Ludwig Wittgenstein himself, John L. Austin, or Peter Strawson began to consider language as a product of human action, built into the social environment, and not just as an abstract, semantic/syntactic structure. The novelty of their account was that they searched for an explanation of the intentionality of language not in its direct relation to the world, but through the mediation of speakers and hearers, who employ that language in their ordinary lives.Thanks to those authors, the way was open for an account of the intentionality of language in terms of the intentional character of human action. By the middle of the twentieth century, two conceptions of this task were proposed, initiated by Grice and Searle, respectively. Grice’s proposal aimed to explain linguistic meaning in 216

Intention

terms of the communicative intentions of speakers, together with the inferential capacities of hearers. In contrast to this, Searle’s proposal was based on Austin’s incipient ideas on the nature of ‘speech acts’ – actions we perform in saying something. The former initiated a branch of pragmatics mostly focused on communicative processes, whereas the latter aims to explain the occurrence of social performances that imply changes in speakers’ normative status (i.e., new rights and obligations arising from the performance of promises, donations, orders, warnings, invitations, and so on). Both conceptions of pragmatics have proved highly influential: Gricean proposals were decisive for the later appearance of relevance theory, giving rise to cognitive pragmatics and intention-based semantics, whereas Searlean speech act theory has been widely applied beyond linguistics, in fields like psychology, artificial intelligence, sociology, legal studies, or literary theory. Despite both being fruitful, the two trends are in fact quite an ‘uneasy couple’ (Dascal 2003: 507). They are sometimes considered as complementary theories, and sometimes as competing accounts. I will show in the next section that the different role accomplished by folk psychological intentions in each of them may be responsible for this tense relationship.

Critical issues and topics Communicative intentions and conversation I have said that Grice’s aim was to reach an explanation of linguistic meaning in terms of communicative intentions, which are those intentions agents have when they try to communicate something. Those intentions, as any other kind of intentions, are defined in terms of goals: the speaker said whatever she said because she was pursuing something.What is specific about those intentions is that speakers rely on the capacity of others to detect their own intentions by means of inference. In this way, the sort of communicative intentions that explain communicative actions are, according to Grice (1989), threefold. A speaker who tries to communicate something intends at the same time: (1) to produce a particular response in her hearer, (2) to make her hearer recognise that she is intending (1), and (3) to fulfil (1) on the basis of her fulfilment of (2). Communicational utterances would thus be performed by speakers with the intention to affect hearers by means of their very recognition of that same intention. Accordingly, understanding would be a matter of utterances being recognised by hearers as being intentionally produced by speakers in order to be recognised as intentional (of course, ‘speaker’ and ‘hearer’ are used here in a wide sense, since Grice’s account is not meant to be restricted to oral communication). At first sight, this account certainly looks stilted. Why could we not simply say that, in communicating, the speaker intends to convey some information to her hearer? The reason this could not suffice may be grasped by means of an example: imagine we have run out of milk at home; I would like my partner to buy some and, in order to attain this, I leave an empty milk bottle on the kitchen table. I don’t mind if she realises that I have done this on purpose, and she does not realise so, but she does discover that we have ran out of milk, and she buys some. Now, even if I succeed in what I intended, could we say that I communicated to her that we had no more milk? Grice’s intuition would be that I did not. I would have managed to convey that piece of information to her, but I would not have communicated it to her because I did not intend her to recognise that I was intending to convey that piece of information to her. In order to have a fully fledged 217

Jesús Navarro

case of communication, the speaker’s performance must be explainable as a fulfilment of complex communicative intentions along the lines of the threefold kind I have sketched. Since Grice’s initial attempt to define communicative intentions as self-referential exhibited fundamental problems (see Avramides 1989), Sperber and Wilson (1986: 58) proposed an alternative account, along the lines of the Gricean proposal, but distinguishing between informative and communicative intentions. In their view, any act of communication would respond to a first-order informative intention (the intention to convey information to others) and, at the same time, to a second-order communicative intention (the intention to make that informative intention manifest to the hearer). Communicative intentions would thus be second order (that is, referred to other intentions). One way or the other, communicative intentions would be the basis of what Grice called non-natural meaning (meaningnn for short), which is the kind of intentionality that characterises language, in contrast to more simple cases of natural meaning (as in the smoke on the horizon, footprints, or the bug’s stripes). Grice’s seminal idea is that language has meaningnn because it is the effect of actions that respond to communicative intentions. Notice that meaningnn was not involved in the previous example, even if it could be a case of successful communication (in case I intended my intention to be recognised, and succeeded in it). In order to have a fully fledged case of linguistic communication, we would additionally need the communicative intention to be fulfilled by means of some linguistic expression.The speaker must produce an utterance that is a token of a linguistic type, that is, an individual occurrence of a meaningful sentence. For instance, instead of leaving an empty bottle on the kitchen table, I could have left a note saying, ‘We’ve run out of milk!’ It may look simple, but difficulties arise for the Gricean project when we make this little change in the story. When communication is based on linguistic means, it seems to rely on the agents’ awareness of the meaning of the linguistic types employed – that is, on their knowledge of what is said by the sentences they utter. I know, for instance, that my partner knows that a note that says, ‘We’ve run out of milk!’ means that we have run out of milk. However, if Grice’s aim was to reach an explanation of the intentionality of language (the aboutness of meaningnn) in terms of the intentions of agents, and he took for granted that the instruments agents employ already have their own intentionality (given that expressions would have some previous meaning), his account would leave its target unexplained. Unlike the case of the empty bottle, whose meaning on the table is fully explained by what I intend to do by leaving it there, the case of the note is more complex. It does not have a meaning because I left it there: I left it there because it already had a meaning! The problem is that the verb ‘mean’ has a tricky double sense here: when we claim that a speaker ‘means’ something, we are talking about what she intends to do, or say, which is a matter of psychological explanations of her actions in terms of reasons; in contrast to this, when we claim that a sentence ‘means’ something, we are talking about what it refers to, which is a matter of the intentionality of language. We are not claiming that a speaker has a ‘meaning’ when we say that he means something, and we are not claiming that a sentence has a ‘reason’ to do what it does when we claim that it means something. This may look like a platitude, but it could be the source of much theoretical confusion in the field. Now, if we want to stick to an account of intentionality in terms of intentional action, the problem is how and why do sentences mean at all, besides or beyond what agents mean by using them? What would explain the intentionality of linguistic expressions themselves? The Gricean project requires some further account of meaningnn in terms of communicative intentions, since otherwise the intentional account of the intentionality of language would be painfully circular. Grice and most post-Gricean accounts have pursued this explanation in terms of conventions: by being systematically employed to perform some communicational purposes, linguistic 218

Intention

expressions acquired some sort of ‘timeless’ meaning that is independent of the ‘occasion’ meaning they may have while being used by any particular speakers (Grice 1989: 145–338). Speakers make use of sentences because there is something that is what is said by them; but, at the same time, the meaningnn of a sentence is an effect of its having become a conventional instrument by which agents intend to say what they say. In that way, linguistic expressions have an intentional content precisely because they are conventional means to fulfil very particular communicative intentions (even if speakers may use them in order to fulfil quite different intentions on particular occasions). In that way, speakers’ linguistic utterances rely on the intentionality of linguistic types but, at the same time, the intentionality of linguistic types stems from those communicative intentions conventionally associated to their utterances. Both what speakers say in their utterances and what is said by the sentences they employ would be explained in terms of communicative intentions (either those of the particular agent, for the utterance, or those conventionally associated to that linguistic type, for the sentence).

Illocutionary intentions and speech acts Let’s now look at the example of the note from a different perspective: the one of speech acts theory, whose main aim is not to account for communication, but for the sort of deeds that we perform by means of language, and the way those deeds transform the social scene. In this case, my note would constitute what is called a speech act: an utterance that would make it the case that agents are committed to new rights and obligations. And, following Austin (1962), my performance could be described as comprising three different nested sub-actions, which are different descriptions of what I actually did: firstly, I wrote something on a piece of paper – something that, as a matter of fact, has some linguistic structure, syntax, and meaning. That is what Austin called the ‘locutionary act’. Secondly, in leaving this note, I said that we had run out of milk, and indirectly requested my partner to buy some. Those are, according to Austin, ‘illocutionary acts’, and they are not just a matter of meaning, but of what he called force – since my utterance actually makes it the case that I performed these deeds. And thirdly, by means of these illocutionary acts, I may have attained some expected goals, like having the milk bought, and obtained some unexpected results, like annoying my partner because it was my turn to buy it. The Austinian perspective is another sort of intentional approach to language, since all three sub-actions may be explained in terms of the intentions agents had while performing them. That is: if we asked for the reasons why I did what I did, my act of leaving the note may be described as a fulfilment of (1) my locutionary intention to write some words; of (2) my illocutionary intention to tell something to my partner, or of the one to indirectly request something from her; and of (3) my perlocutionary intention to have milk bought for tomorrow. Nonetheless, it is important to realise that the role of intentions is crucially different in the case of illocutionary acts. It seems that the description of the utterance as an illocutionary act immediately implies the attribution of the illocutionary intention to the agent, whereas that is not the case with the other acts, which may be unintentionally performed. That is: I may have written some sentence unintentionally, and I may have annoyed my partner unintentionally, but I may not have made a request unintentionally. If I made a request, then I was intending to do it. Illocutions are constituted by intentions in a way locutions and perlocutions are not. 219

Jesús Navarro

This is a feature that is not exclusive of illocutionary acts: there are a good number of actions that are constituted by the intentions to perform them: to communicate, as we saw in the previous section, is a clear case, together with many other non-communicative deeds, such as ‘attacking’, ‘resting’, or ‘achieving’. In the case of both communicative and illocutionary acts, a social trait is added, since their successful performance is constituted not just by the fulfilment of the intention to perform them, but also by the hearer’s recognition of that same intention. In this sense, communicative acts and illocutionary acts seem equal. However, despite this similarity, according to Searle (1969), illocutionary acts are not mere communicative acts, since the former are constituted by rules, whereas the latter are merely regulated by them. Regulative rules apply to activities that are logically prior to such rules, and independent of them. Think about an activity like fishing: there may be many rules on what, when, and how to fish, rules that perhaps indicate the most efficient ways to do it, or ways in which it is allowed. However, the activity of fishing is not itself constituted by those rules: it is merely regulated by them. One may fish not following any rule in particular, and that may still count as ‘fishing’. In contrast to this, constitutive rules are logically prior to the activity in question. They make the activity possible in the first place, and there is no such activity if the agent is not following the rules. Classic examples are games like chess or baseball: people did not first begin to play chess, and then decide to regulate their practice according to rules – rules themselves constitute the practice. In order to perform actions governed by constitutive rules, an agent must know those rules, and that he is acting according to them. Illocutionary intentions, from this perspective, are those agents have when their aim is to follow such constitutive rules. In Searle’s opinion, the Gricean approach to language neglects the fact that speech acts are like chess, not like fishing.The sort of rationality Grice envisaged to account for communication is merely constrained by maxims of a regulative kind: it only requires intelligent agents, able to fulfil their communicative purposes and to guess those of others by inference. But that is not enough for the performance of speech acts, according to Searle, since speakers must additionally know the illocutionary rules in use in their community, and act according to them, in order to perform illocutionary acts. One must know what it is to make a promise in order to make one, or what it is to get married in order to do so. One must have the intention to perform these deeds, that is, to follow the rules that constitute those practices. A theory of speech acts ought to describe and explain, in Searle’s view, those constitutive rules, and the method he proposes for this is to analyse the meaning of illocutionary verbs (like requesting, offering, inviting, promising, or informing), which are explicit indicators of their illocutionary force. A study of those meanings would be the best way to make their rules of use explicit, and thus shed light on the nature of speech acts.This is a move that seems counter to the intentional approach to language, giving conceptual priority to linguistic meaning over agential intentions, and some authors have criticised it as a semanticization of pragmatics. More relevant is the idea that speech acts theory attributes a constitutive role to intentions in the performance of speech acts. Whenever a speech act is performed, the speaker must have had the intention to act according to some instituted rule – otherwise, no speech act was actually performed. However, not all intentions manifested in speech have this constitutive role. The speaker’s act may also respond to other intentions in a contingent way, since we may also express intentions by means of speech that are not constitutive of our speech acts. For instance, by requesting someone to do something we express our desire to have that thing done; by thanking, we express our gratitude. In this sense, most (but not all) speech acts count as expressions of specific mental states, called their ‘sincerity condition’ (Searle 1969, 62). In particular, some speech acts have intentions as their sincerity condition. The case of promises is paradigmatic in this respect, since they have a future intention of the speaker as their sincerity condition (that 220

Intention

is: whenever a speaker promises to do something, she expresses her intention to do it). In these cases, the role of the expressed intention is not constitutive: even if the speaker did not have the intention she expressed, the speech act did occur (her speech act would be insincere, but still committing, giving rise to new rights and obligations). In contrast to this, as I said before, it seems that constitutive, illocutionary intentions must be present whenever the speech act takes place, even in insincere performances. To summarise: on one hand, Gricean communicative intentions are intentions speakers have when they try to influence their hearers by means of their recognition that they are trying to influence them. On the other hand, Austinian/Searlean illocutionary intentions are the intentions to perform speech acts, which are those that modify the normative status (rights and obligations) because they are constituted by socially instituted rules. Those two approaches are not clearly compatible, given the different roles intentions accomplish in each one of them, and the different sort of rationality they attribute to agents. Grasping this difference is essential to navigate the current literature on pragmatics whenever the complex notion of intention is at stake.

Recent work and future directions My aim in this final section is to pinpoint recent debates on the foundations of pragmatics where intentions have an important role to play. Hopefully, understanding on those debates may be earned from the distinctions made in the previous section.

The semantics/pragmatics interface A very influential proposal coming from the Gricean trend is Sperber and Wilson’s relevance theory (1986). Theirs is a work that puts speech acts in a corner, and definitely focuses on the problem of communication. In their view, two main models account for this phenomenon: the code model and the inferential model. The code model would explain communication as the process by which a speaker transfers some message by means of a code, while understanding would consist of the audience decoding that message. In contrast to this, the inferential model claims that communication may be attained without the help of any shared code, just relying on the inferential capacities of the agents. The example of the empty milk bottle (in case it fulfilled my communicative intentions) is a case of purely inferential communication, whereas that of the note implies a code. One basic idea behind Sperber and Wilson’s proposal is that code models are not autonomous, but always depend on inferential capacities.They claim that there is no way to make sense of communication and understanding as a simple process of coding and decoding. One of the main reasons they adduce for this is that codes are not specific enough by themselves to express the thoughts or psychological attitudes of speakers: It is true that a language is a code which pairs phonetic and semantic representations of sentences. However, there is a gap between the semantic representations of sentences and the thoughts actually communicated by utterances. This gap is filled not by more coding, but by inference. (Sperber and Wilson 1986: 9) Such a radical unevenness between mental attitudes and sentence meanings is argued to have two quite disturbing effects, which are the two sides of the same coin: firstly, we could not make the content of a speaker’s intention fully explicit by simply indicating a sentence that expresses 221

Jesús Navarro

it; and secondly, it would be impossible in principle to define the meaning of a given sentence in terms of the content of a speaker’s intentions. There is an important and lively dispute on the interface between semantics and pragmatics that may be interpreted as an effect of this problem. Semantic minimalists defend a traditional frame where semantics determines the content of what is said by a sentence. It does so by allowing us to specify the proposition an utterance expresses in the light of the conventional meanings of words, compositional rules for combining them, and some objective and innocuous features of the context. According to authors like Bach (2004), Cappelen and Lepore (2005), or Borg (2006), pragmatic processes would take all of that as an input and use it inferentially in order to find out the communicative intention of the speaker – and perhaps what speech acts have been actually performed. In contrast to this, defenders of pragmatic intrusion, like Recanati (2010) or Jaszczolt (2005), claim that there is no way to specify the content of what is said in any relevant sense, unless we take into consideration the pragmatic context. And this crucially includes the intentions with which speakers made their utterances. An effect of this sort of position, sometimes labelled as contextualist, is that it implies quite a radical attack on the autonomy of semantic deliberations, which would depend on pragmatic concerns. This debate has significant implications for the very possibility of the Gricean project. If some form of contextualism is right, it seems that the input to pragmatic inference would not be constructed prior to pragmatics itself. If what is said is the starting point for the hearer to find out the speaker’s communicative intentions and, at the same time, pragmatic inference on the agential intentions of speakers is a must to determine what is said, then it seems that we are enclosed in circular reasoning. This is what Levinson (1995; 2000: 186) called Grice’s circle – a problem affecting the very basis of pragmatics that seems to be still on the table. The battlefield of this debate is thus the notion of what is said. According to some authors, its content is some sort of ‘minimal proposition’, which ought later to be specified or qualified by means of pragmatic processes; according to others, we may not attribute any content to it before pragmatic inferences on the intentions of the participants are made. Remembering the two senses of intention exposed at the beginning of this chapter, we may say that it is the autonomy of the intentionality of language that is at stake: semantic minimalists attribute intentionality to linguistic expressions themselves, whereas contextualists and defendants of pragmatic intrusion claim that we may find no fully fledged intentionality in language unless we take the intentional character of agency into consideration.

Between internalism and externalism Many areas of philosophy employ a famous distinction between internalism and externalism, which roughly classifies theories depending on their proneness to rely on mental features of the person or facts about the world around her in order to account for some phenomenon. Applying this distinction to pragmatics, Harnish (2009) has claimed that theories of speech inspired by Austin or Searle are prone to postulate external conditions, since they try to explain the workings of speech acts by making conventions, rules, and norms explicit, while those inspired by Grice are more inclined to internalism, since they understand communication as a process of unveiling communicative intentions. Even if making a clear-cut classification of different pragmatic theories according to this distinction does not seem possible, the tension between internalism and externalism is manifest in the literature. A good example is the discussion of the way the content of demonstratives may be determined. David Kaplan (1977) famously proposed that demonstratives are

222

Intention

incomplete unless accompanied by an overt physical demonstration, since the linguistic rules that govern their use are not sufficient to determine their referents. However, Kaplan himself later argued against this idea, claiming that the ‘directing intention’ of the speaker determines what is the target of the demonstrative (Kaplan 1989). He thus moved from an externalist to an internalist position. This step was criticised by Reimer (1991), who claimed that only manifest, observable, ostensive gestures are semantically significant, regardless of the speaker’s intentions.This, again, is a clearly externalist position – see further discussion in Mount (2008) and King (2014). This is an epistemological issue that would not be deprived of important consequences for research in pragmatics, affecting our methodological strategies to confront empirical data. A radically externalist position would preclude us from including features about the speakers’ minds in our descriptions of their communicative exchanges. And even if we were allowed to describe the conversational scene in cognitive terms, the notion of intention could still be banned for different reasons. For instance, an author like Bach (2006) is known for having resisted admitting intentions as a part of the context on the grounds that context ought be understood as the basis the hearer has when she tries to infer the speaker’s intentions – a reservation that might also be found in Fodor and Lepore (2004). In order to properly assess those refusals, it is crucial to stipulate the notion of context under consideration: on one hand, there is the Austinian idea of context as ‘the total speech act in the total speech situation’ (Austin 1962: 147), where it is hard to deny that intentions are an important part (unless we also wanted to impeach the whole intentional account of action); but, on the other hand, there is an epistemic notion of context, which is the basis for the hearer’s inferential processes. Either understood as some form of ‘common knowledge’ between speaker and hearer (Lewis 1969), or as ‘mutual knowledge’ among them (Schiffer 1972), the epistemic context is what speaker and hearer allegedly share before the communication of new information takes place. It is here where the presence of intentions may seem controversial (Bach 2006: 546). Bach seems right in claiming that it does not make any sense to introduce the speaker’s communicative intention as a part of the epistemic context that allows the hearer to infer that very same intention, since otherwise communication itself would be futile. However, I would say that nothing in his arguments forces us to exclude other intentions the speaker may have as a part of the epistemic context that allows the hearer to infer some current communicative or illocutionary intention. For instance: if other intentions have been previously manifested, then they may be part of the shared or common epistemic context, which allows the hearer to infer the speaker’s specific intentions at some given moment. In this sense, what we need is a deeper understanding of what it is to make an intention public or manifest, allowing it to become a part of the epistemic context. Mitchell Green (2007) is putting forward ideas in this respect that could be ground-breaking. Elaborating on empirical data in the field of biosemiotics, he explains the way in which hidden states and capacities of animals may be made manifest by means of external signs, insofar as those signs have been selected with the function to make those hidden states overt to others. They are even better indicators if they are costly to produce, because then they are hard to imitate by fakers, becoming ‘handicaps’. In Green’s opinion, speech acts ought to be understood along these lines, as a particular kind of ‘handicap’ (2009): they are produced to allow speakers to manifest their intentions in the social scene, and an imbalance between what is manifested and what actually takes place in the speaker’s mind is punished with social exclusion. In the end, the frontier between natural and non-natural meaning (as the one between the internal and the external) may not be as deep as we initially thought it was.

223

Jesús Navarro

Conclusion: Beyond intentions Up to this moment, I have described a model where each individual utterance is explained in terms of its intentional origins, considering each utterance as an independent act motivated by some psychological state of the speaker. However, a recent trend in pragmatics promotes a different approach, focussed on the interactional and collaborative character of conversation. Instead of accounting for speech performance as a sequence of individual and independent utterances chained in serial exchanges, some scholars propose to look at it as a whole process of social interaction, performed simultaneously by all the participants (Bara 2010). In some cases, this theoretical shift is put forward as an abandonment of the intentional account in favour of some alternative possibility (Arundale 2008), whereas in other cases a similar emphasis on the dialogic and negotiated character of communication does not seem to imply such a deep breach with the intentional account (Thomas 1995; Haugh 2009, 2013, Haugh and Jaszczolt 2012). There are serious doubts that a simple causal model of folk psychological explanations, based on the idea that each individual action is the effect of some particular intention in the mind of the speaker, could do all the explanatory work required to account for the complex ways in which agents are held responsive for their words and interact with others by means of speech. In this sense, discussion of the role of agential intentions in communication and speech acts may be illuminated by recent developments in the theory of action, which encourage us to go beyond the cognitive-causal notion of intention that sustains the standard account of action. Besides their mental states being the cause of their behaviour, agents are present in their actions in other complex ways, by means of which they make sense of their deeds as their own. To account for this complex kind of belonging or recognition there are different theoretical instruments at our disposal: not just a standard causal theory of mind (Davidson 1980), but more complex ways to elucidate the rationality of action, such as meta-representational capacities to form second-order beliefs, desires, or intentions about our own first-order mental states, in order to guide them towards what is really important for us (Frankfurt 1988); capacities of prevision and projection, articulated in temporally structured planning strategies oriented towards social coordination (Bratman 1987); or the ability to integrate events in narratives, whose coherence and intelligibility allows us to consider them as a meaningful part of our own biography (Ricœur 1994; Velleman 2009).This seemingly theoretical disquisition could be illuminating in empirical research in order to make sense of the way agents seem to negotiate with others the meaning and force of their utterances, while they try to get along by means of speech. And, although these clues impel us to look beyond intentions, following them could be the only way to attain a really successful intentional account of language.

Suggestions for further reading Grice, H. P. (1989) Studies in the Way of Words. Cambridge, MA: Harvard University Press. Collects the classic papers that initiated the study of language in terms of communicative intentions, and the project of intention-based semantics. Searle, J. R. (1969) Speech Acts: An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Is his systematic development of Austin’s seminal ideas: an account of speech acts theory based on constitutive rules. Green, M. S. (2007) Self-Expression. Oxford: Oxford University Press. Contains a recent development of speech acts theory, well informed on biological and evolutionary facts. Haugh, M. (2008) ‘Intention in pragmatics’, Intercultural Pragmatics, 5(2): 99–110. A collection of recent papers discussing intentions’ role in pragmatics from different perspectives. O’Connor,T. and Sandis, C. (eds.) (2010) A Companion to the Philosophy of Action. Malden:Wiley-Blackwell. A comprehensive collection of papers that addresses the main issues in the theory of action.

224

Intention

Note 1 I would like to thank Anita Avramides,Teresa Bejarano, Adriano Bermudo, María Marbán and Margarita Planelles for insightful discussion and comments. Financial support for this chapter has come from the Spanish government (FFI2011–25131, Ministerio de Investigación e Innovación).

References Anscombe, G. E. M. (1957) Intention. Oxford: Blackwell. Arundale, R. B. (2008), ‘Against (Gricean) intentions at the heart of human interaction’, Intercultural Pragmatics, 5, 229–258. Austin, J. L. (1962) How to Do Things with Words, J. O. Urmson and M. Sbisà (eds.). Oxford: Oxford University Press. Avramides, A. (1989) Meaning and Mind: An Examination of a Gricean Account of Language. Cambridge, MA: MIT Press. Bach, K. (2004) ‘Minding the gap’, in C. Bianchi (ed.) The Semantics/Pragmatics Distinction. Stanford, CA: CSLI Publications. 27–43. Bach, K. (2006) ‘What does it take to refer?’, in E. Lepore and B. C. Smith (eds.) The Oxford Handbook of Philosophy of Language. Oxford: Oxford University Press. 516–556. Bara, B. G. (2010) Cognitive Pragmatics:The Mental Processes of Communication. Cambridge, MA: MIT Press. Borg, E. (2006) ‘Intention-based semantics’, in E. Lepore, and B. C. Smith (eds.) The Oxford Handbook of Philosophy of Language. Oxford: Oxford University Press. 250–266. Bratman, M. (1987) Intention, Plans, and Practical Reason. Cambridge, MA: Harvard University Press. Brentano, F. (1995) Psychology from an Empirical Standpoint. London: Routledge. Cappelen, H. and Lepore, E. (2005) Insensitive Semantics: A Defense of Semantic Minimalism and Speech Act Pluralism. Oxford: Blackwell. Dascal, M. (2003) Interpretation and Understanding. Amsterdam: John Benjamins. Davidson, D. (1980) Essays on Actions and Events. Oxford: Clarendon. Fodor, J. and Lepore, E. (2004) ‘Out of context’, Proceedings and Addresses of the American Philosophical Association, 78(2): 77–94. Frankfurt, H. G. (1988) The Importance of What We Care About. Cambridge: Cambridge University Press. Green, M. S. (2007) Self-Expression. Oxford: Oxford University Press. Green, M. S. (2009) ‘Speech acts, the handicap principle and the expression of psychological states’, Mind and Language, 24(2): 139–163. Grice, H. P. (1989) Studies in the Way of Words. Cambridge, MA: Harvard University Press. Harnish, R. M. 2009. ‘Internalism and externalism in speech act theory’. Lodz Papers in Pragmatics 5: 9–31. Haugeland, J. (1990) ‘The intentionality all-stars’, Philosophical Perspectives, 4: 383–427. Haugh, M. (2009) ‘Intention(ality) and the conceptualization of communication in pragmatics’, Australian Journal of Linguistics, 29(1): 91–113. Haugh, M. (2013) ‘Speaker meaning and accountability in interaction’, Journal of Pragmatics, 48(1): 41–56. Haugh, M. and Jaszczolt, K. M. (2012) ‘Speaker intentions and intentionality’, in K. Allan and K. M. Jaszczolt (eds.) The Cambridge Handbook of Pragmatics. Cambridge: Cambridge University Press. 87–112. Jaszczolt, K. M. (2005) Default Semantics: Foundations of a Compositional Theory of Acts of Communication. Oxford: Oxford University Press. Kaplan, D. (1977) ‘Demonstratives’ in J. Almog, J. Perry and H. Wettstein (eds.) (1989) Themes from Kaplan. Oxford: Oxford University Press. 481–564. Kaplan, D. (1989) ‘Afterthoughts’ in J. Almog, J. Perry and H. Wettstein (eds.) Themes from Kaplan. Oxford: Oxford University Press. 565–614. King, J. C. (2014) ‘Speaker intentions in context’, Noûs, 48(2): 219–237. Levinson, S. C. (1995) ‘Three levels of meaning’, in F. R. Palmer (ed.) Grammar and Meaning: Essays in honour of Sir John Lyons. Cambridge: Cambridge University Press. 90–115. Levinson, S. C. (2000) Presumptive Meanings:The Theory of Generalized Conversational Implicature. Cambridge, MA: MIT Press. Lewis, D. (1969) Convention. Cambridge, MA: Harvard University Press. Millikan, R.G. (2004) Varieties of Meaning. Cambridge, MA: MIT Press. Mount, A. (2008) ‘Intentions, gestures, and salience in ordinary and deferred demonstrative reference’, Mind and Language, 23(2): 145–164. Récanati, F. (2010) Truth-Conditional Pragmatics. Oxford: Clarendon.

225

Jesús Navarro Reimer, M. (1991) ‘Demonstratives, demonstrations and demonstrata’, Philosophical Studies, 63: 187–202. Ricœur, P. (1994) Oneself as Another. Chicago: University of Chicago Press. Schiffer, S. (1972) Meaning. Oxford: Clarendon. Searle, J. R. (1969) Speech Acts: An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Searle, J. R. (1983) Intentionality: An Essay in the Philosophy of Mind. Cambridge: Cambridge University Press. Setiya, K. (2011) ‘Intention’, in E. N. Zalta (ed.) The Stanford Encyclopedia of Philosophy Archive. Available from: http://plato.stanford.edu/archives/spr2011/entries/intention/ (accessed May 2015). Sperber, D. and Wilson, D. (1986) Relevance: Communication and Cognition. Oxford: Blackwell. Thomas, J. (1995) Meaning in Interaction: An Introduction to Pragmatics. London: Longman. Velleman, J. D. (2009) How We Get Along. Cambridge: Cambridge University Press.

226

19 TEMPORAL REFERENCE Kasia M. Jaszczolt

Introduction: Temporal reference, the concept of time and real time Temporal reference in discourse is a topic that is intrinsically connected with the question of how humans conceptualise time. As such, it is also intrinsically connected with the question of linguistic relativity in that the observable differences between natural languages in expressing temporality can be regarded as reflections of the conceptual differences or, alternatively, explained as superficial differences that can be superimposed on the universal construal of temporal reference at the level of atomic concepts. In what follows, I therefore focus on two aspects pertaining to this important question: the views on the human concept of time and crosslinguistic differences in expressing temporality. These two aspects will allow us to propose plausible answers to such questions as: 1. Is the human concept of time universal or language-/culture-specific? 2. What is the relation between the human concept of time and the natural-language devices used to represent temporal reference? 3. What is the division of labour between lexicon, grammar and pragmatics as far as expressing temporal reference is concerned? In the concluding part, I also make tentative comments concerning the relationship between temporality and the metaphysics of time. This chapter is structured as follows. I begin with placing the debates on temporal reference in the context of the interfaces between the metaphysics, epistemology, linguistics, and psychology of time, followed by a brief introduction to two major orientations as far as temporality is concerned, called in the philosophical literature ‘tensed’ and ‘tenseless’ approaches. Next, I move to selected current debates concerning the semantics and pragmatics of time, focusing on the relationship between temporality and modality and on the question of the universal vs. language-specific status of temporal reference. In the next section, I move to the discussion of pertinent cross-linguistic differences in expressing temporality and present and assess explanations of these differences. I conclude with pointing out cutting-edge problems and debates, as well as with an informed speculation on future directions in research on temporal reference.

227

Kasia M. Jaszczolt

Even basic knowledge of Einstein’s theory of relativity equips an average speaker of a natural language with a – not necessarily fully understood – conviction that time is not absolute but varies with the velocity of the moving object. At the same time, an average speaker of a natural language is equipped with a strong conviction that there is some constancy to the placement of events on a timeline, as well as to the classification of these events as past, present or future. Natural languages tend to reflect this conviction by affording us with the expressions for temporal placement such as absolute (although sometimes requiring pragmatic enrichment) adverbials (‘in January 2013’, ‘on 28 February’, ‘on Monday’) or relative temporal adverbials (‘after Christmas’, ‘before the holidays’), as well as ways of expressing the future, present or past-time orientation, be it lexicalised or grammaticalised (‘last week’, ‘two months ago’, ‘he did it’, ‘he will do it’). In view of this fact that languages lexicalise and grammaticalise temporal distinctions that in themselves are said to be relative, the question arises as to where to look for the explanation of the human concept of time: in reality (metaphysics), in cognition (epistemology) or in language use (linguistics and psychology), to name only a few of the relevant disciplines. Searching for the conceptual underpinnings of the linguistic expressions of temporal reference we unavoidably focus on two types: those pertaining to units of time and those pertaining to time flow. To this distinction we then have to add the concept of relative ordering of these units: event e1 may happen before event e2, after e2, or be contemporaneous with e2. Next, we can superimpose this classification into those that have passed (the past), are current (the present) and those that have not yet taken place (the future). Some languages make use of all of these distinctions; some use only a subset. Some add even more finely-grained ones such as the remote and near past, extended present and punctual present. Some languages focus on the temporal orientation; others focus on the source and strength of information with which the events are narrated. To find our way through this significant linguistic diversity, we have to begin with the human concept of time and its relation to ‘real’ time. The Aristotelian conception of time, essentially used in Isaac Newton’s Principia Mathematica towards the end of the seventeenth century, was intuitive and simple: time was represented as a horizontal line that extends infinitely in both directions, into the past and the future. Theological disputes concerning divine intervention notwithstanding, this view happily coincides with the concept of time that we normally adopt. However, already at that time it was suggested that speed may be intertwined with time in that the times of the eclipses of the moons of Jupiter as observed from Earth appeared to be related to the distance of the moon. The ‘ether theory’, according to which the speed of light was to be measured against this enigmatic substance, marks the beginning of the long route to the theses proposed by Henri Poincaré and Albert Einstein that time is not absolute. From then on, the development of the special and then general theory of relativity has been progressively confounding the commonsense picture. We now know that events that appear simultaneous to a stationary observer may not appear as such to a moving agent. A scientifically informed speaker of natural language is likely to know (or believe) that time ‘slows down’ when a star collapses into a black hole, that time has a beginning in the ‘big bang’, and that determinism about the future cannot be correct in that, as Heisenberg’s principle of uncertainty explains, the future position and speed of a particle cannot be predicted: the more accurately we try to measure the position, the more we alter the velocity. The question arises: do these scientific theories have anything in common with the concept of time reflected in natural language discourse, or, rather, do conversational interactants operate in terms of some other, psychological concept of time that is likely to resemble the traditional Newtonian timeline? In what follows, I explore different options and tentatively suggest that human time as expressed in natural language discourse has little to do with the physicists’ debate on the dimensions of space-time but, at the same time, it is compatible with it in that the 228

Temporal reference

spatiotemporal dimensions are perfectly compatible with the psychological dimensions instituted by the fact that human life has a beginning and an end, and is constituted by experiences, memories of experiences and anticipations of experiences. It is this phenomenological dimension that allows the human concept of time and metaphysical time to coexist and to give rise to the lexical, grammatical and pragmatic instantiations. The best way to introduce this ‘psychological time’ is to refer to Edmund Husserl’s (1928) Lectures on the Phenomenology of the Consciousness of Internal Time. The intricacies of this theory cannot be presented here, but, in brief, Husserl distinguishes the apparent time of external objects and the internal time of mental acts through which we cognise these objects. There is time in reality, but it is only useful to us when we reanalyse it in terms of experiences, memories and anticipations. The present, the past and the future are reanalysed as properties of consciousness. Perception of the present moment is called ‘primal impression’, the past is called ‘retention’ (in memory), and expectation (of the future) is called ‘protention’. In brief, for Husserl, time is ‘constituted by consciousness’; time comes ‘from within’ and is imposed on our experiences, and then indirectly on events and states (for a more detailed account, see e.g. Brough’s introduction in Husserl 1928; Bell 1990). What matters for the purpose of the pragmatics of temporal reference is that the phenomenological understanding of time is part of understanding oneself: human life has boundaries, and so has the human concept of time; our life is composed of experiences here and now and experiences remembered as well as those to come – and so is the human concept of time. What different languages ‘choose’ to emphasise in their expressions of temporal reference depends on many factors (cultural, social, historical, geographical, to name a few), but underlyingly these linguistic externalisations are commensurate with the overall concepts of experiencing, remembering and anticipating. As such, they are also compatible with the metaphysical conception of space-time in that the ‘arrow of time’ points in the same direction in thermodynamic, psychological and cosmological explanations: what has happened is irreversible in that (i) the disorder (entropy) increases in the world, just as the future becomes the present and then the past, and just as (ii) the universe is progressively expanding (see e.g. Hawking 1988: 144).

Unreal time: the legacy of John Ellis McTaggart Current debates in the philosophy of time centre around two theories put forward (and subsequently rejected) by Cambridge philosopher John Ellis McTaggart just over a century ago. McTaggart (1908) discusses two approaches to time, called by him A and B theory, and concludes that neither of them is satisfactory and that time is unreal. A theory assumes the existence of the past, the present and the future; B theory assumes the relations of precedence and following among events: I shall speak of the series of positions running from the far past through the near past to the present, and then from the present to the near future and the far future, as the A series.The series of positions which runs from earlier to later I shall call the B series. The contents of a position in time are called events. McTaggart (1908: 111) The A series is necessary to represent change; the B series is necessary to represent ordering of events, but does not suffice: in order to talk about ‘earlier-than’ and ‘later-than’ relations, we have to presuppose the A series and the existence of time. Without it, we would just have a timeless cline which he calls the C series. And yet, the A series cannot be fulfilled in that an event cannot 229

Kasia M. Jaszczolt

simultaneously possess the characteristics of being past, present and future.When we remove the ‘simultaneously’ requirement, however, we presuppose the existence of time and the account becomes circular. McTaggart therefore concludes that the past, the present and the future are not properties of reality, but rather belong to human experience in that we have memories, perceptions and anticipations that create the illusion of time passing and result in the experience of the past, the present and the future. However, the reasoning that since neither A nor B works, time must be unreal, is not correct. Reality of time need not mean the same thing as reality of time flow. For example, for Mellor (1998), time does not flow and events do not turn from being future to being present and then to being past. But we ‘live in time’ and interpret events as if the time flow were real. In this respect, Mellor is a B-theorist and a Kantian; causation is for him an important dimension of psychological time. He argues that there is no objective ‘now’, no objective ‘present moment’. For example, while observing a supernova, we ascribe to it a now that is not the same as the time of the explosion, which in fact took place in the remote past. Nothing in reality is ‘now’ or ‘then’; there is no A theory to rely on (Mellor 1998: 52). Instead, he proposes that it is this causal order of events we experience that gives rise to our concept of time. Hence, our experience relies on the flow of time and on the A series. But this internal flow of time does not supervene on the real time flow. A-theoretic and B-theoretic explanations have engendered heated debates in the literature. They remain at the forefront of the discussions on the metaphysics and epistemology of time because the challenge is their reducibility that makes one of them suffice as a theoretical explanation. The A-theoretic stance can be gleaned from St Augustine’s writings on time and more recently has been represented by Prior (1967, 1968, 2003), Smith (1993), Ludlow (1999) and Øhrstrøm and Hasle (1995). Its most commonly adopted version is the so-called presentism according to which all there is, is the ‘now’: no past or future events exist, although we cannot deny that they did or will happen (but see Rasmussen 2012 for an example of a reduction of A-properties to B-properties in presentism: presentism is compatible with tenseless reality). B theory tends to be more popular among both philosophers and linguists. Its followers include, among others, Bertrand Russell, Albert Einstein, Hans Reichenbach (1948), Hugh Mellor (1998), Joshua Mozersky (2001) and Robin Le Poidevin (2007). The problems with the reality of the A series notwithstanding, it seems that the ‘arrow of time’ Hawking discussed can still be real in the ontological sense: memories create the past, and anticipations the future, and the arrow relies on the relation of causation. Quite independently of the question of the relation of psychological time to real time, there is a fundamental question concerning the status of the concept of time with respect to its universality or cultural relativity, as well as with respect to the primitive/compositional nature. In other words, the concept of time may presuppose real time or not, it may be universal or culture-specific, and it may constitute an atomic entity or be composed out of more primitive units. In what follows, I briefly address the view that both real time and the concept of time can be translated into a more primitive notion of modality. Real time is founded on modality in the sense of metaphysical possibility, while psychological time is founded on modality in the form of epistemic possibility. As a result, arguably, modality obviates the need for a linear sequence of events as well as the need for causation as an explanans. In everyday parlance, we have the experience of perfect quantifiability of time. On one hand, there are eventualities (states, events and processes, but see also more detailed classifications such as Vendler’s [1967]); on the other, there are intervals and moments in time that are normally easily expressible in natural language by means of temporal adverbials and, where available, aspectual distinctions. If we adopt the B theory, we can locate a given eventuality in relation to others. 230

Temporal reference

We can also adopt an ‘eclectic’ approach on which the experience, memory and anticipation superimpose a certain now on the scale, ending up with a form of presentism that presupposes B rather than A theory (Rasmussen 2012). Distinguishing between the essential B-theoretic perspective and the subjective superimposition of now is important in that it allows us to combine the metaphysical and the epistemic (and thereby also linguistic, conversational) perspectives on time in that, as Davies puts it, The labels ‘past’ and ‘future’ may legitimately be applied to temporal directions, just as ‘up’ and ‘down’ may be applied to spatial directions, but talk of the past and the future is as meaningless as referring to the up and the down. (Davies 2012: 11) In other words, temporal distinctions are ‘real’, but only in the sense of reality imposed by the experiencing agent. However, the question remains how to represent the intuition of the qualitative difference between the past (memories of experiences) and the future (anticipations of experiences) when the tools we have available are quantitative in the form of distances, intervals and points on the scale. We can, for example, assume that while the present is real, actual and existing, the past is only real insofar as it bears some effect on the now (Łukasiewicz 1961), where the effect can be achieved with or without agency. If the past event does not have present effects, it is degraded to the status of mere possibility. The future can be conceptualised analogously: it exists when it can be projected, anticipated, from the present. This anticipation is intrinsically connected with the agent’s personal time, viewed in the perspective of his/her own finite life, bounded by the events of birth and death – a theory Heidegger (1953) developed in his Sein und Zeit. Ample linguistic evidence supports this relative status of the past and the future. Faller (2002) points out that languages have hierarchies of epistemic modals out of which speakers can choose in order to convey a proposition with an adequate degree of commitment to its truth. In possible-worlds semantics, modality is conceived of as strongly interacting with time. Modal statements are analysed in terms of the existence of a possible world that belongs to the speaker’s set of worlds (creating the speaker’s epistemic state), such that a certain property is instantiated in this world at a certain time (represented as an interval). Kratzer (1981) argues that these worlds can be ordered, producing a gradation of modality. Different interactions between modal and other operators (such as present or perfect) or grammatical features of constituents (such as the verb class) produce different temporal distinctions as in (1)–(3). (1) Aly may/might go to London (tomorrow). (2) Aly may/might be in London (now). (3) Aly may/might have been in London (yesterday). In order to explain the asymmetry of time (which standardly amounts to claiming that the past and the present are real and the future is not, see e.g. Tooley 1997), we can either correlate the types of such constructs with the past and the future orientation, respectively (which I called elsewhere, Jaszczolt 2013, the modal-contextualist view), or index the instances of the use of the modal operator with the specification with respect to the hic et nunc (with ‘n’ standing for ‘now’) as in mayn, mayn+1, mayn-2, etc., depending on the accepted temporality scale (the directquantitative view). Arguments for the modal-contextualist view would include the fact that some modals have a default temporal orientation (e.g. the Thai modal auxiliary d1ay1II that by default assumes the past-time reference, see Srioutai 2006). Also, assuming that the English will 231

Kasia M. Jaszczolt

has retained (or regained, see Fleischman 1982 on bidirectional semantic shift) its modal status, the future-time reference also testifies to this solution. Needless to say, when analysed as supervenient on modality, time seems to be well represented by a B theory, with the phenomenological proviso that the human lifespan always establishes a point of reference n. This outcome allows us to bring together B-theoretic metaphysical time and the essentially B-theoretic but qua underlying modal and qua situated in n, human concept of time.

Selected problems and debates In order to represent temporal reference in truth-conditional semantics, the metalanguage of predicate logic had to be extended. Propositions had to be relativised to times, which means in practice that models against which the truth values are checked render extensions relativized to times. One way to proceed is to adopt Prior’s (e.g. 1967, 1968, 2003) tense logic in which four sentential operators are added to represent temporal location, among them F and P for the future and the past, respectively, as in (4b). (4) a.  Fido barked. (4) b.  P Bark (Fido). The truth conditions for Pp and Fp are as in (5) and (6) respectively: (5) Pp is true in M at t iff ∃t' (t' < t ∧ p is true in M at t') (6) Fp is true in M at t iff ∃t' (t < t' ∧ p is true in M at t') where t stands for time instant, ‘iff ’ for ‘if and only if ’, ‘∧’ for the logical connective of conjunction, and ‘∃’ for the logical operator of existential quantification (‘there is a time t'. . . . ’). It has to be pointed out at this juncture that the unit of time is variously conceptualised in the literature with respect to duration: it is taken to be an instant (moment) or an interval, with the interval as the methodologically and metaphysically preferred option (see e.g. Dowty 1979; Kamp and Reyle 1993 for a discussion). The elegance of the representations in (5)–(6) notwithstanding, one of the main current debates in the philosophy of time revolves round the notion of ‘truth-makers’: are truth-makers for past and future propositions instantiated in the present, or are they themselves in the past or the future? If the latter, the problem of the reality of the past and the future reappears with full force; if the first, then the problem lies with the metaphysical status of these truth-makers (see e.g. Dummett 2004, 2006; Tallant and Ingram 2012). In Discourse Representation Theory (DRT, e.g. Kamp and Reyle 1993), events and states are represented with temporal location with respect to the present in that time can precede or follow the deictically construed now (as in t < n) and an event can be temporally included in t (e ⊆ t) or a state can overlap with the temporal reference given by t (s ο t). These semantic representations clearly make use of the B-theoretic concept of temporal ordering, but they also make use of the A-theoretic concept of the present, and in consequence of the past (t < n) and the future (t > n). An alternative way of semantically representing temporal reference is construing a temporal argument. For example, Enç (1987) represents tenses as referential expressions denoting intervals with the temporal argument for the verb (see also Hornstein 1990 on tenses as adverbs). But this discussion is tangential to the pragmatics of temporal reference.

232

Temporal reference

Now, while the semantics of temporal reference is relatively well researched and the solutions adhere to one of the previously discussed models, the pragmatics of temporal reference has to rely on more diversified sources in utterance processing.This is so independently of the question as to whether we want the semantics to (more or less) mirror syntax, as in minimalist approaches (see Borg 2004, 2012; Cappelen and Lepore 2005), or, rather, we want it to represent the intuitive content that can be remote from the content provided by the logical form of the sentence, as in contextualist approaches (see e.g. Recanati 2004, 2010; Jaszczolt 2005, 2010). Note that whatever kind of content semantic theory represents, the minimal-sentential or the contextualintuitive, it can fit squarely into the DRT-like representations. In other words, when we take, for example, tense-time mismatches such as in (7) and (8), truth-conditional, model-theoretic semantics offers tools to represent the content both of the sentence and of the utterance. (7) I can’t meet you tomorrow. I am rowing for my college. (8) This is what happened to me last weekend. I go to the library, sit down in the reading room, and hear this voice behind me saying, ‘How are you, old chap’. . . . It is clear from the context that (7) has future-time reference and (8) past-time reference, and this information can easily be included in the representation of the truth-conditional content. The theory of Default Semantics, which is a radical contextualist truth-conditional semantic theory, analyses many examples of tense/time mismatches, as well as instances of communication where there is no overt marker of temporal reference, and shows how information from different sources available in conversation can be merged on the level of such a pragmatics-rich semantic representation. I return to this topic in the following section. Analogously, the fact that metaphorical expressions employed for the purpose of temporal reference exhibit universal (or at least widely cross-culturally shared) as well as language-specific characteristics points in the direction of their common conceptual core with the culture-relative overlay. But debates still continue on such topics as the status of the conceptualisation of time along the horizontal axis with the left-right orientation pertaining to the past-future (see e.g. Boroditsky 2001 and a response from January and Kako 2007; Vallesi et al. 2008; Ulrich and Maienborn 2010) and the status of the spatial metaphorical schemas on which events or speakers are moving objects (Lakoff and Johnson 1980, 1999; see Stern 2000 for criticism).These debates, albeit all directly pertinent, have to be left out of this introduction as they belong more in the domains of cognitive science, psychology and cognitive semantics than in pragmatics sensu stricto.

Cross-linguistic differences in temporal reference: some examples Referring to the past, present and future can be accomplished in a variety of ways, both within one selected language and cross-linguistically. These ways include the lexicon, the grammatical markers of time, automatic assignment of a salient interpretation to an overtly tenseless expression and conscious pragmatic inference of the temporality of the situation referred to in the utterance. In other words, the means range from lexical, through grammatical, to pragmatic. To repeat, this linguistic diversity in expressing temporal reference is represented within a single language and also delineates some cross-linguistic differences. Some languages do not have grammatical tenses, for example Mandarin Chinese, Navajo, Yucatec Maya, West Greenlandic. Yucatec Maya lacks both tenses and temporal adverbials; temporal reference is achieved through the interaction of aspectual marking with default pragmatic interpretations (see Bohnemeyer 2002; Saussure 2012). In English, temporal ordering of events is normally achieved pragmatically

233

Kasia M. Jaszczolt

by the linear juxtaposition of sentences referring to events, sometimes linked by the conjunctive connective as in (9). (9) Tom practised for six months and passed his Grade 7 piano exam. In Swahili, the meaning ‘and then’ is rendered by grammatical means in the form of the socalled consecutive tense: the marker ka replaces the tense marker as in (10) where the past time reference is marked only in the first phrase by li, after which the consecutive ka is used. (10) a. . . . . wa-Ingereza  wa-li-wa-chukua  wa-le   maiti, 3pl-British    3pl-past-3pl-take 3pl-dem corpses ‘. . . . then the British took the corpses, b. wa-ka-wa-tia    katika bao  moja, 3pl-cons-3pl-put.on on   board one put them on a flat board, c. wa-ka-ya-telemesha maji-ni  kwa  utaratibu w-ote.  .  .  . 3pl-cons-3pl-lower  water-loc with order   3pl-all and lowered them steadily into the water. . . . ’ (after Jaszczolt 2009: 90, adapted from Givón 2005: 154). It has to be remembered that the grammatical and the pragmatic means of expressing temporal ordering are equally adequate from the point of view of communication of information. Ambiguities created by the sequence of tense (SOT) phenomena such as in example (11) further corroborate the idea that grammatical and pragmatic means of expressing temporality have to be regarded on an equal footing. In (11), it is the pragmatic inference (or default assumptions) that assign the temporal reference to Mary’s state as located at the time of speaking or in the future. (11) John will think that Mary is pregnant.(from Hornstein 1990: 86). Next, it is pertinent to consider the evidentiality/temporality distinction. Matses, a Panoan language spoken in the Amazon region, has an evidential system that requires that the source of information is overtly specified when the speaker reports on a past event on the basis of inferential evidence (see Fleck 2007). The sentence has to state overtly how long ago the event took place, as well as how long ago the speaker obtained evidence. This is done by means of inflection that Fleck calls double tense: the verbal inflectional suffix combines temporal and evidential information. This is exemplified in (12); ERG stands for ‘ergative’, and DIST.PAST.INF-REC. PAST.EXP for ‘distant past inferential’ combined with ‘recent past experiential’. The overall temporal information is that the speaker discovered the hut a short time ago, while it was made a long time ago. (12) mayu-n          bëste-wa-nëdak-o-s¸h. non.Matses.Indian-ERG hut-make-DIST.PAST.INF-REC.PAST.EXP-3 ‘Non-Matses Indians (had) made a hut.’ (after Jaszczolt 2012: 103, from Fleck 2007: 590). 234

Temporal reference

Matses utilises three past tenses: recent, distant and remote. There are also three evidential distinctions: direct experience, inference and conjecture. According to Fleck, nëdak expresses the distant past from between about a month ago to the speaker’s infancy, supported by inferential evidence, while o is used for recent past, from immediately before the time of the utterance to, normally, up to a month ago, supported by experiential evidence. As a result, nëdak-o combines two foci of temporal reference: the distant past of the making of the hut, and the recent past of obtaining information about it. Matses also allows other combinations, such as that for recent past inferential or distant past experiential, which adds up to nine markers for the past time reference. The temporal reference expressed as double tense in Matses would require in English either the combination of the grammar and the lexicon or, when the temporal reference is obvious from the context, the combination of the grammatical and the pragmatic means. Next, tense-time mismatches that are well exemplified in English discourse testify to the grammar/pragmatics mix in representing temporality. (7), repeated below, with an overtly present continuous tense, is an example of the so-called futurate progressive. (7) I can’t meet you tomorrow. I am rowing for my college. Sentence (13), with the simple present form, is an example of the so-called tenseless future. (13) Tomorrow I go to London. In (8), repeated below, we have an instance of the use of the simple present with past-time reference, called the past of narration (or, less appropriately from the point of view of the temporal reference it expresses, vivid or historic present). (8) This is what happened to me last weekend. I go to the library, sit down in the reading room, and hear this voice behind me saying, ‘How are you, old chap’. . . . All in all, the verb forms in English need not map onto the temporality of the expressed eventuality. Further, as is well known, there are languages in which temporal reference is not grammatically marked (as in Mandarin), or it can be marked but this marking is optional (as in Thai). St’àt’imcets (Lillooet Salish) of British Columbia has only the future (marked by ‘kelh’) – nonfuture distinction (Matthewson 2006). In Central Pomo, a North American language, the future can be expressed as either realis or irrealis, depending on the degree of probability the speaker attaches to the event (see Haan 2006: 42). Another native American language, Caddo, expresses the future as realis (Haan 2006: 41). This is only a handful of pertinent examples of how the expression of temporal reference differs cross-linguistically as well as intralinguistically. Faced with this diversity, one ought to pose a question as to whether the human concept of time is universal or language- (or culture-) relative. On one hand, the presence of intralinguistic relativity seems to suggest that the linguistic diversity need not reflect the conceptual diversity. On the other, the fact that some languages make certain temporal concepts overtly expressible by grammaticalising or lexicalising them signals that the concept itself may have cultural roots. The debate between universalism and relativity, rekindled recently in the domain of space (see e.g. Levinson 2003; Levinson and Burenhult 2009), is also far from resolved in the domain of time. Like in the domain of space, so in the domain of time it seems that there are some correlations between expressions of a particular language and concepts, but at the same time it seems that these differences can be brought down to universal basic building 235

Kasia M. Jaszczolt

blocks. In Representing Time (Jaszczolt 2009), I put forward a view that the human concept of time, as well as metaphysical time itself, supervenes (depends in terms of definitional characteristics) on modality. In particular, expressing temporality means underlyingly expressing a certain degree of detachment from certainty – a detachment triggered by the concept of detachment from the here and now. Commonsense judgement suggests that the past and the future are removed from such certainty while present experiences come with the ‘zero’ degree of detachment. However, as linguistic evidence suggests, even the present allows for such a gradation; just as the past can give rise to stronger or weaker representations in the memory, and the future to stronger or weaker anticipations, so the present can result in strongly or weakly held representations, often associated with the nature of evidence. The use of modals for the present, as in (14b)–(14d), testifies to this gradation. The acceptability of the tenet that Tom is in London is weaker than that expressed in (14a). (14)  (14)  (14)  (14) 

a.  b.  c.  d. 

Tom is in London now. Tom may be in London now. Tom might in London now. Tom should be in London now.

It has been well documented and demonstrated through theoretical arguments that temporality and modality are closely intertwined. The English will combines both functions, with different degrees of modal colouring in different periods of its historical development. In many languages, the past-tense form is also used to express ‘unreality’ in conditionals and this association appears to be systematic (see e.g., Portner 2009: 242–243). In addition to ample supporting evidence from languages where modality and temporality are expressed by the same constructions, the proposal that temporality can be explained in terms of modal building blocks has an important advantage. As discussed earlier, languages exhibit lexicon/syntax/pragmatics trade-offs in expressing temporal reference. These different sources of information about time can be most adequately represented when we dissociate the semantic representation of temporality from grammatical tense and from the temporal adverbials but instead regard the latter two on a par with the pragmatic means. Only by adopting this perspective can we adequately account for the tense-time mismatches discussed previously, giving due consideration to the fact that they are not exceptions but rather common, well-entrenched ways of speaking. The implications for formal semantic accounts are that relying on formal devices to represent temporality, such as the feature TENSE in DRT (Kamp and Reyle 1993), will not suffice. Equally, adopting this assumption we can represent temporal reference in utterances without any temporal markers and thereby treat the pragmatic means of expressing time on a par with the overt linguistic markers such as adverbials, tense and aspect (see e.g., Saussure 2012, for example on the interaction of pragmatic inference with semantic aspect). Default Semantics (DS) is a radical contextualist theory of discourse interpretation that allows for such equal treatment of the sources of information (Jaszczolt, e.g., 2005, 2009). For example, in Thai, (15) can express any temporal reference with any degree of epistemic detachment, such as those in (15a)–(15i) (from Srioutai 2006: 45). (15) m3ae:r3i:I kh2ian n3iy3ai: Mary

write novel

(15) a. Mary wrote a novel. (15) b. Mary was writing a novel. (15) c. Mary started writing a novel but did not finish it. 236

Temporal reference

Figure 19.1  Representation of the pragmatically inferred reading of (15) as (15a) in Default Semantics

(15) d. (15) e. (15) f. (15) g. (15) h. (15) i.

Mary has written a novel. Mary has been writing a novel. Mary writes novels. / Mary is a novelist. Mary is writing a novel. Mary will write a novel. Mary will be writing a novel.

The past-time reference can be represented in DS as in Figure 19.1 where Σ stands for the representation of the meaning intended by the speaker and recovered by the addressee (the socalled merger representation), ‘x’ and ‘y’ for the discourse referents pertaining to ‘Mary’ and ‘novel’ respectively, the subscripts stand for the type of process used in utterance interpretation that contributes to the resulting reading (CD = ‘cognitive default’; WS = ‘processing of word meaning and sentence structure’; CPI = ‘conscious pragmatic inference’), the square brackets stand for the material on which the process operates, and ACCΔrp ├ Σ for the modal operator ‘it is acceptable to the degree Δ, here pertaining to the regular past (rp), that Σ' is the case’ (for the details of the latest version of the theory, see Jaszczolt 2010). Temporal concepts lexicalised or grammaticalised in various languages, such as those mentioned earlier, can also be represented in Σs of DS in that, to repeat, the theory can represent all lexical, grammatical and pragmatic sources of information as well as the interaction of the resulting processes of interpretation (see Jaszczolt 2012 for more examples of accounting for linguistic diversity and Jaszczolt 2013 for the ways of translating the future-past asymmetry into the values of Δ in ACCΔrp ├ Σ). All in all, the commonsense hypothesis that the conceptualisation of time has both universal and culture-dependent characteristics finds a ready explanation when we subscribe to the theory that what is universal is the modal building blocks and what is language-specific is the realisation of these as temporal adverbs and adverbial phrases, tense, aspect, evidentials, default temporal interpretations of tenseless expressions or temporal interpretations resulting from pragmatic inference as governed by Gricean (or neo-Gricean) principles of rational communication.

Summary and future directions It is evident from this brief overview that lexicon and the grammar alone are not reliable means of expressing the human concept of time. Instead, language users rely on the interactions and 237

Kasia M. Jaszczolt

trade-offs between various means, including grammar and lexicon, but also pragmatic inference, social and cultural assumptions (a.k.a. salient meanings, conversational defaults) and other sources, depending on how finely-grained we want our analysis to be. These interactions and trade-offs are present within a single natural language, testifying to a variety of choices afforded to language users. As such they constitute a pragmatic universal of pragmatic, interactive, compositionality of meaning. But they also help explain cross-cultural differences in speaking about time, at the same time allowing for the strong possibility that there is a common underlying conceptual level where time is essentially represented as an epistemic modal notion, the departures from the now – the now that is in constant flux for every language user.The essence of human temporal reference is precisely this modal basis, expressed through these diversified means. Thanks to the universal principles of human communication, best captured as some rationality heuristics of a Gricean kind (Grice 1975; Horn 1984, 2004; Levinson 2000), this diversity does not hinder but rather enriches communication and normally results in the match between the intended and the recovered content – in temporal reference and across the board.

Suggestions for further reading Callender, C. (ed.) (2011) The Oxford Handbook of Philosophy of Time. Oxford: Oxford University Press. This handbook provides a comprehensive introduction to various aspects of philosophy of time. Filipović, L. and Jaszczolt, K. M. (eds.) (2012a) Space and Time in Languages and Cultures: Linguistic Diversity. Amsterdam/Philadelphia: John Benjamins. The articles in this collection offer examples of cutting-edge research on how different languages represent location in time (and space) and what aspects of temporal (and spatial) reference pose a problem for a language learner. Filipović, L. and Jaszczolt, K. M. (eds.) (2012b) Space and Time in Languages and Cultures: Language, Culture, and Cognition. Amsterdam/Philadelphia: John Benjamins. The articles in this collection complement the previous volume in discussing linguistic and conceptual representation of events and theories thereof, as well as various cultural perspectives on time (and space). Jaszczolt, K. M. (2009) Representing Time: An Essay on Temporality as Modality. Oxford: Oxford University Press. The book is devoted to the semantics and pragmatics of temporal reference in natural language. It presents an overview of problems and solutions as well as a proposal that the human concept of time depends in terms of definitional characteristics on epistemic modality (the epistemic detachment from the ‘here and now’), accompanied by semantic representations in the theory of Default Semantics. Saussure, L. de (2012) ‘Temporal reference in discourse’, in K. Allan and K. M. Jaszczolt (eds.) The Cambridge Handbook of Pragmatics. Cambridge: Cambridge University Press. 423–446. The article concerns pragmatic inference of information about temporality and aspectuality in discourse.

References Bell, D. (1990) Husserl. London: Routledge. Bohnemeyer, J. (2002) The Grammar of Time Reference in Yucatec Maya. Munich: Lincom Europa. Borg, E. (2004) Minimal Semantics. Oxford: Oxford University Press. Borg, E. (2012) Pursuing Meaning. Oxford: Oxford University Press. Boroditsky, L. (2001) ‘Does language shape thought? Mandarin and English speakers’ conceptions of time’, Cognitive Psychology, 43: 1–22. Cappelen, H. and Lepore, E. (2005) Insensitive Semantics: A Defense of Semantic Minimalism and Speech Act Pluralism. Oxford: Blackwell. Davies, P. (2012) ‘That mysterious flow’, Scientific American Special, 21(1): 8–13. Dowty, D. R. (1979) Word Meaning and Montague Grammar: The Semantics of Verbs and Times in Generative Semantics and in Montague’s PTQ. Dordrecht: D. Reidel. Dummett, M. (2004) Truth and the Past. New York: Columbia University Press. Dummett, M. (2006) Thought and Reality. Oxford: Clarendon Press.

238

Temporal reference Enç, M. (1987) ‘Anchoring conditions for tense’, Linguistic Inquiry, 18: 633–657. Faller, M. (2002) ‘Remarks on evidential hierarchies’, in D. Beaver, L. Casillas Martínez, B. Clark and S. Kaufmann (eds.) The Construction of Meaning. Stanford: CSLI Publications. 89–111. Fleck, D. W. (2007) ‘Evidentiality and double tense in Matses’, Language, 83: 589–614. Fleischman, S. (1982) The Future in Thought and Language: Diachronic Evidence from Romance. Cambridge: Cambridge University Press. Givón,T. (2005) Context as Other Minds:The Pragmatics of Sociality, Cognition and Communication. Amsterdam: J. Benjamins. Grice, H. P. (1975) ‘Logic and conversation’, in P. Cole and J. L. Morgan (eds.) Syntax and Semantics, (vol. 3). New York: Academic Press. Reprinted in H. P. Grice (1989) Studies in the Way of Words. Cambridge, MA: Harvard University Press. 22–40. Haan, F. de (2006) ‘Typological approaches to modality’, in W. Frawley (ed.) The Expression of Modality. Berlin: Mouton de Gruyter. 27–69. Hawking, S. W. (1988) A Brief History of Time: From the Big Bang to Black Holes. London: Bantam Press. Heidegger, M. (1953) Sein und Zeit. Tübingen: Max Niemeyer; trans. J. Stambaugh (1996) Being and Time. Albany: State University of New York Press. Horn, L. R. (1984) ‘Toward a new taxonomy for pragmatic inference: Q-based and R-based implicature’, in D. Schffrin (ed.) Georgetown University Round Table on Languages and Linguistics. Washington, DC: Georgetown University Press. 11–42. Horn, L. R. (2004) ‘Implicature’, in L. R. Horn and G. Ward (eds.) The Handbook of Pragmatics. Oxford: Blackwell. 3–28. Hornstein, N. (1990) As Time Goes By:Tense and Universal Grammar. Cambridge, MA: MIT Press. Husserl, E. (1928) Vorlesungen zur Phänomenologie des inneren Zeitbewusstseins. Jahrbuch für Philosophie und phänomenologische Forschung IX. Halle: Max Niemeyer; trans. J. B. Brough (1991) ‘Lectures on the phenomenology of the consciousness of internal time’, in On the Phenomenology of the Consciousness of Internal Time (1893–1917), (part A). Dordrecht: Kluwer. January, D. and Kako, E. (2007) ‘Re-evaluating evidence for linguistic relativity: Reply to Boroditsky (2001)’, Cognition, 104: 417–426. Jaszczolt, K. M. (2005) Default Semantics. Oxford: Oxford University Press. Jaszczolt, K. M. (2009) Representing Time: An Essay on Temporality as Modality. Oxford: Oxford University Press. Jaszczolt, K. M. (2010) ‘Default Semantics’, in B. Heine and H. Narrog (eds.) The Oxford Handbook of Linguistic Analysis. Oxford: Oxford University Press. 193–221. Jaszczolt, K. M. (2012) ‘Cross-linguistic differences in expressing time and universal principles of utterance interpretation’, in L. Filipović and K. M. Jaszczolt (eds.) Space and Time in Languages and Cultures: Linguistic Diversity. Amsterdam: John Benjamins. 95–121. Jaszczolt, K. M. (2013) ‘Temporality and epistemic commitment: An unresolved question’, in K. M. Jaszczolt and L. de Saussure (eds.) Time: Language, Cognition, and Reality. Oxford: Oxford University Press. 193–209. Kamp, H. and Reyle, U. (1993) From Discourse to Logic: Introduction to Modeltheoretic Semantics of Natural Language, Formal Logic and Discourse Representation Theory. Dordrecht: Kluwer. Kratzer, A. (1981) ‘The notional category of modality’, in H.-J. Eikmeyer and H. Rieser (eds.) Words,Worlds and Contexts – New Approaches to Word Semantics. Berlin: W. de Gruyter. 38–74. Rev. ed. in A. Kratzer (2012) Modals and Conditionals. Oxford: Oxford University Press. 27–69. Lakoff, G. and Johnson, M. (1980) Metaphors We Live By. Chicago: University of Chicago Press. Lakoff, G. and Johnson, M. (1999) Philosophy in the Flesh: The Embodied Mind and Its Challenge to Western Thought. New York: Basic Books. Le Poidevin, R. (2007) The Images of Time: An Essay on Temporal Representation. Oxford: Oxford University Press. Levinson, S. C. (2000) Presumptive Meanings:The Theory of Generalized Conversational Implicature. Cambridge, MA: MIT Press. Levinson, S. C. (2003) Space in Language and Cognition: Explorations in Cognitive Diversity. Cambridge: Cambridge University Press. Levinson, S. C. and Burenhult, N. (2009) ‘Semplates: A new concept in lexical semantics?’ Language, 85: 153–174. Ludlow, P. (1999) Semantics,Tense, and Time: An Essay in the Metaphysics of Natural Language. Cambridge, MA: MIT Press.

239

Kasia M. Jaszczolt Łukasiewicz, J. (1961) ‘O determinizmie’, in J. Słupecki (ed.) Jan Łukasiewicz. Z zagadnień logiki i filozofii. Pisma wybrane. Warszawa: Państwowe Wydawnictwo Naukowe; trans. Z. Jordan (1970) ‘On determinism’ in L. Borkowski (ed.) Jan Łukasiewicz: Selected Works. Amsterdam: North-Holland. 110–128. Matthewson, L. (2006) ‘Temporal semantics in a superficially tenseless language’, Linguistics and Philosophy, 29: 673–713. McTaggart, J. E. (1908) ‘The unreality of time’, Mind, 17. Reprinted in J. E. McTaggart (1934) Philosophical Studies. London: E. Arnold. 110–131. Mellor, D. H. (1998) Real Time II. London: Routledge. Mozersky, J. M. (2001) ‘Smith on times and tokens’, Synthese, 129: 405–411. Øhrstrøm, P. and P. F. V. Hasle (1995) Temporal Logic: From Ancient Ideas to Artificial Intelligence. Dordrecht: Kluwer. Portner, P. (2009) Modality. Oxford: Oxford University Press. Prior, A. N. (1967) Past, Present and Future. Oxford: Clarendon Press. Prior, A. N. (1968) Papers on Time and Tense. Oxford: Clarendon Press. Prior, A. N. (2003) Papers on Time and Tense: New Edition. Oxford: Oxford University Press. Rasmussen, J. (2012) ‘Presentists may say goodbye to A-properties’, Analysis, 72: 270–276. Recanati, F. (2004) Literal Meaning. Cambridge: Cambridge University Press. Recanati, F. (2010) Truth-Conditional Pragmatics. Oxford: Clarendon Press. Reichenbach, H. (1948) Elements of Symbolic Logic. New York: Macmillan. Saussure, L. de (2012) ‘Temporal reference in discourse’, in K. Allan and K. M. Jaszczolt (eds.) The Cambridge Handbook of Pragmatics. Cambridge: Cambridge University Press. 423–446. Smith, Q. (1993) Language and Time. Oxford: Oxford University Press. Srioutai, J. (2006) ‘Time conceptualization in Thai with special reference to D1ay1II, Kh3oe:y, K1aml3ang, Y3u:I and C1a’, unpublished PhD thesis, University of Cambridge. Stern, J. (2000) Metaphor in Context. Cambridge, MA: MIT Press. Tallant, J. and Ingram, D. (2012) ‘Time for distribution?’, Analysis, 72: 264–270. Tooley, M. (1997) Time,Tense, and Causation. Oxford: Clarendon Press. Ulrich, R. and Maienborn, C. (2010) ‘Left-right coding of past and future in language:The mental timeline during sentence processing’, Cognition, 117: 126–138. Vallesi, A., Binns, M. A. and Shallice, T. (2008) ‘An effect of spatial-temporal association of response codes: Understanding the cognitive representations of time’, Cognition, 107: 501–527. Vendler, Z. (1967) Linguistics in Philosophy. Ithaca, NY: Cornell University Press.

240

20 FORMAL AND NATURAL LANGUAGES What logic tells us about natural language1 Jacques Moeschler

Introduction There is an old tradition in semantics and pragmatics calling for a sound relation between logic and natural language (NL). One classical issue is the possible meaning relation between logical connectives and NL connectives. Observations about the discrepancies between the syntactic and semantic behavior of logical connectives (especially negation, conjunction, disjunction, and conditional) have led to different proposals. One of the classical outcomes can be summarised as a non-formalist approach: in that perspective, there is either an indirect relation between logic and language – this is mainly the position of most philosophers of languages, such as Strawson (1952) – or there is clearly no relation between logic and NL – this is for instance the position of the French semanticist Oswald Ducrot (Ducrot 1980, 1989; Anscombre and Ducrot 1983). On the opposite side are explicit proposals arguing for a clear and direct connection between logic and NL that can be defined as comprising the formalist position (Gazdar 1979; McCawley 1981). Hence, the most important challenge of the formalist position is to explain why natural languages do not behave like formal languages. Since Grice’s seminal article ‘Logic and Conversation’ (1975), which is the real foundation of inferential and cognitive pragmatics, a third possible relation between logic and NL has been considered: the difference between logic and natural language is not a question of semantics, but of pragmatics.This program explains how NL connectives, as well as negation, can receive a more precise pragmatic meaning. In this chapter, we address from a Gricean perspective the relation between logic and language, with a special focus on logical connectives and quantifiers. Before addressing these issues, we introduce the relation between logic and semantics and between propositions (the object of logic), sentences (the objects of linguistics), and utterances (the object of pragmatics). The last section addresses the general issue of how pragmatic enrichment can be computed from logical properties, and will be illustrated through the question of the pragmatic meaning of linguistic negation.

241

Jacques Moeschler

Logic as a universal tool in formal and lexical semantics Let us start with a French sentence such as (1), which is semantically ambiguous since it has two possible meanings, depending on the scope of negation (wide vs. narrow), of which the wide scope reading is preferred: (1) Tous les étudiants n’ont pas réussi. All the students Neg-Cl have Neg passed a.  Wide scope reading: Not all the students passed b.  Narrow scope reading: All the students did not pass One way of expressing these two meanings is to use a simple formal language such as predicate calculus, which is a first-order logic language. Predicate calculus contains quantifiers (universal ∀, existential ∃), predicates, and terms (proper names and variables), as well as a set of logical connectives (negation ¬, conjunction ∧, disjunction ∨, conditional →, equivalence ↔). Such a formal language has a syntax, defined by the set of expressions of the language and a set of rules, and a semantics, defined by the way expressions receive their denotation in a model, and by the corresponding semantic rules. In other words, each string of symbols expressing a meaning in that formal language receives an interpretation depending on the model in which the expressions and sentences are interpreted (cf. Dowty et al. 1981; Chierchia and McConnell-Ginet 1990; Moeschler 2007a for more precise presentations). Sentence (1) can receive different translations in predicate calculus, which give rise to different semantic interpretations. Figures 20.1 and 20.2 depict the two readings set-theoretically, whereas the formulae in (2) are said to be the logical forms of the sentence, and their meanings are represented as in (3): (2) a. wide scope reading (preferred): ¬[∀x [student(x) → pass(x)] ‘it is not the case that for every x who is a student, x passed’ b. narrow scope reading (dispreferred): ∀x [student(x) → ¬pass(x)] ‘for all x who is a student, it is not the case that x passed’ (for reasons of simplification, tenses are not semantically translated in these logical forms)

Student (X)

Pass (X)

Figure 20.1  Wide scope reading (‘not all students passed’)

Student (X)

Pass (X)

Figure 20.2  Narrow scope reading (‘no students passed’)

242

Formal and natural languages

(3) a. ⟦ ¬[∀x [student(x) → pass(x)] ⟧ = 1 if, and only if, the intersection of the denotation of student (a set of individuals) is not included in the set the individuals who passed – there must be a subset of students who did not pass (⟦α⟧ = 1 means ‘the denotation of α is true’). b. ⟦ ∀x [student(x) → ¬pass(x)] ⟧ = 1 if, and only if, the intersection between the set of students and the set of individuals who passed is the empty set – there is no subset as a result of the intersection of sets. This approach has given rise to a general theory for quantification, under the label of generalised quantifiers (Barwise and Cooper 1981). In this formal set-theoretical description, quantifiers are defined as relations between sets of individuals. For instance, the English quantifiers all, some, and no receive the following descriptions: (4) ⟦all P are Q⟧= {P} ⊆ {Q}

‘the denotation of all P are Q means that the set of P is included in the set of Q’

(5) ⟦some P are Q⟧ = {P} ∩ {Q} ≠ ø

‘the denotation of some P are Q means that the intersection of the set of P and the set of Q is not empty’

(6) ⟦no P is Q⟧ = {P} ∩ {Q} = ø

‘the denotation of no P is Q means that the intersection of the set of P and the set of Q is empty’

Hence, the first utility of logic, as far as logical words like negation and quantifiers are concerned, is to translate the meanings of sentences into a formal and unambiguous language, and to give their interpretation within a model in terms of truth-values. What is required, moreover, is that the formal devices can translate a natural language sentence into a formal language and then assign it an interpretation, as Figure 20.3 shows: Sentence in NL  Translation into Predicate Calculus  Interpretation in a Model Figure 20.3  From sentence to meaning

The translation step requires that sentences are generated via syntactic rules of predicate calculus, and then, in order to obtain a semantic interpretation, their logical forms are interpreted via semantic rules. As regards sentences containing quantifiers, a higher-order language must be used in order to give an expression for all words in natural languages, that is, for all lexical and non-lexical categories. For instance, the French linguistic quantifier tous les in (1) receives a translation by using the λ-operator: this operator is a functional operator, allowing defining sets from propositions. For instance, tous les defines a relation between two sets, given respectively by the restrictor (the common noun in the subject position, étudiant), represented by the variable Q, and the nucleus (the VP of the sentence réussir), represented by the variable P: (7) a. b. c. d.

tous les: λQ[λP[∀x[Q(x) → P(x)]]] étudiant: student ne. . .pas: ¬ réussir: pass 243

Jacques Moeschler

(8) gives the result of applying compositional syntactic rules, in which (b) and (c) are the result of applying a λ-conversion rule: this rule states that a predicate variable introduced by the λ-operator can be substituted by a predicate constant (student in (b), pass in (c)), respectively, for the predicate variables Q and P. The wide scope reading is obtained in (8): (8) a. ¬ [λQ[λP[∀x[Q(x) → P(x)]]] (student)(pass)] ‘it is not the case that, for the sets of predicates Q and P, for all x, if x is true of Q, then x is true of P, Q being here a student and P pass’ b. ¬ [λP[∀x[student(x) → P(x)]](pass)] ‘it is not the case that for the set of predicates P, for all x, if x is a student, then x is true of P, P being here pass’ c. ¬[∀x[student(x) → pass(x)]] ‘It is not the case that for all x, if x is a student, then x pass’. In other words, what (8) shows is how a simple negative sentence can be syntactically composed with formal expressions associated to every linguistic item of a natural language (respectively, tous les, étudiants, ne. . . . pas, réussir). Such a syntactic composition (8a) allows the application of lambdaconversation rules, which yields as a result a logical expression in predicate calculus (8c). The last step of the analysis will be the application of semantic rules producing as a result an interpretation in terms of truth values: as the model gives denotations to étudiant and réussir, a sentence as tous les étudiants n’ont pas réussi will be declared as true or false relative to the model of interpretation. A second example of a classical use of formal logic is meaning relations in lexical semantics. Some meanings are connected by entailment relations: in this case, if a proposition P(x) is true, then the proposition Q(x) is true too. For instance, if x is married, then x is not a bachelor. This relation can be stated with the help of the universal quantifier, meaning that there is no exception, that is, there is no x that can be married and a bachelor at the same time: (9) ∀x [married(x) → ¬bachelor(x)] These entailment relations are the strongest semantic relations in natural language (NL) and are the basis for conceptual organisation. For instance, hyponymy (a relation between a superordinate concept and its subordinate one, such as Chow and dog) and meronymy (a part-whole relation, such as finger and hand) are based on entailment: the relation between being a Chow and being a dog, a classical hyponymy relation, is indeed by entailment.This can be shown with the following examples: (10) Nath bought a Chow → Nath bought a dog (11) Nath did not buy a Chow → Nath bought a dog or Nath did not buy a dog Now, we resume our first argument for connecting logic and NL: logic, as a formal language, can be used to describe sentence and lexical meanings. Note, however, that this first argument for the relation between logic and NL does not say a lot about the genuine relation between them. Why not use Pascal, Oberon, Prolog, C, or Java as a formal language? The historical argument (‘logic was the first formal language to be used, so let us continue using this very robust formal language’) is weak, because a lot of studies on natural language processing use such formal programming languages (see, for instance, the series of books by Gazdar and Melish (1989) for programming languages such as Prolog, LISP, and POP-11). Therefore the systematic use of logic in language description and meaning theory, from antiquity to contemporary times, must be grounded in some other way than history. 244

Formal and natural languages

Propositions, sentences, and utterances Before giving serious and sound arguments for a logical analysis of connectives, let me make the concepts of proposition, sentences, and utterance more precise. Maximal linguistic units are sentences, whereas logic is about propositions. So the first question to ask is what could be the relation between sentences and propositions. In the formal semantics program, as presented in the section ‘logic as a universal tool in formal and lexical semantics’, it is admitted that the semantic content of a sentence is a proposition, which is evaluated in terms of truth-values. Generally speaking, two additional assumptions are taken for granted in linguistics: uttered sentences are speech acts, and in a classical analysis of speech acts (Austin 1962; Searle 1969, 1979; Searle and Vanderveken 1985), propositions are embedded under an illocutionary force operator, which can describe all types of forces for declarative, interrogative, and imperative sentences. Sentences (12) receive the speech act analysis given in (13): (12) a. b. c. d. e.

I accept to support your application. I promise you to support your application Can I ask you to support my application? Who phoned today? What a wonderful day!

(13) a. b. c. d. e.

ACCEPTATION (the speaker supports his addressee’s application) PROMISE (the speaker supports his addressee’s application) REQUEST (the addressee supports speaker’s application) QUESTION (someone phoned today) EXPRESSION (the day is wonderful)

Next, sentences are uttered in contexts, yielding utterances. Since utterances have effects in contexts, one part of the content of the utterance is its explicit content, the other its implicit one. In pragmatics, a difference is generally made between non-logical implicit meaning, defined as implicature by Grice (1975), and non-logical explicit meaning, that is, explicature (Sperber and Wilson 1986). For instance, taking into account these two layers of meaning (Moeschler 2013a), utterance (14B) receives the analysis given in (15): (14) A: B: (15) Explicature:  Implicature: 

How was your pragmatics exam? Some of my students passed. At least some of B’s students passed the pragmatics exam. Not all of B’s students passed the pragmatics exam.

At the explicit level, (14B) means that in order to be true, at least two students must have passed, and the implicature restricts the lower bound reading (where some is compatible with all) to ‘not all’ (see Noveck and Sperber 2007 and Moeschler 2017 for an alternative analysis at the level of explicature; Moeschler 2012 for a general introduction). It is important to acknowledge that these aspects of the semantic and the pragmatic contents of an uttered sentence correspond to propositions and, as such, have logical properties. We will come back to the logical properties of semantic and pragmatic meaning later on.

How to obtain a pragmatic reading from logical meaning The first argument for strengthening the relation between logic and NL is linked to logical words and their meaning in logic and NL. The discrepancy between these meanings 245

Jacques Moeschler

has given rise to three different positions: the formalist, the non-formalist, and the Gricean (pragmatic). The formalist approach is based on a very restricted definition of what could be a truth-functional connective. Formally, truth-functional connectives (TFCs) are ‘functions which take a set of truth values as their sole argument’ (Gazdar 1979: 75). The theoretical load of this definition is that only three connectives are candidates: conjunction, inclusive disjunction, and exclusive disjunction, because of their truth-conditional meaning and a pragmatic criterion, the principle of confessionality: a TFC must confess the falsity of its argument, that is, yields truth-value false when its arguments are false. For instance, the conditional connective (if. . . . then) is not confessional because it yields false arguments to true conditional sentences: if Peter comes, Mary will leave is true when both propositions are false (if Peter does not come, Mary will not leave). On the contrary, a disjunction is confessional, because it yields a false proposition when both disjuncts are false: cheese or dessert is false if there is no cheese and there is no dessert. On the other hand, the non-formalist approach, as represented, for instance, by Ducrot (1989), argues that non-logical uses of connectives, as the conditional, for instance, are explained by referring to a more complex semantic structure. For instance, a conditional in (16a) does not receive as a reading (16b), whereas (17a) receives the analyses given in (17b): (16) a. If you give him a whisky and some water, he will be happy. b. If you give him a whisky, he will be happy, and if you give him some water, he will be happy. (17) a. If you are thirsty, there is some beer in the fridge. b. If you are thirsty, I inform you that there is some beer in the fridge. What conclusions have been drawn from these arguments? It has been shown that, from a logical point of view, the meaning of connectives cannot be explained with the classic tools of logic: the formalist view (Gazdar) is so reductionist that no generalisation can be drawn to explain the linguistic meaning of TFCs.The non-formalist view (Ducrot), however, leads to a non-­reductionist approach that also fails to lead to any generalisations. At this stage, it would appear that the problem is too difficult to solve. If the non-formalists are right, then the semantics of natural language should not yield to the temptation of using the tools of logic to solve it. Unfortunately, the side effect of this radical position is very strong and devastating for semantics: the traditional views of Frege, Russell, and other philosophers of language might have to be abandoned. This was the situation before Grice’s proposal in his article ‘Logic and Conversation’. Initially, Grice’s philosophical concern was to explain the difference in meaning between logical operators such as ¬, ∧, ∨, ⊃, ∀, ∃ and their linguistic counterparts not, and, or, if, all, and some. The originality of Grice’s approach was to avoid the classic philosophy of language debate between the formalist and the non-formalist approaches. For Grice, one way of avoiding the mistake brought about by the formalist/non-formalist alternatives lay in inquiring ‘into the general conditions that, in one way or another, apply to conversation as such, irrespective of its subject matter’ (Grice 1975: 43). The main purpose of Grice’s logic of conversation was therefore to understand how and why logical words have such different uses in natural language. But in his seminal article, Grice made a larger proposition that explains how and why logical connectives in natural languages receive different pragmatic meanings in language use: he proposed a complete theory of non-truth-conditional meaning with a whole rationale on how pragmatic meaning can be derived from utterances and speaker’s meanings. Let us start with the simple logical meanings and the ordinary use of logical connectives in NL. Table 20.1 gives the truth table for the classical logical connectives: 246

Formal and natural languages Table 20.1 Truth conditions for logical connectives

P 1 1 0 0

Q 1 0 1 0

Conjunction

Disjunction

Conditional

Equivalence

Negation

P∧Q 1 0 0 0

P∨Q 1 1 1 0

P→Q 1 0 1 1

P↔Q 1 0 0 1

¬P 0 1

Let us examine some striking divergences between the logical and linguistic meanings of connectives (cf. Allwood et al. 1977 for an introduction to the meaning of logical connectives, and Mauri and Van der Auwera 2012 for a synthesis on the relation between logical connectives and NL ones). 1. Negation: First of all, negation in logic is a propositional connective, that is, it has wide scope on propositions. Its best paraphrase is ‘it not the case that. . . .’. For instance, sentence (18a) receives as its logical meaning (18b): (18) a. It is not raining. b. It is not the case that it is raining. Second, linguistic negation is mainly a constituent negation, with narrow scope. The first challenge for a logic-based meaning for negation is to explain how a wide scope operator can receive a narrow scope interpretation in NL. For instance, how can we explain that (19a) receives one of the four possible readings, implying that negation is a constituent negation, and not a sentential negation? (19)

a. Mary did not kiss John. b. Mary did not kiss John, Susan did. c. Mary did not kiss John, but Peter. d. Mary did not kiss John, she insulted him. e. Mary did not kiss John, Susan insulted Peter.

Third, with regard to the negation truth table, the use of linguistic negation is restricted to situations where the negated proposition is false, yielding a true proposition (0 → 1). So the truth table is too informative or under-informative and does not capture the main linguistic properties of negation. As an example, asserting (19a) does not mean that the negative sentence is false, because linguistic negation cannot apply to a true proposition. 2. Conjunction: One strong logical property of logical conjunction is its commutativity: (P ∧ Q) is logically equivalent to (Q ∧ P); in effect, the truth-conditions of a conjunction states that both conjuncts have to be true, which implies that the order is irrelevant with respect to truth-conditions. However, conjunction raises a new issue at the semantic and pragmatic levels: how to explain that in utterance (20), P and Q and Q and P do not have the same truth-conditions, that is, they describe different situations in the world: (20) a. Mary pushed John and he fell. b. John fell and Mary pushed him. 247

Jacques Moeschler

In other words, with events, the meaning of and is temporal and/or causal (see Wilson and Sperber 2012, chapter 8, and Moeschler 2000 for a synthesis): temporal and causal relations are not captured with the truth-conditions of and. 3. Disjunction: Logical disjunction is said to be inclusive, allowing both disjuncts to be true. The main use of or in NL is not inclusive, however: inclusive meaning is expressed by and/or (Ariel 2013). Or generally receives an exclusive reading, that is, excluding that both disjuncts can be true together. As a result, the use of ordinary or in NL conveys by implicature the ignorance of the speaker about the truth of the disjuncts. (21) is a classical Gricean example, implicating that the speaker does not know where the mother is: (21) Daughter:  Dad, where is mom? Father: In the cellar or in the kitchen. Menus in restaurants give another classical difference between or-interpretations. A menu with cheese or dessert is generally cheaper than a menu with cheese and dessert, because in the first case, the client cannot choose both. Again, the challenge for a logic-based meaning theory is to explain why natural languages restrict logical meanings. The classical pragmatic analysis assigns to the intended meaning the status of implicature: P or Q, in the context of the restaurant menu, implicates the negation of the upper-bound expression, that is, not(P and Q). The rationale of this analysis is that a rational speaker, being cooperative, is presumed to respect the first maxim of quantity (‘Make your contribution as informative as is required’, Grice 1975: 34). 4. Conditional: The use of conditionals in NL presents many differences with logical meaning, as has been shown in examples (16) and (17). One of the most sophisticated differences is represented by counterfactuals, which imply the falsity of both the antecedent and the consequent as in (22): (22) If I were rich, my 2CV would be a Rolls Royce. The most salient divergence in meaning is that NL conditionals are interpreted as bi-conditionals. This is an interesting property, because it logically means that by saying if P, then Q, the speakers means if not-P, then not-Q, which is not logically entailed by a conditional proposition, as example (23) shows: (23) a. If you mow the lawn, I will give you 10€. b. I you do not mow the lawn, I do not give 10€. c. If I do not give you 10€, you will not mow the lawn. In effect, the only logically valid inference from if P, then Q is if not-Q, then not-P, known as contraposition, which means that the only logical meaning of (23a) is given by (23c).The question to be answered is then how to explain the pragmatic inference if not-P, then not-Q, analysed as an invited inference by Geis and Zwicky (1971). These examples can be interpreted in two directions: first, as an argument against the relation between logical and linguistic meaning of connectives; second, as an argument for a pragmatic interpretation of the meaning of logical words (cf. Carston 1996, 2002; Moeschler 2010 for a post-Gricean analysis of negation; Horn 1972; Levinson 1983, 2000; Moeschler 2007b for a Gricean approach of logical words and quantifiers; Moeschler et al., forthcoming for a general introduction to implicatures). In what follows, we are going to follow this second path (see ‘A logical-semantic meaning for pragmatically enriched interpretations?’). 248

Formal and natural languages

The mystery of the non-lexicalisation of negative particulars What could be the relation between the semantics and pragmatics of NL logical words? To answer this question, I would like to discuss the general issue of what type of logical meaning can or cannot be represented in NL. Take the example of quantifiers and negation, which have been described in a logical square since Aristotle. The logical or Aristotelian square is made of four vertices, and two axes: positive vs. negative, universal vs. particular.These vertices are labeled A (positive and universal, Affirmo), I (positive and particular, affIrmo), E (negative and universal, nEgo), and O (negative and particular, negO). Figure 20.4 is the classical representation of the logical square. Affirmation Universals

Negation A

Contraries

E

Contradictories

Particulars

I

Subcontraries

O

Figure 20.4  The logical square

The vertices in this Figure entertain the following logical relations: (24)

a. b. c. d.

A and E are contraries: they cannot be true together. I and O are sub-contraries: they cannot be false together. A and O, as well as I and E are contradictories: only one can be true. A entails I, and E entails O: if A is true, then I is true; if E is true, O is true.

The following logical forms are associated with the vertices: (25) a. A ∀x [Fx → Gx] all F are G b. I ∃x [Fx ∧ Gx] some F are G c. E ¬∃x [Fx ∧ Gx] no F is G d. O ∃x [Fx ∧ ¬Gx] some F are not-G ¬∀x [Fx → Gx] not all F are G (these two expressions are logically equivalent) Linguistic examples in French and English are given in (26): (26) a. b. c.

Toutes les étudiantes sont intelligentes. All students are smart. Quelques étudiantes sont intelligentes. Some students are smart. Aucune étudiante n’est intelligente. No student is smart. 249

Jacques Moeschler Table 20.2  Lexical realisation of the logical square vertices A

I

E

O

French

English

French

English

French

English

French

English

tous toujours les deux

all always both

some sometimes one (of them)

*nitous *nitoujours *nideux

*nall *nallways *noth

and

aucun jamais ni l’un. . .ni l’autre ni

no never neither

et

quelques quelquefois l’un (des deux) ou

not

*niet

*nand

or



d.  Quelques étudiantes ne sont pas intelligentes/Toutes les étudiantes ne sont pas intelligentes. Some students are not smart/ Not all the students are smart. Now, two issues are linked to the logical square, both of which raise the question of the relation between logic and NL: First, do all vertices of the logical square receive a linguistic translation? Second, what is the relation between particulars and universals, from a pragmatic point of view? Let us first answer the question of linguistic translation. As Horn’s hypothesis in many works on the logical square states (for instance Horn 2004, 2012), there is no lexical realisation in NL for O. The reason Horn gives is the complexity of its semantic value. ‘Given that languages tend not to lexicalise complex values that need not to be lexicalised, particularly with closed categories like quantifiers, we predict that some. . . . not will not be lexicalised, and this is precisely what we find’ (Horn 2004: 11). Indeed, if we take the example of French and English, there are no lexical realisations for O as regards quantifiers, temporal adverbs, duals, plurals, or logical connectives (cf. Table 20.2). From a logical point of view, negative vertices are the result of the composition of negation and a quantifier, and the question is to know why ¬∃x is realised linguistically (by no), while ¬∀x is not (*nall). In fact, there is a logical answer to the problem of semantic complexity (Moeschler 2017). A, I, and E are defined by logical relations between sets: as stated in (4) to (6), the semantics of all, some, and no are the results of inclusion and intersection between sets. In the case of some and no, the computation of intersection is easy: for some, look for the intersection between two sets, and for no, check that the intersection is the empty set. But what about some. . . . not (not all)? Let us start with a classical semantic description of the denotation of (26d), repeated in (27): (27) Some students are not smart. What does it mean? What the quantifier and the negation target is the complement-set of the intersection obtained by some. So what an addressee must do is compute (i) the intersection set and (ii) its complement-set. Figure 20.5 depicts this computation, the grey subset being the denotation of (40). From a conceptual point of view, it means that natural languages do not like to attach complex meaning into a single lexical item. The only solution for NL is compositional, that is, syntactic: take a quantifier, take a negation, and use compositional syntactic rules to express a complex meaning, resulting from two logical operations.

250

Formal and natural languages

Student (X)

Smart (X)

Figure 20.5  ⟦Some . . . not ⟧ as a complement set

Now, we have to address a second question, which is about the relations between universals and particulars. From a logical point of view, universal quantifiers (positive or negative) unilaterally entail their corresponding particulars. So, for instance, all x entails some x, as no x entails not all x. But on the other hand, as this implication is unilateral, what is the relation between some and all, and not all and no? This question has been addressed with pragmatic tools, that is, the notion of implicature. Let us take the some/all contrast. Some has a lower-bound interpretation, meaning at least some (if not all), as made explicit in (28): (28) Some students, if not all, are smart. Now, some has also an upper-bound meaning: if the speaker says some, it is because he cannot say more (Grice 1975), as the first maxim of quantity states (‘Make your contribution as informative as is required (‘for the current purpose of the exchange’), Grice 1975: 45), in order not to rule out the supermaxim of quality (‘Try to make your contribution one that is true’, Grice 1975: 46). So a speaker saying some students are smart cannot commit himself to all students are smart. As the speaker is presumably respecting the first maxim of quantity, the addressee is expected to understand that some implicates not all. In other terms, all semantically entails some, and some pragmatically implicates not all. This meaning has been called a scalar implicature (this is the classical picture of scalar implicatures, as in Horn 1972, Gazdar 1979, Horn 1989, Levinson 2000, Geurts 2010). So, the logical square is not only a way of representing logical relations such as entailment, contrariety, contradiction, and sub-contrariety: it is also an explanation of one pragmatic relation, that is, scalar implicatures, triggered by linguistic material and belonging to what Grice (1975) called generalised conversational implicatures. As such, scalar implicatures are defined as nontruth-conditional meaning, being cancellable, as shown in (29): (29) Some students, in fact all, are smart. This general pragmatic property has led Gazdar (1979) to differentiate potential implicatures (im-plicatures) from actual implicatures. Implicatures are triggered by linguistic material, whereas (actual) implicatures have to be confirmed, either by linguistic or by contextual information. In conclusion, the first relation between logic and NL is pragmatic meaning. An implicature, being a non-truth-conditional meaning, is what NLs have been able to construct to ensure verbal communication that is efficient and rapid at exploiting inferential abilities. Moreover, implicit communication, going beyond entailments and other logical properties in building pragmatic relations, can be thought as one recent novelty in language evolution

251

Jacques Moeschler

and communication (see Reboul 2007, 2013, and 2017 for a complete argument). In a nutshell, logical and pragmatic meanings are entrenched in such a way they are part of human cognition.

A logical-semantic meaning for pragmatically enriched interpretations? Up until now, I have given arguments to defend examining the relation between logic and NL: there is a strong connection, because pragmatic meaning is connected to logical meaning. If this proposal makes sense, then it should be demonstrated that logical meaning is not only connected to pragmatics, but also plays a role in computing pragmatic meaning. This section aims to give such an argument, and will be based on Robyn Carston’s work on negation (Carston 1996, 2002; see also Moeschler 2010 for a more general presentation). There are two main uses of negation (Horn 1985, 1989): descriptive and metalinguistic. Descriptive uses are truth-conditional and take narrow scope, whereas metalinguistic ones are non-truth-conditional and take wide scope. Examples (30) and (31) are typical examples of descriptive (downward) and metalinguistic (upward) readings: (30) a. Abi is not beautiful (she is unsightly). b. Abi is not beautiful; she is gorgeous. The reason (30b) takes wide scope is because the corrective clause (COR) entails the positive one (POS), whereas in descriptive negation (with narrow scope), COR entails the negative clause (NEG), as in (31) (Moeschler 2013b for a detailed analysis): (31) a. Descriptive negation: COR → NEG b. Metalinguistic negation: COR →POS In descriptive negation, as COR entails NEG, negation scopes over a specific constituent of the sentence. In (30a), the contrast is between beautiful and unsightly, which are antonyms. In antonymy, a positive term entails the negation of the other, so unsightly semantically entails not beautiful. In metalinguistic negation by contrast, with its upward meaning, COR predicate is higher in the same semantic scale, and entails the predicate of POS: if Abi is gorgeous, then she is beautiful. This raises a very awkward question: can negation be bi-directional, going downward in its ordinary use (30a) and upward, in more specific context, as (30b), that is, when the corrective clause entails the positive one? This question presupposes that negation has the same meaning both in (30a) and in (30b), but this is not the case, as (31) shows. Moreover, what is strange in (30b) and its description in (31b) is that NEG should be compatible with its positive counterpart POS in case COR entails it. Fortunately, there is an answer to this paradox: negation, in (30b), cannot target POS predicate, as in descriptive negation, since the utterance would be contradictory: the speaker would assert that Abi is gorgeous, entail that she is beautiful, and deny, because of negation, that she is beautiful, leading to a contradiction. The only way of understanding negation is thus a wide scope reading: negation does not target the proposition (32a), but its scalar implicature (32b): (32) a. Abi is beautiful. b. Abi is not more than beautiful.

252

Formal and natural languages

So, by uttering (30b), the speaker is no more contradictory: she just refuses to assert a proposition that is not strong enough, which (33) makes explicit: (33) I cannot affirm that Abi is beautiful, because she is gorgeous. So the pragmatic meaning can be translated in (34): (34) not(Abi is not more than beautiful) ∧ (Abi is gorgeous) A last issue is to understand how (34) is obtained.The standard view requires pragmatic machinery only if semantic entailments lead to contradiction (e.g. Burton-Roberts 1989 and Carston 2002 for a discussion). In other words, negation is taking narrow scope by default, and will receive a more specific scope (implicature) to avoid contradiction: (35) Semantics: x is not-P Pragmatics: not (x is not more than P) The same analysis has been given for presupposition, showing that the first reading must be cancelled because of the corrective clause, which functions as an explanation: (36) a. The king of France is not bald; there is no king of France. b. I haven’t given up smoking; I’ve never smoked in my life. However, there is a good argument to question this analysis: metalinguistic negation can be uttered in a second position, which implies that there is no revision of a first reading (Carston 2002: 310). Moreover, it is more suitable to reserve a reparatory reading in case such a second treatment is intended, as in humor, for instance. (37) stands for the metalinguistic negation second position, and (38) as an example of humoristic exploitation of metalinguistic negation: (37) a. There is no king of France, so the king of France is not bald. b. I have never smoked, so I have not stopped smoking. (38) Marguerite Duras, qui n’a pas écrit que des conneries. Elle en a aussi filmé. (Pierre Déproges, Textes de scènes, Paris, Seuil, 71) ‘Margerite Duras, who has not written only shit. She also has filmed some.’ The alternative proposal is thus to reverse the order of scope computation. At a semantic level, negation takes wide scope (negation is propositional, that is, underspecified), and takes narrow scope only at a second stage. In that perspective, when negation is descriptive, negation scopes over the predicate (39), whereas when it is metalinguistic, it scopes over the proposition and all its presuppositions or only over its implicature (40): (39) (40)

a. The king of France is NOT-bald. b. Marguerite Duras has written NOT-ONLY shit. a. NOT [‘the king of France is bald’], because there is no king of France. b. NOT [‘Marguerite Duras has written shit’], because she also has filmed some.

253

Jacques Moeschler

When no corrective sentence is linguistically realised, negation is attracted by the predicate, for reasons of relevance and manner (Grice’s maxims of relation ‘Be relevant’ and brevity ‘Be brief ’). When COR is present and entails NEG, narrow scope reading is computed; however, when COR entails POS, negation does not move in the logical form and triggers a metalinguistic reading, by negating the targeted semantic and pragmatic materials. So the two steps in analysis are the following for example (30): (41) Semantics: not (x is P) Pragmatics: descriptive negation x is not-P metalinguistic negation not (x is not more than P) The advantages of this pragmatic derivation are clear: (i) negation is no more ambiguous, since it is semantically a propositional operator; (ii) there is no mismatch between the semantics meaning of negation and its logical meaning; (iii) metalinguistic uses are not necessarily given by a second processing, occurring only when a contradiction arises between NEG and COR. In other words, the metalinguistic issue shows that contrary to appearance, negation has as its semantics a simple propositional meaning, taking wide scope, which will be adjusted at the pragmatic level depending on the context in which the negative utterance has occurred (cf. Moeschler 2013b for a description of such contexts).

Conclusion In this chapter, I raised the issue of the relation between logic and NL. I showed that it could be argued that a sound relation exists between logic and NL if logic is understood in a minimalist sense: providing a semantics for logical words, which needs to be enriched by pragmatic processes. This conclusion has to be understood not only from a formal point of view. It has some consequence in related domains, mainly the language cognition interface (Anderson et al. 2013 for a general survey). This relation is now crucial for any type of research on language, whatever the theoretical presupposition and background are. But it has also some consequence in related topics, mainly the acquisition, not of linguistic competence, but of pragmatics (Zufferey 2015). Studies in experimental pragmatics, mainly initiated by Ira Noveck (Noveck 2001; Noveck and Sperber 2007; Noveck and Reboul 2010 for a synthesis), give empirical and experimental arguments in favor of this hypothesis. So it is now clearly demonstrated that pragmatic meanings of logical words – as scalar implicatures of some, exclusive meaning of or – are acquired later by children, who tend to behave more logically than adults. In a nutshell, logical meaning is not only grounded in pragmatic meaning because semantic and pragmatic meaning relations are based on logical properties of lexical items, but also because those linguistic expressions which have a logical semantics play a crucial role in the development of cognition by children and in the acquisition of pragmatics.

Suggestions for further reading Bach, E. (1989) Informal Lectures on Formal Semantics. New York: State University of New York Press. An elegant and comprehensible introduction to formal philosophy and semantics. Written for a Chinese audience, the book makes explicit a lot of assumptions that usual textbooks do not. Cann, R., Kempson, R., and Gregoromichelaki, E. (2009) Semantics: An Introduction to Meaning in Language. Cambridge: Cambridge University Press. A recent introduction to formal approaches to meaning, with a variety of topics, including logical inference, quantification, anaphora and discourse, time and events, word meaning, and ellipsis.

254

Formal and natural languages Grice, H. P. (1989) Studies in the Way of Words. Cambridge, MA: Harvard University Press. The collection of Grice’s papers include not only his seminal article ‘Logic and conversation’, but also ‘Meaning’ and ‘Utterer’s meaning and intention’, among others. A necessary reading for the comprehension of the current development of pragmatics. Gazdar, G. (1979) Pragmatics: Implicature, Presupposition, and Logical Form. New York: Academic Press. Chapter 4, ‘Logical functors’, is a precise argument in favour of the formalist analysis. This reading is a very good introduction to the formalist view of pragmatics. Geurts, B. (2010) Quantitative Implicatures. Cambridge: Cambridge University Press. An in-depth investigation to scalar implicatures, from a Gricean perspective, with some extensions to psychological plausibility, free choices, and embedded implicatures.

Note 1 This chapter has been written under the Swiss National Scientific Foundation research project LogPrag (Semantics and Pragmatics of Logical Words, project n° 100012_146093). Many thanks to Joanna Blochowiak and Karoliina Lohiniva for their help, comments, and suggestions.

References Allwood, J., Andersson, L. G., and Dahl, Ö. (1977) Logic in Linguistics. Cambridge: Cambridge University Press. Anderson, S. R., Moeschler, J., and Reboul, F. (eds.) (2013) The Language-Cognition Interface: Actes du 19e Congrès International des Linguistes. Geneva: Librairie Droz. Anscombre, J. C. and Ducrot, O. (1983) L’argumentation dans la langue. Brussels: Mardaga. Ariel, M. (2013) ‘Or: questioning meaning and use’, oral presentation at Workshop Semantics and Pragmatics of Logical Words, 19th International Congress of Linguists, Geneva. Austin, J. L. (1962) How to Do Things with Words. Cambridge: Cambridge University Press. Barwise, J. and Cooper, R. (1981) ‘Generalized quantifiers and natural language’, Linguistics and Philosophy, 4: 159–219. Burton-Roberts, N. (1989) The Limits to Debate. Cambridge: Cambridge University Press. Carston, R. (1996) ‘Metalinguistic negation and echoic use’, Journal of Pragmatics, 25: 309–330. Carston, R. (2002) Thoughts and Utterances:The Pragmatics of Explicit Communication. Oxford: Blackwell. Chierchia, G. and McConnel-Ginet, S. (1990) Meaning and Grammar: An Introduction to Semantics. Cambridge, MA: MIT Press. Dowty, D. R., Wall, R. E., and Peters, S. (1981) Introduction to Montague Semantics. Dordrecht: Reidel. Ducrot, O. (1980) Les Échelles Argumentatives. Paris: Minuit. Ducrot, O. (1989) Logique, Structure, Énonciation. Paris: Minuit. Gazdar, G. (1979) Pragmatics: Implicature, Presupposition, and Logical Form. New York: Academic Press. Gazdar, G. and Melish, C. (1989) Natural Language Processing in PROLOG: An Introduction to Computational Linguistics. Wokingham: Addison-Wesley. Geis, M. and Zwicky, A. (1971) ‘On invited inferences’, Linguistic Inquiry, 2: 561–566. Geurts, B. (2010) Quantitative Implicatures. Cambridge: Cambridge University Press. Grice, H. P. (1975) ‘Logic and conversation’, in P. Cole and J. L. Morgan (eds.) Syntax and Semantics: Speech Acts. New York: Academic Press. 41–58. Horn, L. R. (1972) On the Semantic Properties of Logical Operators in English. Bloomington: Indiana University Linguistic Club. Horn, L. R. (1985) Metalinguistic negation and pragmatic ambiguity’, Language, 61(1): 121–174. Horn, L. R. (1989) A Natural History of Negation. Chicago: University of Chicago Press. Horn, L. R. (2004) ‘Implicature’, in L. R. Horn and G. Ward (eds.) The Handbook of Pragmatics. Oxford: Blackwell. 3–28. Horn, L. R. (2012) ‘Histoire d’O: Lexical pragmatics and the geometry of opposition’, in J.Y. Béziau and G. Payette (eds.) New Perspectives on the Square of Opposition. Bern: Peter Lang. 383–346. Levinson, S. C. (1983) Pragmatics. Cambridge: Cambridge University Press. Levinson, S. C. (2000) Presumptive Meanings.The Theory of Generalized Conversational Implicature. Cambridge, MA: MIT Press.

255

Jacques Moeschler Mauri, C. and Van der Auwera, J. (2012) ‘Connectives’, in K. M. Jaszczolt and K. Allan (eds.) Cambridge Handbook of Pragmatics. Cambridge: Cambridge University Press. 377–401. McCawley, J. (1981) Everything That Linguists Have Always Wanted to Know about Logic*. . . . . . *But Were Ashamed to Ask. Chicago: University of Chicago Press. Moeschler, J. (2000) ‘Representing events in language and discourse’, in M. Coene, W. de Mulder, P. Dendale, and Y. D’Hulst (eds.) Triani augusti vestigia pressa sequamur. Studia linguistica in honorem Liliane Tasmowski. Padova: Unipress. 461–479. Moeschler, J. (2007a) ‘Introduction to semantics’, in M. Rajman (ed.) Language and Speech Engineering. Lausanne: EPFL Press. 27–49. Moeschler, J. (2007b) ‘Why are there no negative particulars? Horn’s conjecture revisited’, Generative Grammar in Geneva, 5: 1–13. Moeschler, J. (2010) ‘Negation, scope and the descriptive/metalinguistic distinction’, Generative Grammar in Geneva, 6: 29–48. Moeschler, J. (2012) ‘Conventional and conversational implicatures’, in H. J. Schmid (ed.) Cognitive Pragmatics. Berlin: Mouton de Gruyter. 407–434. Moeschler, J. (2013a) ‘Is a speaker-based pragmatics possible? Or how can a hearer infer a speaker’s commitment?’, Journal of Pragmatics, 43: 84–97. Moeschler, J. (2013b) ‘How “logical” are logical words? Negation and its descriptive vs. metalinguistic uses’, in M. Taboada and R. Trnavac (eds.) Nonveridicality, Evaluation and Coherence Relations. Leiden: Brill. 76–110. Moeschler, J. (2017) ‘Back to negative particulars. A truth-conditional pragmatic account’, in S. Assimakopoulos (ed.) Pragmatics and its Interfaces. Berlin: Mouton de Gruyter. 7–32. Moeschler, J., Zufferey, S. and Reboul, A. (forthcoming) Implicature. Cambridge: Cambridge University Press. Noveck, I. (2001) ‘When children are more logical than adults: Investigations of scalar implicature’, Cognition, 78(2): 165–188. Noveck, I. and Reboul, A. (2010) ‘Experimental pragmatics: A Gricean turn in the study of language’, Trends in Cognitive Sciences, 12(11): 425–431. Noveck, I. and Sperber, D. (2007) ‘The why and how of experimental pragmatics:The case of “scalar implicatures” ’, in N. Burton-Roberts (ed.) Pragmatics. London: Palgrave. 184–212. Reboul, A. (2007) Langage et Cognition Humain. Grenoble: PUG. Reboul, A. (2013) ‘The social evolution of language and the necessity of implicit communication’, in S. R. Anderson, J. Moeschler, and F. Reboul (eds.) The Language-Cognition Interface. Geneva: Librairie Droz. 253–273. Reboul, A. (2017) Communication and Cognition in the Evolution of Language. Oxford: Oxford University Press. Searle, J. R. (1969) Speech Acts. Cambridge: Cambridge University Press. Searle, J. R. (1979) Expression and Meaning. Cambridge: Cambridge University Press. Searle, J. R. and Vanderveken, D. (1985) Foundation of Illocutionary Logic. Cambridge: Cambridge University Press. Sperber, D. and Wilson, D. (1986) Relevance: Communication and Cognition. Oxford: Blackwell. Strawson, P. F. (1952) Introduction to Logical Theory. London: Methuen. Wilson, D. and Sperber, D. (2012) Meaning and Relevance. Cambridge: Cambridge University Press. Zufferey, S. (2015) Acquiring Pragmatics. London: Routledge.

256

21 PRESUPPOSITION AND ACCOMMODATION Jacopo Romoli and Uli Sauerland

Introduction A basic discovery of research in semantics, pragmatics and their interface is that the meaning of an utterance is not always a single, unified whole, but can be divided into different components of meaning. This chapter addresses one such division: the division between presupposed and asserted content. Whether this division aligns with that of semantics and pragmatics is a matter of controversy, with some theories pulling presuppositions more towards the semantic side while others consider them more as part of the pragmatic realm.1 All theories of presuppositions, however, involve some combination of semantic and pragmatic elements in their account, making the presupposed/asserted content distinction one of the most emblematic phenomena of the semantic/pragmatic interface. The divide between presupposed and asserted aspects of meaning, originally observed by the German philosopher Gottlob Frege (Frege 1892), plays a role in the interpretation of an utterance in two ways. First, presupposed and asserted content can interact differently with other utterances in a discourse. Second, the presupposed and asserted content of parts of a complex utterance can also interact differently with other parts in the compositional build-up of utterance meaning. The division between these two types of content is, however, not as rigid as we would expect. In particular, in some cases, it is possible for presupposed content to acquire properties of asserted content. For these cases, an ‘accommodation’ operation has been hypothesised, and this is the second topic of this chapter. Evidence for accommodation as well as for the division between presupposed and asserted content comes both from the discourse properties of complete utterances, and their compositional interpretation of complex utterances. In the remainder of this section, we introduce data from the discourse level. In the second section, we turn to evidence from complex utterances, and at the same time introduce the framework of Dynamic Semantics which arguably represents the most influential account of these data. In the final section, we discuss some important challenges to the Dynamic Semantics framework. As these later developments show, the research on presuppositions and accommodation remains one of the most debated topics in the semantics/pragmatics literature. Moreover, recently presuppositions have been investigated more and more with psycholinguistic methods, also in comparison to other semantic/pragmatic inferences like implicatures. This experimental

257

Jacopo Romoli and Uli Sauerland

work has brought new important data into the theoretical debate and contributed even more to make this an exciting time to study these phenomena. Consider a first piece of evidence for the division of presupposed and asserted content at the utterance level. As participants of conversations, we draw a variety of inferences from utterances. Imagine, for instance, a situation in which you do not know that you have a sister; imagine further that you are told the sentence in (1a) or that in (1b). It is clear that in both cases you would equally conclude that you have a sister. (1) a. You have a sister and she is awesome. b. Your sister is awesome. We also have the intuition, however, that this same piece of information is conveyed differently in (1a) versus (1b). For instance, it appears to be a natural and legitimate conversational move to object to (1b) using (2). However, (the first part of) (2) is intuitively an unnatural response to (1a), as observed by Shannon (1976) and Von Fintel (2008) among others. That is, the type of discourse move in (2) appears appropriate only in response to (1b). (2) Hey wait a minute! I didn’t know I have a sister! This difference tells us that the same inference − that the hearer has a sister − can have a different discourse status if drawn from a sentence like (1a) versus one like (1b). And this difference is one of the arguments for treating this inference as having different natures in (1a) versus (1b): if we can felicitously use a discourse move like (2), the relevant inference is considered a presupposition (but see Chemla 2008 and Romoli 2014 among others). To illustrate this difference further, consider a sentence like (3a): from (3a) we typically derive both the inference in (3b) and that in (3c). Again, however, the pieces of information in (3b) and (3c) appear to be associated with the utterance in (3a) in different ways. (3) a. Jack stopped coming to class. b. Jack doesn’t come now. c. Jack used to come. (3c) represents presupposed content and (3b) does not. The inference in (3b) is simply an entailment of (3a) as shown by the fact that a sentence like ‘Jack stopped coming to class and he comes to class now’ is contradictory (which in turn shows that when (3a) is true, then (3b) must also be). One piece of evidence in this respect is that, in the same way as earlier, (4a) is a felicitous response to (3a), but (4b) is not. As is standard in the literature, we will indicate unnaturalness/ oddness with the diacritic ‘#’. (4) a. Hey wait a minute! I didn’t know that he used to come. b. #Hey wait a minute! I didn’t know that he doesn’t come now. A second test can also be applied. This is based on explicitly asserting (3b) or (3c) before (4a): in the case of (3c) the result is felicitous, (5a), but that of (3b) is unnatural, (5b). (5) a. Jack used to come to class and stopped coming. b. #Jack doesn’t come to class now and stopped coming.

258

Presupposition and accommodation

This again shows us that the inferences in (3b) and (3c) have different discourse status, despite looking similar on the surface. Information arising as a presupposition can be asserted before its corresponding presuppositional sentence, while that arising as entailment cannot. In other words, presuppositions, but not entailments, can be ‘old information’ in this sense.2 As just seen, the presence of words like your and stop in a sentence appears to trigger a presupposition. Beyond these two cases, presuppositions arise from a variety of other ‘triggers’. One case discussed from very early on is that of definite descriptions: the observation is that a sentence like (6a) presupposes something along the lines of (6b) (Strawson 1950 and much subsequent work). (6) a. The current president of Harvard is an historian. b. Harvard has currently a president. Other triggers include words like win, also, again or bound morphemes like that associated with number, gender, tense and syntactic constructions like it-clefts. Statements (7) through (10) represent a very partial set of presuppositional sentences along with their presuppositions. (7) (8) (9) (10)

a. Jack won the Boston Marathon. b. Jack participated in the Boston Marathon. a. Yesterday, Jack showed up late again. b. Jack showed up late before yesterday. a. It is Jack who broke the vase. b. Somebody broke the vase. a. The person over there introduced herself to the president. b. The person over there is female.

As seen earlier, presupposed content can represent old information that is already known to the discourse participants. This observation is the starting point for the traditional approach to presuppositions. Specifically, this approach conceives presuppositions as not only allowed to be old information, in the sense given previously, but as required to be so. In particular, the idea is that a presuppositional sentence like (6a) is only felicitous in contexts in which it is already assumed to be believed by the conversation participants that there is a president of Harvard. This approach, which stems from the work of Stalnaker (1973, 1974, 1978), Karttunen (1973, 1974) and Heim (1982, 1983), characterises presuppositions as constraining the class of contexts to which the presupposing sentence can be uttered felicitously. In other words, the presupposition of a sentence S must be ‘satisfied’ in a context for S to be felicitously asserted in that context. For this reason, this approach is sometimes called ‘the satisfaction theory of presuppositions’.3 To be more precise about how presuppositions are conceived in the satisfaction approach, we need to be more precise about how to model contexts. Following Stalnaker (1973, 1974, 1978, 2002) (see also Stalnaker 2014), we can think of a context as determining, among other things, a common ground, a set of propositions that are believed to be believed by the conversation participants (if only for the sake of the conversations). A common ground is, in turn, associated with a set of worlds: the worlds compatible with each of the propositions in the common ground. In Stalnaker’s terminology, this set of worlds is the ‘context set’. For our purposes, we can identify a context with the context set it determines, so we will simply use the term ‘context’ in this

259

Jacopo Romoli and Uli Sauerland

way from now on. We can now go back to the idea discussed earlier and define presuppositions as in (11). (11) A sentence S presupposes a proposition p if for any context c in which S can be uttered felicitously, c entails p. So for instance, (6a) presupposes that there is a president of Harvard. In the satisfaction approach, this becomes a condition on the assertability of a sentence like (6a): (6a) is only felicitous in a context in which it is known (to be commonly known) by the participants of the conversation that there is a president of Harvard. As you can imagine, one immediate question for this traditional approach is what happens if a sentence like (6a) is uttered in a context in which it is not taken for granted that Harvard has a president. Indeed, we have the intuition that at least in some cases we could utter (6a) in a context in which this information is not already commonly known. How can a case like this be reconciled with the satisfaction approach? The standard response, going back to Lewis (1979), is that this information, if not particularly controversial or relevant for the conversation, could just be accommodated in the context. Accommodation is a central component of the satisfaction approach to presuppositions, and it has been conceived and implemented in different ways. For illustration here, we make use of an implementation based on presupposition theories like Krahmer (1998). Essentially, the idea is that accommodation is an operation, which we will call ACC, that can apply to utterances or parts of utterances and transforms the presupposed content into asserted content (Krahmer 1998; Beaver and Krahmer 2001; Fox 2008, 2012; George 2008, among others). According to this implementation, a sentence like (6a) is ambiguous: it could be analysed as in (12a) and require the context to entail its presuppositions, but it could also be analysed as in (12b), where ACC transforms the presuppositions of its arguments into entailments. When analysed as (12b), (6a) could be paraphrased as: ‘There is a current president of Harvard and the current president of Harvard is an historian.’4 (12) a. The current president of Harvard is an historian. b. ACC[The current president of Harvard is an historian.] In sum, we saw that certain inferences that we draw from utterances appear to have a characteristic discourse status; these inferences constitute the presupposed content of these utterances and are distinguished from inferences which arise from the asserted content. We also saw that when we conceive of presuppositions as requirements on the context of utterance, like in the satisfaction approach, we need to assume a process of accommodation, which allows presuppositions to be accommodated in case they happen not to be old information in the context. We sketched a possible implementation of the accommodation process as insertion of an accommodation operator, which we called ACC. In the next section, we turn to the other argument for the division between presupposed and asserted content, which has to do with the compositional interaction of presuppositions in complex sentences.

Projection and dynamic semantics Projecting, suspending and filtering In the previous section, we considered the distinction between presupposed and asserted content at the utterance level. Now, we address the fact that presupposed and asserted content also

260

Presupposition and accommodation

exhibit distinct behaviour in complex sentences. Recall that presuppositions are generally tied to the presence of a specific lexical trigger, for example, the verb stop in (13) (repeated from (3a)). The data we turn to in this section concern sentences where such lexical triggers occur embedded in the scope of semantic operators. We will now introduce three distinct outcomes from such a case, following and adapting traditional terminology: the presupposition can (1) Project, that is, remain unaffected; (2) Be filtered, that is, be modified; or (3) Be suspended, that is, become part of the asserted content. To illustrate first a case of projection, consider the sentence in (13) and its inferences in (14a) and (14b). (13) Jack stopped coming to class. (14) a. Jack doesn’t come to class. (asserted content) b. Jack used to come to class. (presupposed content) The observation is as follows: when we embed (13) under negation, (15a), the antecedent of a conditional, (15b), a question, (15c) or a possibility modal (15d), the inference in (14a) does not survive, while that in (14b) does. In other words, while (14a) is only an inference of (13), (14b) is an inference of both (13) and all of (15a) to (15d). In the traditional metaphor, (14b) projects out when (13) is embedded in complex sentences like (15a) to (15d), while (14a) does not. (15)

a. Jack didn’t stop coming to class. b. If Jack stopped coming to class, Mary won’t be happy. c. Did Jack stop coming to class? d. It’s possible that Jack stopped coming to class.

Explaining the behaviour of presuppositions in complex sentences is also called the ‘projection problem’ of presuppositions.The projection problem is one of the main problems in the presupposition literature.The task is complicated by the fact that presuppositions do not always project out of embeddings. As mentioned earlier, there are in particular two cases in which projection does not happen, and we illustrate them in turn.The first type of non-projection, suspension, can be seen with cases like (16). What (16) shows is that the presupposition that Jack used to come to class does not project here, or it would be in contradiction with the continuation. In other words, (16) does not mean what the contradictory (17) means. (16)   Jack didn’t stop coming to class, because he never came! (17)  Jack used to come to class and didn’t stop, because he never came! The second type of non-projection, filtering, is illustrated by the cases in (18) and (19). While (18) is another case of projection out of embeddings (i.e., the presupposition of the consequent, that Mary has a brother, appears to project and become an inference of the conditional in (18)), this is not the case for (19) (see section 3.2 for issues arising from presupposition projection from the consequent of conditional sentences). In other words, while you would conclude that Mary has a brother upon hearing (18), you certainly would not infer that from (19). (18)  If Mary comes to visit, she will stay with her brother. (19)  If Mary has a brother, she will stay with her brother.

261

Jacopo Romoli and Uli Sauerland

Analogous patterns can be reproduced with other connectives like conjunction. Consider for instance (20a) to (20b): while we conclude that Mary has a brother from both (20a) and (20b), only in the case in (20a) this inference is a presupposition. (20) a. Mary will come to visit and she will stay with her brother. b. Mary has a brother and she will stay with her brother. To illustrate the difference between (20a) and (20b), consider the contrast in (21a) versus (21b): in line with our observations about the discourse status of presuppositions and entailments, we can see that the inference that Mary has a brother behaves as a presupposition in (21a), but not in (21b).This is because this inference is allowed to be old information in (21a), but not in (21b). (21) a. Mary has a brother and she will come to visit and she will stay with her brother. b. #Mary has a brother and she has a brother and she will stay with her brother. In sum, the question for presupposition theories is how to account for the cases in which presuppositions do not project, that is for cases in which they are suspended or filtered.

Accounting for projection and non-projection Suspension as local accommodation The concepts of accommodation and non-projection are related. More specifically, cases of what we called suspension could be thought of as the process of accommodation happening locally, when presupposition triggers are embedded in the scope of some semantic operators. As we saw, a sentence like (15a), repeated in (22), has a reading in which the presupposition appears suspended, as in (16), repeated in (23). (22) Jack didn’t stop coming to class. (23) Jack didn’t stop coming to class because he never came! In the implementation of accommodation presented earlier, complex sentences containing a non-projecting presuppositional trigger can be analysed with an ACC operator in an embedded position. So for instance a sentence like (22) could be associated with either of the three structures in (25). (24) a. not[Jack stopped coming to class] b. ACC[not[Jack stopped coming to class]] (global accommodation) c. not[ACC[Jack stopped coming to class]] (local accommodation) Whilst (24a) involves simple projection of the presupposition that Jack used to come to class, the different scopes of ACC in (24b) and (24c) represent different ways in which the accommodation process can happen. The first case in (24b) is a case of what we can call ‘global’ accommodation: since ACC applies after negation in (24b), negation doesn’t apply to the presupposition that Jack used to come class. Therefore the unnegated presupposition becomes part of the asserted content as in the paraphrase (25). (25) Jack used to come to class and didn’t stop coming to class. 262

Presupposition and accommodation

On the other hand, in (24c), accommodation happens ‘locally’. As a consequence, negation applies after ACC, therefore the presupposition that Jack used to come to class is part of the asserted content that negation applies to. The interpretation of representation (24c) is therefore one that is compatible with Jack never having come to class as paraphrased in (26). This interpretation captures presupposition suspension and can be brought out if (15a) is followed by ‘. . . . he never came!’ as in (23). In this approach, the problem of presupposition suspension becomes the problem of where the ACC operator should be merged. (26) It’s not true that Jack used to come to class and stopped. However, this approach does not straightforwardly extend to the other case of non-projection, what we called filtering.To illustrate, let us go back to the type of cases in (27a) vs. (27b), repeated from previously: where (27a) involves projection of the presupposition of the consequent, but (27b) does not. (27) a. If Mary comes to visit, she will stay with her brother. b. If Mary has a brother, she will stay with her brother. To account for the contrast in (27) in terms of accommodation, one could stipulate that (27a) should be analysed as (28a), involving global accommodation, and therefore be equivalent to (28b). Analogously, one could analyse (27b) as in (29a), involving local accommodation this time, so that it would mean roughly (29b) and would correctly not give rise to the inference that Mary has a brother. (28) (29)

a. Mary has a brother and if she comes to visit, she will stay with him. b. ACC[If Mary comes to visit, she will stay with her brother] a. If Mary has a brother, she has a brother and will stay with him. b. If Mary has a brother, ACC[she will stay with her brother]

The question is, of course, why not the other way round.That is, the challenge for this approach is to tell us what prevents one to analyse (27b) as (30a), incorrectly predicting the meaning in (30b) and the inference that Mary has a brother. (30) a. ACC[If Mary has a brother, she will stay with her brother] b. Mary has a brother and if she has a brother, she will stay with him. This question has been explored in various ways, and for the time being we know of no fully satisfactory answer (see Gazdar 1979; Van der Sandt 1992; Geurts 1996, 1999; Heim 1990; Beaver 2010; among others). A second problem for an attempt to explain presupposition filtering by means of accommodation is that the putatively accommodated content retains properties of old information. For instance, one question for this approach would be why (29a) is felicitous contrary to (31), which is its corresponding meaning after ACC has transformed the presupposed into asserted content. (31) #If Mary has a brother, she has a brother and will stay with her brother. In sum, the idea of accommodating at different levels can account for the presupposition suspension cases, but not the cases of filtering, or at least not in a straightforward way. 263

Jacopo Romoli and Uli Sauerland

In the next section, we discuss how filtering and projection are accounted for in Dynamic Semantics, a particularly successful theory, which is part of the satisfaction approach described earlier. Finally, we discuss some of the problems that have been raised for Dynamic Semantics at the beginning and in more recent years. Notice that, with respect to the initial Stalnakerian pragmatic characterisation of presuppositions, Dynamic Semantics pulls presupposition more into the semantics side of the semantics/pragmatics interface – a move which remains controversial (for criticism and discussion, see Schlenker 2008a, 2009; Chierchia 2009; Rothschild 2011; Stalnaker 2014, among others).

Updating the context compositionally Dynamic Semantics was built on the idea that the context of a conversation does not remain immutable, but changes dynamically as the discourse enfolds. In particular, any utterance, if accepted by the conversation participants, narrows down the set of possibilities that are open in the context. More precisely, following Stalnaker, we can model the way an utterance modifies a context with set intersection: for any context c and sentence S, the (felicitous) utterance of S in c produces a context c’ which is equivalent to the intersection of c with the meaning associated to S (i.e. the set of worlds in which S is true). Dynamic semantics captures directly the notion of old information that we introduced earlier: after the successful utterance of a sentence S, any information S contributed is part of the new context c’. In a sequence of sentences such as (32), then, the information contributed by the first counts as old information for the second sentence, and thereby satisfies the presupposition. (32) John used to come to class. Now he stopped. Dynamic Semantics extends the idea that a prior sentence provides a context for the evaluation of a later sentence to sentence internal interpretation. For example, it seeks to develop an account of the two examples in (33) where just as in (32) the presupposition of stop is satisfied by the initial clause. (33) a. John used to come to class and now he stopped. b. If John used to come to class, he stopped now. Developing this intuition, Heim (1982, 1983) proposed to identify the meaning of sentences with the way they can change a context – that is, their context change potentials. Using this approach, she provided an account of when presupposition projects and when it is filtered. In the following, we illustrate the idea for the cases of conjunction and negation, but generalisations of the approach to several other connectives and operators are discussed by Heim (1992), Beaver (2001) and others. First, Dynamic Semantics assumes that an embedded presuppositional sentence has no truthvalue if uttered in a context that doesn’t satisfy its presuppositions. Following, for example, Heim and Kratzer (1998), we model presuppositions as partial functions within the framework of model-theoretic semantics. Partial functions are neither true nor false in certain contexts, but undefined in the contexts where a presupposition is not satisfied. So for instance, the meaning of a presuppositional sentence like (35a), repeated from previously, is a partial function from contexts which satisfy (35a)’s presupposition (35b) to contexts in which it is true that Jack stopped coming to class. Therefore, the result of updating a context 264

Presupposition and accommodation

c with (35a) is shrinking c to the subset of c-worlds in which it is true that Jack stopped coming to class, provided that c entailed that Jack used to come to class in the first place. (35) a. Jack stopped coming to class. b. Jack used to come to class. This basic condition in (35) captures the idea of presupposition as conditions on contexts by Stalnaker. So far we have only looked at atomic sentences; when we move to complex sentences, we need an account of how the context change potentials of complex sentences are determined from those of their parts. The initial idea in Dynamic Semantics is that once we define, for each connective and operator, how they contribute to the composition of context change potentials, this will automatically derive when presuppositions project and when they do not. In other words, by defining how the assertion parts of sentences compose, we would derive how the presupposition parts do.We will come back to this claim in the section on Asymmetry of and. For now let’s consider how, in the case of negation and conjunctions, this is accomplished. Consider negation first. The effect of updating a sentence of the form ‘not S’ to a context c is to subtract from c the worlds in which S is true. It is the way in which this is obtained, however, that is crucial for presuppositions: first we interset the context c with S and then we subtract the worlds in the intersection from c. In other worlds, the operation in set theoretic terms is the following: c − (c ∩ S). What is crucial here is that the update of the context with not S contains in its specification the update of the context with S, in symbols: |S|(c).The latter, in turn, will only be possible if the context presupposes all of S’s presuppositions. This last bit, therefore, correctly predicts that not S will inherit all of S’s presuppositions. The task for Dynamic Semantics is then to define the context change potential of not S in a way that adequately reflects our intuitions about its truth-conditions and presuppositions: (36) |not Sp|(c) = c − |Sp|(c) (36) correctly predicts that (37a) inherits the presuppositions of its positive counterpart, (37b): that Jack used to come to class. Again, this is because updating the context c with |not S| requires that we first update c with S. But as we know, this latter update requires S’s presuppositions to be satisfied. In this way, a negated sentence like (37a) inherits all the presuppositions of its positive counterpart. (37)  a. Jack didn’t stop coming to class.   b. Jack stopped coming to class. Consider now the case of conjunction and let us go back to cases like (33a), repeated from earlier. The relevant generalisation about (38) was that the presuppositions of the second conjunct are not projected to the whole conjunction if they are entailed by the information of the first conjunct. In other words, (38) doesn’t presuppose what its second conjunct presupposes, that John used to come to class, because this is entailed by the first conjunct. (38) John used to come to class and now he stopped. The task for Dynamic Semantics is to define a context change potential of conjunction which can account for the just mentioned generalisation. In order to do so, the following definition of the context change potential of a conjunctive sentence is standardly assumed: (39) |S1 and S2|(c) = |S2|(|S1|(c)) 265

Jacopo Romoli and Uli Sauerland

Definition (39) requires that in updating a context c with a conjunction S1 and S2 we first update c with S1 and then we update the resulting context with S2. This immediately predicts that while S1 is uttered in the context of c, S2 is (as if it is) uttered in a derived context, call it c’. In turn, (39) correctly predicts that the presupposition of S1 will have to be satisfied in c for S1 to be defined, but that of S2 only needs to be satisfied in c’. To illustrate these predictions with a concrete example, consider again the sentence in (38): according to (39), (38) can be uttered in any context, despite the fact that the second conjunct, ‘John stopped coming to class,’ is presuppositional. This is because any context c updated with the first conjunct, ‘John used to come to class,’ will entail the presupposition of the second conjunct (that John used to come to class). In other words, the presupposition of the second conjunct is correctly predicted to be filtered if the initial context updated with the first conjunct entails it. This brief reconstruction was just a sketch, and there are a variety of details that one can implement in different ways. Moreover, one has to define systematically the context change potentials of all connectives, modals and quantifiers in a way that links the projection properties of sentences and the way their assertion parts are composed. Generalised in this way, however, the dynamic approach has been one of the most successful approaches to account for the projection problem of presuppositions. Before going to some challenges to this approach, it is important to notice that while it provides an account of presupposition projection and presupposition filtering, it still requires an extra mechanism to account for the cases of presupposition suspension like (40), repeated from earlier. (40) John didn’t stop coming to class, because he never came! The mechanism adopted is an implementation of Lewis’ idea of accommodating presuppositions. However, rather than changing the sentence, like the ACC operator introduced earlier does, the idea is changing the context to which the sentence is updated. That is to say, the accommodation process acts on the context by modifying it so that it winds up entailing the relevant presupposition. Moreover, in parallel to how ACC can be merged at different levels in a sentence, this accommodation process can also apply at different levels. In particular, in the case of (40), a change is required not in the global context against which the entire sentence is uttered, but rather in a more local context, that of the sentential argument of negation. To illustrate, remember that the context change potential of negation for a sentence like not Sp was c − |Sp|(c). This local accommodation process is assumed to apply before the |Sp| is updated, so that the presupposition is satisfied in that local context. We can formulate this as in (41). (41) |not[John stopped coming to class]|(c) = c − |John stopped coming|(c) meaning of neg = c − |John stopped coming|(c ∩ λ w[[John used to come]]w) local acc = c − (c ∩ λw[[John used to come]]w ∩ λw[[John stopped coming]]w) More concretely, what happens is the sentence embedded in the scope of negation, John stopped coming to class, is updated to a context that is modified to entail the presupposition that John used to come. Negation then applies, excluding worlds in which John used to come and stopped, so resulting in a context in which John either never came to class or used to come and didn’t stop. This context is then compatible with the continuation that John never came to class. Through this process of (local) accommodation, Dynamic Semantics can then account for data like (40). See Von Fintel 2008 for a critical discussion of local accommodation in Dynamic Semantics. 266

Presupposition and accommodation

Extensions and challenges to Dynamic Semantics Though Dynamic Semantics has arguably been the most influential account of presuppositions over the past thirty years, a variety of problems have been pointed out with the approach. In this section we review some of them and we point to the responses proposed in the literature.

Asymmetry of ‘and’ One important problem for Dynamic Semantics regards the relationship between how in a complex sentence the assertion and presupposition component of its parts are composed – which is just another way of saying when presuppositions project and when they do not. In particular, the problem has to do with the claim reviewed previously that the projection of presuppositions can simply be derived by giving a truth-conditionally adequate formulation of how the assertion components of a sentence are put together compositionally. To illustrate the issue, let us go back to conjunction. Consider the sentences in (42a) and (42b), which presuppose that John has a sister and that she is in town, respectively. Now, consider the conjunction of (41a) and (41b) in (42). (42) a. John’s sister is in town. b. John discovered that she is in town. (43) John’s sister is in town and John discovered that she is in town. Recall that in defining the context change potential of (43) in terms of its parts in (42a) and (42b) we want to obtain two results. First, we want the net result to reflect our intuitions that a context updated with a conjunction of two sentences should be narrowed in such a way that any world in which either of the two sentences is not true should be discarded. In other words, we want the update of a context c with (43) to be just those worlds of c in which both (42a) and (42b) are true; that is, those worlds in which John’s sister is in town and that John has come to know this. Moreover, intuitively, (43) still presupposes that John has a sister (the presupposition of [42a]), but not that she is in town (the presupposition of [42b]). So we want our definition of the context change potential to reflect this intuition. The initial claim in Dynamic Semantics was that the presupposition projection properties of a complex sentence like (43) follow directly from defining a context change potential that is truth-conditionally adequate. For instance, the context change potential proposed for a conjunction of the form S1 and S2 is as in (44). (44) |S1 and S2| = |S2|(|S1(c)) As seen earlier, (44) requires that we first update the context c with the first conjunct and correctly predicts that the presuppositions of the first conjunct have to be satisfied in c. As a consequence, the presuppositions of the first conjunct are correctly predicted to project to the whole conjunction. On the other hand, the second conjunct does not update the initial context c, but it rather updates c intersected with the first conjunct, call it c’.Thus this predicts that this derived context c’, not c, has to satisfy the presuppositions of the second conjunct. As a consequence, if the presuppositions of the second conjunct are entailed by the first conjunct, we predict that they will always be satisfied in c’, no matter what the initial context c is. This explanation appears to elegantly link how truth-conditions are obtained compositionally and presupposition projection. As Soames (1989), Irene Heim herself (Heim 1990) and, 267

Jacopo Romoli and Uli Sauerland

more recently, Schlenker (2008a, 2009) have pointed out, however, when we look closely the link is not tight enough.This is because we can define a context change potential for a conjunction like (43) that is as truth-conditionally adequate as (44) (i.e. it narrows down the context in the same way), but it does not make the right prediction for presupposition projection. One example of this is (45): (45) is a simple variant of (44), but the two conjuncts are updated in a different order. The crucial point is that (45) still predicts that, in the end, the worlds of c that remain are only those compatible with both S1 and S2. This makes (45) truth-conditionally adequate. (45) |S1 and S2| = |S1|(|S2(c)) However, the predictions of presuppositions are not adequate and are as follows: c should presuppose what S1 and S2 presuppose unless S2 entails the presuppositions of S1. If these predictions were correct, a sentence like (46) should be natural and non-presuppositional, as the presupposition of the first conjunct should be filtered by the second one. (46) #John’s sister is in town and John has a sister. On the contrary, (46) is odd. Its symmetric version in (47), on the other hand, is natural and presuppositionless. (47) John has a sister and John’s sister is in town. This contrast is an argument for adopting (44) rather than (45). The problem is that they are truth-conditionally equivalent, therefore it is not true that presupposition projection is derived in Dynamic Semantics simply by adequately defining how the assertion component of sentences are to be put together compositionally. This explanatory challenge for Dynamic Semantics has sparked an intense debate in the literature in recent years and triggered a variety of responses, departing more or less radically from the dynamic approach (see Fox 2008, 2012; George 2008; Schlenker 2008a, 2009; Chemla 2010; Rothschild 2011; see also Schlenker 2008b for a summary of the debate). Within the dynamic approach, Rothschild (2011) and Schlenker (2009) have proposed a way of constraining context change potentials that does not suffer from the explanatory problem. We refer the reader to their papers for details, but in a nutshell the idea is that a sentence like (43) can be thought of as associated with both (44) and (45), but on top of this there is a general independent ordering constraint which favours the former over the latter. This constraint can be implemented in different ways, but usually it is conceived of as based on linear order. The idea is that a sentence is processed incrementally as it is heard, and it is this that creates an asymmetry between the conjunct coming first, the first conjunct, and the one coming last, the second conjunct (Schlenker 2008a, 2009).This, in turn, affects how presuppositions are filtered, in a more general and indirect way than in the dynamic approach, where the asymmetry is encoded directly in the meaning of connectives and quantifiers.

Conditional presuppositions In this section, we put aside the problem of asymmetry, and we focus on the prediction of the dynamic approach for cases in which the presupposition trigger appears in the consequent of a conditional. We saw that Dynamic Semantics predicts that the presuppositions of the second 268

Presupposition and accommodation

conjunct of a conjunction project unless they are entailed by the first conjunct. In other words, we could express the predicted presupposition of a conjunction of the form in (48) as the conditional presupposition: ‘if q, then p.’ (48) q and Sp As we saw, this correctly predicts that (49a), repeated from earlier, presupposes nothing. More precisely, it only presupposes the tautological (49b), hence it does not impose any restriction on the class of contexts in which it can be uttered. (49) a. Mary has a brother and she will come with her brother. b. If Mary has a brother, Mary has brother. On the other hand, the dynamic account predicts that (50a) presupposes (50b). But (50b) appears too weak as it stands: intuitively, we would want to infer from (50a) that Mary has a brother. The problem in this particular case, however, does not arise because the asserted conjunction in (50a) entails that Mary doesn’t want to come alone, therefore from the latter and the conditional presupposition in (50b) we can indeed conclude that Mary has a brother. The combination of an entailment of the assertion and the conditional presupposition could be the source of our intuition that (50b) gives rise to the inference that Mary has a brother. (50) a. Mary doesn’t want to come alone and she will come with her brother. b. If Mary doesn’t want to come alone, Mary has brother. When we turn to conditionals, however, the issue just sketched for conjunctions reemerges in a more problematic way. To illustrate, consider the corresponding conditionals of the two conjunctions in (51a) and (52a). One can show that the dynamic account predicts also in this case the conditional presuppositions in (51b) and (52b). (51) (52)

a. If Mary has a brother, she will come with her brother. b. If Mary has a brother, Mary has a brother. a. If Mary doesn’t want to come alone, she will come with her brother. b. If Mary doesn’t want to come alone, she has a brother.

While the prediction appears entirely correct for (51a), it is clearly wrong for (52b). Moreover, contrary from the case of conjunction, in the case of conditionals there is no entailment of the assertion that comes to the rescue. That is to say, (52a) does not entail that Mary doesn’t want to come alone. One can also construct a similar contrast involving a case in which the conditional presupposition appears either intuitively correct, as in (53a), or not, as in (54a) (Perez-Carballo 2008).5 (53) (54)

a. If Paul is a devout Catholic, he has a Bible. b. If Paul is a devout Catholic, he will read his Bible tonight. a. If Paul is tired tonight, he has a Bible. b. If Paul is tired tonight, he will read his Bible tonight.

We are left with the problem that the predicted projected presupposition from the consequent of a conditional sentence appears correct in some cases, (53a), but too weak in others, (54a).This 269

Jacopo Romoli and Uli Sauerland

challenge to the dynamic approach is usually referred to as ‘the proviso problem’ – a term introduced by Geurts (1996, 1999). The task the dynamic approach faces is to account for when a conditional sentence of the form ‘if q and Sp’ should give rise to the weak conditional inference q → p and when it can be associated to the stronger unconditional inference p. The main attempt in response to this problem is to derive the non-conditional presupposition on top of the conditional one, as a separate mechanism (see Karttunen and Peters 1979; Heim 2006; Singh 2007, 2008;Van Rooij 2007; Perez-Carballo 2008; Von Fintel 2008; Schlenker 2011, among others; see also Romoli et al. 2011 and Chemla and Schlenker 2011 for related experimental work on presupposition projection in conditional sentences).

Differences among triggers Theories of presuppositions like Dynamic Semantics traditionally treat all presuppositions uniformly. And indeed presuppositions do behave uniformly in most cases. For instance, consider the case of win. As seen previously, a sentence with win like (55a), its negation in (55b), its questioned version in (55c) and a conditional or a possibility modal embedding (55a) like (55d) and (55e), all give rise to the inference that Bill participated in the Boston Marathon. (55)

a. Bill won the Boston Marathon. b. Bill didn’t win the Boston Marathon. c. Did Bill win the Boston Marathon? d. It’s possible that Bill won the Boston Marathon. e. If Bill won the Boston Marathon, he will celebrate tonight.

If we consider any other presupposition trigger, we can reproduce the same projection behaviour. For instance, if we consider it-clefts, we can see that each of (56a) to (56e) gives rise to the inference that somebody broke the computer in the same way. (56) a. It is Mary who broke that computer. b. It isn’t Mary who broke that computer. c. Is it Mary who broke that computer? d. It’s possible that it is Mary who broke that computer. e. If it is Mary who broke that computer, she should repair it. In these examples, we observe that presuppositions behave uniformly in their projection behaviour through propositional connectives, modals, and questions as most approaches would lead us to expect. Despite their similarity, however, as discussed since Karttunen (1971) and Stalnaker (1974), there appear to be differences in the behaviour of certain triggers. One first example of a difference in projection between triggers was discussed by Heim (1991), Sauerland (2008) and Percus (2006), who showed that in a number of cases two members of a pair with complementary presuppositions exhibit differences in their projection behaviour. For instance, Sauerland (2003, 2008) discusses the difference between singular and plural marking of nouns. (57a)/(57b) show the presuppositions without projection: the singular variant presupposes that John has exactly one sister, while the indefinite variant presupposes that he has more than one. (57) a. John invited his sister. John has exactly one sister. 270

Presupposition and accommodation

b. John invited his sisters. John has more than one sister. (58a)/(58b) show the corresponding projection of the two triggers from the scope of a universal quantifier. Projection of the presupposition of the singular results in the universal presupposition that every student has exactly one sister. The plural, however, projects to a weaker, existential presupposition, namely that at least one student has more than one sister. (58) a. Every student invited his sister. Every student has exactly one sister. b. Every student invited his sisters. At least one student has more than one sister. Similar contrasts obtain for other pairs of presupposition triggers that yield complementary presuppositions when they are not embedded. A second example was discussed by Abusch (2002, 2010), who introduced a distinction between two classes of presupposition triggers that she refers to as ‘soft’ and ‘hard’ triggers. This distinction is based on two differences.The first difference has to do with the ease of suspension of the presupposition associated with a trigger. This can be brought about via creating a context in which the speaker is evidently ignorant about the presupposition; those triggers that do not give rise to infelicity in such contexts are soft triggers. Consider the following two examples modeled on Abusch (2010) that show that, according to this diagnostic, win and it-clefts are soft and hard triggers, respectively. (59) I don’t know whether Bill ended up participating in the Marathon yesterday. But if he won, he will celebrate tonight. (60) I don’t know whether anybody broke that computer. #But if it is Mary who did it, she should repair it. The second difference has to do with the projection behaviour in quantificational sentences. While, as we saw, the presuppositions of soft and hard triggers appear to pattern alike with respect to the projection behaviour when embedded in connectives, modals and questions, they behave differently when embedded in quantificational sentences. The gist of the observation, due to Charlow (2009), is that while hard presuppositions appear to project universally in quantificational sentences in a uniform way, the force of the projection inference of soft presuppositions depends on the quantifier involved. So, for instance, if we consider also, which patterns like it-clefts in explicit ignorance contexts, we observe that all of (61a) to (61d) give rise to the inference that each of these students smoke something other than Marlboro. (61) a. Each of these students also smokes [Marlboro]F. b. None of these students also smokes [Marlboro]F. c. More/fewer than three of these students also smoke [Marlboro]F. d. Some of these students also smokes [Marlboro]F. On the other hand, when we turn to stop, which patterns with win in the suspension contexts, we find a non-uniform behaviour depending on the quantifier involved. In other words, (62a) and (62b) give rise to the inference that each of these students used to smoke; (63a) and (63b), 271

Jacopo Romoli and Uli Sauerland

on the other hand, do not give rise to the same inference about all individuals in the domain. This pattern has also been shown experimentally by Chemla (2009a). (62) a. Each of these students stopped smoking. b. None of these students stopped smoking. (63) a. More/fewer than three of these students stopped smoking. b. Some of these students stopped smoking. While the boundaries of the soft vs. hard distinction are controversial (see Abbott 2006 and Klinedinst 2010 for discussion), the division between weak and hard presupposition triggers is widely accepted (but see Abrusán 2014 for a recent criticism of the distinction).The challenge for Dynamic Semantics is to account for this difference in behaviour among presupposition triggers. Several researchers have attempted to explain the division by bringing soft triggers closer to implicatures. To illustrate the gist of the idea, consider the sentence in (64a) and its negative counterpart in (64b). As seen previously, the relevant point to explain here is that both (64a) and (64b) give rise to the same inference in (65). In a traditional approach to presuppositions like Dynamic Semantics, this is explained by assuming that (65) is a presupposition of the sentence in (64a), which then projects from under negation in the case of (64b). (64) a. Bill won the Marathon. b. Bill didn’t win the Marathon. (65) Bill participated in the Marathon. The recent approaches mentioned earlier would instead explain the fact that (65) is both an inference of (64a) and (64b) very differently: the idea is that first (65) is just an entailment of (64b) and, second, (65) is an implicature of (64b). In a nutshell, the implicature is derived by assuming that (64b) and (66) compete with each other, in such a way that upon hearing (64b) the hearer will conclude that (66) is false, thereby concluding (65) (= it’s false that Bill didn’t participate in the Marathon). (66) Bill didn’t participate in the Marathon. By characterising the inferences of soft triggers like win in this way, these recent approaches hope to capture the differences between soft and hard triggers mentioned earlier; for discussion, see Chierchia and McConnell-Ginet (2000); Simons (2001); Abusch (2002, 2010); Chemla (2009b, 2010); Romoli (2014); for experimental work challenging this approach, see Chemla and Bott (2013); Romoli and Schwarz (2014); Bill, Romoli, Schwarz and Crain (2016); for different takes on the soft-hard distinction, see Abbott (2006), Klinedinst (2010), Fox (2012) among others.

Differences among quantifiers In addition to differences between quantifiers, quantification and presupposition have been shown to interact in more complex ways. Consider the phenomenon Sauerland (2013) refers to as ‘weakened projection’. Gender marking on bound pronouns generally exhibits a projection behaviour similar to presuppositions, as Stechow (2003), Heim (2008), Sauerland (2008) and Sudo (2012) discuss. For example, (68) leads to an inference that all the students are female. (68) None of the students except Sue embarrassed herself. 272

Presupposition and accommodation

However, the same pattern doesn’t obtain in (69) where the focus sensitive particle only is used to express quantification. (69) Only Sue embarrassed herself. This pattern has led to accounts where focus particle quantification differs in its interaction with presupposition from other quantificational operators.

Conclusion In this short chapter, we could only scratch the surface of presupposition and accommodation. They both remain widely studied phenomena in the semantics/pragmatics literature and related fields.The following suggestions for further reading provide additional material on both research areas.

Suggestions for further reading Beaver, D. and Geurts, B. (2011) ‘Presuppositions’, in C. Maienborn, K. Von Heusinger and P. Portner (eds.) Semantics: An International Handbook of Natural Language Meaning (vol. 3). Berlin: De Gruyter Mouton. This chapter is an update on more recent theories of presupposition projection as compared to various classical accounts of the problem. It also discusses various other notions and problems related to presuppositions. Beaver, D. and Zeevat, H. (2007) ‘Accommodation’, in G. Ramchand and C. Reiss (eds.) The Oxford Handbook of Linguistic Interfaces. Oxford: Oxford University Press. 503–538. Von Fintel, K. (2008) ‘What is accommodation, again?’, Philosophical Perspectives, 22(1): 137–170. These two chapters survey the notion of presupposition accommodation and the different implementations it has been given in the literature. Heim, I. (1990) ‘Presupposition projection’, in R. Van der Sandt (ed.) Reader for the Nijmegen Workshop on Presupposition, Lexical Meaning, and Discourse Processes. University of Nijmegen. This chapter introduces the main problem of presupposition projection and surveys the various classical accounts of the problem. Schwarz, F. (2014) ‘Introduction: Aspects of meaning in context-theoretical issues and experimental perspectives’, in F. Schwarz (ed.) Experimental Perspectives on Presuppositions. Cham: Springer. 1–37. Presuppositions have recently started being investigated through psycholinguistic methods more and more; this chapter is a useful survey of the different studies.

Notes 1 The distinction between semantics and pragmatics is, of course, in itself controversial. A simple way to characterise the distinction is in terms of meaning encoded in words and morphemes and the way they are put together, on one hand, and further inferences derived from the hearer by reasoning over the speaker’s communicative intentions, on the other; see Schlenker (forthcoming) and Stalnaker (2014) among many others for discussion. 2 Notice, in passing, that an account for the oddness of sentences like (5b) is also needed. The general idea in the literature is assuming an assertability condition on sentences, which requires a speaker not to assert something that is already (assumed to be) known by the participants of the conversation. In the case of (5b), this condition is violated because the speaker asserts in the first part of the sentence what the second part also asserts (that Jack doesn’t come to class at the moment). For different implementations of this idea, see Stalnaker (1978); Singh (2007); Fox (2008); Schlenker (2009); Katzir and Singh (2013); Meyer (2013); Mayr and Romoli (2016) among others. 3 It is standard to include in the satisfaction approach all theories that impose a condition requiring that the context entail the presuppositions of the asserted sentence.Therefore this approach includes a variety of other more recent accounts like Schlenker (2008a, 2009) and Fox (2008, 2012), among others. 4 Notice that while one could conceive of the ACC operation as a purely semantic operation, the decision as to whether to interpret the sentence with or without ACC is pragmatic in nature. So here too we see a close interaction between semantic and pragmatic ingredients.

273

Jacopo Romoli and Uli Sauerland 5 The conditional inference in (53b) is a genuine presupposition in that it shows the discourse and projective properties outlined in the introduction. First, it can be old information in the context, (i), and it appears to project from further embeddings like (ii) (Schlenker 2011). (i) If Paul is a devout Catholic, he has a Bible, and if he is a devout Catholic, he will read his Bible tonight. (ii) If Paul is a devout Catholic, will he read his Bible tonight?

References Abbott, B. (2006) ‘Where have some presuppositions gone?’, in B. J. Birner and G. Ward (eds.) Drawing the Boundaries of Meaning: Neo-Gricean Studies in Pragmatics and Semantics in Honor of Laurence R. Horn. Amsterdam/Philadelphia: John Benjamins. 1–33. Abrusán, M. (2014) ‘Disappearing acts of presuppositions: Cancelling the soft-hard distinction’, unpublished manuscript, CNRS. Abusch, D. (2002) ‘Lexical alternatives as a source of pragmatic presupposition’, in B. Jackson (ed.) Semantics and Linguistic Theory: Proceedings of SALT 12. Ithaca, NY: CLC Publications. 1–19. Abusch, D. (2010) ‘Presupposition triggering from alternatives’, Journal of Semantics, 27(1): 1–44. Beaver, D. (2001) Presupposition and Assertion in Dynamic Semantics. Stanford, CA: CSLI Publications. Beaver, D. (2010) ‘Have you noticed that your belly button lint colour is related to the colour of your clothing?’, in R. Bäuerle, U. Reyle and T. E. Zimmerman (eds.) Presuppositions and Discourse: Essays Offered to Hans Kamp. Bingley: Emerald Group Publishing Limited. 65–100. Beaver, D. and Geurts, B. (2011) ‘Presuppositions’, in C. Maienborn, K. Von Heusinger and P. Portner (eds.) Semantics: An International Handbook of Natural Language Meaning. Berlin: Mouton de Gruyter. Beaver, D. and Krahmer, E. (2001) ‘A partial account of presupposition projection’, Journal of Logic, Language and Information, 10(2): 147–182. Bill, C., Romoli, J., Schwarz, F. and Crain, S. (2016) ‘Scalar implicatures vs. presupposition: The view from acquisition’, TOPOI, 25(1): 57–71. Charlow, S. (2009) ‘ “Strong” predicative presuppositional objects’, Unpublished ms. (available from http:// simoncharlow.com/papers/presupp.pdf ) Chemla, E. (2008) ‘An anti-introduction to presuppositions’, in P. Egré and G. Magri (eds.) Presuppositions and Implicatures: Proceedings of MIT-France Workshop on Scalar Implicature and Presupposition. Cambridge: MIT Working Papers in Linguistics. 60. Chemla, E. (2009a) ‘Presuppositions of quantified sentences: Experimental data’, Natural Language Semantics, 17(4): 299–340. Chemla, E. (2009b) ‘Presuppositions of quantified sentences: Experimental data’, unpublished previous version of Chemla 2009a. Available from: www.emmanuel.chemla.free.fr (accessed May 2015). Chemla, E. (2010) ‘Similarity: Towards a unified account of scalar implicatures, free choice permission and presupposition projection’, unpublished manuscript. Chemla, E. and Bott, L. (2013) ‘Processing presuppositions: Dynamic semantics vs pragmatic enrichment’, Language and Cognitive Processes, 38(3): 241–260. Chemla, E. and Schlenker, P. (2011) ‘The reality of conditional presuppositions’, unpublished manuscript ENS. Chierchia, G. (2009) ‘On the explanatory power of dynamic semantics’, Keynote presentation, Sinn und Bedeutung 14, University of Vienna. Chierchia, G. and McConnell-Ginet, S. (2000) Meaning and Grammar: An Introduction to Semantics. Cambridge, MA: MIT Press. Fox, D. (2008) ‘Two short notes on Schlenker’s theory of presupposition projection’, Theoretical Linguistics, 34(3): 237–252. Fox, D. (2012) ‘Presupposition projection from quantificational sentences:Trivalence, local accommodation, and presupposition strengthening’, manuscript, Hebrew University of Jerusalem and MIT. Frege, G. (1892) ‘Über Sinn und Bedeutung’, Zeitschrift für Philosophie und Philosophische Kritik, 100: 25–50. Gazdar, G. (1979) Pragmatics: Implicature, Presupposition, and Logical Form. New York: Academic Press. George, B. (2008) ‘Presupposition repairs: A static, trivalent approach to predicting projection’, unpublished MA thesis, University of California, Los Angeles. Geurts, B. (1996) ‘Local satisfaction guaranteed: A presupposition theory and its problems’, Linguistics and Philosophy, 19(3): 259–294.

274

Presupposition and accommodation Geurts, B. (1999) Presuppositions and Pronouns. Oxford: Elsevier. Heim, I. (1982) ‘The semantics of definite and indefinite noun phrases’, unpublished PhD dissertation, University of Massachusetts. Heim, I. (1983) ‘On the projection problem for presuppositions’, in D. P. Flickinger (ed.) Proceedings of WCCFL 2. Stanford, CA: CSLI Publications. 114–125. Heim, I. (1990) ‘Presupposition projection’, in R. Van der Sandt (ed.) Reader for the Nijmegen Workshop on Presupposition, Lexical Meaning, and Discourse Processes. Nijmegen, Netherlands: University of Nijmegen. http://semanticsarchive.net/Archive/GFiMGNjN/Presupp%20projection%2090.pdf Heim, I. (1991) ‘Artikel und Definitheit’, in A. Von Stechow and D.Wunderlich (eds) Semantik: Ein Internationales Handbuch der Zeitgenössischen Forschung. Berlin: Mouton de Gruyter. 487–535. Heim, I. (1992) ‘Presupposition projection and the semantics of attitude verbs’, Journal of Semantics, 9: 183–221. Heim, I. (2006) ‘On the proviso problem’, talk given at the Milan Meeting, Gargnano. Heim, I. (2008) ‘Feature on bound pronouns’, in D. Harbour, D. Adger and S. Béjar (eds.) Phi Theory. Oxford: Oxford University Press. 35–56. Heim, I. and Kratzer, A. (1998) Semantics in Generative Grammar. Malden, MA: Blackwell. Karttunen, L. (1971) ‘Some observations on factivity’, Papers in Linguistics, 4: 55–69. Karttunen, L. (1973) ‘Presupposition of compound sentences’, Linguistic Inquiry, 4(3): 169–193. Karttunen, L. (1974) ‘Presupposition and linguistic context’, Theoretical Linguistics, 1: 181–194. Karttunen, L. and Peters, S. (1979) ‘Conventional implicature’, Syntax and Semantics, 11: 1–56. Katzir, R. and Singh, R. (2013) ‘Hurford disjunctions: Embedded exhaustification and structural economy’, Proceedings of Sinn und Bedeutung 18,Vitoria-Gasteiz. 201–216. Klinedinst, N. (2010) ‘Totally hardcore semantic presuppositions’, unpublished manuscript. Krahmer, E. (1998) Presuppositions and Anaphora. Stanford, CA: CSLI Publications. Lewis, D. (1979) ‘Scorekeeping in a language game’, Journal of Philosophical Logic, 8(3): 339–359. Mayr, C. and Romoli, J. (2016) ‘A puzzle for theories of redundancy: exhaustification, incrementality, and the notion of local context’. Semantics & Pragmatics 9 (early access). Meyer, M. C. (2013) ‘Ignorance and grammar’, unpublished PhD dissertation, MIT. Percus, O. (2006) ‘Antipresuppositions’, in A. Ueyama (ed.) Theoretical and Empirical Studies of Reference and Anaphora: Toward the Establishment of Generative Grammar as an Empirical Science. Japan Society for the Promotion of Science. 52–73. Pérez-Carballo, A. (2008) ‘Toward a dissolution of the proviso problem’, in P. Egré and G. Magri (eds.) Presuppositions and Implicatures: Proceedings of MIT-France Workshop on Scalar Implicature and Presupposition 60. Cambridge: MIT Working Papers in Linguistics. Romoli, J. (2014) ‘The presuppositions of soft triggers are obligatory scalar implicatures’, Journal of Semantics, 32(2): 1–47. Romoli, J. and Schwarz, F. (2014) ‘An experimental comparison between presuppositions and indirect scalar implicatures’, in F. Schwarz (ed.) Studies in Theoretical Psycholinguistics: Experimental Perspectives on Presuppositions. Dordrecht: Springer. 215–240. Romoli, J., Sudo,Y. and Snedeker, J. (2011) ‘An experimental investigation of presupposition projection in conditional sentences’, in N. Ashton, A. Chereches and D. Lut (eds.) The 21st Semantics and Linguistics Theory Conference. Rutgers University. 67–88. Rothschild, D. (2011) ‘Explaining presupposition projection with dynamic semantics’, Semantics and Pragmatics, 4(3): 1–43. Sauerland, U. (2003) ‘A new semantics for number’, in R. B.Young and Y. Zhou (eds.) Proceedings of Semantics and Linguistic Theory 13. Ithaca, NY: CLC Publications. 258–275. Sauerland, U. (2008) ‘On the semantic markedness of Phi-features’, in D. Harbour, D. Adger and S. Béjar (eds.) Phi Theory. Oxford: Oxford University Press. 57–82. Sauerland, U. (2013) ‘Presuppositions and the alternative tier’, in T. Snider (ed.) Proceedings of Semantics and Linguistic Theory Conference 23. University of California, Santa Cruz. 156–173. Schlenker, P. (2008a) ‘Be articulate: A pragmatic theory of presupposition projection’, Theoretical Linguistics, 34(3): 157–212. Schlenker, P. (2008b) ‘Presupposition projection: the new debate.” Semantics and Linguistic Theory [Online], 18 (2008): 655-693. Available at: . Date accessed: 21 nov. 2016. doi:http://dx.doi.org/10.3765/salt.v18i0. 2503. Schlenker, P. (2009) ‘Local contexts’, Semantics and Pragmatics, 2(3): 1–78. Schlenker, P. (2011) ‘A note on the proviso problem’, Natural Language Semantics, 19(4): 395–422.

275

Jacopo Romoli and Uli Sauerland Schlenker, P. (forthcoming) ‘The semantics - pragmatics interface’, in M. Aloni and P. Dekker (eds.) The Cambridge Handbook of Formal Semantics. Cambridge: Cambridge University Press. Shannon, B. (1976) ‘On the two kinds of presuppositions in natural language’, Foundations of Language, 14: 247–249. Simons, M. (2001) ‘On the conversational basis of some presuppositions’, in R. Hastings, B. Jackson and Z. Zvolenszky (eds.) Proceedings of Semantics and Linguistic Theory Conference 11. Ithaca, NY: CLC Publications. 431–448. Singh, R. (2007) ‘Assertability constraints and absurd assertions’, unpublished manuscript, MIT. Singh, R. (2008) ‘Modularity and locality in interpretation’, dissertation, MIT. Soames, S. (1989) ‘Direct reference and propositional attitudes’, in J. Almog, J. Perry and H. Wettstein (eds.) Themes from Kaplan. Oxford: Oxford University Press. 393–420. Stalnaker, R. (1973) ‘Presuppositions’, Journal of Philosophical Logic, 2: 447–457. Stalnaker, R. (1974) ‘Pragmatic presuppositions’, in M. Munitz and D. Unger (eds.) Semantics and Philosophy. New York: New York University Press. 197–213. Stalnaker, R. (1978) ‘Assertion’, Syntax and Semantics, 9: 315–332. Stalnaker, R. (2002) ‘Common ground’, Linguistic and Philosophy, 25: 701–721. Stalnaker, R. (2014) Context. Oxford: Oxford University Press. Strawson, P. F. (1950) ‘On referring’, Mind, 59: 320–344. Sudo,Y. (2012) ‘On the semantics of Phi features on pronouns’, unpublished PhD dissertation, MIT. Van der Sandt, R. (1992) ‘Presupposition projection as anaphora resolution’, Journal of Semantics, 9: 333–377. Van Rooij, R. (2007) ‘Strengthening conditional presuppositions’, Journal of Semantics, 24(3): 289–304. Von Fintel, K. (2008) ‘What is accommodation, again?’, Philosophical Perspectives, 22(1): 137–170. Von Stechow, A. (2003) ‘Binding by verbs’, in Proceedings of North-Eastern Linguistic Society 33. University of Massachusetts at Amherst: GLSA.

276

22 GRAMMATICALISATION1 María José López-Couso and Elena Seoane

Introduction Grammaticalisation, also known as ‘grammaticization’ (cf. Hopper 1991; Himmelmann 2004) or ‘grammatization’ (cf. Matisoff 1991), is broadly understood as the development along the cline of grammaticality ‘lexical > grammatical > more grammatical’. The growing popularity of the field over the past three decades is reflected in a vast recent literature on grammaticalisation and grammaticalisation phenomena, including acknowledged canonical references (Lehmann 1982, 1985; Traugott 1982; Heine et al. 1991; Hopper 1991), textbooks (Hopper and Traugott 2003), handbooks (Narrog and Heine 2011), lexicons (Heine and Kuteva 2002), special issues in journals (Language Sciences, Campbell 2001 and Kranich and Breban 2015), monographs (e.g., Heine 1993), and collective volumes (e.g., Traugott and Heine 1991; Giacalone Ramat and Hopper 1998; Wischer and Diewald 2002; Fischer et al. 2004; López-Couso and Seoane 2008; Seoane and López-Couso 2008; Stathi et al. 2010; Traugott and Trousdale 2010; Davidse et al. 2012). There are also a great many journal articles dealing with grammaticalisation and related processes of language change, including lexicalisation, degrammaticalisation, analogy, and (inter-) subjectification, these in different domains, different languages and language families, and from different approaches and perspectives. The past few decades have also seen a proliferation of workshops and conferences devoted to grammaticalisation, most notably the New Reflections on Grammaticalization (NRG) series (Potsdam 1999, Amsterdam 2002, Santiago de Compostela 2005, Leuven 2008, and Edinburgh 2012) and From Ideational to Interpersonal: Perspectives from Grammaticalization (FITIGRA, Leuven 2005). As will become clear in the sections that follow, studies on grammaticalisation place special emphasis on the relevance of the semantic and pragmatic aspects of language change (e.g., Bybee et al. 1994; Hopper and Traugott 2003), rather than on formal evidence (but see Van linden et al. 2010). It is not surprising, therefore, that pragmatics has established itself as a major area of research in grammaticalisation studies (consider, for example, the number of publications devoted to the development of pragmatic markers since the appearance of Brinton’s classic monograph in 1996). In this context, and due to limitations of space, no attempt is made here at completeness. For a more comprehensive discussion of grammaticalisation and its relation to pragmatics, the reader is referred to the selective bibliography mentioned throughout this chapter and references therein. 277

María José López-Couso and Elena Seoane

A condensed history of grammaticalisation Although the term grammaticalisation seems to have been coined by the French Indo-Europeanist Antoine Meillet at the beginning of the twentieth century (see later in this chapter), interesting insights into the origin of grammatical forms can also be found in the work of eighteenth- and nineteenth-century philosophers. As early as 1746, Étienne Bonnot de Condillac claimed that certain verbal inflections (e.g., those marking tense) had their origins in independent items. A century later, John Horne Tooke (1857) argued that adverbs, prepositions, and conjunctions derived from nouns and verbs (‘necessary words’ in his terminology) through processes of ‘abbreviation’ and ‘mutilation’. Central figures in the early study of grammaticalisation are the nineteenth-century German linguists Franz Bopp, who included the change from lexical item to grammatical form among his principles of comparative grammar (1816),Wilhelm von Humboldt, who maintained that grammar evolves from concrete ideas through a four-stage development (1825), and Georg von der Gabelentz (1891), for whom grammatical forms arise from the competition between the tendency towards ease of articulation (which wears down linguistic forms) and the tendency towards distinctness (which requires the appearance of new forms to replace the worn-out ones). The publication in 1912 of Meillet’s article ‘L’evolution des formes grammaticales’ marked a turning point in grammaticalisation studies. Although some of Meillet’s ideas on grammaticalisation are strongly reminiscent of von Humboldt’s and Gabelentz’s observations on the origin and development of grammatical forms, it was Meillet who coined the term itself and who first recognised grammaticalisation as the most important process for the development of new grammatical material in languages. He defines grammaticalisation as ‘le passage d’un mot autonome au rôle d’élément grammatical’ [‘the shift of an autonomous word to the status of grammatical element’] (1912: 131). Although grammaticalisation understood in this way is restricted to having its effect on individual words only, in several parts of his article Meillet hints at the possibility of interpreting the process as also involving phrases and even word-order types. As for many later practitioners of grammaticalisation theory, the cause of grammaticalisation for Meillet is a loss of expressivity of a frequently used word or collocation, usually accompanied by attrition in its phonological substance and a bleaching in its meaning. An example is the development of the modern Greek future tense morpheme tha from an earlier construction, thelô ina ‘I wish that’. He also admits, however, that weakening can be followed by reinforcement, with the emergence of new expressions to convey the same idea as the grammaticalised forms (‘renouvellement’). Much subsequent work on grammaticalisation has adopted Meillet’s approach to the topic, so that many definitions of the process offered over the final part of the twentieth century remind us of Meillet’s first account in the 1910s. Thus, for example, Croft (1990: 230) characterises grammaticalisation as ‘the process by which full lexical items become grammatical morphemes’. Trask takes a similar stand in one of the leading historical textbooks: An ordinary lexical item with an ordinary meaning comes to be used in some particular context; it is then bleached of its original meaning and becomes a mere grammatical marker in a syntactic construction. Finally, it is reduced to a bound morpheme, an affix, a piece of morphology. (Trask 1996: 146) Given its strong diachronic dimension, grammaticalisation became a neglected topic for several decades after Meillet, in the context of the dominant theoretical paradigms of structuralism and 278

Grammaticalisation

generativism and their emphasis on synchrony. A notable exception was Jerzy Kuryłowicz, an Indo-Europeanist who added an important refinement to Meillet’s definition of the process. For Kuryłowicz, ‘grammaticalization consists in the increase of the range of a morpheme advancing from a lexical to a grammatical or from a less grammatical to a more grammatical status’ (Kuryłowicz 1965: 52).The idea that grammaticalisation involves not only a change from lexical to grammatical, but also from less grammatical to more grammatical, is echoed in many later descriptions, including those of Heine and colleagues (1991: 3) and Lehmann (1982: v), where grammaticalisation is seen as ‘a process which turns lexemes into grammatical formatives and renders grammatical formatives still more grammatical’. Givón (1991: 305) proposes referring to the second part of Kuryłowicz’s bipartite definition (less grammatical > more grammatical) as ‘secondary grammaticalization’. In line with Givón’s terminology, Traugott (2002: 26–27) suggests the label ‘primary grammaticalization’ to refer to the first half of the definition (lexical > grammatical). Interestingly, secondary grammaticalisation and the distinction between secondary and primary grammaticalisation have become themes of central concern in current grammaticalisation theory (cf. the Language Sciences special issue, 2015, edited by Kranich and Breban, entirely devoted to it). It was only in the 1970s that the period of what Lehmann (1982) aptly called ‘amnesia’ about grammaticalisation ended and the topic began to enjoy a new, growing wave of interest among linguists, encouraged in particular by the work of scholars like Talmy Givón, Charles N. Li and Sandra A.Thompson, and Joseph H. Greenberg. Givón’s paper ‘Historical syntax and synchronic morphology: An archaeologist’s field trip’, with the well-known slogan ‘Today’s morphology is yesterday’s syntax’ (Givón 1971: 413), represents a major landmark in this new phase of grammaticalisation research, with studies now extending beyond the original diachronic domain to include synchronic and typological approaches. A prolific period of activity followed in the field, as noted in our introductory section, and this literature would come to form the basis of modern grammaticalisation theory. The abundance of work on grammaticalisation produced over the past three decades also provides us with alternative and more specific characterisations of the term. For example, Paul Hopper (1987) uses ‘grammaticalization’ as a synonym (or near synonym) of ‘grammar’. In his view, grammar is not static, but is constantly emerging from its use in discourse. As a consequence, ‘there is, in other words, no “grammar”, but only “grammaticization” – movements towards structure’ (Hopper 1987: 148). Recent grammaticalisation literature has also noted that items which undergo grammaticalisation do so only in specific morphosyntactic contexts and under certain pragmatic circumstances, and that in a way it is the whole construction that becomes grammaticalised. Definitions of grammaticalisation along these lines (the ‘grammaticalisation of constructions’) can be found in Lehmann (1992: 406), Bybee (2003: 602), Traugott (2003a: 645), and Brinton and Traugott (2005: 99), among many others. Especially after the advent of Construction Grammar (Goldberg 1995, 2006; Croft 2001), the focus of grammaticalisation research has indeed shifted from lexemes to constructions. Grammaticalisation shares with constructionalisation (cf. Traugott and Trousdale 2010; Trousdale 2010, 2012b) the view of grammar as a unified whole where pragmatics, semantics, syntax, morphology, and phonology go hand in hand, so that when a construction grammaticalises it changes its pragmatics, meaning, and function, and accordingly changes its grammatical structure and behaviour. In other words, whereas traditional approaches to grammaticalisation focus on the formation of individual elements (e.g., the development of the English modal auxiliary can from the Old English lexical verb cunnan ‘know’), the constructional approach considers the emergence and development of whole constructional types and how they are optimised in constructional networks (Traugott 2007, 2008; Trousdale 2012b). 279

María José López-Couso and Elena Seoane

Looking closer into grammaticalisation As can be gathered from the foregoing discussion, grammaticalisation can be regarded as a ‘crosscomponential’ (McMahon 1994: 161) or global change which affects different components of linguistic analysis, from phonology and morphosyntax to semantics and pragmatics. Suitable starting points for the description of the basic features of grammaticalisation at these different levels are Lehmann’s (1982, 1985) parameters and Hopper’s (1991) principles, which cover advanced and incipient grammaticalisation, respectively, and which are thus complementary. Hopper (1991: 22) proposes a catalogue of five principles of grammaticalisation operating in the early stages of grammaticalisation. These are: (a) Layering: within a specific functional domain, new layers are constantly emerging and come to coexist and interact with the older layers, which are not necessarily discarded. For example, the newer Latin periphrastic future (cantare habeo ‘I have to sing’ > ‘I will sing’) coexisted at one stage with the older inflected future (cantabo ‘I will sing’). (b) Divergence: a lexical item undergoing grammaticalisation in a specific context may remain as an autonomous element in its other contexts, so that both grammaticalised and nongrammaticalised forms exist side by side. An example is the French negative supportive particle pas, which coexists with the noun pas ‘step’ from which it derives. (c) Specialisation: as grammaticalisation advances, the variety of choices in a given functional domain is reduced once certain forms assume a more general grammatical meaning. Thus, out of a wide array of forms (e.g., pas ‘step’, point ‘point’, mie ‘crumb’, gote ‘drop’, etc.) which could be used to reinforce negation in Old French, only pas has become fully grammaticalised. (d) Persistence: traces of the source meaning of an item undergoing grammaticalisation may survive and affect the distribution of the grammaticalised form. For example, the original purposive meaning of be going to is reflected in its use as the marker of the future of schedule or plan. (e) Decategorialisation: items undergoing grammaticalisation tend to lose or neutralise the morphosyntactic properties characteristic of the source forms. A clear example is the development of the Old English noun hwil ‘time’ into the present-day conjunction while, which no longer shows the grammatical features of the category ‘noun’ (e.g., it cannot take an article or be modified by an adjective). On the other hand, in Lehmann’s model, grammaticalisation is characterised as affecting the autonomy (or degree of freedom) of a linguistic sign as regards its weight, its cohesion, and its variability. In his view, grammaticalisation can be measured diachronically by means of the following six processes (examples from Lehmann 1985: 308–310): (a) Attrition: reduction of phonological and semantic substance; for example, Latin ad > Romance a, where loss of the consonant (erosion) is accompanied by the loss of the local feature in ad (bleaching, desemanticisation). (b) Paradigmaticisation: stronger integration of a grammatical formative into a paradigm; for example, Latin primary prepositions ad, de > French à, de, which form the paradigm of oblique cases. (c) Obligatorification: loss of paradigmatic variability, reduction of choice within a paradigm; for example, while the Latin preposition de could be replaced by other members from the same paradigm, such as ab or ex, French de cannot. 280

Grammaticalisation

(d) Condensation: reduction of scope; for example, French avoir as an auxiliary only has scope over the past participle with which it combines in the perfect construction vs. Latin habere, which takes a noun phrase complement even in its auxiliary use. (e) Coalescence: increase in bondedness; for example, Latin habere becomes cliticised to the verb in the Romance future: cantare habeo ‘I have to sing’, ‘I will sing’ > French je chanterai ‘I will sing’. (f) Fixation: loss of syntagmatic variability; for example, French à and de occupy a fixed syntactic slot in the noun phrase vs. Latin ad and de, which were more mobile. Despite the wide recognition of the importance of Lehmann’s parameters of grammaticalisation, several aspects of his proposal have been extensively criticised over the past few years. While most of his processes imply a loss in (phonological, morphosyntactic, and/or semantic) properties, grammaticalisation clearly involves some gains as well. Thus, although grammaticalisation typically entails bleaching of referential and lexical meaning (desemanticisation), it also implies the acquisition of abstract meaning (pragmatic enrichment or strengthening). For example, in the grammaticalisation of have-perfects, have loses its possessive meaning and acquires the sense of perfect tense and aspect. Similarly, go in be going to loses the lexical meaning of motion and gains the grammatical meaning of futurity and intentionality. The acquisition of new pragmatic meanings and functions will be discussed later in detail when we examine grammaticalisation and pragmaticalisation. Lehmann’s parameters of obligatorification and condensation have also been called into question. As regards the latter, for instance, it has been shown that scope reduction does not adequately account for certain types of grammaticalisation processes, such as the development of discourse markers (see later) or that of the English modals (cf. Nordlinger and Traugott 1997; Tabor and Traugott 1998), where grammaticalisation correlates with scope expansion. Besides the characteristics of grammaticalisation discussed earlier, frequency stands out as another central feature of the process, not only as a key factor of change (frequency contributes to grammaticalisation, as the trigger for some of the phonological, morphosyntactic, and semantic-pragmatic changes associated with it), but also as a result of change (the shift from lexis to grammar is typically accompanied by an increase in frequency, as a consequence of the extension of the grammaticalised item/construction to further contexts [see later]). Although the relevance of frequency in grammaticalisation had already been recognised in works such as Traugott and Heine (1991: 9), it subsequently came to receive increasing attention in the literature, to the point of being incorporated into some definitions of the process: I will argue for a new definition of grammaticization, one which recognizes the crucial role of repetition in grammaticization and characterizes it as the process by which a frequently used sequence of words or morphemes becomes automated as a single processing unit. (Bybee 2003: 603) The pervasive and overarching influence of frequency on grammaticalisation is also seen in the fact that not all types of lexical items represent valid input for (primary) grammaticalisation; rather, different forms have different potential for becoming grammaticalised. Cross-linguistic evidence suggests that, among other features, source concepts must be items belonging to the core vocabulary of the language and of frequent and general use, representing the basic properties of their respective semantic domains (e.g., ‘come’ and ‘go’, but not ‘run’) (cf. Givón 1979: 222; Heine et al. 1991: 34; Traugott and Heine 1991: 7–8). This would explain why certain 281

María José López-Couso and Elena Seoane

concepts are very frequently chosen to be developed into grammatical forms and categories across languages. Thus, for instance, nouns denoting bodily parts (e.g., ‘back’, ‘heart’, ‘foot’) very frequently grammaticalise as adpositions (e.g., behind, in front of), and basic processes (e.g., ‘go’, ‘have’, ‘finish’) commonly evolve into categories related to verbs, such as tense, aspect, and mood (cf., for example, the development of futurity markers from verbs of motion [e.g., be going to] and that of perfective markers from verbs of completion [e.g., Chinese liao ‘finish’ > le]).

Grammaticalisation ‘deconstructed’: unidirectionality, degrammaticalisation, and lexicalisation Although grammaticalisation has become a particularly attractive and fruitful topic in linguistic research over the past couple of decades, many of its basic tenets have been subject to heavy criticism. In a chapter titled ‘Deconstructing grammaticalization’ in his book Language Form and Language Function, Frederick Newmeyer claims that ‘there is no such thing as grammaticalization, at least in so far as it might be regarded as a distinct grammatical phenomenon requiring a distinct set of principles for its explanation’ (1998: 226). Unsurprisingly, Newmeyer’s chapter stimulated a heated debate about grammaticalisation theory. This reached its climax with the publication of the 2001 special issue of Language Sciences edited by Lyle Campbell, which provides a ‘critical assessment’ of grammaticalisation with contributions from Janda, Joseph, Newmeyer, and Norde which are all now considered classic. The unidirectionality hypothesis is probably the most commonly discussed of all the problems raised by critics of grammaticalisation theory.2 As shown earlier, since the times of Condillac grammaticalisation (viewed diachronically) has been taken to be a unidirectional process (lexical > grammatical > more grammatical). “Unidirectionality applies on all levels of language. At the semantic-pragmatic level, grammaticalization goes from full referential meanings to bleached, grammatical meanings, from less to more subjective, and from concrete to abstract, not vice versa. Similarly, at the syntactic level, the change proceeds from lexical to grammatical, from a major category (noun, verb) to a minor one (conjunction, preposition), and from less bound to more bound. Finally, at the phonological level, full phonological forms become reduced forms.” Though many consider unidirectionality a fundamental property of grammaticalisation (built into its very definition), various potential counterexamples to unidirectionality have been identified, and the hypothesis has been strongly criticised on both theoretical and methodological grounds (cf., among others, Newmeyer 1998; Lass 2000; Janda 2001; Joseph 2004). However, no indisputable instances of a complete reversal of grammaticalisation (i.e., moving all the way along the cline of grammaticality) have thus far been proposed. Moreover, many claimed counterexamples to unidirectionality have proved illegitimate or spurious (cf. Traugott 2001), and those which remain (e.g., Pennsylvania German wotte; cf. Burridge 1998) are too few, in comparison to the large number of cases which confirm the hypothesis, to affect its cross-linguistic statistical validity (cf. Haspelmath 1999, 2004). It seems, then, that despite the existence of examples where the directionality of change appears to be reversed, the robustness of the principle of unidirectionality is beyond reasonable doubt. As pointed out in Brinton and Traugott (2005: 97), a common tendency in the grammaticalisation literature is to classify examples not conforming to the unidirectional model of grammaticalisation as instances of either degrammaticalisation or lexicalisation. In what follows we explain the differences between these three processes of change. Norde (2009: 120) defines degrammaticalisation as ‘a composite change whereby a gram in a specific context gains in autonomy or substance on more than one linguistic level (semantics, morphology, syntax or phonology)’. Prototypical examples of degrammaticalisation are the 282

Grammaticalisation

emergence of the noun handicap from the syntactic phrase hand in cap, and the fused compound mildew from Old English mele ‘honey’ and deaw ‘dew’ (cf. Brinton and Traugott 2005: 98). All examples of degrammaticalisation share four basic characteristics: (1) counterdirectionality, which involves a decrease in bondedness; (2) novelty, since the outcome of degrammaticalisation is the emergence of a new gram and not mere semantic or frequency changes in an already existing gram; (3) infrequency, especially as compared to grammaticalisation processes (see earlier); and (4) discontinuity, since in most cases degrammaticalisation involves one single shift from right to left in Hopper and Traugott’s (2003: 7) cline of grammaticality (content item > grammatical word > clitic > inflectional affix) with no subsequent domino effect. In this sense, degrammaticalisation is not a mirror-image reversal of grammaticalisation, since it does not go all the way from affix to content item with the same intermediate steps (cf. Norde 2010: 128–129). A finer-grained analysis of degrammaticalisation distinguishes three different types, depending on the level at which it takes place: degrammation, deinflectionalisation, and debonding (cf. Norde 2009: 133–134; 2010: 137–138). Degrammation occurs at the content level (cf. Ander­ sen’s 2008 model) and involves resemanticisation, that is, a shift from grammatical content to lexical content (e.g., the emergence of the Welsh noun eiddo ‘property’ from the possessive pronoun eiddo ‘his’; cf. Norde 2009: 145–148). Deinflectionalisation, in turn, involves a shift from more to less grammatical at the content-syntactic level (e.g., the Swedish inflectional genitive -s, which evolved from inflectional affix to clitic; cf. Norde 2010: 140–142). Finally, debonding, at the morphosyntactic level, is the process whereby a bound morpheme becomes a free morpheme (e.g., the development of the Irish pronoun muid ‘we’ from a first person plural verbal suffix; cf. Norde 2009: 204–207). Lexicalisation, according to Brinton and Traugott, (2005: 96–97) is a historical change that produces new lexical forms primarily out of syntactic constructions (e.g., run-of-the-mill ‘ordinary’) and word formations (e.g., woman from the Old English compound wifmann ‘wife.man’). As Trousdale points out, the crucial difference between grammaticalisation and lexicalisation concerns the function of the new item: in grammaticalisation it has a procedural or indexical (i.e., grammatical) function, while in lexicalisation it has a referential or conceptual (lexical) function (2012a: 541). Therefore, lexicalisation, unlike grammaticalisation, does not entail pragmatic enrichment, but rather an increase in semantic specificity and contentfulness. Despite the obvious differences between the two processes, they nevertheless share a number of traits, the most important being fusion (i.e., the fixing of collocations, which involves the shift from compositional to holistic; e.g., childhood from Old English cildhad ‘child.state’) and coalescence (the reduction of phonological segments, as in barn from Old English bere ‘barley’ and ærn ‘house’; cf. Brinton and Traugott 2005: 105; Trousdale 2012a: 543).

Grammaticalisation contexts, (inter-)subjectification, and pragmaticalisation Recent literature has shown that the semantic changes involved in grammaticalisation, such as desemanticisation or bleaching, are the result of perceiving and using pragmatic implicatures that arise in context. If we hear ‘I have the fish boiled, would you like some?’, we can infer that the speaker has boiled the fish, a perfect-resultative meaning; if we say ‘I am going to visit my aunt’, we imply motion but also futurity, a pragmatic implicature which is perceived by the hearer. This pragmatic conditioning of grammaticalisation is what Heine et al. (1991) refer to as ‘context-induced reinterpretation’: it is the hearer who, receiving a construction in a particular pragmatic context, reinterprets it accordingly. By contrast, scholars favouring a cognitiveinferential conceptualisation of pragmatics prefer to use the term ‘invited inferencing’: it is the 283

María José López-Couso and Elena Seoane

speaker who produces not only literal meaning, but also further meanings for the hearer to infer (cf. Traugott and König 1991; Traugott and Dasher 2002; Jucker 2012). Both context-induced reinterpretation and invited inferencing assume pragmatic polysemy and/or ambiguity to take place in context. In other words, new grammatical forms do not develop in all uses of an item serving as the potential input for grammaticalisation (see earlier). On the contrary, grammaticalisation is to a great extent context-dependent. Drawing on data from various languages, Heine (2002) proposes a four-stage scenario of contextual requirements for grammatical meanings to evolve. At stage I (‘initial stage’) the item shows its original source meaning, unconstrained by the context. Stage II, in turn, corresponds to the ‘bridging context’ (cf. Evans and Wilkins 1998), which gives rise to an inference (cf. Grice 1975) in favour of a target meaning, which is foregrounded. Although the target meaning offers the most plausible interpretation at this stage, the source meaning cannot be completely ruled out in a bridging context. At a later stage (Stage III), however, the context is no longer compatible with the original meaning. ‘Switch contexts’ of this kind play a particularly relevant role in grammaticalisation (cf. Heine 2002: 83, 86). Finally, at the ‘conventionalization stage’ (Stage IV) the target meaning is no longer contextually constrained and can therefore be extended to further contexts. The development of the modern High German concessive connective dabei ‘although’ from its source temporal meaning ‘at that occasion, during that event’ serves as an example (cf. Heine 2002: 91–94). (1) a. Karl geht schlafen; dabei trägt er einen Schlafanzug.   [Stage I]

b. Karl geht schlafen; dabei ist er gar nicht müde.   [Stage II]



‘Karl is going to bed; still, he is not tired at all.’

c. Karl geht schlafen; dabei geht er um diese Zeit nie schlafen.   [Stage III]



‘Karl is going to bed; (at that occasion) he is wearing a pyjama.’

‘Karl is going to bed; although he never goes to bed at this time.’

d. Karl geht schlafen; dabei war er eben noch überhaupt nicht müde.   [Stage IV]



‘Karl is going to bed, although a moment ago he was not tired at all.’

At Stage I (1a), dabei receives a temporal interpretation, the situations in the two clauses being described as occurring within the same time span (simultaneity). In turn, (1b) is an example of a bridging context, in which the new meaning of concessivity is foregrounded (the information in the dabei clause contradicts common experience: one does not normally go to bed unless one is tired), although the source temporal reading cannot be completely ruled out. Stage III (switch context) is illustrated in examples such as (1c), where the temporal meaning is backgrounded and the only reasonable interpretation for the dabei clause is a concessive one. Finally, the context in (1d) shows that the simultaneity reading is no longer possible (cf. the presence of the adverbial eben noch ‘a moment ago’), so that dabei is fully conventionalised as a concessive connective. Diewald (2002) has proposed an alternative (though closely related)3 scenario. Using the grammaticalisation of German modal verbs as an example, Diewald (2002) suggests that the development of grammatical forms can be divided into three successive stages corresponding to different context types: ‘untypical’, ‘critical’, and ‘isolating’ contexts. While the untypical context simply provides the semantic and structural preconditions for grammaticalisation to occur, the critical context represents the beginning of the grammaticalisation process, since it allows 284

Grammaticalisation

alternative interpretations, including the new grammatical meaning. Consider in this sense the highly ambiguous Middle High German example in (2), with the modal können ‘can’, which allows (at least) three different readings in present-day German in the absence of contextual or situational clues (cf. Diewald 2002: 110–111). Finally, in isolating contexts only one of the competing interpretations is possible, so that the process of grammaticalisation is completed. Diewald’s (2002) model of relevant context types in grammaticalisation has been extended by Diewald and Smirnova with the introduction of a fourth developmental stage: the stage of ‘paradigmatic (re-)integration’, defined as ‘the estabilizing process of a new grammatical(ized) sign, which turns it into an integral part of a paradigm’ (2012: 128). In their view, this final stage, where the new form becomes integrated into a paradigm, serves as a decisive criterion for differentiating grammaticalisation from lexicalisation (see earlier) and from other types of language change. (2) der kunde se baz gelobet hân.

he can her better praised have

a. ‘He could have praised her better.’ b. ‘He was able to have her as a praised one better.’ c. ‘Perhaps, he has praised her better.’

To further illustrate the pragmatic effects of context, we should also mention subjectivity, the kind of stance most widely studied in relation to grammaticalisation. Subjectification and intersubjectification are ubiquitous processes of semantic change cross-linguistically, presumably because speakers attempt to incorporate into the communicative event the relevance of what is being said, and, for that reason, they project onto it their own subjective attitude and beliefs. According to Traugott (2010b: 32), expressions undergoing semantic and pragmatic change can often be organised in a continuum of (inter-)subjectivity: non-/less subjective > subjective > intersubjective. Though no overarching definition of subjectification and intersubjectification can be found in the literature (consider, for example, the rather different views of Traugott and Langacker, discussed in Traugott 1995 and López-Couso 2010), subjectification can be broadly described as a semantic-pragmatic process involving a shift in meaning from the objective description of an external situation to the expression of the speaker’s internal view, belief, or attitude towards the proposition. The semantic-pragmatic changes involved in the grammaticalisation of modal must, which originally expressed deontic modality (obligation) and later became an epistemic modal expressing deduction and conclusion, illustrate the correlation between subjectification and grammaticalisation (Brinton and Traugott 2005: 108–109). As this example also shows, subjectification normally takes place in primary grammaticalisation (see earlier), when the weakening of external meaning and the strengthening of pragmatic inferences usually occur. As for intersubjectification, it involves a semantic change towards the expression of ‘social deictic purposes’ (Traugott 2003b: 134), that is, to express attention to the roles of speaker and addressee both in an epistemic sense (their attitudes and views) and in a social sense (their social stance and identity; Traugott 2003b: 128). Examples of intersubjectification which reflect the social stance of meaning are euphemisms (e.g., pass away and depart for die) and politeness markers, such as the use of terms of endearment (sweetheart, sugar) as markers of positive politeness (Nevala 2010: 431). Intersubjectification, however, rarely intervenes in grammaticalisation processes; the few cases that have been attested, such as the development of pragmatic markers (well, like, right), are controversial, since for some authors pragmatic markers are not the result of 285

María José López-Couso and Elena Seoane

grammaticalisation, but of pragmaticalisation, an independent process to which we will devote the remainder of this section. The development of pragmatic/discourse markers differs from standard cases of grammaticalisation in crucial aspects: they can emerge from a variety of sources, including subject + verb matrix clauses (you know, you see), adverbial or relative clauses (as it seems), adverbial prepositional phrases (indeed), adverbs (well), and verbal forms (Mexican Spanish dale), among others. Pragmatic markers, therefore, do not constitute a formal grammatical class and do not become part of a paradigm (Brinton 2010: 285). Moreover, they are optional (there is no obligatorification), they are loosely connected with the sentence, often constituting a separate intonation unit (no syntactic fixation), and they normally have scope beyond the clause level (no scope reduction). Semantically, they have no referential meaning and do not affect the truth value of the sentence; instead, they have a cognitive function (epistemic, illocutionary force meanings), an interactional function (hedging, turn-taking), and a metatextual function (structuring, refocusing), and most of them are polyfunctional elements. These peculiar features of discourse markers have led some authors to maintain that their development belongs outside the realm of grammaticalisation, as mentioned previously in relation to intersubjectification. For these authors, discourse markers are the result of an independent process, pragmaticalisation, a term first introduced by Erman and Kotsinas (1993) and embraced by many others (Aijmer 1997; Günthner and Mutz 2004; Frank-Job 2006; Claridge and Arnovick 2010).4 Claridge and Arnovick’s (2010: 187) definition captures the essence of pragmaticalisation as the process ‘by which a lexico-grammatical sequence or word form, in a given context, loses its propositional meaning in favour of an essentially metacommunicative, discourse interactional meaning and/or (an already pragmatic element) continues to develop further pragmatic functions or forms’. Thus, the output of pragmaticalisation are pragmatic/discourse markers (e.g., well), hedges (e.g., as it were), interjections (e.g., Jesus/gee), swearing expressions, politeness markers (e.g., pray (you)/prithee), and conversational/textual routines (e.g., good-bye, bless you) (Claridge and Arnovick 2010: 167). In all these cases of pragmaticalisation, there is a loss of truth-conditional and content meaning, the scope is no longer within the proposition but expands over it, and the meaning becomes subjective or intersubjective, thus following the paths of directionality in semantic change outlined in Traugott and Dasher (2002: 40). By way of illustration, let us briefly sketch the emergence of the pragmatic marker well, as described in Jucker (1997) and Jucker and Taavitsainen (2013: 138–142). Originally an adjective and adverb with full lexical meaning, well began to develop a number of pragmatic marker functions in late Middle English. The first of these was as a so-called frame marker, a pragmatic use with discourse structuring function which emerged in the context of introducing direct speech, as in (3). (3) ‘Ye say well’, seyde the kynge. ‘Aske what ye woll and ye shall have hit and hit lye in my power to gyff hit.’ ‘Well,’ seyde thys lady, ‘than I aske the hede of thys knyght that hath wonne the swerde.’ (Helsinki Corpus, cmmalory 1470) The semantic change involved here is from propositional ‘fittingly, properly’ to textual ‘o.k. then’, representing acceptance of some preceding point in discourse. Another pragmatic use emerged in early Modern English, that of a face-threat mitigator, illustrated in (4). (4) I wis I wis, she is more your friend, then your are your owne. Well let her be what she will sayd her husband: but if shee come any more in my house, shee were as good no. (Helsinki Corpus, cefict2b 1619) 286

Grammaticalisation

Here the speaker starts with well in order to mitigate disagreement with his wife’s earlier statement, which is clearly an interpersonal, subjective use.

Conclusion As the foregoing discussion has shown, pragmatic factors in grammaticalisation have attracted much attention and are currently under active consideration in several areas (see Traugott 2010a: 109–116, for a full account). A particularly productive domain is the development of pragmatic markers from the point of view of pragmaticalisation and historical pragmatics (e.g., English indeed, anyway; see Brinton 1996; Traugott and Dasher 2002: 95, 164; clause-final kong ‘say’ in Hokkien; see Chappell 2012). Contrastive studies of the development of pragmatic markers have also generated considerable interest, as shown by Fanego’s (2010) study of Spanish de hecho compared to English in fact and the studies in Degand and Simon-Vandenbergen (2011), as well as López-Couso and Méndez-Naya (2015), which compares the development in English and Romance of evidential/epistemic markers originating in combinations of a verb and a complementiser (e.g., English looks like, Spanish dizque). Another fertile area of study is that of the emergence of scalar expressions, such as so, pretty (Tagliamonte 2008), own (König and Vezzosi 2008), totally (Aijmer 2011), absolutely (Núñez-Pertejo 2013), full, highly, very, well, and out-intensifiers (Méndez-Naya and Pahta 2010; Méndez-Naya 2014). In terms of methodology, current directions in work on grammaticalisation and pragmatics involve a combination of corpus research and qualitative analyses of pragmatic factors, such as pragmatic inferences, context, and genre. Traditionally, corpus-based pragmatic studies adopted by necessity a semasiological approach, the mapping of form to function, but recently semantic tagging has made it possible to identify some pragmatic functions automatically. However, in most cases it is the linguist who needs to analyse the pragmatic factors, which do not appear in conventional lexical or grammatical forms in the early stages of grammaticalisation (Kytö 2010: 51).

Suggestions for further reading Brinton, L. J. and Traugott, E. C. (2005) Lexicalization and Language Change. Cambridge: Cambridge University Press. This concise volume provides an excellent overview of earlier approaches to lexicalisation as well as the relation between lexicalisation and grammaticalisation. The authors propose a unified model of these two processes of language change, conveniently illustrated with examples (mostly) from the history of English. Hopper, P. J. and Traugott, E. C. (2003) Grammaticalization. Cambridge: Cambridge University Press. This textbook constitutes a solid introduction to the basic tenets of grammaticalisation, illustrated with a wealth of examples drawn from various languages and language families. It provides students with the necessary background on fundamental theoretical and methodological issues relevant to grammaticalisation studies, including a chapter on the crucial role that pragmatic factors play in grammaticalisation processes. Jucker, A. H. and Taavitsainen, I. (2010) Handbooks of Pragmatics: Historical Pragmatics. Berlin: De Gruyter Mouton. The twenty-two chapters of this handbook together represent a comprehensive overview of central areas of interest in historical pragmatics, and include those relevant to the present chapter, among them grammaticalisation (chapter 4 by E. C. Traugott), (inter-)subjectification (chapter 5 by M. J. LópezCouso), pragmaticalisation (chapter 6 by C. Claridge and L. Arnovick), and discourse markers (chapter 10 by L. Brinton). Jucker, A. H. and Taavitsainen, I. (2013) English Historical Pragmatics. Edinburgh: Edinburgh University Press. This textbook introduces students to key issues in historical pragmatics, offering critical evaluations of methodological approaches and data sources, together with selected case studies on various research

287

María José López-Couso and Elena Seoane topics (e.g., discourse markers, interjections, politeness markers). It contains a chapter devoted to grammaticalisation and pragmaticalisation (chapter 8), which includes several exercises and a section on further reading. Traugott, E. C. and Dasher, R. B. (2002) Regularity in Semantic Change. Cambridge: Cambridge University Press. This monograph focuses on historical semantics and pragmatics, with data drawn mainly from the histories of English and Japanese. It pays due attention to several key issues and topics discussed in the present chapter, among them inferencing, unidirectionality, grammaticalisation and (inter-)subjectification, and the development of pragmatic markers and of modal verbs.

Notes 1 We would like to thank the editors of this volume, especially Gerard J. Steen, for helpful comments on an earlier version of this chapter. For generous financial support, we are grateful to the following institutions: the European Regional Development Fund, the Spanish Ministry of Economy and Competitiveness (grants FFI2014–52188-P, FFI2014–53930-P, and FFI2014–51873-REDT), and the Directorate General for Scientific and Technological Promotion of the Regional Government of Galicia (grants GPC2014/004 and R2014/016). These grants are hereby gratefully acknowledged. 2 Other frequent criticisms of grammaticalisation include the following: (i) the mechanisms which characterise grammaticalisation (e.g., desemanticisation, reanalysis, phonetic reduction, subjectification, etc.) are not unique to grammaticalisation; (ii) not all instances of language change involving some aspect of grammar are due to grammaticalisation; and (iii) ‘grammaticalisation theory’ is not truly a theory. 3 The two proposals are, in principle, compatible, though they do not match exactly, given that they emphasise different aspects of the process of grammaticalisation. Thus, Heine (2002) focuses on the semantic changes taking place in the different context types, while Diewald (2002) pays considerable attention to morphological and structural aspects, as well as to paradigmatic relations in the target category. For a brief comparison of the two models, see Diewald (2002: 117). 4 For many other authors, pragmaticalisation is simply a subtype of grammaticalisation, involving a movement towards discourse (Traugott 1995; Wischer 2000; Brinton 2010).

References Aijmer, K. (1997) ‘ “I think” – an English modal particle’, in T. Swan and O. Jansen Westvik (eds.) Modality in Germanic Languages: Historical and Comparative Perspectives. Berlin/New York: Mouton de Gruyter. 1–47. Aijmer, K. (2011) ‘ “Are you totally spy?” A new intensifier in present-day American English’, in S. Hancil (ed.) Marqueurs Discursifs et Subjectivité. Rouen: University of Rouen and Havre. 155–172. Andersen, H. (2008) ‘Grammaticalization in a speaker-oriented theory of change’, in T. Eythórsson (ed.) Grammatical Change and Linguistic Theory:The Rosendal Papers. Amsterdam/Philadelphia: John Benjamins. 11–44. Bopp, F. (1816) Über das Conjugationssystem der Sanskritsprache in Vergleichung mit Jenem der Griechischen, Lateinischen, Persischen, und Germanischen Sprache. Frankfurt: Andreäischen. Brinton, L. J. (1996) Pragmatic Markers in English. Grammaticalization and Discourse Functions. Berlin/New York: Mouton de Gruyter. Brinton, L. J. (2010) ‘Discourse markers’, in A. H. Jucker and I. Taavitsainen (eds.) Handbooks of Pragmatics: Historical Pragmatics. Berlin/New York: Mouton de Gruyter. 285–314. Brinton, L. J. and Traugott, E. C. (2005) Lexicalization and Language Change. Cambridge: Cambridge University Press. Burridge, K. (1998) ‘From modal auxiliary to lexical verb: The curious case of Pennsylvania German “wotte” ’, in R. M. Hogg and L. Van Bergen (eds.) Historical Linguistics 1995: German Linguistics. Amsterdam: John Benjamins. 19–31. Bybee, J. (2003) ‘Mechanisms of change in grammaticization:The role of frequency’, in B. D. Joseph and R. D. Janda (eds.) The Handbook of Historical Linguistics. Oxford: Blackwell. 602–623. Bybee, J., Perkins, R., and Pagliuca, W. (1994) The Evolution of Grammar. Chicago/London: University of Chicago Press. Campbell, L. (ed.) (2001) ‘Grammaticalization: A critical assessment’ [Special issue], Language Sciences, 23(2–3).

288

Grammaticalisation Chappell, H. (2012) ‘ “Say” in Sinitic: From “verba dicendi” to attitudinal discourse markers’, in J. Van der Auwera and J. Nuyts (eds.) Grammaticalization and (Inter-)Subjectification. Brussels: Contactforum. 81–110. Claridge, C. and Arnovick, L. (2010) ‘Pragmaticalisation and discursisation’, in A. H. Jucker and I. ­Taavitsainen (eds.) Handbooks of Pragmatics: Historical Pragmatics. Berlin: Mouton de Gruyter. 165–192. Condillac, E. B. de. (1746) Essai sur l’Origine des Connaissances Humaines. Amsterdam: Pierre Mortier. Croft, W. (1990) Typology and Universals. Cambridge: Cambridge University Press. Croft, W. (2001) Radical Construction Grammar. Oxford: Oxford University Press. Davidse, K., Breban,T., Brems, L., and Mortelmans,T. (eds.) (2012) Grammaticalization and Language Change: New Reflections. Amsterdam: John Benjamins. Degand, L. and Simon-Vandenbergen, A. M. (eds.) (2011) ‘Introduction: Grammaticalization and (inter-)subjectification of discourse markers’, Linguistics, 49(2): 287–293. Diewald, G. (2002) ‘A model for relevant types of contexts in grammaticalization’, in I.Wischer and G. Diewald (eds.) New Reflections on Grammaticalization. Amsterdam/Philadelphia: John Benjamins. 103–120. Diewald, G. and Smirnova, E. (2012) ‘ “Paradigmatic integration”: The fourth stage in an expanded grammaticalization scenario’, in K. Davidse, T. Breban, L. Brems, and T. Mortelmans (eds.) Grammaticalization and Language Change: New Reflections. Amsterdam: John Benjamins. 111–133. Erman, B. and Kotsinas, U. B. (1993) ‘Pragmaticalization: The case of “ba’ ” and “you know”, Studier i ­Modernspråkvetenskap, 10: 76–93. Evans, N. and Wilkins, D. (1998) ‘The knowing ear: An Australian test of universal claims about the semantic structure of sensory verbs and their extension into the domain of cognition’, Arbeitspapier 32, Institut für Sprachwissenschaft, Universität Köln. Fanego,T. (2010) ‘Paths in the development of elaborative discourse markers: Evidence from Spanish’, in K. Davidse, L. Vandelanotte, and H. Cuyckens (eds.) Subjectification, Intersubjectification and Grammaticalization. Berlin/New York: Mouton de Gruyter. 197–237. Fischer, O., Norde, M., and Perridon, H. (eds.) (2004) Up and Down the Cline:The Nature of Grammaticalization. Amsterdam/Philadelphia: John Benjamins. Frank-Job, B. (2006) ‘A dynamic-interactional approach to discourse markers’, in K. Fischer (ed.) Approaches to Discourse Particles. Amsterdam: Elsevier. 359–374. Giacalone Ramat, A. and Hopper, P. J. (eds.) (1998) The Limits of Grammaticalization. Amsterdam/Philadelphia: John Benjamins. Givón, T. (1971) ‘Historical syntax and synchronic morphology: An archaeologist’s field trip’, Papers of the Chicago Linguistic Society, 7: 394–415. Givón, T. (1979) On Understanding Grammar. New York: Academic Press. Givón, T. (1991) ‘The evolution of dependent clause morpho-syntax in Biblical Hebrew’, in E. C. Traugott and B. Heine (eds.) Approaches to Grammaticalization: Types of Grammatical Markers. Amsterdam: John Benjamins. 257–310. Goldberg, A. E. (1995) Constructions: A Construction Grammar Account of Argument Structure. Chicago, IL: University of Chicago Press. Goldberg, A. E. (2006) Constructions at Work:The Nature of Generalization in Language. Oxford: Oxford University Press. Grice, H. P. (1975) ‘Logic and conversation’, in P. Cole and J. L. Morgan (eds.) Syntax and Semantics: Speech Acts. New York: Academic Press. 41–58. Günthner, S. and Mutz, K. (2004) ‘Grammaticalization vs pragmaticalization? The development of pragmatic markers in German and Italian’, in W. Bisang, N. P. Himmelmann, and B. Wiemer (eds.) What Makes Grammaticalization? A Look from Its Fringes and Its Components. Berlin/New York: Mouton de Gruyter. 77–107. Haspelmath, M. (1999) ‘Why is grammaticalization irreversible?’, Linguistics, 37: 1043–1068. Haspelmath, M. (2004) ‘On directionality in language change with particular reference to grammaticalization’, in O. Fischer, M. Norde, and H. Perridon (eds.) Up and Down the Cline: The Nature of Grammaticalization. Amsterdam/Philadelphia: John Benjamins. 17–44. Heine, B. (1993) Auxiliaries: Cognitive Forces and Grammaticalization. New York/Oxford: Oxford University Press. Heine, B. (2002) ‘On the role of context in grammaticalization’, in I. Wischer and G. Diewald (eds.) New Reflections on Grammaticalization. Amsterdam/Philadelphia: John Benjamins. 83–101. Heine, B., Claudi, U., and Hünnemeyer, F. (1991) Grammaticalization: A Conceptual Framework. Chicago/ London: University of Chicago Press. Heine, B. and Kuteva,T. (2002) World Lexicon of Grammaticalization. Cambridge: Cambridge University Press.

289

María José López-Couso and Elena Seoane Himmelmann, N. P. (2004) ‘Lexicalization and grammaticization: Opposite or orthogonal?’, in W. Bisang, N. P. Himmelmann, and B. Wiemer (eds.) What Makes Grammaticalization: A Look from Its Components and Its Fringes. Berlin/New York: Mouton de Gruyter. 21–42. Hopper, P. J. (1987) ‘Emergent Grammar’, Berkeley Linguistic Society, 13: 139–157. Hopper, P. J. (1991) ‘On some principles of grammaticization’, in E. C. Traugott and B. Heine (eds.) Approaches to Grammaticalization. Amsterdam/Philadelphia: John Benjamins. 17–35. Hopper, P. J. and Traugott, E. C. (2003) Grammaticalization. Cambridge: Cambridge University Press. Horne Tooke, J. (1857) Epea Pteroenta, or the Diversions of Purley. London: J. Johnson’s. Janda, R. D. (2001) ‘Beyond “pathways” and “unidirectionality”: On the discontinuity of language transmission and the counterability of grammaticalization’, Language Sciences, 23(2–3): 263–340. Joseph, B. D. (2004) ‘Rescuing traditional (historical) linguistics from grammaticalization theory’, in O. Fischer, M. Norde, and H. Perridon (eds.) Up and Down the Cline: The Nature of Grammaticalization. Amsterdam/Philadelphia: John Benjamins. 45–71. Jucker, A. H. (1997) ‘The discourse marker “well” in the history of English’, English Language and Linguistics, 1: 91–110. Jucker, A. H. (2012) ‘Linguistic levels: Pragmatics and discourse’, in A. Bergs and L. J. Brinton (eds.) Handbook of Linguistics and Communication Science: Historical Linguistics of English. Berlin: Mouton De Gruyter. 197–212. Jucker, A. H. and Taavitsainen, I. (eds.) (2010) Handbooks of Pragmatics: Historical Pragmatics. Berlin/New York: Mouton De Gruyter. Jucker, A. H. and Taavitsainen, I. (2013) English Historical Pragmatics. Edinburgh: Edinburgh University Press. König, E. and Vezzosi, L. (2008) ‘Possessive adjectives as a source of intensifiers’, in E. Seoane and M. J. López-Couso (eds.) Theoretical and Empirical Issues in Grammaticalization. Amsterdam/Philadelphia: John Benjamins. 183–206. Kranich, S. and Breban, T. (eds.) (2015) ‘The late stages of grammaticalization’ [Special issue], Language Sciences, 47. Kuryłowicz, J. (1965) ‘The evolution of grammatical categories’, Diogenes, 13(51): 55–71. Kytö, M. (2010) ‘Data in historical pragmatics’, in A. H. Jucker and I. Taavitsainen (eds.) Handbooks of Pragmatics: Historical Pragmatics. Berlin/New York: Mouton De Gruyter. 33–67. Lass, R. (2000) ‘Remarks on (uni)directionality’, in O. Fischer, A. Rosenbach, and D. Stein (eds.) Pathways of Change. Grammaticalization in English. Amsterdam/Philadelphia: John Benjamins. 207–227. Lehmann, C. (1982) Thoughts on Grammaticalization: A Programmatic Sketch, vol. I. Cologne: Universität zu Köln, Institut für Sprachwissenschaft. [Pub. as Thoughts on Grammaticalization. Munich: Lincom Europa, 1995]. Lehmann, C. (1985) ‘Grammaticalization: Synchronic variation and diachronic change’, Lingua e Stile, 20(3): 303–318. Lehmann, C. (1992) ‘Word order change by grammaticalization’, in M. Gerritsen and D. Stein (eds.) Internal and External Factors in Syntactic Change. Berlin/New York: Mouton De Gruyter. 395–416. López-Couso, M. J. (2010) ‘Subjectification and intersubjectification’, in A. H. Jucker and I. Taavitsainen (eds.) Handbooks of Pragmatics: Historical Pragmatics. Berlin/New York: Mouton De Gruyter. 127–163. López-Couso, M. J. and Méndez-Naya, B. (2015) ‘Epistemic/evidential markers of the type verb + complementizer: Some parallels from English and Romance’, in A. D. M. Smith, G. Trousdale, and R. ­Waltereit (eds.) New Directions in Grammaticalization Research (Studies in Language Companion Series 166). Amsterdam/Philadelphia: John Benjamins. 93–120. López-Couso, M. J. and Seoane, E. (eds.) (2008) Rethinking Grammaticalization: New Perspectives (Typological Studies in Language 76). Amsterdam/Philadelphia: John Benjamins. Matisoff, J. A. (1991) ‘Areal and universal dimensions of grammatization in Lahu’, in E. C. Traugott and B. Heine (eds.) Approaches to Grammaticalization: Focus on Theoretical and Methodological Issues. Amsterdam/ Philadelphia: John Benjamins. 383–453. McMahon, A. M. S. (1994) Understanding Language Change. Cambridge: Cambridge University Press. Meillet, A. (1912) ‘L’évolution des formes grammaticales’, Scientia (Rivista di Scienza), 12–26(6): 384–400. Méndez-Naya, B. (2014) ‘Out of the spatial domain: “Out”-intensifiers in the history of English’, Folia Linguistica Historica, 35: 241–274. Méndez-Naya, B. and Pahta, P. (2010) ‘Intensifiers in competition: The picture from early English medical writing’, in I. Taavitsainen and P. Pahta (eds.) Early Modern English Medical Texts: Corpus Description and Studies. Amsterdam/Philadelphia: John Benjamins. 193–215.

290

Grammaticalisation Narrog, H. and Heine, B. (eds.) (2011) The Oxford Handbook of Grammaticalization. Oxford: Oxford University Press. Nevala, M. (2010) ‘Politeness’, in A. H. Jucker and I. Taavitsainen (eds.) Handbooks of Pragmatics: Historical Pragmatics. Berlin/New York: Mouton De Gruyter. 419–450. Newmeyer, F. J. (1998) Language Form and Language Function. Cambridge, MA: MIT Press. Norde, M. (2009) Degrammaticalization. Oxford: Oxford University Press. Norde, M. (2010) ‘Degrammaticalization: Three common controversies’, in K. Stathi, E. Gehweiler, and E. König (eds.) Grammaticalization: Current Views and Issues. Amsterdam/Philadelphia: John Benjamins. 123–150. Nordlinger, R. and Traugott, E. C. (1997) ‘Scope and the development of epistemic modality: Evidence from “ought to” ’, English Language and Linguistics, 1: 295–317. Núñez-Pertejo, P. (2013) ‘From degree adverb to response token: “Absolutely” in Late Modern and Contemporary British and American English’, Neuphilologische Mitteilungen, 114: 245–273. Seoane, E. and López-Couso, M. J. (eds.) (2008) Theoretical and Empirical Issues in Grammaticalization. ­Amsterdam/Philadelphia: John Benjamins. Stathi, K., Gehweiler, E., and König, E. (eds.) (2010) Grammaticalization: Current Views and Issues. Amsterdam/Philadelphia: John Benjamins. Tabor, W. and Traugott, E. C. (1998) ‘Structural scope expansion and grammaticalization’, in A. Giacalone Ramat and P. J. Hopper (eds.) The Limits of Grammaticalization. Amsterdam/Philadelphia: John ­Benjamins. 229–272. Tagliamonte, S. (2008) ‘So different and pretty cool! Recycling intensifiers in Toronto, Canada’, English Language and Linguistics, 12(2): 361–394. Trask, R. L. (1996) Historical Linguistics. London/New York/Sydney/Auckland: Arnold. Traugott, E. C. (1982) ‘From propositional to textual and expressive meanings: Some semantic-pragmatic aspects of grammaticalization’, in W. P. Lehmann and Y. Malkiel (eds.) Perspectives on Historical Linguistics. Amsterdam/Philadelphia: John Benjamins. 245–271. Traugott, E. C. (1995) ‘Subjectification in grammaticalization’, in D. Stein and S. Wright (eds.) Subjectivity and Subjectivisation. Cambridge: Cambridge University Press. 31–54. Traugott, E. C. (2001) ‘Legitimate counterexamples to unidirectionality’, paper presented at Freiburg University, October 17, Freiburg, Germany. Available from: http://web.stanford.edu/~traugott/papers/ Freiburg.Unidirect.pdf (accessed November 2016). Traugott, E. C. (2002) ‘From etymology to historical pragmatics’, in D. Minkova and R. Stockwell (eds.) Studies in the History of English: Millennial Perspectives. Berlin: Mouton De Gruyter. 19–49. Traugott, E. C. (2003a) ‘Constructions in grammaticalization’, in B. D. Joseph and R. D. Janda (eds.) The Handbook of Historical Linguistics. Oxford: Blackwell. 624–647. Traugott, E. C. (2003b) ‘From subjectification to intersubjectification’, in R. Hickey (ed.) Motives for Language Change. Cambridge: Cambridge University Press. 124–139. Traugott, E. C. (2007) ‘The concepts of constructional mismatch and type-shifting from the perspective of grammaticalization’, Cognitive Linguistics, 18: 523–557. Traugott, E. C. (2008) ‘The grammaticalization of NP of NP patterns’, in A. Bergs and G. Diewald (eds.) Constructions and Language Change. Berlin/New York: Mouton De Gruyter. 23–45. Traugott, E. C. (2010a) ‘Grammaticalization’, in A. H. Jucker and I.Taavitsainen (eds.) Handbooks of Pragmatics: Historical Pragmatics. Berlin/New York: De Gruyter Mouton. 97–126. Traugott, E. C. (2010b) ‘Revisiting subjectification and intersubjectification’, in K. Davidse, L.Vandelanotte, and H. Cuyckens (eds.) Subjectification, Intersubjectification and Grammaticalization. Berlin/New York: De Gruyter Mouton. 29–70. Traugott, E. C. and Dasher, R. B. (2002) Regularity in Semantic Change. Cambridge: Cambridge University Press. Traugott, E. C. and Heine, B. (eds.) (1991) Approaches to Grammaticalization. Amsterdam/Philadelphia: John Benjamins. Traugott, E. C. and König, E. (1991) ‘The semantics-pragmatics of grammaticalization revisited’, in E. C. Traugott and B. Heine (eds.) Approaches to Grammaticalization: Theoretical and Methodological Issues. Amsterdam/Philadelphia: John Benjamins. 189–218. Traugott, E. C. and Trousdale, G. (2010) ‘Gradience, gradualness and grammaticalization – how do they intersect?’, in E. C. Traugott and G. Trousdale (eds.) Gradience, Gradualness and Grammaticalization. Amsterdam/Philadelphia: John Benjamins. 19–44.

291

María José López-Couso and Elena Seoane Trousdale, G. (2010) ‘Issues in constructional approaches to grammaticalization in English’, in K. Stathi, E. Gehweiler, and E. König (eds.) Grammaticalization: Current Views and Issues. Amsterdam/Philadelphia: John Benjamins. 51–72. Trousdale, G. (2012a) ‘Grammaticalization, lexicalization and constructionalization from a cognitive-pragmatic perspective’, in H. J. Schmid (ed.) Cognitive Pragmatics. Berlin: De Gruyter Mouton. 535–560. Trousdale, G. (2012b) ‘Grammaticalization, constructions and the grammaticalization of constructions’, in K. Davidse, T. Breban, L. Brems, and T. Mortelmans (eds.) Grammaticalization and Language Change: New Reflections. Amsterdam/Philadelphia: John Benjamins. 167–198. Van linden, A.,Verstraete, J. C., and Davidse, K. (eds.) (2010) Formal Evidence in Grammaticalization Research. Amsterdam/Philadelphia: John Benjamins. Von der Gabelentz, G. (1891/1901) Die Sprachwissenschaft: Ihre Aufgaben, Methoden und Bisherigen Ergebnisse. Leipzig: Weigel Nachfolger. Von Humboldt, W. (1825) ‘Über das Entstehen der Grammatikalischen Formen und Ihren Einfluß auf die Ideenentwicklung’, Abhandlungen der Königlichen Akademie der Wissenschaften zu Berlin: 401–430. Wischer, I. (2000) ‘Grammaticalization versus lexicalization: “Methinks” there is some confusion’, in O. Fischer, A. Rosenbach, and D. Stein (eds.) Pathways of Change. Grammaticalization in English. Amsterdam/ Philadelphia: John Benjamins. 355–370. Wischer, I. and Diewald, G. (eds.) (2002) New Reflections on Grammaticalization. Amsterdam/Philadelphia: John Benjamins.

292

Cognition and pragmatics

23 METAREPRESENTATION Nicholas Allott

Introduction Humans and other thinking beings can represent the world. My cat knows which cupboard her food is in: she has a mental representation of that state of affairs. Given that I believe that she knows where her food is, I have a mental representation of her mental representation. I also know where it is, and I know that I know that: I have a mental representation of my own mental representation about the food’s location.What is more, using language I can describe all of these facts about representations. The capacity to think and talk about representations, that is, to represent representations or metarepresent, is a species characteristic of human beings. All developmentally normal human beings, across all cultures, metarepresent, and with considerable facility; and metarepresentation is a central property of both human language use and human thought. Quotation, both direct and indirect, as in (1a and 1b), respectively, is metarepresentational: one representation is embedded in another. (1) a. John said, ‘Mary’s arrived!’ b. John said that Mary had arrived. Some metarepresentational thoughts are described in (2). In each case the embedded sentence (the ‘that’-clause) represents the world as being a certain way, and is embedded under an ‘attitude’ verb (these include think, believe, doubt, desire, intend, wish) which describes the relation that the referent of the subject of the sentence (John, in these cases) has to the lower-level representation. (2) a. John believes that Mary has arrived. b. John doubts that Mary has brought wine. Human communication is essentially metarepresentational according to the prevailing, broadly Gricean view. When a speaker produces an utterance, the hearer’s task is to work out what the speaker intended to convey. Thus the audience of the utterance starts with a mental

295

Nicholas Allott

metarepresentation like the one in (3a) and on that basis arrives at a mental metarepresentation like the one in (3b). (3) a. John said, ‘It’s snowing here.’ b. John means (i.e. intends me to think that he intends me to think) that it is snowing in Oslo now. In the sense that we are concerned with here, a representation is defined as a state that is about or of something. This property of ‘aboutness’ – called intentionality in philosophy – is generally agreed to be possessed by some or all mental states. Previously, a metarepresentation has been defined as a representation of a representation. Another definition in the literature is that a metarepresentation is a higher-order representation with a lower-order representation embedded within it (Wilson 2000: 411).

Higher-order metarepresentation The metarepresentations in (1), (2) and (3a) are first-order: they consist of one level of metarepresentation of a representation. However, metarepresentations can themselves be embedded in other representations to form higher-order metarepresentations, as in (3b), and in (4), where numbered parentheses have been used to mark the different levels. (4) a. [2Mary wrote [1that John said [0that she had arrived]]]. b. [3Mary said [2that she once wrote [1that John said [0that she had arrived]]]]. c. [2Mary knows that [1John knows [0that she has arrived]]]. d. [3John suspects [2that Mary knows [1that he doubts [0that she has brought wine]]]]. Language supports metarepresentations of arbitrarily many orders, although processing higherorder metarepresentations is subject to capacity limitations on memory and attention: ‘Human limits on embedding are not impressive: only about four steps make our species uncomfortable’ (Premack and Woodruff 1978: 515–516).

Outline of this chapter This chapter is concerned with metarepresentation as it relates to communication and language use. As the section ‘Metarepresentation and communication’ explains, the leading pragmatic theories inherit from the work of the philosopher Paul Grice an account of communication as inference about what the speaker intended to convey, based on what the speaker uttered. Such accounts presuppose that both speaker and hearer have considerable facility with higherorder metarepresentations like (3b). This picture of communication, most thoroughly explored in Sperber and Wilson’s relevance theory, construes it as a kind of metapsychology: an ability to infer others’ mental states. The ability to infer others’ mental states (beliefs, desires, intentions etc.) based on their ordinary, non-communicative behaviour is known as ‘mindreading’ or ‘theory of mind’ and has been extensively studied in recent decades. The relation between mindreading and utterance interpretation is discussed in this section of this chapter. ‘Metarepresentational use of language’ provides a brief survey of the extensive cross-disciplinary literatures on quotation and metalinguistic negation, the clearest cases of language used to metarepresent. It also describes the now widely accepted claim made by Sperber and Wilson that verbal irony is a tacitly metarepresentational use of language. Finally it sets out the attempt made 296

Metarepresentation

in relevance theory to provide a unified account covering interrogative sentences, in addition to quotation, metalinguistic negation and verbal irony, in terms of the interpretive use of language: that is, use of language to represent an utterance or a thought other than the speaker’s own.

Metarepresentation and communication It is a fundamental assumption of modern pragmatics that on the basis of an utterance the hearer attributes to the speaker an intention to convey a certain meaning.The hearer starts from a metarepresentation of what the speaker said and ends with a metarepresentation of the communicator’s meaning, as in (3).This model of communication derives primarily from Grice.The role of metarepresentation in his work on meaning and conversation is explained in ‘Speaker meaning’ and ‘Implicatures’. Relevance theory has developed the most comprehensive picture of the role of metarepresentation in communication, putting it in the context of work in psychology on ‘theory of mind’ or ‘mindreading’, the human ability to attribute thoughts to others on the basis of their behaviour. Given that utterance interpretation is the attribution of thoughts to the speaker on the basis of her communicative action, it is a form of mindreading. The sections ‘Infinite regresses of metarepresentations?’ and ‘The psychology of communication: relevance theory and mindreading’ set out the role of metarepresentation in communication according to relevance theory; ‘Theory of mind and communication’ discusses the relation between mindreading and communication.

Speaker meaning Two strands of Grice’s work are concerned with communication: his work on speaker meaning and his theory of conversation. Both make essential use of metarepresentations. Grice defined speaker meaning in terms of three intentions of the utterer, U: ‘By uttering x, U meant something’ is true if for some audience A, U uttered x intending: i) A to produce some particular response r. ii) A to recognise that U intends (i), and iii) A’s recognition that U intends (i) to function, in part, as a reason for (i). (Grice 1969: 151; see also Strawson (1964) on Grice’s (1957) original proposal.) Consider the first intention, (i), to get A to produce a certain response, r. Utterances of indicative sentences such as (5a) aim at getting the audience to entertain a belief, while utterances of imperatives such as (5b) aim at getting the hearer to intend to perform a certain action (Grice 1989: ch. 6; Neale 1992: 546, fn 53). So it seems that the intended response r is (or at least includes) a mental state (belief or intention) and thus intention (i) is always a metarepresentation. (5) a. Bertrand’s book is on the table. b. Shut the window! The second and third intentions are also metarepresentational, if ‘recognising’ that a person has intention X entails mentally representing her as intending X. On that assumption, (ii) entails that the speaker intends the hearer to represent the speaker as intending (i). That is, the utterer 297

Nicholas Allott

intends the audience to think that she (U) intends him (A) to have response r (i.e. to entertain a certain belief or intention). Why should an utterer not aim at getting the audience to believe a certain proposition directly, without all of this metarepresentation? Grice envisaged such cases and argued that they were not examples of speaker meaning, but of a kind of manipulation. Suppose that the audience, A, is a detective investigating a crime and U knows the identity of the culprit. Wanting to keep herself out of it, U might contrive a false clue to the real criminal’s identity, dropping his handkerchief at the crime scene for A to find. In such a case, U intends A to come to believe p, but does not intend that A come to believe that she, U, intends him to believe p. Grice’s point was that such examples lack something that is essential to speaker meaning: namely that the utterer openly intends the audience to come to a certain conclusion. That is why he defines speaker meaning in terms of an intention that the speaker intends the hearer to recognise.

Implicatures Grice’s theory of conversation attempts to explain how it is possible for speakers to mean something more than, or different from, what the words they utter literally mean, and for hearers to understand them when they do this, as in John’s indirect reply in (6a) and the attested example of verbal irony in (6b): (6) a. Mary:  Would you like a bit of this mutton stew? John:  You’re forgetting that I’m a vegetarian. b. Germany also decided to shut down its nuclear power plants after the Fukushima crisis, due to the imminent risk of tsunamis in Bavaria. (From an article by George Monbiot in The Guardian, 5 February 2013) In cases like these, Grice proposed that the speaker intends to convey a conversational implicature, that is, something i) other than what she literally says, which ii) she intentionally implies in (or by) saying what she does. Grice’s introduction of conversational implicatures includes an account of how the hearer could derive them, on the assumption that the speaker is rational and cooperative, and thus is conforming with certain expectations about rational cooperative speakers: that they be relevant, do not say things they believe to be false and so forth (Grice 1967). Grice’s ‘working-out schema’ is given later, where p is the proposition expressed by the sentence uttered, and q is the proposition implicated. Obviously, it is thoroughly metarepresentational: A general pattern for the working out of a conversational implicature might be given as follows: ‘He has said that p; there is no reason to suppose that he is not observing the maxims, or at least the CP [Cooperative Principle]; he could not be doing this unless he thought that q; he knows (and knows that I know that he knows) that I can see that the supposition that he thinks that q is required; he has done nothing to stop me thinking that q; he intends me to think, or is at least willing to allow me to think, that q; and so he has implicated that q.’ (Grice 1967: 31) The psychological reality of the working out schema has often been challenged. Is it plausible that hearers reason like this, or that speakers expect them to? In making and understanding

298

Metarepresentation

utterances, we are not aware of having to engage in explicit, reflective reasoning or to entertain such complex metarepresentational thoughts, although we may do so in some cases. Such criticisms miss their mark, because Grice’s concern was not that speakers and hearers explicitly reason like this, but that we could. That is, it must be possible that the hearer could work out that an implicature is present (Grice 1967: 31). Elsewhere Grice argues that in reasoning we can either go ‘the hard way’ or ‘the quick way’, where the quick way, ‘made possible by habituation and intention’, is to intuitively grasp the conclusion, skipping intermediate steps needed to spell out the inference fully (Grice 2001: 17). He did not claim that hearers have to consciously or explicitly entertain such complex metarepresentational thoughts as, for example, ‘he knows (and knows that I know that he knows) that I can see that the supposition that he thinks that q is required.’

Infinite regresses of metarepresentations? Another point on which Grice’s views have been thought psychologically unrealistic is the implication that speaker and hearer have to entertain infinite series of increasingly higher-order metarepresentations of speaker’s intentions or of each other’s knowledge. Grice’s definition of speaker meaning tries to capture the idea that communication is overt or transparent in a certain sense. As discussed earlier, this requires not only i) that the speaker intend the hearer to entertain p, but also ii) that the speaker intends the hearer to recognise this intention. But it may be that the speaker also has to iii) intend the hearer to recognise the intention in (ii), and iv) to intend the hearer to recognise the intention in (iii) and so on, with no limit (Strawson 1964; Grice 1969: 156–157; Schiffer 1972: ch. 2; Neale 1992: 549–550). The argument for this requirement turns on certain complex counterexamples (from Schiffer and Strawson) to Grice’s definition of meaning, which aim at showing that there are cases where intentions (i)–(iii) are present but which should not count as cases of meaning. One response to these counterexamples was Schiffer’s (1972) suggestion that intentions (i)– (iii) must be mutually known to hearer and speaker. That is, the hearer and speaker must both know them, each know that the other knows them, and so on, without limit: an infinite series of metarepresentations. Background knowledge might also need to be mutually known. Clark and Marshall (1981) argued that mutual knowledge is required for speaker and hearer to correctly coordinate on a referent. If the speaker does not know that the hearer knows a certain film is showing, she can’t be sure that when she says ‘The film that is showing’ she will be correctly understood. But the speaker’s knowing that the hearer does know which film is showing is not enough. If the hearer does not know that the speaker knows that he (the hearer) knows which film is showing, he can’t be sure that she is referring to Monkey Business rather than some other film which she wrongly thinks that he thinks is showing. By similar arguments it can be shown that there is no principled limit to the order of the metarepresentations required. In early linguistic pragmatics, considerable effort was expended on a psychologically realistic solution to the problem of mutual knowledge (Clark and Carlson 1981; Smith 1982; Sperber and Wilson 1986, 1990; Perner and Garnham 1988). Interest in these problems has since faded, although Sperber and Wilson’s solution (1986: 38–50), that communication depends not on mutual knowledge, but on a weaker condition, mutual manifestness, which does not require speaker or hearer to entertain an infinite series of metarepresentations, became one of the foundations of relevance theory.

299

Nicholas Allott

The psychology of communication: relevance theory and mindreading Relevance theory treats metarepresentation as central to communication in two ways (Noh 2000: 4; Wilson 2000: 424). First, it claims that utterance interpretation is inference from a metarepresented utterance to a metarepresentational intention attributed to the speaker, and it has made proposals about the order of metarepresentations involved in communication, and about the relation of this capacity to general mindreading abilities. These are the topics of the remainder of this section. Second, relevance theory tries to give a unified account of various types of use of language as interpretive: that is, representing thoughts or utterances attributed to others. This is dealt with in the next section: ‘Metarepresentational use of language’. Like most current theories of pragmatics, relevance theory differs from Grice in proposing a much greater role in utterance interpretation for inference, extending to disambiguation, reference resolution and other processes involved in inferring the proposition expressed. Thus, in contrast to Grice’s theory of conversation, where a metarepresented proposition – what the speaker asserted – is the starting point for inference, relevance theory proposes that utterance interpretation starts from a metarepresentation of the speaker’s utterance, as in (3a). According to relevance theory, the communicative faculty is a module specialised for inferring from an utterance the speaker’s informative and communicative intentions, where these are defined as follows: The informative intention: The intention to inform an audience of something. The communicative intention: The intention to inform the audience of one’s informative intention. (Wilson and Sperber 2004: 611) Suppose that a speaker utters ‘It’s snowing here’, with the following informative intention, which is a first-order metarepresentation: an attitude to a belief (cf. Sperber 1994): (7) John should believe [that it is snowing in Oslo]. Therefore the hearer, John, if he grasps the informative intention, entertains a second-order metarepresentation: (8) She intends [me to believe [that it is snowing in Oslo]]. Note that if the hearer grasps the informative intention, then communication has succeeded, since he has arrived at what it was that the speaker wanted to convey to him. This, then, is what speakers generally aim at. So, in general, a speaker has a third-order intention that her hearer recognise her informative intention. This is the communicative intention. In the example, this is the metarepresentation in (9): (9) John should believe [that [I intend [him to believe [that it is snowing in Oslo]]]]. If the hearer becomes aware of the speaker’s communicative intention, he forms a fourth-order metarepresentation – in this case, the one in (10): (10) She intends [me to believe [that she intends [me to believe [that it is snowing in Oslo]]]]. 300

Metarepresentation

Theory of mind and communication Do [we] have a meta-representational ability? Do birds fly? Do fish swim? Humans can no more refrain from attributing intentions than they can from batting their eyelids. (Sperber 1994: 187)

Theory of mind or ‘mindreading’ (Baron-Cohen 1995) is the ability all developmentally normal adult human beings possess to attribute to others mental states such as beliefs, desires and intentions – that is, to form metarepresentations of others’ mental states – on the basis of their behaviour.We do it automatically and mostly without explicit reasoning. Suppose you see a man walk towards a door, pulling a key from his pocket. He puts it into the keyhole and applies turning force. It does not turn. He pulls it out of the keyhole and looks at it with a puzzled expression. We cannot help but metarepresent his thoughts: he wanted to open the door and wrongly thought that the key he had was the right one. Sperber and Wilson have pointed out that on a broadly Gricean view of communication, utterance interpretation is a type of mindreading: the hearer attributes a communicative intention to the speaker on the basis of her utterance. Support for this view is provided by evidence that mindreading and utterance interpretation pattern together to some extent both in childhood development and in the way that they break down due to pathology (Sperber and Wilson 2002: 7–8). In autistic spectrum disorders, there are deficits in both general mindreading and nonverbal communication (Perner et al. 1989; Langdon et al. 2002). (Although Chevallier et al. 2010 and Chevallier et al. 2011 show that some communicative impairment in ASD is not due to ToM deficit.) Mindreading ability correlates with communicative abilities including word learning (Bloom 2000; Happé and Loth 2002) and reference resolution (Mitchell et al. 1999). Standard tests for mindreading assess the ability to attribute false beliefs, since this reveals the ability to attribute a different belief from one’s own. For some years, it was thought that children below about four years old could not mindread since they failed the standard (verbal) false belief task. That seemed problematic for the view that utterance interpretation is a type of mindreading, given that younger children can make and understand utterances. However, more sophisticated experiments have demonstrated that children attribute beliefs well before they pass the verbal task (Onishi and Baillargeon 2005; Southgate et al. 2007; Surian et al. 2007). There is some evidence that utterance interpretation uses a different mechanism from general mindreading (Wilson 2000). Adults have difficulty processing metarepresentations of more than around four orders, but we are quite capable of talking and thinking about embedded speaker meanings, as in (11). If each layer of speaker meaning is processed as a fourth-order metarepresentation, then the sentence in (11) should be unprocessable, since it would require entertaining eighth- and twelfth-order metarepresentations. (11) I didn’t mean that you said that the weather in Oslo is always bad; what I meant was that you strongly implied it, or that what you meant was that that was what Mary meant. It also seems introspectively implausible that all these layers of metarepresentation are actively entertained every time we understand an utterance. Grice’s definition of speaker meaning that implies that (e.g.) (12b) is (part of) the correct analysis of (12a) was a considerable theoretical achievement, not a simple consultation of intuitions. (12) a. John means that it is snowing in Oslo now. b. John intends me to think that he intends me to think that it is snowing in Oslo now. 301

Nicholas Allott

All of this suggests that we normally understand speakers without unpacking the notion of meaning. That is, the mental representation of a speaker meaning is more like (12a) than (12b). This in turn suggests that hearers do not work out for each utterance that it is an attempt to get them to think that the speaker intends them to think something. It seems more likely that we have a specialised procedure that works out speaker meaning and makes representations like (12a) available to the rest of cognition.

Metarepresentational use of language The previous section has explored the idea that the intentions underlying communication are metarepresentational. The content of the speaker’s meaning, that is, what the speaker intends to communicate, can also be metarepresentational. This section looks at forms of language and types of language use that have been analysed as metarepresentational, from the obvious cases of quotation and metalinguistic negation, to verbal irony, which is now generally seen as essentially allusive, and concludes with relevance theory’s attempt to provide a unified account of these phenomena, together with interrogatives, as instances of interpretive use.

Quotation In both direct and indirect quotation, as in (13a) and (13b), respectively, one representation is embedded in another and the embedded sentence is a representation of something uttered or thought: thus the whole is a metarepresentation. Direct quotation highlights the form of what is quoted, indirect, the semantic content. (13) a. John said to me, ‘You’ll surely come up with something to save the company.’ b. John said that I would think of a plan. It is now usual to distinguish at least two further types of quotation: free indirect discourse (McHale 1978) and ‘mixed’ quotation, as in (14a) and (14b), respectively. There is also mention or ‘pure’ quotation, in which an abstract type is represented, as in (14c). (14) a. John’s irrationally optimistic, as usual. I’m surely going to come up with a plan to save the company! b. According to John, I’ll ‘surely come up with’ a plan. (Cf Wilson 2000: 413) c. ‘come up with’ is a phrasal verb which means devise or produce an idea, plan or similar and has the main stress on the second word. (Cf Noh 2000: 7, 11; Wilson 2000: 413) ‘Pure’ and ‘mixed’ quotation are explicitly metarepresentational. Free indirect discourse is not. In (14a), the hearer has to infer that the speaker does not herself endorse what her second sentence seems to say, and that she is tacitly attributing a thought or utterance to John. Another type of tacitly metarepresentational quotation, free direct speech, is defined as direct quotation where the verb of saying is not present (Noh 2000: 15). This is also known as ‘zero quotation’ (Mathis and Yule 1994: 63; see also Buchstaller, Chapter 30, this volume). There are several overlapping literatures on quotation. It has attracted attention in philosophy (e.g., Davidson 1979/1984), linguistic formal semantics (e.g., Partee 1973), psycholinguistics (e.g., Wade and Clark 1993), literary studies (e.g., Leech and Short 1981: ch. 10) and corpus linguistics and stylistics (Semino and Short 2004), as well as pragmatics (Coulmas 1986; Clark and Gerrig 1990; Noh 2000: ch. 1; Wilson 2000: 424–437). Cappelen and Lepore (2012) is a 302

Metarepresentation

recent survey from the point of view of philosophy, while Noh (2000: ch. 1), although slightly out of date, is the best general overview. Noh (2000) and Wilson (2000) are the best guides to the connections between quotation and metarepresentation. Philosophy and formal semantics have focussed on direct quotation and pure mention, plus indirect quotation seen as a propositional attitude, while pragmatics and literary studies have been interested in all types except pure mention. Recently there has been some dialogue between these camps, with debate over whether a ‘semantic’ theory of quotation is possible, that is, one that treats the various forms of quotation as contributing compositionally to truth-conditions. Quotation (other than pure quotation) is often called reported speech, but these terms are not synonymous, since not all quoted utterances are reports. That is, quotation is always metarepresentational, but the attitude to what is metarepresented can vary. Consider the examples in (15): (15) a. John: Mary: b. John: Mary:

Are you a student? Am I a student? (Noh 2000: 16) Leave me alone. Leave you alone! (Noh 2000: 16)

Mary’s utterances are not reports but echoes of John’s utterances. In (15a), she might be wondering aloud, reposing John’s question to herself; or she might be echoing John’s utterance in order to ridicule it.The examples in (15) are free indirect speech, but indirect speech and mixed quotation can also be used echoically (Noh 2000: 16–17) as in (16a) and (16b), respectively: I’m the best damn cook in England. (16) Peter: a. Mary: [You say/You think] you’re the best cook in England. (Cf. Noh 2000: 17) b. Mary: You say you’re ‘the best damn cook’ in the country. This links quotation to verbal irony, understood as echoic allusion (see later).

Metalinguistic negation Metalinguistic negation involves the representation of an utterance or another representation such as a thought and its rejection on grounds other than its falsity, as in the examples in (17). The phenomenon was noted by Fillmore (1971) and Ducrot (1974) and brought to wider attention by Horn (1985; 1989: ch. 6). In uttering (17a), the speaker objects to an utterance or thought that represents Mary as sometimes late; in (17b), the negation targets the speech sounds used; in (17c), the objection is to representing Poul Anderson as merely ‘one of the best’. As this example demonstrates, notwithstanding its name, metalinguistic negation does not require the use of ‘not’ nor indeed any morphologically overt negation (Noh 2000: 117–118; Carston 2002: 300–301). (17) a. Mary isn’t sometimes late – she is always late. (Noh 2000: 109) b. These aren’t [təmɑ:təuz]; they’re [təmeɪɾouz]. (Cf. Carston 2002: 300) c. Everybody knows that Poul Anderson is one of the best: they’re wrong – he’s better! (Jerry Pournelle, quoted on the back cover of a 1983 edition of Poul Anderson’s Mirkheim: Langford 2010) Several diagnostics have been suggested to distinguish metalinguistic negation from ordinary negation. The most clear-cut is lexical incorporation (e.g. not happy → unhappy). Ordinary 303

Nicholas Allott

negation can lexically incorporate, as in (18a), but metalinguistic negation cannot, as (18b) demonstrates. Another hallmark of metalinguistic negation is that there is often contrastive stress on the locus of the objection and correction, as indicated by the italics in (17a). (18) a. She’s [not happy/unhappy]. b. She’s [not happy/*unhappy] – she’s ecstatic. Metalinguistic negation has been the focus of several connected debates. Is negation linguistically ambiguous between truth-conditional and metalinguistic senses, or is there just one negation operator used for both ‘ordinary’ and metalinguistic negation? Is metalinguistic negation truth-functional? Can it operate only on actual prior utterances; or possible utterances; or utterances and thoughts, as assumed in the definition given earlier (see Carston 2002: 296–297)? The most developed account of metalinguistic negation as a form of metarepresentational use of language is found in a series of publications by Carston and Noh (Carston 1996, 1998, 1999, 2002: 291–302; Carston and Noh 1996; Noh 2000: ch. 3). Their main claim is that it involves echoic use: part of a potential or actual utterance or thought is metarepresented and the result is what is negated. For example, the linguistic material ‘these aren’t [təmɑ:təuz]’ in (17b) is pragmatically enriched to be understood as something like these aren’t what is properly called [təmɑ:təuz]. This is propositional, so the ordinary, truth-functional negation operator can be applied to it. With a similar enrichment of ‘they’re [təmeɪɾouz]’ we then have (19) as what is expressed by a (typical) use of (17b). (19) It is not the case that (these are what is properly called [təmɑ:təuz]); they are what is properly called [təmeɪɾouz]. This account of metalinguistic negation has no need to assume that ‘not’ is ambiguous between a sense that operates on propositions and a sense that operates on other types of material. This means that the account requires no extra stipulations to give an analysis of cases like (17c) where the word ‘not’ is not present. A further advantage is the theoretical unification of metalinguistic negation with other cases of metarepresentational use that do not involve negation but that are otherwise similar. An example is the use of the sentence in (20) to mean the person who has been described as/thought of as ‘the intelligent bloke’ certainly is stupid: (20) The intelligent bloke certainly is stupid. (cf. Carston 2002: 301)

Irony Sperber and Wilson’s account of verbal irony (Sperber and Wilson 1981; Wilson and Sperber 1992, 2012; Wilson 2006, 2009) and recent alternative theories analyse irony as metarepresentational use of language. It is not possible to define verbal irony in a theory-neutral way, but there is considerable agreement about paradigm examples, such as (6b) and insincere utterances of the sentences in (21): (21) a. What lovely weather! [said in a hailstorm] b. I love your tie. The classical characterisation of verbal irony, still given by most dictionaries, is the use of words or a sentence to mean the opposite of what they normally, or conventionally, mean. Thus (21a) 304

Metarepresentation

might be taken to express ‘What horrible weather’ and (21b) to express ‘I hate your tie.’ On this account, there is nothing essentially metarepresentational about irony. But the classical definition is problematic. Grice pointed out that one cannot freely mean the opposite of what one says. His example in (22) is an utterance which, intuitively, is not successful as irony: (22) Look, that car has all its windows intact. [said of a car with broken windows] (Grice 1967: 53) Sperber and Wilson propose that in irony i) the words uttered are used to represent a thought or another utterance (actual or possible) and ii) to express a negative (‘dissociative’) attitude to the thought or utterance represented. An additional feature of fully ironic utterances is iii) that both (i) and (ii) are tacit, in the sense that they are not explicitly spelled out in the linguistic form of what is uttered. According to this analysis, an ironic utterance of (21a), if echoing and mocking an expectation, could be roughly glossed as ‘It was ridiculous to expect lovely weather,’ or – if mocking a prior utterance – as ‘It was ridiculous to say that there would be lovely weather.’ This account ‘treats ironical utterances as forming a natural class with other types of interpretive, attributive or echoic use’ (Wilson 2006: 1736).That fits well with irony’s affinity for quotation, semi-quotation and allusion. On this understanding of irony, not only the utterances in (21) are ironic, but also the use of ‘intelligent bloke’ in (20) and the second sentence in (14a). The account correctly predicts that (22) will be felicitous in a context where it can be understood as echoing an expectation or utterance (Sperber and Wilson 1986: 240–241; Wilson 2006: 1728, 1732, 1735). The echoic account implies that irony makes greater demands of the hearer’s metarepresentational abilities than non-ironic utterances that are otherwise similar; and evidence indicates that the ability to comprehend irony correlates with fairly sophisticated metarepresentational ability (as measured by a second-order false belief task) and makes greater demands than the interpretation of metaphor (Happé 1993). For a summary of the experimental work and discussion in the light of recent developments in mindreading research, see Wilson (2009: 188–190). Rivals to Sperber and Wilson’s theory include the pretence account, a development of a suggestion made by Grice (1967:54) and further developed by Clark and Gerrig (1984). The claim is that the speaker of an ironic utterance is only pretending to perform a speech-act – for example, the speaker of (20b) is pretending to praise the hearer’s tie without really doing so – and generally there is in addition a negative or mocking attitude towards the speech act that is simulated. Recent pretence accounts of irony adopt from Sperber and Wilson the view that it is metarepresentational. These are allusional pretence accounts (Walton 1990; Kumon-Nakamura et al. 1995; for discussion, see Wilson 2009: 205–210). Thus the consensus recognises that irony is metarepresentational use of language, since: Both [the echoic and allusional pretence accounts] recognise that irony involves the attribution of a thought (or perspective, or point of view) to a specific person or type of person, or to people in general, and the expression of a dissociative attitude to the attributed thought. (Wilson 2006: 1736)

Interpretive and echoic use Sperber and Wilson (1986: 224–254) (cf. also Wilson and Sperber 1988; Wilson 2000) and Noh (2000) propose a unified account of metarepresentational uses of language. They claim that each 305

Nicholas Allott

utterance is relevant by representing another representation: every utterance resembles a thought of the speaker’s. That resemblance may be relevant because the thought resembles a state of affairs (they call this ‘descriptive use’), or because it resembles another thought or utterance (‘interpretive use’) (1986: 230–231). Echoic utterances, reported speech and irony are all cases of interpretive use. A second claim is that the relation between the utterance and the representation it represents is one of resemblance (not, in general, identity). Resemblance comes in degrees. Sperber and Wilson define interpretive resemblance as greater the more implications are shared in context by the two representations (1986: 224–254). Identity is the limiting case in which all implications are shared. Resemblances may be of different kinds. Wilson suggests that a difference between interpretive and metalinguistic resemblance underlies the different types of quotation: [T]ypically, direct quotation. . . . increases the salience of formal or linguistic properties, and indirect quotation. . . . increases the salience of semantic or logical properties. We might call these resemblances metalinguistic, on the one hand, and interpretive, on the other. . . . . Mixed quotation. . . . exploits both metalinguistic and interpretive resemblances, while reports of thought, and metarepresentations of thought in general, are typically interpretive. (Wilson 2000: 426) Previous sections have set out echoic-use theories of irony and metalinguistic negation. Noh also develops metarepresentational accounts of echo questions (1998; 2000: ch. 4) and metarepresentational uses in conditionals (2000: ch. 5), and Papafragou (1996) analyses metonymy as interpretive use. Wilson and Sperber (1988) give an account of imperative and interrogative sentences based on the two claims presented earlier, with one further assumption. In descriptive use, a thought may describe either an actual state of affairs or a desirable one; equally, in interpretive use a thought can be an interpretation of either an attributed thought or utterance or a desirable one. They argue that the imperative sentence type encodes a descriptive relation to a desirable state of affairs, while the interrogative sentence type encodes an interpretive relation to desirable thoughts or utterances, that is, relevant answers.

Conclusion Both thought and language have metarepresentational resources. There is consensus that human language and our metarepresentational abilities are closely linked, but no agreement about which is more fundamental. Debate has mostly focussed on evolutionary priority. Sperber argues that mindreading ability was a prerequisite for the emergence of human language, since our use of language depends on mindreading. Human language profoundly differs from animal communication systems such as bee dances and vervet monkey calls in that it provides only partial clues to the speaker’s meaning. If human linguistic communication was once purely a matter of coding and decoding, then ‘there is no reason to assume that our ancestors had the resources to become aware of the representational character of their signals, any more than bees or vervet monkeys do’ (Sperber 2000: 122). On the other hand, if human language always provided only partial clues to the speaker’s meaning, then interpretation has always depended on mindreading ability, so language cannot have preceded the ability to metarepresent. 306

Metarepresentation

However, an argument can be made for the priority of language, if we assume that complex thought is enabled by language and that metarepresentation requires recursion, the ability which enables us to construct embedded representations. Chomsky and colleagues have argued that recursion is the essential property of language (Hauser et al. 2002) and that natural language is the language of thought (Chomsky 2007: 22–26). It may be that there were no recursive thoughts before humans had language and a fortiori, no metarepresentation.

Suggestions for further reading Carruthers, P. and Smith, P. K. (eds.) (1996) Theories of Theories of Mind. Cambridge: Cambridge University Press. A useful collection of papers on mindreading. Noh, E. J. (2000) Metarepresentation: A Relevance-Theory Approach. Amsterdam: John Benjamins. An ecumenical and interesting guide to the metalinguistic resources of natural language, including quotation and metalinguistic negation. Papafragou, A. (2002) ‘Mindreading and verbal communication’, Mind and Language, 17: 55–67. A technical but still accessible paper on the role of mindreading in communication, with a focus on children’s abilities. Sperber, D. (1995) ‘How do we communicate?’, in J. Brockman and K. Matson (eds.) How Things Are: A Science Toolkit to the Mind. New York: W. Morrow. 191–199. A brief, accessible article in which Sperber argues that ‘human communication is a by-product of human meta-representational capacities’. Wilson, D. (2000) ‘Metarepresentations in linguistic communication’, in D. Sperber (ed.) Metarepresentations: A Multidisciplinary Perspective. Oxford: Oxford University Press. 411–448. A key paper which discusses both the role of mindreading in communication and the metarepresentational resources of natural language.

References Baron-Cohen, S. (1995) Mindblindness: An Essay on Autism and Theory of Mind. Cambridge, MA: MIT Press. Bloom, P. (2000) How Children Learn the Meanings of Words. Cambridge, MA: MIT Press. Cappelen, H. and Lepore, E. (2012) ‘Quotation’, in E. N. Zalta (ed.) The Stanford Encyclopedia of Philosophy Archive. Available from: http://plato.stanford.edu/archives/spr2012/entries/quotation/ (accessed May 2015). Carston, R. (1996) ‘Metalinguistic negation and echoic use’, Journal of Pragmatics, 25(3): 309–330. Carston, R. (1998) ‘Negation, “presupposition” and the semantics/pragmatics distinction’, Journal of Linguistics, 34(2): 309–350. Carston, R. (1999) ‘Negation, “presupposition” and metarepresentation: A response to Noel Burton-Roberts’, Journal of Linguistics, 35(2): 365–389. Carston, R. (2002) Thoughts and Utterances:The Pragmatics of Explicit Communication. Oxford: Blackwell. Carston, R. and Noh, E. J. (1996) ‘A truth-functional account of metalinguistic negation, with evidence from Korean’, Language Sciences, 18(1–2): 485–504. Chevallier, C., Noveck, I., Happé, F. and Wilson, D. (2011) ‘What’s in a voice? Prosody as a test case for the Theory of Mind account of autism’, Neuropsychologia, 49(3): 507–517. Chevallier, C.,Wilson, D., Happé, F. and Noveck, I. (2010) ‘Scalar inferences in Autism Spectrum Disorders’, Journal of Autism and Developmental Disorders, 40(9): 1104–1117. Chomsky, N. (2007) ‘Of minds and language’, Biolinguistics, 1: 9–27. Clark, H. H. and Carlson,T. B. (1981) ‘Context for comprehension’, in J. Long and A. Baddeley (eds.) Attention and Performance IX. Hillsdale, NJ: Lawrence Erlbaum. 313–330. Clark, H. H. and Gerrig, R. J. (1984) ‘On the pretense theory of irony’, Journal of Experimental Psychology, 113: 121–126. Clark, H. H. and Gerrig, R. J. (1990) ‘Quotations as demonstrations’, Language, 66(4): 764–805. Clark, H. H. and Marshall, C. R. (1981) ‘Definite reference and mutual knowledge’, in A. K. Joshi, B. L. Webber and I. A. Sag (eds.) Elements of Discourse Understanding. Cambridge: Cambridge University Press. 10–63.

307

Nicholas Allott Coulmas, F. (ed.) (1986) Direct and Indirect Speech. Berlin: Mouton de Gruyter. Davidson, D. (1979/1984) ‘Quotation’, in D. Davidson (ed.) Inquiries into Truth and Interpretation. Oxford: Clarendon Press, Oxford University Press. 79–92. Ducrot, O. (1974) ‘Les échelles argumentatives’, in O. Ducrot (ed.) La preuve et le dire: langage et logique. Paris: Mame. 225–285. Fillmore, C. (1971) ‘Types of lexical information’, in D. D. Steinberg and L. A. Jakobovits (eds.) Semantics: An Interdisciplinary Reader in Philosophy, Linguistics and Psychology. Cambridge: Cambridge University Press. 370–392. Grice, H. P. (1957) ‘Meaning’, The Philosophical Review, 66(3): 377–388. Grice, H. P. (1967) ‘Logic and conversation’, William James lectures delivered at Harvard University. Reprinted in H. P. Grice (1989) Studies in the Way of Words. Cambridge, MA: Harvard University Press. 1–143. Grice, H. P. (1969) ‘Utterer’s meaning and intentions’, The Philosophical Review, 78(2): 147–177. Grice, H. P. (1989) Studies in the Way of Words. Cambridge, MA: Harvard University Press. Grice, H. P. (2001) Aspects of Reason. Oxford: Clarendon Press. Happé, F. (1993) ‘Communicative competence and theory of mind in autism: A test of relevance theory’, Cognition, 48(2): 101–119. Happé, F. and Loth, E. (2002) ‘ “Theory of mind” and tracking speakers’ intentions’, Mind and Language, 17(1–2): 24–36. Hauser, M. D., Chomsky, N. and Fitch,W.T. (2002) ‘The faculty of language:What is it, who has it, and how did it evolve?’, Science, 298(5598): 1569–1579. Horn, L. R. (1985) ‘Metalinguistic negation and pragmatic ambiguity’, Language, 61(1): 121–174. Horn, L. R. (1989) A Natural History of Negation. Chicago: University of Chicago Press. Kumon-Nakamura, S., Glucksberg, S. and Brown, M. (1995) ‘How about another piece of pie: The allusional pretense theory of discourse irony’, Journal of Experimental Psychology: General, 124(1): 3–21. Langdon, R., Davies, M. and Coltheart, M. (2002) ‘Understanding minds and understanding communicated meanings in schizophrenia’, Mind and Language, 17(1–2): 68–104. Langford, D. (2010). ‘Blurb masterclass’, Ansible, 279. Available from: http://news.ansible.uk/a279.html (accessed May 2015). Leech, G. N. and Short, M. (1981) Style in Fiction: A Linguistic Introduction to English Fictional Prose. London: Longman. Mathis, T. and Yule, G. (1994) ‘Zero quotatives’, Discourse Processes, 18(1): 63–76. McHale, B. (1978) ‘Free indirect discourse: A survey of recent accounts’, PTL: A Journal for Descriptive Poetics and Theory of Literature, 3: 249–287. Mitchell, P., Robinson, E. J. and Thompson, D. E. (1999) ‘Children’s understanding that utterances emanate from minds: Using speaker belief to aid interpretation’, Cognition, 72(1): 45–66. Neale, S. (1992) ‘Paul Grice and the philosophy of language’, Linguistics and Philosophy, 15(5): 509–559. Noh, E. J. (1998) ‘Echo questions: Metarepresentation and pragmatic enrichment’, Linguistics and Philosophy, 21(6): 603–628. Noh, E. J. (2000) Metarepresentation: A Relevance-Theory Approach. Amsterdam: John Benjamins. Onishi, K. H. and Baillargeon, R. (2005) ‘Do 15-month-old infants understand false beliefs?’, Science, 308(5719): 255–258. Papafragou, A. (1996) ‘On metonymy’, Lingua, 99(4): 169–195. Partee, B. (1973) ‘The syntax and semantics of quotation’, in S. R. Anderson and P. Kiparsky (eds.) A Festschrift for Morris Halle. San Francisco: Holt, Reinhart and Winston. 410–418. Perner, J., Frith, U., Leslie, A. M. and Leekam, S. R. (1989) ‘Exploration of the autistic child’s theory of mind: Knowledge, belief, and communication’, Child Development, 60(3): 689–700. Perner, J. and Garnham, A. (1988) ‘Conditions for mutuality’, Journal of Semantics, 6(1): 369–385. Premack, D. and Woodruff, G. (1978) ‘Does the chimpanzee have a theory of mind?’, Behavioral and Brain Sciences, 1(4): 515–526. Semino, E. and Short, M. (2004) Corpus Stylistics: Speech, Writing and Thought Presentation in a Corpus of English Writing. London: Routledge. Schiffer, S. R. (1972) Meaning. Oxford: Oxford University Press. Smith, N.V. (ed.) (1982) Mutual Knowledge. London: Academic Press. Southgate, V., Senju, A. and Csibra, G. (2007) ‘Action anticipation through attribution of false belief by 2-year-olds’, Psychological Science, 18(7): 587–592.

308

Metarepresentation Sperber, D. (1994) ‘Understanding verbal understanding’, in J. Khalfa (ed.) What is Intelligence? Cambridge: Cambridge University Press. 179–198. Sperber, D. (2000) ‘Metarepresentations in an evolutionary perspective’, in D. Sperber (ed.) Metarepresentations: A Multidisciplinary Perspective. New York: Oxford University Press. 117–137. Sperber, D. and Wilson, D. (1981) ‘Irony and the use-mention distinction’, in P. Cole (ed.) Radical Pragmatics. New York: Academic Press. 295–318. Sperber, D. and Wilson, D. (1986) Relevance: Communication and Cognition. Oxford: Blackwell. Sperber, D. and Wilson, D. (1990) ‘Spontaneous deduction and mutual knowledge’, Behavioral and Brain Sciences, 13(1): 179–184. Sperber, D. and Wilson, D. (2002) ‘Pragmatics, modularity and mind-reading’, Mind and Language, 17(1–2): 3–23. Strawson, P. F. (1964) ‘Intention and convention in speech acts’, The Philosophical Review, 73(4): 439–460. Surian, L., Caldi, S. and Sperber, D. (2007) ‘Attribution of beliefs by 13-month-old infants’, Psychological Science, 18(7): 580–586. Wade, E. and Clark, H. H. (1993) ‘Reproduction and demonstration in quotations’, Journal of Memory and Language, 32(6): 805–819. Walton, K. L. (1990) Mimesis as Make-Believe: On the Foundations of the Representational Arts. Cambridge, MA: Harvard University Press. Wilson, D. (2000) ‘Metarepresentations in linguistic communication’, in D. Sperber (ed.) Metarepresentations: A Multidisciplinary Perspective. Oxford: Oxford University Press. 411–448. Wilson, D. (2006) ‘The pragmatics of verbal irony: Echo or pretence?’, Lingua, 116(10): 1722–1743. Wilson, D. (2009) ‘Irony and metarepresentation’, UCL Working Papers in Linguistics, 21: 183–226. Wilson, D. and Sperber, D. (1988) ‘Mood and the analysis of non-declarative sentences’, in J. Dancy, J. M. E. Moravcsik and C. C. W. Taylor (eds.) Human Agency: Language, Duty, and Value; Philosophical Essays in Honor of J. O. Urmson. Stanford, CA: Stanford University Press. 77–101. Wilson, D. and Sperber, D. (1992) ‘On verbal irony’, Lingua, 87(1–2): 53–76. Wilson, D. and Sperber, D. (2004) ‘Relevance theory’, in L. R. Horn and G. L. Ward (eds.) The Handbook of Pragmatics. Malden, MA: Blackwell. 607–632. Wilson, D. and Sperber, D. (2012) ‘Explaining irony’, in D. Wilson and D. Sperber (eds.) Meaning and Relevance. Cambridge: Cambridge University Press.

309

24 RELEVANCE1 Stavros Assimakopoulos

Introduction The domain of inferential pragmatics owes a lot of its existence to the pioneering work of Paul Grice (1957, 1975), which emphasised that successful verbal communication requires not only the knowledge of some linguistic code, but also a general ability to draw inferences. As he famously showed, it is because of this ability that we manage to spontaneously communicate more than we actually say in everyday conversation. For example, consider the following exchange: (1) John:  Excuse me, do you have a lighter? Mary:  I don’t smoke. Here, Mary is, strictly speaking, saying only that she is not a smoker, but John can take her utterance quite straightforwardly to mean that she does not have a lighter. Indeed, in this particular context it would be quite hard – if not impossible – for John to think that Mary intended to communicate anything different to him. In the Gricean picture, such inferences, dubbed ‘particularised conversational implicatures’, arise due to the manipulation of a set of norms that we typically abide by as rational human beings in our effort to be cooperative with our interlocutors. These maxims, as Grice called them, include various categories, namely, quality (truthfulness), quantity (informativeness), relation (relevance) and manner (clarity) (for an overview, see Davis 2014). In our particular example in (1), it seems to be the maxim of relation that ensures the success of Mary’s communicative act. In this respect, it is by assuming that Mary is a cooperative interlocutor who therefore observes all the maxims, that John can go on to infer that Mary could not have said that she does not smoke in response to his request, unless she wanted him to think that she does not have a lighter; otherwise her answer would have been irrelevant to his original question. Even though Grice provided a description of all the other maxims he put forth in his account, the analysis of the one that called for speakers to ‘be relevant’ during their communicative interactions remained notoriously elusive to him: ‘Though the maxim itself is terse, its formulation conceals a number of problems that exercise me a good deal. . . . . I find the treatment of such questions exceedingly difficult, and I hope to revert to them in a later work’ (Grice 310

Relevance

1975: 46). However, he unfortunately never returned to this issue.The years to come saw several attempts to capture the notion of relevance both within the greater Gricean picture (e.g., Dascal 1977; Werth 1981; Holdcroft 1987) and outside of it (e.g., Van Dijk 1979; Allwood 1984), but unfortunately none of them managed to remain in the spotlight of pragmatic theorising for too long. It was not until the development of relevance theory in the 1980s that the term ‘relevance’, in the technical sense Dan Sperber and Deirdre Wilson assigned to it, gained its place as a fundamental concept in the study of inferential pragmatics.

Relevance as a psychological property Relevance theory may have indeed been motivated by Grice’s seminal work, but, as we will see, it also departs quite substantially from it in important respects. For one, even though Sperber and Wilson share Grice’s conviction that inference has a central role to play in an account of communication, they do not base their argumentation on systems of social norms or maxims that underlie cooperative conversational behaviour. Therefore, their definition of relevance is not meant to fill in the aforementioned gap in the Gricean system. Instead, they focus on the mental processes that enable us to communicate, and in particular those processes that mediate the interpretation of some communicative stimulus by its audience.This outlook gives relevance theory an essentially cognitive perspective, which can be contrasted with the mainly philosophical one that both Grice and the neo-Griceans after him seem to have adopted; hence, the usual reference to the relevance-theoretic framework as a post-Gricean, rather than a neo-Gricean approach to pragmatics: What is needed is an attempt to rethink, in psychologically realistic terms, such basic questions as: What form of shared information is available to humans? How is shared information exploited in communication? What is relevance and how is it achieved? What role does the search for relevance play in communication? (Sperber and Wilson 1986/1995: 38) Against this background, Sperber and Wilson propose that the inferential comprehension of communicative stimuli is carried out by a dedicated mental module.2 What triggers this module is the detection in our environment of ostensive stimuli, that is, stimuli that make a peer’s intention to communicate a piece of information obvious to us; as, for example, when someone waves at us, calls our name or even persistently stares at us. Once such a communicative intention is recognised and our attention is drawn to what that person wants to communicate, it is up to the proposed module to inferentially calculate her3 intended meaning. Since successful communication is never guaranteed and miscommunication can and often does occur during our everyday interactions, the inference at play in this calculation is patently non-demonstrative, that is, a type of inference that is best described as ‘suitably constrained guesswork’ (Sperber and Wilson 1986/1995: 69) rather than a failsafe logical process. This kind of guesswork effectively involves forming and evaluating hypotheses about the meaning that is communicated through the use of some particular ostensive stimulus as well as the right set of background assumptions that the addressee needs to consider in order to reach this meaning. This set of assumptions, which comprises not only immediately observable information, but also general beliefs about the world as well as dispositions towards the communicator in that particular instance, constitutes the context against which the content of the communicative stimulus will be processed with a view to recovering the communicator’s intended meaning or, as relevance theorists call it, her informative intention. In our familiar exchange in (1), for example, John can only reach the 311

Stavros Assimakopoulos

conclusion that Mary does not have a lighter if he combines her explicitly expressed message that she is not a smoker with the contextual assumption that ‘people who do not smoke do not normally carry a lighter on them’. It is when we reach what relevance theory views as orchestrating the operation of this specialised inferential module, or the ostension processor, as it is commonly referred to in the ­framework, that we come across the definition of relevance it puts forth. Using the term in a strictly technical sense, Sperber and Wilson characterise relevance as a property exhibited by some input to mental processing. Such input can include a stimulus that needs to be perceived, an utterance that calls for interpretation, or even a contextual assumption that needs to be taken into account for some processing to have the desired results. In this setting, relevance is described as a trade-off between what relevance theorists call positive cognitive effects and processing effort. From the relevance-theoretic perspective, an input generates cognitive effects when it manages to bring about a non-trivial change in an individual’s system of beliefs. For such a change to be non-trivial, it needs to come in one of two forms. The first one is dubbed a contextual implication and signifies a new conclusion derived from the unification of a communicative stimulus with the context of its processing, as in our now familiar case of John inferring Mary’s implicature in (1). The second is the modification of the status of some existing assumptions in the context of this processing, which corresponds to either strengthening them, by backing them up with further evidence, or contradicting and erasing/revising them. For relevance theorists, the achievement of cognitive effects, and in particular positive ones, which are bound to lead us to true beliefs (cf. Sperber and Wilson 1995: 263–266), is viewed as the aim that our cognitive system has when engaging in the processing of some input. This effectively means that, other things being equal, the greater positive cognitive effects the processing of some input brings about to the individual, the more relevant this input is deemed for him/her at that moment. Processing effort, on the other hand, which obviously refers to the relative difficulty of the computations involved in the processing of this input, works in quite the opposite direction; that is, other things being equal, the more effort that the individual has to spend in processing some input, the less relevant this input is deemed for him/her at that moment.

Relevance in cognition and communication We have so far identified the cognitive mechanism that is, according to relevance theorists, responsible for the pragmatic processing that mediates the comprehension of communicative stimuli and described relevance as a balance between effort and cognitive gain. It is now time to turn to the two main generalisations that are put forth in their framework. As we will see, it is on the basis of these two generalisations, or principles of relevance, as they are commonly called, that relevance-theoretic pragmatics operates. Following mainstream research in the domain of evolutionary psychology, relevance theorists maintain that, in much the same way as other biological systems, the current state of human cognition cannot but be an adaptation that has evolved through Darwinian processes of natural selection. Therefore, the argument goes, much like ‘we can expect that the structure, placement and mode of operation of a muscle will tend to minimise the energy costs of performing the bodily movement it is its function to produce,’ we should also ‘expect to find a tendency towards maximal efficiency in the design of cognitive mechanisms’ (Sperber and Wilson 1995: 262, emphasis as in original). In combination with the aforementioned characterisation of relevance then, the case can be made that our cognitive system functions automatically aiming at maximal relevance, that is, the achievement of the greatest possible amount of positive cognitive effects for the smallest possible amount of processing effort. This generalisation, which is essentially a 312

Relevance

hypothesis about the evolution of the human mind into an all the more efficient mechanism through processes of natural selection, is explicitly spelled out in the First or Cognitive Principle of Relevance, according to which, human cognition tends to be geared to the maximisation of relevance. Clearly, the implications of this particular outlook to cognition are far-reaching in their own right, both in terms of how our cognitive resources are organised and the way in which mental computations of any kind take place. But one particular area in which this rationale finds a straightforward application in relevance theory is that of inferential communication. This is what the Second or Communicative Principle of Relevance, according to which, every act of ostensive communication communicates a presumption of its own optimal relevance, aims at describing. In order to make sense of this principle, we first need to identify what communicating a presumption of optimal relevance entails.The starting point for the discussion here is the observation that ostensive acts of communication seem to typically pre-empt our attention. For example, once we notice someone in our environment who is trying to get our attention, we instinctively realise that she wants to communicate something to us. We may of course choose to ignore her before the actual communicative act takes place, but if we do not do so, we cannot help but process what it is that she wishes to communicate. From the relevance-theoretic perspective, this suggests that the recognition of a communicative intention instantly triggers the ostension processor, which will in turn automatically start formulating hypotheses about our communicator’s informative intention on the basis of her chosen stimulus. Given, however, the vast amount of information that surrounds us at every moment of our waking lives, why is it that communicative stimuli always manage to pre-empt our attention? If our cognitive system is indeed oriented towards maximal relevance, as the cognitive principle of relevance suggests, one would expect that it would not automatically engage in the computation of ostensive stimuli, unless it had evolved to treat their processing as worthwhile in all instances. For relevance theorists, this suggests that such stimuli always come with a manifest guarantee that they will generate enough positive cognitive effects to offset the effort required to process them. This creates a first condition under which the processing of a communicative stimulus is to take place, that is, the condition that, once recognised as ostensive, a stimulus will always be treated by our cognitive system as at least relevant enough to be worth our processing effort. The second condition that underlies the presumption of optimal relevance has to do with the addressee’s confidence in the communicator’s choice of stimulus. An initial observation in this respect is that during the spontaneous comprehension of some ostensive stimulus our cognitive system does not pause to assess whether that particular stimulus is the best one the communicator could have used for her purposes, but just goes on to uncover her informative intention straightaway. Clearly, such an assessment of the communicator’s choice of stimulus can – and often does – take place later on in the processing of her communicative act, but the calculation of her informative intention always precedes such further processing. For relevance theorists, this suggests that, once our cognitive system recognises a stimulus as being ostensively communicated, it will automatically engage in its comprehension, assuming that it is the most relevant one compatible with communicator’s abilities and preferences. Against this background, the second principle of relevance aims to address the inferential comprehension of communicative stimuli from a cognitive perspective, indicating what it necessarily involves in its premises. And even though it suspiciously resembles a Gricean maxim in its formulation, it is emphatically not meant to be treated as such. As Sperber and Wilson characteristically note, ‘Communicators and audience need no more know the [communicative] principle of relevance to communicate than they need to know the principles of genetics in order to reproduce. Communicators do not “follow” the principle of relevance; and they could not violate it even if they wanted to’ (1986/1995:162). The prediction made in this regard then is 313

Stavros Assimakopoulos

that, once we recognise a peer’s communicative intention, we have no choice but to engage in the comprehension of her informative intention automatically assuming that it is optimally relevant. In this respect, the notion of optimal relevance, which is to be distinguished from that of maximal relevance that, as we saw earlier, is taken to govern cognition in general, signifies the search for cognitive effects that can adequately substantiate a communicator’s informative intention for no unjustifiable processing effort on behalf of the addressee.

Relevance and the pragmatics of linguistic communication It should be evident from common experience that not all kinds of ostensive stimuli are capable of revealing an individual’s informative intention with the same amount of precision. For example, suddenly pointing towards a passer-by while talking to a friend might not be as successful in communicating that this passer-by is your old schoolteacher as it would be if coupled with the utterance ‘This is my old schoolteacher.’ In general, as Sperber and Wilson point out, verbal communication ‘gives rise to the strongest possible form of communication’ (1986/1995: 60), especially since once you recognise that an utterance that is spoken in a language you master is directed towards you, it is practically impossible to filter it out as background noise. In light of this, the study of linguistic communication provides a particularly fruitful ground for exemplifying relevance-theoretic pragmatics and discussing its differences as compared to competing (neo-)Gricean accounts. Since the driving force behind Grice’s original argumentation was the question of how the study of pragmatics needs to be distinguished from that of semantics, this seems like an appropriate place to start the discussion here too. Given its cognitive outlook, relevance theory draws the distinction between an utterance’s semantic content and pragmatically inferred meaning by distinguishing between the two mental processes that it takes to be responsible for their determination in the hearer’s mind. In this setting, the recognition of an ostensive stimulus as linguistic triggers in the first instance a module that belongs to our linguistic system proper, and whose task is to decode utterances into context-independent conceptual representations, or logical forms, as relevance theorists commonly call them. This semantic decoding process, during which each uttered word is translated into its direct conceptual counterpart, is necessary so that the linguistic stimulus is converted into a format that can undergo further, pragmatic this time, processing. During this inferential process, the ostension processor that relevance theorists postulate takes the output of decoding and performs the tasks of (a) developing it into a full proposition that corresponds to the one the speaker intended to communicate in the first place, (b) selecting an appropriate set of contextual assumptions against which it will calculate the speaker-intended cognitive effects, and (c) calculating these cognitive effects. Of these three tasks, the one that has generated the most controversy is that of developing an utterance’s logical form into a full proposition that matches the one the speaker had in mind when engaging in her communicative act. This is hardly surprising, given the fact that in the mid-1980s, when relevance theory was originally developed, most scholars in the field followed Grice in assuming that inferential reasoning is necessary for the construction of conversational implicatures alone.4 Sperber and Wilson, however, argued that such reasoning is also involved in the derivation of an utterance’s explicitly expressed meaning. In the picture they envisaged, an utterance’s decoded logical form is typically an incomplete propositional schema that needs to be inferentially enriched in order to gain full propositional status. It is therefore only this enriched proposition, which, by analogy with implicature, they called explicature, that can be evaluated for its truth or falsity in the context of the communication at hand, and can thus be of any real use to the hearer during the interpretation of an utterance. 314

Relevance

Since then, and following further research on the matter, most notably by Robyn Carston (cf. 1988, 2002), relevance theorists have not only stood their ground, but have also strengthened their position, eventually adopting a strong version of the linguistic underdeterminacy thesis, according to which, ‘linguistically encoded meaning never fully determines the intended proposition expressed’ (Carston 2002: 49, emphasis as in original). By way of example, consider the following: (2) They said they would go to the bank in the morning. (3) Apples are red. vs Watermelons are red. (4) Sam:  Do you want to go grab something for lunch? Tina:  I’ve eaten. (5) Pilots shouldn’t drink. It should be obvious that in order to determine the truth conditions of the proposition explicitly expressed by all of the utterances in (2) to (5), at least some contextual enrichment is required. In (2), this requirement seems to be triggered by lexical items within the utterance itself; in order to be able to assign a truth value to the proposition expressed by it, the hearer would need to assign a referent to the pronominal expression ‘they’ as well as identify which of the two encoded senses of the ambiguous lexical item ‘bank’ (‘bank building’ versus ‘riverbank’) the speaker intended to communicate to him in the first place. In the remaining cases, on the other hand, the required enrichment seems to be context-driven rather than linguistically mandated. In (3), the semantic content of the two juxtaposed utterances seems to be underdetermined in relation to at least the property of each fruit that ‘red’ applies to; that is, the peel of the apple and the edible part of the watermelon. In (4), even though the proposition ‘Tina has eaten’ can be typically assigned a truth value as it stands, it constitutes such an obvious truism that it cannot in this context (and indeed most other contexts) be taken to be what Tina intended to express. Instead, Sam will have to enrich it into an explicature which specifies that Tina has eaten something for lunch on that particular day in order to reach her implicature that she will not be joining him. Similarly, in order for the hearer to reconstruct the explicature of the utterance in (5), he would need to pragmatically adjust the encoded atomic5 concept DRINK into the more specific ad hoc concept DRINK*, whose denotation includes only alcoholic beverages, rather than any other kind of liquid drink. Clearly, such a superficial overview of only a handful of the sources of linguistic underdeterminacy that have been examined within relevance theory cannot do justice to the breadth of research that has led relevance theorists to adopt the radical position they have in relation to the pragmatics of explicit communication (for much more detailed discussion, see Hall, Chapter 26, this volume). However, it should help show that, in the communicative setting, the determination of the truth conditions of the proposition explicitly expressed by an utterance is not as straightforward as traditional semanticists have for long taken it to be. All in all, the relevancetheoretic argumentation in this vein has considerably helped pave the way for the eventual recognition of truth-conditional pragmatics as a legitimate theoretical enterprise in the study of linguistic meaning even outside of the framework’s remit. Moving beyond an utterance’s basic explicature, relevance theory also treats illocutionary force as part and parcel of explicitly communicated information, suggesting that it forms part of an utterance’s higher-level explicatures. To this effect, consider the following familiar motto: (6) Drink responsibly! For relevance theorists, the basic explicature this utterance communicates would be the same as the one the corresponding declarative form conveys, that is, something like (a): ‘the hearer 315

Stavros Assimakopoulos

drink(s) (alcohol) responsibly’. Through the use of the imperative mood, however, the speaker is also communicating that her informative intention is not merely (a) but rather some higherlevel explicature of the type: ‘The speaker requests that (a)’. Apart from mood indicators and intonation, higher-level explicatures can also be communicated by other means too, such as propositional attitude parentheticals – as in (7) Fortunately, Mary found her lost ring. where the speaker not only communicates the explicature that ‘Mary found her lost ring’, but also the higher-level explicature that ‘the speaker finds it fortunate that Mary found her lost ring.’ Even interjections are amenable to a similar analysis. For example, in (8), the speaker expresses her surprise or relief that ‘the hearer has found his mobile phone’ by means of an interjection. (8) Oh! You’ve found your mobile phone. Turning to the implicit side of verbal communication now, we are effectively entering the domain of implicature, a type of communicated meaning traditionally addressed following Grice’s rationale. In this respect, before we see how relevance theory approaches the notion of implicature, a small digression into the original Gricean account of it seems in order. In the discussion up to this point, we have encountered only one example of implicature in (1), which we then called a particularised conversational implicature. Apart from this kind, which is obviously dependent on maxim-driven inference over what is said in conjunction with the context in which an utterance is produced, Grice also identified another kind of conversational implicature, which he called ‘generalised’. Generalised conversational implicatures are again maxim-driven, but this time largely context-independent. Consider, for example, the following exchange: (9) Tom:  How was your sister’s party? Susan:  Some guests had a really good time. In this case, Susan’s response to Tom’s question can be quite straightforwardly taken to communicate (at least) the following two implicatures to him: a. ‘Not all guests had a really good time.’ b. ‘The party wasn’t a big success.’ The second of these implicatures is clearly an example of a particularised conversational implicature, since it can only be derived in the context of this particular exchange. On the other hand, the first one seems to be communicated by the use of the lexical item ‘some’, rather than by the context in which this utterance was produced. Along these lines, it is conversational implicatures that generally arise through the use of certain lexical items like the partitive ‘some’, which are dubbed generalised in the Gricean picture. The only effect that the context can have on such implicatures is that it can on occasion cancel them, as would happen if Susan had instead uttered, ‘Some guests had a really good time. Actually, all of them did.’ Finally, the last category of implicature Grice distinguished was the one he called ‘conventional’ implicature. Conventional implicatures are also communicated through the use of specific lexical items, customarily discourse markers/connectives, such as ‘but’, ‘moreover’ and ‘therefore’; yet they differ from generalised conversational ones in that they are not cancellable, nor can they be affected by contextual considerations. 316

Relevance

Going back to the relevance-theoretic framework, its approach to conversational implicature again departs considerably from the original Gricean description, as well as its updated versions proposed by the neo-Griceans.To begin with, relevance theorists argue that the distinction between generalised and particularised conversational implicatures is unmotivated. In their account, implicatures of the former kind are taken as context-driven, in much the same way as those of the latter are typically considered to be (cf. Carston 1995). As Carston characteristically discusses, from the relevance-theoretic perspective, all conversational implicatures6 belong to a ‘continuum of cases from the very frequent to the one-off ’ (2002: 142), a point further vindicated by extensive experimental research on the processing of scalar inferences (cf. Noveck and Sperber 2007). Even so, relevance theorists still distinguish between two kinds of implicature in the conversational setting, but they do so on the basis of the particular comprehension task that they take as responsible for each kind’s derivation.The most transparent one is that of implicated conclusions, which are effectively speaker-intended contextual implications, that is, inferences derived from the unification of an utterance’s explicitly expressed meaning with the contextual assumptions selected for its processing, such as the implicature that ‘Mary does not have a lighter’ in our familiar example in (1). Apart from implicated conclusions though, relevance theorists also postulate the category of implicated premises, which are again speaker-intended inferences, but this time belong to the task of selecting a context against which the hearer will attempt to recover the speaker-intended meaning. Consider, for example, the following exchange: (10) Mark:  Peter told me he likes you. Would you go out with him? Kate:  I don’t date ugly guys. Obviously, an implicated conclusion that Mark can reasonably assume to have been communicated to him by Kate’s utterance in (10) is that she does not want to go out with Peter. However, this conclusion is warranted only if the assumption that ‘Peter is an ugly guy’ is used as a premise in its derivation. And since this is an assumption that Kate demonstrably intends Mark to include in his context of interpretation, it is a premise that she has essentially implicated. Finally, with respect to the cases that Grice described as conventional implicatures, relevance theorists have argued that they should not be thought of as implicatures per se, as they are by definition the non-cancellable meanings that certain lexical items carry. In this vein and following Blakemore’s original argumentation (1987), most Gricean conventional implicatures have been re-analysed in relevance theory as procedures, that is, semantic constraints on the derivation of pragmatic inferences. To use a well-worn, and necessarily simplified example, ‘but’ (on its ‘denial of expectation’ use) is taken in the relevance-theoretic account to encode an instruction for the ostension processor to process the clause it introduces as contradicting and therefore eliminating an assumption that should be readily available to the hearer at that processing stage. So, the incorporation of a clause introduced with ‘but’ in Kate’s utterance in the exchange in (10) can easily eliminate the original implicated conclusion that ‘Kate does not want to go out with Peter’, as (11) exemplifies: (11) Mark:  Peter told me he likes you. Would you go out with him? Kate:  I don’t date ugly guys, but Peter is such a sweetheart. Even though the notion of procedural encoding was originally meant to serve as a cognitive alternative to that of conventional implicature, relevance theorists eventually recognised that even more linguistic devices, such as tense, intonational distinctions, illocutionary-force indicators or even pronouns, are amenable to a procedural analysis. In light of this, the relevance-theoretic category of procedural meaning is no longer co-extensive with that of non-truth-conditional 317

Stavros Assimakopoulos

meaning that motivated the postulation of conventional implicatures by Grice in the first place, and procedural information has been shown to constrain not only the derivation of implicated premises and implicated conclusions, but that of (basic and higher-level) explicatures as well (cf. Wilson and Sperber 1993; Blakemore 2002).

Utterance comprehension: the relevance-theoretic heuristic Having discussed the tasks that the ostension processor, which relevance theorists postulate, is responsible for during interpretation, we are now in a position to see how they form part of the overall comprehension process, by looking at an example step-by-step description of how Ken manages to understand Ann’s utterance in the following exchange: (12) Ken:  Won’t your dad join us for lunch? Ann:  He’s blue because of last night’s game.

(a) Ann has uttered, ‘He’s blue because of last night’s game.’

[Ann’s utterance is decoded into a logical form, which is incomplete, since at least the referent of ‘he’ needs to be identified and ‘blue’ needs to be disambiguated between its BLUE1(= coloured blue) and BLUE2(= sad) sense.] (b) Ann’s utterance will be automatically processed as optimally relevant to Ken.



[Expectation raised by the recognition of Ann’s communicative intention, as the communicative principle of relevance prescribes.] (c) Ann’s utterance will achieve relevance by providing an answer to Ken’s question.



[More specific expectation raised by (b), together with the fact that questions typically call for an answer.] (d) The fact that one is BLUE2 is a good reason for not wanting to have lunch.



[Highly accessible contextual assumption which might help satisfy expectation (c). Accepted as an implicated premise of Ann’s utterance.] (e) Ann’s dad is BLUE2 because of the result of last night’s football game.



[Enrichment of the logical form of Ann’s utterance which could combine with (d) to lead to the satisfaction of (c). Accepted as the basic explicature of Ann’s utterance.] (f) Ann’s dad will not join Ken and her for lunch.



[Inferred from (d) and (e), and satisfying (c). Accepted as an implicit conclusion of Ann’s utterance.] (g) Ann’s dad will not join Ken and Ann for dinner and the reason for this is that he is BLUE2 because of the result of last night’s football game.



[Further implicated conclusion inferred from (d) and (e), which alongside (f) satisfies (b).] (h) Ann’s dad has no personal issues with Ken and Ann, Ann’s dad still enjoys Ken’s cooking, and so forth.



[Further implicated conclusions inferred from (g) plus background knowledge, which also contribute to the satisfaction of (b).] 318

Relevance

As follows from this example, Ken’s ostension processor inferentially develops the logical form of Ann’s utterance in step (12e), derives an implicated premise in step (12d), and some implicated conclusions in steps (12f), (12g) and (12h). In relation to this description, however, three further clarifying remarks are in order. First of all, apart from the general expectation of optimal relevance that all ostensive stimuli raise, an addressee might also on occasion have some even more specific expectations of relevance that arise from the setting in which a particular communicative act takes place. For example, as noted in (12c), having asked a question, Ken can reasonably expect Ann’s utterance to give a response to it. Second, relevance theory assumes that both implicatures and explicatures can be communicated with varying degrees of strength, depending on the confidence with which a hearer can assume that they form part of the speaker’s informative intention. So, it may be the case that Ann only intended her response to lead Ken to the implicatures in (f) and (g), but Ken could also derive the weaker implicatures in (h) and still assume that they were part of Ann’s original informative intention (perhaps because the other day they were discussing whether Ann’s father enjoys Ken’s cooking or because Ann knows that Ken gets easily offended or for any other reason for that matter). Then, despite what this step-by-step description might appear to suggest, relevance theorists do not take all the processing involved in utterance comprehension to be sequentially ordered. In the Gricean picture, inference is assumed to generate conversational implicatures on the basis of ‘what is said’ and only after this has been fully recovered. On the contrary, in relevance theory, an utterance’s communicated meaning is taken to be simultaneously processed at both the explicit and the implicit levels, while at the same time the context of interpretation gets updated with new assumptions as needed. In this respect, an utterance’s implicatures are not calculated only after its explicitly expressed meaning has been derived, nor does the ostension processor wait for the relevant grammatical module to finish decoding the utterance’s logical form before it starts to construct this explicature. For example, in our exchange, Ken will have probably derived the implicature in (12f) before he finishes decoding and inferentially developing Ann’s utterance – most likely by the time he has disambiguated the intended meaning of ‘blue’. What is more, as follows from the discussion towards the end of the previous section, should Ann have uttered ‘He’s blue because of last night’s game, but I’m sure he will’ instead of what she did, the implicature in (12f) would again have been derived, but then eliminated. All in all, relevance theorists insist that pragmatic processing takes place in parallel with decoding, developing its output on the fly; in other words, all the interpretive hypotheses that are assessed with a view to recovering the speaker’s informative intention are arrived at by a process of mutual parallel adjustment, which may require several backwards and forwards inferences to be made. This finally brings us to the way in which, according to relevance theory, the ostension processor goes about recovering this speaker-intended meaning; that is, by means of the following relevance-based comprehension procedure: a. Follow a path of least effort in computing cognitive effects: Test interpretive hypotheses (e.g., disambiguations, reference resolutions, enrichments of the decoded logical form, identification of implicated premises and conclusions, etc.) in order of accessibility. b. Stop when your expectations of relevance are satisfied (or abandoned, as in the case of miscommunication). This procedure can be shown to follow directly from the two principles that relevance theorists propose. For one, if our cognitive system exhibits a tendency towards maximal efficiency, as the cognitive principle of relevance stipulates, there is no reason to suppose that we need to consider 319

Stavros Assimakopoulos

all possible interpretive routes before choosing which one to follow during interpretation. After all, even if we wanted to, we would not have the cognitive resources to do so in the few milliseconds during which such spontaneous comprehension typically takes place. Since, then, according to the communicative principle of relevance, all ostensive stimuli are automatically treated as able to yield the desired cognitive effects for no unjustifiable effort expenditure, it makes sense for the ostension processor to use the most cost-effective strategy while processing them; the assessment of interpretive hypotheses in the order in which they become available being the most obvious way of doing this. All this processing will then be concluded once the addressee’s expectations of relevance are satisfied, that is, once he reaches the first set of cognitive effects that can be accepted as the one that the communicator intended him to arrive at in the first place.

Conclusion Ultimately, the comprehension procedure that relevance theorists propose is nothing more than a ‘fast and frugal heuristic’ in the sense of Gigerenzer and colleagues (1999), and as such it does not guarantee that the cognitive effects eventually derived in the mind of the hearer will indeed be the speaker-intended ones. This on its own underlines the psychological plausibility of the relevance-theoretic account, since, contrary to most traditional pragmatic theories, it provides an explanation of how successful human communication is probable rather than guaranteed. In this way, relevance theory shifts the focus from the theoretical convenience that ties many researchers of pragmatics to a particular philosophical outlook in relation to the study of communication and opens up the way for the fruitful interaction of philosophical and psychological research. After all, relevance theorists have through the years made a number of testable predictions (cf. Van der Henst and Sperber 2004), some of which have been instrumental in the increasing growth of the field of experimental pragmatics.

Suggestions for further reading Wilson, D. and Sperber, D. (2002) ‘Relevance theory’, UCL Working Papers in Linguistics, 14: 249–287. Available from: www.phon.ucl.ac.uk/publications/WPL/02papers/wilson_sperber.pdf (accessed 10 October 2015). This is a much lengthier and more detailed overview of relevance theory that is freely available online. Carston, R. and Powell, G. (2005) ‘Relevance theory – new directions and developments’. UCL Working Papers in Linguistics, 17: 279–299. Available from: www.ucl.ac.uk/pals/research/linguistics/publications/ wpl/05papers/carston_powell (accessed 10 October 2015). This comprehensive paper, which is also freely available online, situates some of the latest developments in Relevance theory within the context of current research in philosophy of language and cognitive science. Wilson, D. and Sperber, D. (2012) Relevance and Meaning. Cambridge: Cambridge University Press. This collection comprises some of the most influential papers that the proponents of Relevance theory have produced, providing detailed argumentation on some of the most recent developments within the framework. Clark, B. (2013) Relevance theory. Cambridge: Cambridge University Press. This textbook provides a thorough and particularly accessible overview of most topics that relevance theorists have extensively dealt with, and whose treatment has been inevitably omitted or just mentioned in passing in this chapter due to space restrictions. Yus, F. (2015) Relevance theory Online Bibliographic Service. Available from: http://personal.ua.es/francisco. yus/rt.html (accessed 10 October 2015). This neatly organised list of research publications on Relevance theory reveals the breadth of work that it has inspired in a variety of fields, from philosophy and cognitive psychology to translation and literary studies, and is an indispensable reference point for anyone interested in the framework.

320

Relevance

Notes 1 I would like to thank the editors of the Routledge Handbook of Pragmatics for their kind invitation, guidance and valuable advice regarding the contents of this chapter. I am also grateful to Billy Clark, Albert Gatt, Alexandra Vella and Becky Vella Muskat for their extensive and insightful comments on earlier versions of it. Needless to say, all errors remain mine. 2 The proposal that the inferential processing of a communicative stimulus is effected by a specialised module, and more specifically a module pertaining to our more general mindreading/theory of mind cognitive capacity, is a relatively recent development in the relevance-theoretic framework (cf. Sperber and Wilson 2002). In their early work, Sperber and Wilson followed Fodor’s (1983) distinction between informationally encapsulated mental modules and non-modular central systems, assuming that pragmatic processing belongs to the latter. 3 For ease of exposition and to avoid confusion, I will follow the common practice of referring to the communicator/speaker as a female and the addressee/hearer as a male. 4 In all fairness, it should be acknowledged that Grice also saw disambiguation and reference resolution as necessary for the determination of ‘what is said’. However, he never discussed how these two processes are to be approached in his framework, which makes it reasonable to assume that he did not think that his maxims had anything to do with them. 5 Relevance theory has from its emergence adopted Fodor’s position on conceptual structure – or better, lack thereof (cf. Fodor 1998) – in much the same way as it has been sympathetic to his version of the representational theory of mind (cf. Fodor 1975, 2008; but for a point of disagreement in this regard, see Sperber and Wilson 1998). 6 Here, note that some paradigm cases of generalised conversational implicature in the Gricean picture, such as the ‘and’-conjunction or negation (cf. Carston 2002: chapters 3 and 4, respectively), as well as several scalar inferences (cf. Carston 1995, 1998), have been re-analysed in relevance theory as pragmatic contributions to an utterance’s basic explicature rather than implicatures per se.

References Allwood, J. (1984) ‘On relevance in spoken interaction’, in S. Bäckman and G. Kjellmer (eds.) Papers on Language and Literature. Gothenburg: Acta Universitatis Gothoburgensis. 18–35. Blakemore, D. (1987) Semantic Constraints on Relevance. Oxford: Blackwell. Blakemore, D. (2002) Relevance and Linguistic Meaning: The Semantics and Pragmatics of Discourse Markers. Cambridge: Cambridge University Press. Carston, R. (1988) ‘Implicature, explicature, and truth-theoretic semantics’, in R. Μ. Kempson, (ed.) Mental Representations: The Interface between Language and Reality. Cambridge: Cambridge University Press. 155–181. Carston, R. (1995) ‘Quantity maxims and generalised implicatures’, Lingua, 96(4): 213–244. Carston, R. (1998) ‘Informativeness, relevance and scalar implicature’, in R. Carston and S. Uchida (eds.) Relevance theory: Applications and Implications. Amsterdam: John Benjamins. 179–236. Carston, R. (2002) Thoughts and Utterances:The Pragmatics of Explicit Communication. Oxford: Blackwell. Dascal, M. (1977) ‘Conversational relevance’, Journal of Pragmatics, 1: 309–328. Davis, W. (2014) ‘Implicature’, in E. N. Zalta (ed.) The Stanford Encyclopedia of Philosophy Archive. Available from: http://plato.stanford.edu/archives/fall2014/entries/implicature/ (accessed 10 October 2015). Fodor, J. A. (1975) The Language of Thought. Cambridge, MA: Harvard University Press. Fodor, J. A. (1983) The Modularity of Mind: An Essay on Faculty Psychology. Cambridge, MA: MIT Press. Fodor, J. A. (1998) Concepts:Where Cognitive Science Went Wrong. Oxford: Oxford University Press. Fodor, J. A. (2008) LOT 2:The Language of Thought Revisited. Oxford: Oxford University Press. Gigerenzer, G., Todd, P. M. and the ABC research group (1999) Simple Heuristics That Make Us Smart. Oxford: Oxford University Press. Grice, H. P. (1957) ‘Meaning’, Philosophical Review, 66(3): 377–388. Grice, H. P. (1975) ‘Logic and conversation’, in P. Cole and J. L. Morgan (eds.) Syntax and Semantics: Speech Acts. New York: Academic Press. 41–58. Holdcroft, D. (1987) ‘Conversational relevance’, in J. Verschueren, J. and M. Bertuccelli-Papi (eds.) The Pragmatic Perspective: Selected Papers from the 1985 International Pragmatics Conference. Amsterdam: John Benjamins. 477–496.

321

Stavros Assimakopoulos Noveck, I. and Sperber, D. (2007) ‘The why and how of experimental pragmatics: The case of “scalar inferences” ’, in N. Burton-Roberts (ed.) Pragmatics. Basingstoke: Palgrave Macmillan. 184–212. Sperber, D. and Wilson, D. (1986/1995) Relevance: Communication and Cognition. Oxford: Blackwell. Sperber, D. and Wilson, D. (1998) ‘The mapping between the mental and the public lexicon’, in P. Carruthers and J. Boucher (eds.) Language and Thought: Interdisciplinary Themes. Cambridge: Cambridge University Press. 184–200. Sperber, D. and Wilson, D. (2002) ‘Pragmatics, modularity and mind-reading’, Mind and Language, 17(1–2): 3–23. Van der Henst, J. B. and Sperber, D. (2004) ‘Testing the cognitive and the communicative principles of relevance’, in I. Noveck and D. Sperber (eds.) Experimental Pragmatics. London: Palgrave. 141–169. Van Dijk, T. A. (1979) ‘Relevance assignment in discourse comprehension’, Discourse Processes, 2: 113–126. Werth, P. (1981) ‘The concept of “relevance” in conversational analysis’, in P. Werth (ed.) Conversation and Discourse: Structure and Interpretation. London: Croom Helm. 129–154. Wilson, D. and Sperber D. (1993) ‘Linguistic form and relevance’, Lingua, 90: 1–25.

322

25 METAPHOR IN PRAGMATICS Miriam Taverniers

Introduction An overview of the pragmatic perspective on metaphor can be constructed in multiple ways. First of all, it is not automatically clear what ‘pragmatics’ is, and second, metaphor is an extremely vast topic, and not a pragmatic topic per se, in contrast to, for instance presupposition, intention, relevance, pragmatic enrichment. As to the latter complication, it is safe to say that metaphor is not dealt with systematically and comprehensively in pragmatics.When it is mentioned, it is often as an intermediate category: it is subsumed under more general phenomena (non-literal language, loose language), and sub-types of metaphor, such as conventional vs. novel metaphor and other types of distinctions, are often not taken into account. As a result, there is no clear ‘problem set’ of metaphor in relation to pragmatics. As to the first complication: this chapter could be on metaphor in pragmatics, or on the pragmatics of metaphor. The former implies pragmatics in the narrow sense (so-called theoretical or Anglo-American pragmatics), the type of pragmatics that has thematised topics such as relevance, intention and pragmatic enrichment. The latter implies pragmatics in the broader sense (so-called social or Continental European pragmatics), as a general perspective on language, focusing on all aspects of communicative language use in context (cf. Jucker and Taavitsainen 2013) on these two types of ‘pragmatics’. The first perspective is taken here. Let us consider an initial example of metaphor, in order to introduce some basic terminology and determine what dimension(s) of metaphor are highlighted in the pragmatic perspective: (1) For some people, this task is a real breeze. A type of minimal definition of metaphor (and this is about the maximum level of detail that scholars of metaphor agree about) is that it is a phenomenon in language through which something is seen as something else: here, a ‘task’ is seen as a ‘breeze’. There is thus a relation between two things (called source and target, here ‘breeze’ and ‘task’). This relation is not straightforward. Rather, there is some tension, and this is how we recognise a metaphor: as a tension between the literal meaning of breeze and a metaphorical meaning. When this tension is rendered in a definition of metaphor as a clash between ‘saying one thing’ and ‘meaning something else’, then we are squarely within the domain of pragmatics, which deals with distinctions between ‘what 323

Miriam Taverniers

is said’ and ‘what is communicated by the speaker in a specific context’. From this perspective, then, it will be seen that metaphor is a pragmatic phenomenon par excellence. If we now return to the ‘problem set’ of metaphor in pragmatics, it is clear that the tension between a literal and a metaphorical interpretation will play a major role, and more specifically, the idea that metaphorical meaning is in a way indirect will definitely be on the agenda. A major question, which will form a running thread throughout, is: how is the metaphorical meaning launched in relation to the literal meaning? This chapter is organised in terms of the unfolding story of how metaphor has been looked upon in theoretical pragmatic accounts of language. The starting point is an exploration of the birth of the first pragmatic models of metaphor against the background of the then prevailing views of metaphor in linguistics. After that we will see how distinct dimensions of metaphor have been thematised in the course of the development of modelling metaphor in a pragmatic theory.This chapter ends with an overview of issues that have emerged in the story of metaphor in pragmatics and are not yet solved, and other challenges that are likely to play a role in future research of metaphor from a pragmatic perspective.

Historical foundations Theoretical context It is useful to briefly consider the theoretical context of the origin of pragmatic models of metaphor. They developed against the background of older and concurrent (broadly semantic, linguistic and/or philosophical) views of linguistic meaning in general and metaphor in particular. In the 1960s–1970s, there were broadly three such views on ‘meaning’: (1) A lexicalist view, which focuses on lexical/semantic features (of words) and selection restrictions (of words in the context of a syntagm, such as a clause). Metaphor is here seen as an anomaly (a word does not have its ‘ordinary’ lexical features) or a violation of selection restrictions (a word occurs in a clause where it is normally not expected). The interpretation of a metaphor is based on a cancellation of selection restrictions or transfers of lexical features (e.g., Matthews 1971; Levin 1977). In this perspective, example (1) is regarded as semantically anomalous, because breeze applies to types of wind or air current, and not to human activities. Thus the selection restriction [+ applies to air current] is violated. In the interpretation of the metaphor, this restriction is cancelled, and other selection restrictions of breeze are transferred to the target lexeme task, for example [+ moves quickly/smoothly/ easily]. (2) A referentialist view, which focuses on reference (of words) and truth value (of sentences). From this perspective, a metaphorical expression is seen as untrue/false (it is not literally true), and the interpretation of a metaphor is based on broader, encyclopedic knowledge of the characteristics of the empirical referents of lexical items (e.g., Sanders 1973). In this view, example (1) is seen as empirically false: it is not empirically true that ‘this task is a breeze.’ In interpreting the metaphor, language users draw on their non-linguistic knowledge about breezes, for example, that they are light, gentle winds, and then transfer this knowledge onto the target referent, so that for instance ‘light’ becomes an empirical feature of the entity that this task refers to. (3) An interactional view, which also brings in broader knowledge (also called knowledge of gestalts, associations), and focuses on an interaction between gestalts or concepts in the interpretation of a metaphor. In creating or understanding metaphors, assumptions about a source concept 324

Metaphor in pragmatics

and a target concept are combined (interact), and this leads to (new) metaphorical meaning (e.g., Black 1954, 1979; Reddy 1969). With regard to example (1), what is emphasised in this view is that conceptual knowledge of both breezes and tasks is combined and aspects of the two concepts are integrated in the interpretation of the metaphor. For instance both tasks and breezes have a certain speed; they have a certain degree of strength (and ‘strength’ can then mean ‘wind force’ in the case of the breeze and ‘difficulty’ in the case of the task): such pieces of conceptual knowledge are combined in order to come to an understanding of ‘a task that is a breeze’. As will become clear, these distinct views on meaning will be important in order to understand the various pragmatic proposals that have been made.

The birth of the pragmatic approach to metaphor Within pragmatics largely two distinct models can be recognised which have been foundational for further developments with regard to language in general, and metaphor in particular, viz. the Gricean programme and speech act theory. Both frameworks start from the idea that the meaning of an expression is not just its sentence meaning (or word meaning), understood as linguistic meaning in the traditional sense, hence as having to do with denotation/sense relations and/or reference/truth value of words and clauses (however these are defined, which is a fundamental theoretical problem of its own). There is a second dimension to meaning, which has come to be called utterance meaning (or speaker’s meaning; speaker utterance meaning): this is the speaker’s intended meaning, a meaning which does not arise solely on the basis of sentence meaning, but depends on how the utterance is taken in a particular context (cf. Grice 1975; Searle 1975). A distinction is thus made between ‘what is said’ and ‘what is meant by the speaker’, and the task of pragmatics, as this approach has come to be called, is to account for the meaning the speaker intended in the specific context, and the processes hearers use to get at this meaning. In more general terms, this pragmatics is about speakers’ intentions and hearers’ needs. Although early pragmatic views of metaphor are set off from a (more traditional) semantic approach, it should at the same time be recognised that they naturally follow from it (cf. Leezenberg 2001): starting from the view that metaphor can be seen as a falsity (referentialist view) or anomaly (lexicalist view) at the semantic level, what has to be accounted for is the fact that the speaker who utters a ‘false’ or ‘anomalous’ expression still wants to communicate a meaning: the utterance meaning. The initial proposals to theorise ‘utterance meaning’ within a more general theory of intentions and needs are specifications of (various types of) meanings which are beyond and above sentence meaning. Austin (1962) and Searle (1975) define different types of speech acts as different from ‘what is said’ or locutionary meaning (direct and indirect speech acts; illocutionary and perlocutionary meanings) (see also Chapter 18 on Intention by Navarro [this volume]). Grice (1975) introduces various types of implicature (particularised and generalised conversational implicature, and conventional implicature), which are inferences the hearer makes on the basis of the assumption that the speaker’s utterances are meant to be efficient (an assumption referred to as the Cooperative Principle, and modeled by Grice in terms of a number of maxims of conversation). Metaphor is then characterised in relation to those types of utterance meaning. Grice (1975) regards metaphor, together with other ‘figurative language’, viz. irony, meiosis (understatement) and hyperbole, as a conversational implicature arising from the speaker’s (deliberate) flouting of the maxim of quantity: ‘what is said’ is not true,1 and the hearer to whom this is clear infers further meaning, beyond the literal meaning and in accordance with the context. Some authors 325

Miriam Taverniers

Figure 25.1  Early pragmatic views on the relation between literal and non-literal meaning

(e.g., Loewenberg 1975; Mack 1975) working in the framework of speech act theory regard uttering a metaphorical expression, or ‘metaphoring’, as a type of speech act. The literal expression is here regarded as a locutionary meaning, and the metaphorical one as illocutionary. Another possibility is to see metaphor as an indirect speech act (cf. Searle 1979). In this case the literal expression is an illocution (e.g., statement), and the metaphorical meaning is a perlocution (it has an effect on the hearer: it makes the hearer think about other aspects so that a meaning beyond the illocution is arrived at). In these studies, metaphor is a second-order phenomenon, in which sentence meaning and utterance meaning are separated, and the utterance meaning is reached through implicature or (indirect) speech act (see Figure 25.1 for a visualisation). One study which deserves special attention is Searle’s (1979) ‘Metaphor’. In addition to characterising metaphor as an indirect speech act, Searle pays attention to two other dimensions of metaphor: how the metaphorical meaning is resolved, and different degrees of conventionality. In contrast to Grice, for whom a metaphor is resolved through pragmatic inferential processes (implicatures) on the part of the hearer, Searle also refers to the role of shared, stereotypical associations in the interpretation of metaphors. With reference to the example in (2), he refers to associations of pigs with ‘filthiness’, ‘sloppiness’ and the like (Searle 1979: 105). (2) Sam is a pig. (Searle 1979: 98) Searle distinguishes three degrees of conventionality. (i) In a simple metaphorical utterance, the speaker says one thing but means something else; (ii) in an open ended metaphorical utterance, the speaker says one thing but ‘means metaphorically an indefinite range of meanings’ (Searle 1979: 110). In these two types, the metaphorical utterance meaning is arrived at by going through literal sentence meaning. This is in contrast to a third type, (iii) dead metaphor: here the literal meaning is ‘bypassed’, and a former metaphorical meaning becomes the new literal meaning.

Features of the pragmatic perspective on metaphor that appear from early studies The pragmatic perspective on metaphor that appears in early studies such as the Gricean programme and early speech act theory are characterised by the following features: (a) ‘Literal meaning’ is taken as a starting point: in general, ‘what is meant’ is defined in relation to what is said, and in metaphor, hearers infer a secondary, metaphorical meaning because they realise that the literal meaning ‘does not work’ (the sentence is false or anomalous). (b) Metaphor is seen as one type of a more general phenomenon, viz. figurative or non-literal language, which is placed on a par with (indirect) speech acts and other things that fall under the heading of ‘conversational implicature’ (irony, understatement, hyperbole).

326

Metaphor in pragmatics

(c) The focus in general appears to be on how the hearer recognises metaphor as something ‘special’. That ‘something special’ is either seen as a (special type of) (indirect) speech act, or as based on a conversational implicature: by using general pragmatic processes such as inferential reasoning the hearer realises that he2 must look for a meaning beyond what is literally said.The approach of metaphor as requiring inferential processes can be seen as the pragmatic alternative to the narrow semantic conception of meaning and metaphor. It will be referred to as a contextual-inferential view. (d) Less is said on how the hearer construes the interpretation of a metaphor, although Searle, who also takes into account shared conventions, is an exception in this respect (cf. also Searle 1989, who focuses on this aspect in his analysis of differences between Gricean pragmatics and speech act theory). In terms of the distinction made earlier, it is clear that Searle allows for a conceptual-associative dimension in characterising metaphor, although he does not specify how conceptual aspects come into play. (e) No distinction is made between novel and conventional metaphors. Because the pragmatic perspective, at least in early studies, focuses on how metaphor is recognised as a conversational implicature, it is relatively creative metaphors that pragmatists have in mind (because otherwise, there is nothing peculiarly ‘metaphorical’ to recognise). As we saw, Searle’s (1979) pragmatic account is an exception in this respect too. As we will see, most of these features will become tendencies in later pragmatic conceptions of metaphor.

Outlook: the Gricean gap and how to fill it After the foundational models sketched earlier, later pragmatic accounts are to a large extent concerned with filling the Gricean gap between ‘what is said’ and ‘what is communicated’. Most of these build on Grice’s programme and are guided by the idea that the bifurcation between on one hand ‘sentence meaning’ and on the other ‘utterance meaning’ as arrived at through conversational implicature is too coarse. Conversational implicature3 is generally kept as a key pragmatic process in deriving the speaker’s intended meaning and pragmatics becomes largely concerned with modelling what happens in between ‘what is is said’, which is taken as the starting point, and arriving at ‘what is communicated’ through conversational implicature. As we will see, several proposals are made to recognise a step in between, which will be referred to as an intermediate level of meaning. In the pragmatic literature the issue of what step(s) is/are needed in accounting for metaphor and non-literal language in general has been thematised in terms of two questions (which are not independent, as we will see): (i) What is the role of (conversational) implicature, or pragmatic inferential processes in general? and (ii) How is metaphorical meaning construed (i.e. how is the tension resolved)? The next two sections take each of these questions up in turn.

What is the role of implicature, and inferential processes in general? The first pragmatic elaboration of Grice’s programme focuses on the notion of implicature and on what role it plays, at various points, in getting from ‘what is (literally) said’ to ‘what is (metaphorically) meant’. Considerable attention will be paid to this stage in the development of the pragmatics of non-literal meaning, because the distinctions proposed here form the barebones of later refinements.

327

Miriam Taverniers

Implicature and explicature: relevance-based pragmatic inferences are everywhere Sperber and Wilson (1986a) collapse Grice’s assumptions about efficient communication (his maxims) into one principle, namely the Principle of Relevance: the speaker intends to convey relevant meaning. In their model, which forms the basis of Relevance theory (henceforth RT) (see Chapter 24 on Relevance by Assimakopoulos [this volume]), one level of inferential processes is added just beyond ‘what is said’, namely, explicature. Explicature, as an intermediate level of meaning, is the adaptation of ‘what is said’ to the immediate context of the utterance, by resolving reference, interpreting deictic elements (such as tense; see also Chapter 19 on Temporal Reference by Jaszczolt [this volume]), and by disambiguating expressions. The explicature of ‘what is said’ then forms the basis for further implicatures which make the expression relevant in the given context. Figure 25.2 visualises explicature as Sperber and Wilson’s (1986a) proposal to insert an intermediate level of meaning between ‘what is said’ and ‘what is communicated’. Sperber and Wilson (1986a, b) discuss metaphor only very briefly. On one hand, metaphor, as in Grice’s theory, is seen as a conversational implicature. On the other hand, the interpretation of a (conventional) metaphor by a hearer, which is a disambiguation of the expression as having a figurative (and not a literal) meaning, and thus an explicature of ‘what is said’, is described as based on ‘stereotypical assumptions’ (Sperber and Wilson 1986b: 167) that the hearer derives from his encyclopedic knowledge of the concept that is used metaphorically. With regard to (3), which Sperber and Wilson adduce as an example of a ‘highly standardised’ metaphor, they refer to assumptions of filthiness and untidiness. These assumptions arise because the encyclopedic information for the concept encoded by pigsty is put together with that encoded by room: assumptions which are activated are thus relevant in the context (of talking about a room). Important, the derivation of (metaphorical) meaning through retrieving stereotypical assumptions which are relevant in the context is seen by Sperber and Wilson as based on implicatures. (3) This room is a pigsty. (Sperber and Wilson 1986a: 236) Note that the way in which Sperber and Wilson (1986b) draw on encyclopedic information in describing how metaphorical meaning arises is very reminiscent of the interaction view of metaphor (cf. earlier). However, apart from mentioning that the metaphorical meaning is generated on the basis of implicatures, they do not specify how this meaning, as a combination of encyclopedic assumptions, is constructed.

Figure 25.2  Sperber and Wilson’s (1986a) explicature and its relation to implicature

328

Metaphor in pragmatics

In their analysis of the example in (3), Sperber and Wilson (1986a: 236) go on to indicate that the meaning ‘this room is filthy and untidy’ forms the basis for a further implicature: the hearer infers that the meaning the speaker intends must be something else, because otherwise she would literally have said ‘This room is filthy and untidy.’ The further implicature, in Sperber and Wilson’s analysis, is that the room is filthy and untidy ‘beyond the norm’.

Grice’s circle and why it is really a spiral Levinson ([1988] 2000) criticises the notion of explicature. According to him, it leads to a circularity: conversational implicatures come into play when the sentence meaning is explicated (they do not work on ‘what is said’, but on the explicature), and the explicature of meaning depends on the same inferential processes as implicature, so essentially these are the same processes. This circularity is a problem which Levinson ([1988] 2000) dubs ‘Grice’s circle’ (see the visualisation in Figure 25.3). This perception of a cycle is due to a lack of recognising that there are two sides to an implicature (see Figure 25.4): on one hand the implicature itself as a detection that something more must be meant/intended, on the basis of general assumptions about communication (cooperation, relevance or whatever principles), and on the other hand the resolution of the implicature, viz. how exactly this ‘richer meaning’ is grasped. I referred to this distinction earlier as that between the recognition vs. interpretation of non-literal meaning. Resolution of implicature in many cases requires something more than just the knowledge that the speaker will be cooperative or give relevant information and hence that the hearer must look for another meaning. One example was mentioned previously: Searle (1979), who refers

Figure 25.3  Grice’s circle

Figure 25.4  Two sides of an implicature: detection and resolution

329

Miriam Taverniers

to conceptual/conventional assumptions that are drawn on in interpreting metaphors. As we saw, Sperber and Wilson (1986a) equally refer to encylopedic schemata from which assumptions about the source concept are retrieved. An early work which pays explicit attention to the distinction between implicature and resolution as distinct processes is Horn (1984). Horn (1984) shows how implicatures generated by lexemes can be resolved by narrowing and broadening lexical senses. A type of broadening Horn mentions (1984: 35) is the use of a trade name to convey a more generic meaning, of which some examples of his are given in (4). The implicature here is that the speaker will not have intended to give too much information, so information about the trade name itself is left out of consideration. (4) a. kleenex → any paper tissue b. thermos → any vacuum flask Once it is recognised that an implicature is a process of detecting a need and resolving it, it becomes clear that Grice’s circle is a fallacy: there is no circularity sur place, but rather a cycle of implicature detection and resolution. Re-inserting the notions of ‘what is said’ (as the starting point), ‘what is meant’, and explicature as an intermediate stage, we arrive at the picture shown in Figure 25.5. Note that this idea of successive cycles of implicatures is also implied (no pun intended!) in Sperber and Wilson’s (1986a) analysis of the room is a pigsty described earlier. With regard to this example, successive cycles may include the following: (1) (2)

‘what is said’ the room is a pigsty → (explicature by disambiguation:) ‘the room is untidy (or filthy)’; the explicature of the room is a pigsty is ‘the room is untidy (or filthy)’ → (the speaker did not literally say the room is filthy/untidy, so she must have meant something beyond literally ‘untidy or filthy’:) ‘the room is untidy or filthy beyond the norm’

Figure 25.5  Implicature as a recursive process

330

Metaphor in pragmatics

Two further notes are in order: [1] It is through the recurrent cycles of implicature that what is interpreted as a conventional metaphor at one level (level (1)) can be the input for a further implicature (level (2)) which opens up the (conventional) metaphor and adds a further aspect to it.4 [2] Notice also that the cycle of implicature can continue to recur, beyond the level of resolving a metaphorical meaning. In certain contexts, there might be a further implication (cf. Thibault 1999: 587): (3)

the intended meaning is ‘the room is untidy beyond the norm’ → (untidiness beyond the norm is not desirable or acceptable (probably, in the eyes of the speaker), so she must have meant something extra:) ‘the speaker wants the hearer to clean the room’.

As Sperber and Wilson (1986a) indicate, how many implicatures are involved in metaphor is dependent on how creatively metaphorical an expression is (and, on the basis of the third cycle indicated earlier, it can be added that it also depends on whether the utterance is also intended or can be interpreted as a perlocutionary act (i.e. an expression that has an effect on the hearer, for instance, requiring action, such as cleaning the room in our example)). This brings us to the notion of degrees of conventionality.

Degrees of conventionality: adding a dynamic dimension Although, as said earlier, the treatment of metaphor in Sperber and Wilson (1986a) is very brief, they also pay attention to the difference between conventional (‘highly standardised’ in their terminology) and novel or creative metaphors. The pigsty example is given as an illustration of the former. The conventional metaphor is resolved at the level of explicature: it is a disambiguation of one sense from another, both of which are part of the code of language. The conventional metaphorical sense is arrived at on the basis of one or two assumptions about pigsties (‘untidy’, ‘dirty’) which immediately come to mind as relevant in the given context. Sperber and Wilson (1986a) call these types of conventional assumptions (note the parallel with Searle (1979)) strong implicatures. Strong implicatures give meanings that are almost certainly intended to be conveyed by the speaker: it is relatively clear that these implicatures apply, because they are relatively standardised. The conventional metaphorical meaning ‘an untidy place’ may be just one of different senses a word has. In this sense then arriving at the meaning of a conventional metaphor is disambiguating a polysemous word. Relatively creative metaphors, such as (5) given by Sperber and Wilson (1986a), are different in that their explicature is indeterminate, because several weak implicatures arise: in this case, they may include that John is stubborn, persistent, obstinate, insensitive. Weak implicatures are only potential implicatures, and it is not certain whether the speaker has intended the meanings that they give rise to. (5) Robert is a bulldozer. (Sperber and Wilson 1986a) Thus in Sperber and Wilson (1986a) the distinction between conventional and creative metaphor is analysed as one between (i) few strong implicatures resulting in a disambiguated explicature on one hand, and (ii) a range of weak implicatures which give an indeterminate explicature on the other. Poetic metaphor is precisely based on the generation of such a wide range of weak implicatures: ‘the wider the range of potential implicatures and the greater the hearer’s responsibility for constructing them, the more poetic the effect, the more creative the metaphor’ (Sperber and Wilson 1986a: 236). 331

Miriam Taverniers

It is clear how Sperber and Wilson’s (1986a) interpretation of relatively standardised and relatively creative to poetic metaphors ties in with Searle’s (1979) views on degrees of conventionalisation. The opposite of highly creative metaphor is dead metaphor. We saw how Searle describes this in terms of the metaphor of bypassing the original literal meaning. It is useful to bring in, at this point, another metaphor that has been put forward (initially by Morgan 1978, see also Horn 1984), which resonates with Searle’s but which characterises conventionalisation in relation to implicature. Morgan (1978) interprets conventionalised uses in language as shortcircuited implicatures. These implicatures can still be calculated (for this reason, Goatly 1997) replaces the term dead metaphor by sleeping metaphor), but after short-circuiting they are not usually calculated during conversation.

The place of explicature in relation to implicatures We can now return to the issue of the place of explicature in relation to implicature. Much pragmatic work in the past three decades focuses on how explicature (or another intermediate level of meaning)5 and implicature can be kept apart, and/or how the division of labour between them can be understood. Different types of distinctions in the zone between ‘what is said’ and ‘what is communicated’ have been suggested, and most of these remain debated to the current day (e.g., Horn 2004; Carston 2008). What is clear is that explicature occurs at the beginning of any cycles of implicature between ‘what is said’ and ‘what is meant’: in other words, it is reserved to refer to a first level of ‘what is meant’ beyond ‘what is said’. The question is: what makes this level so special that it deserves a separate term? This can be looked at in two ways which are complementary (and not incompatible, as has sometimes been assumed, cf. also Carston 2005), and a third way which takes into account the diachronic dimension. Note that these possibilities, which are visualised in Figure 25.6 and summarised later, are three different ways in which the first step in proceeding from ‘what is said’ to ‘what is meant’ can be theorised. Taking example (3) as an illustration, this first step refers to the interpretation of this room is a pigsty as ‘this room is dirty’ (before further inferences are made, such as ‘it is dirty beyond the norm’ and ‘it should be cleaned’). (1) On one hand, a distinct type of implicature is at stake. This is the view taken by Levinson (2000) and neo-Gricean pragmatics, the framework of which he is one of the major proponents: the implicatures that lead to a first level of meaning, which in this framework is called presumptive meaning or utterance-type meaning, are generalised conversational implicatures, rather than particularised conversational implicatures, which are the further (higher, in terms of the visualisation proposed in this chapter) types of implicatures that arise on the basis of presumptive meaning. (2) On the other hand, a distinct type of implicature resolution is at stake: the resolution of implicatures at the level of explicature interacts with what is coded in the language, that is, with the syntax and semantics of ‘what is meant’. This is what happens in processes of disambiguating senses, resolving reference and fixing deictic elements. This is the view that is highlighted in more recent Relevance theory, which has developed the idea of implicature resolution at the level of explicature beyond the three types of processes just mentioned (disambiguation, reference and deixis resolution), as we will see in the next section. (3) There is a third way in which the first level of inferences, or an intermediate level of meaning, is different from the other, higher levels of implicature: as the level closest to the language code (‘what is said’), it is the level which is by definition at risk of conventionalisation.When implicatures are so strong that they are conventionally conveyed, they are prone to absorption in the language code, that is, in ‘what is said’ (and thus to being short-circuited). 332

Metaphor in pragmatics

Figure 25.6 Three ways in which a first intermediate level of meaning between ‘what is said’ and ‘what is meant’ is ‘special’

How is the interpretation of metaphorical meaning construed? A change in focus Most current pragmatic theories recognise that there is an intermediate level of meaning, interpreted as a level of pragmatically enriched ‘what is said’: it is ‘what is said’ enriched with information that derives from pragmatic inferences (e.g., disambiguation, reference resolution). Pragmatic processes which pertain to this level are called primary pragmatic processes; secondary pragmatic processes is the term reserved for further inferences, once the explicature or pragmatically enriched content is derived. Burning issues remain on how this level of enriched meaning is to be delineated, and we will turn to some of those later. Pragmatic enrichment is due to certain processes that make ‘what is said’ more relevant in the given context: they add meaning to ‘what is said’, which is itself seen as indeterminate, viz. linguistic code not yet adapted to the specific situation. These processes depend on pragmatic information about the context. In RT they are seen as inferential, more specifically, they are regarded as generated by implicatures (a point of contention which we will turn to later).Within the more general view of cycles of implicatures as sketched earlier, these processes of pragmatic enrichment, when they are seen as related to implicatures, pertain to the resolution side of implicatures. This means that there is a shift in focus compared to earlier models which highlighted the role of implicature and how it arises. Next to the question of how, through implicatures, the hearer infers that ‘another meaning’ rather than the literal one must be intended in a non-literal expression (he has to resolve the implicature), what becomes (equally) important in the pragmatic modelling of metaphor is the question of how exactly this other, richer, nonliteral meaning is grasped (i.e. how the implicature is resolved). Thus, what is thematised is the question of how the interpretation of metaphor is construed, against the background that the recognition of metaphor is due to implicatures.

Loose language and ad hoc concept creation through broadening and narrowing lexical concepts The focus on how the interpretation of a metaphor is construed necessarily brings in conceptual information. We saw how in traditional RT terms, the interpretation of metaphor is based on information retrieved from the hearer’s encyclopedic knowledge about the source concept. In conventional metaphors, there are one or two assumptions about the source concept that immediately come to mind as relevant in the given context (i.e. they are strong implicatures) and that 333

Miriam Taverniers

lead to lexical disambiguation. Creative metaphors lead to a richer range of assumptions, which do not disambiguate the expression (except that it is to be taken as a metaphor, and not literally) and leave the interpretation open to the hearer. What is still not accounted for (compare earlier models such as Searle 1979 briefly described earlier) in this theory is how exactly this conceptual information comes into play (how is it generated or computed?). Modern RT accounts (e.g., Wilson and Sperber 2002; Wilson and Carston 2006; Sperber and Wilson 2008) regard metaphor and other types of non-literal language as examples of loose language. Metaphorical words are used loosely in that the meaning they get in a specific context diverges from their standard meaning, that is, the meaning encoded in the language. It is clear that what is intended here is relatively novel metaphor, since conventional metaphors’ meanings are encoded in language and can be retrieved through disambiguating polysemy, as we saw previously. Such context-specific meanings are called ad hoc concepts (Wilson and Carston 2007; Carston 2010). An ad hoc concept is thus a pragmatic adjustment (or modulation) of a linguistically encoded concept (a lexical concept). In keeping with general relevance-theoretic insights, ad hoc concepts are created on the basis of implicatures deriving assumptions from the hearer’s encyclopedic/conceptual knowledge about the lexical concept. This basic idea is then further refined in three ways: (1) About the assumptions retrieved from the encyclopedic knowledge of the lexical concept, it is argued that some are more easily accessible (i.e. more at risk of being activated to resolve the implicature) because of their higher relevance to the specific context, which appears from ‘other concepts in the sentence uttered’ and ‘the wider discourse context’ (Carston and Wearing 2011: 287). In the case of metaphor, this context thus includes expressions that refer to the target concept, and the understanding of the metaphor depends on bringing together encyclopedic assumptions of the source and target concepts. This is very much in vein with Sperber and Wilson’s (1986a) model of metaphor, and, as we saw, the older interaction theory. What is new is that these assumptions are further characterised as emergent properties. Taking the bulldozer example ([5]) as an illustration: ‘stubbornness’, or ‘persistence’, or ‘insensitiveness’ are not properties that bulldozers (as lexical concept) standardly have (Carston 2002: 350). Instead, these are human properties that emerge, ad hoc, by putting together bulldozer and Robert. This is essentially carrying through the interactionist model of metaphor: there is not just a bringing together of source and target assumptions, but new assumptions appear on the basis of their interaction, that is, emergent properties. (2) The notion of ad hoc concept is introduced to tackle exactly how these emergent properties come into play. The pragmatic modulation of a lexical concept through which an ad hoc concept is created is based on processes of narrowing and broadening (loosening) the lexical concept. Emergent properties are ‘carried over’ into the construction of an ad hoc concept. In this way, the lexical concept is at the same time broadened (it contains more information, namely the newly emerged properties) and narrowed (some information that is conveyed in the linguistically encoded concept is not relevant and is left out of consideration in the ad hoc concept). With regard to the bulldozer example, the broadening consists of the addition of the newly emerged properties (‘stubbornness’, ‘persistence’ and so on). The narrowing pertains to the fact that other assumptions about bulldozers are left out of the picture (e.g. their efficiency in land clearing). (3) The construction of an ad hoc concept on the basis of a lexically encoded concept is a process of ‘mutual adjustment of explicit content and implicit import’ (Carston 2002: 146), where the former refers to the derivation of explicatures as enriched meaning and the latter to inferences due to implicatures. From Figure 25.7 it can be seen how these refer to the two processes which serve as input to the level of enriched ‘what is said’. Both of these processes are 334

Metaphor in pragmatics

Figure 25.7  Ad hoc concept formation as based on a mutual adjustment of two inputs

seen as providing ‘interpretive hypotheses’ about the communicated content. The view is that multiple such hypotheses arise, rapidly and in parallel, in order to arrive at a meaning that is adapted to the context. Within the model of ad hoc concept creation, metaphor is thus accounted for at the level of primary pragmatic processes because the processes of broadening and narrowing pertain to lexical concepts encoded by words metaphor is now a topic in lexical pragmatics (cf. Wilson and Carston 2007).

Metaphor interpretation at the semantics/pragmatics interface, and the role of implicature again: challenges for further research We can now return to the question of why in recent RT pragmatic enrichment is regarded as a distinct level of meaning in getting from ‘what is said’ to ‘what is meant’. The reason is more fleshed out than just the view that primary pragmatic processes are unique in that they interact with the linguistic code (which, as we have just seen, pertains to a modulation of coded meaning). In RT primary pragmatic processes are seen as leading to an enrichment of the propositional content of the utterance, whereas higher implicatures are non-committal (they can be cancelled). However, this is a highly controversial view, and the discussion of it is completely embedded within the wide-ranging theoretical discussion of the semantics-pragmatics interface. Many open questions are bound to be on the agenda of future pragmatic research. Some of those that pertain to metaphor are summarised next: [1] A first issue is the very nature of primary pragmatic processes required in the interpretation of metaphor, which remains a bone of contention. More specifically, the question is whether the retrieval of encyclopedic assumptions (feeding into the creation of ad hoc concepts) should be seen as an inferential process, as is the case in RT, where it is dependent on implicatures. An alternative view, advocated by Récanati (2004), is that these processes are not inferential, but purely associative (see Mazzone 2011 for a discussion of the difference). In this view, implicature has no role in the retrieval of these assumptions, and only comes into play as a secondary pragmatic process (after the enriched meaning is derived, which is called intuitive content in Récanati’s framework). A related issue, also discussed by Mazzone (2011), is the role of schemata (systems of associations) in the construction of a metaphorical meaning. It is possible that both types of processes (inferential, drawing on implicatures and associative, drawing on the activation of schemata) are complementary in the use of metaphor. If this turns out to be the case, it will be desirable for pragmatic theories of metaphor to engage with cognitive linguistic models of metaphor, in which conceptual schemata play a major role. 335

Miriam Taverniers

[2] The issue in [1] pertains to the question of what knowledge a language user needs in order to use metaphors, and the types of knowledge which are suggested as complementary in [1] are on one hand, knowledge of the rules of communication (i.e. pragmatic knowledge in the narrow sense, such as the assumption that the speaker acts intentionally and strives to make her contribution relevant in the given context); on the other hand, conceptual knowledge about the world and how humans interact with it.   However, the question of what knowledge is needed to use metaphors can be opened up further: what also comes into play is our knowledge of social-semiotic aspects of metaphor use. This is cursorily hinted at when relevance theorists mention that the retrieval of assumptions from encyclopedic/conceptual knowledge depends on how the metaphorical word is used in the sentence and the discourse. RT appears to have a narrow view on this or at least currently does not seem interested in pursuing this observation about the linguistic embedding of metaphors in sentences and discourse. It is striking, in this respect, that most metaphors that are adduced and analysed as examples are predicational in form (of the type This room is pigsty, Robert is a bulldozer). A challenge for pragmatic theories of metaphor is to break loose of this restriction which is based in the paradigm of predication logic (recall the referentialist view of linguistic meaning and how pragmatics was set up to depart from this: a figurative sentence is not true, that is, it is not a true proposition or predication). In doing so, pragmatics should engage with work on the language of metaphor and on the issue of how metaphors are signalled in text through the use of various types of expressions (cf. Goatly 1997; Taverniers 2002; Steen 2007: chapter 11), and with ‘social-semiotic’ studies (in the sense of Halliday’s (e.g., 1978) social-semiotic theory of language and register/ genre) which have provided insights on how metaphors are used and function in registers and genres (cf. Thibault 1997). [3] Another remaining challenge, as hinted at earlier, is the delineation of the level of enriched ‘what is said’ (‘explicature’ in RT), that is, the level of primary pragmatic processes. Carston mentions, at distinct places, two issues, each of which pertains to one of the two inputs to the level of enriched meaning (recall Figure 25.7 in which these two inputs are highlighted): [3.1] Delineation of enriched meaning in relation to implicature cycles: For any pragmatic process that is seen as contributing to utterance meaning, ‘the issue remains whether it contributes to explicature. . . . or is an independent, implicitly communicated, proposition’ (Carston 2002: 185). [3.2] Delineation of enriched meaning in relation to a minimal ‘what is said’: another main issue concerns the question of ‘whether or not there is need for a minimal level of “what is said” in addition to this pragmatically-imbued level of truth-conditional content’ (Carston 2005: 308). [4] It is striking that the distinction between conventional and creative metaphor is not systematically dealt with in RT. However, it can be argued that the issues mentioned in [3] are to be seen as completely relativised in relation to the time scale of language evolution. For instance, when implicatures become so strong as to become conventionalised (i.e. to be short-circuited), the distinction between the explicature thus derived and ‘what is said’ becomes minimal, so much so that the enriched meaning due to the original implicature can be absorbed in the linguistic code, and thus can become part of ‘what is said’. Carston (2004: 653) states that conventional implicature, which is ‘a category which simply does not arise within relevance theory’ is ‘in need of a radical reworking’. It is possible that conventional implicature may play a role in clarifying the interaction between the instantial creation of (metaphorical) meanings on the basis of what is encoded in language, and how 336

Metaphor in pragmatics

these meanings can become codified themselves. In general, an important challenge for pragmatic models of metaphor is the issue of how metaphor interacts with different time scales (within language, and within the unfolding of discourse). For promising insights in terms of dynamic open systems theory, see Gibbs and Santa Cruz (2012).   An important sub-theme within this issue is the conventionalisation (or codification) of metaphor. In order to come to a better understanding of this process and what this means for the pragmatic conception of metaphor, pragmatics will have to connect with work on lexicalisation and grammaticalisation, and especially insights on processes such as pragmatic strengthening and semantic bleaching (see Chapter 9 Variation and Change: Historical Pragmatics by Jucker and Landert [this volume] and also Chapter 22 Grammaticalization by López-Couse and Seoane [this volume]). [5] Experimental pragmatics is still relatively young, and it is likely that empirical testing of hypotheses about metaphor processing will play a growing role in future research. With regard to metaphor processing, there are broadly two competing hypotheses (see Bambini and Resta 2012 for an overview). (1) According to the literal-first hypothesis (or sequential view), metaphor comprehension occurs in sequential stages: accessing literal meaning, rejecting it, and searching for a new metaphorical interpretation. (2) According to the directaccess hypothesis (or parallel view), metaphorical meaning is accessed directly, simultaneously with the literal meaning. So far only scattered evidence has been yielded (cf. Bambini and Resta 2012). Cognitive linguistic theories of metaphor and psycholinguistic models of language processing make similar or comparable claims about the online processing of metaphors, and experimental pragmatics will have to connect to work in those frameworks. [6] A final issue is the raison d’être of metaphor as a distinct linguistic phenomenon. As we saw throughout this chapter, within pragmatic theories, metaphor is usually dealt with as one type of a more general phenomenon of language, viz. non-literal language use or loose language. A remaining question is: what aspects are peculiar to metaphor to warrant its recognition as a linguistic phenomenon in its own right? (cf. Carston and Wearing 2011 for some initial proposals to recognise metaphor as a unique phenomenon).

Conclusion In this chapter we have seen how the current pragmatic conception of metaphor developed in broadly three stages. In a first stage (Searle, Grice), metaphor is recognised as a secondary phenomenon, viz. a speech act or implicature. In the latter view, metaphor is based on flouting a maxim of conversation: through this the hearer recognises that there must be something more than just the literal meaning. In a second stage (e.g., early Relevance theory, Sperber and Wilson) there is no longer a flouting of maxims, but the use of metaphor is seen as guided by a general principle of communication, viz. the assumption that the speaker gives relevant information. Attention is paid, albeit not systematically and in brief terms, to degrees of conventionality in metaphor, interpreted as a distinction between a limited number of strong (in conventional metaphor) vs. a larger range of weak (in creative metaphor) implicatures that contribute to the explicature of what is said. The focus here is on how metaphor can be recognised in relation to the process of implicature. In the most recent stage, pragmatics thematises the level of explicature and the way in which ‘what is said’ is modulated to receive a meaning that is relevant in the context (primary pragmatic processes).The focus here is on how the interpretation of metaphor is construed, and this is modelled as a broadening and narrowing of the lexical meaning encoded in language, a process referred to as ad hoc concept formation. 337

Miriam Taverniers

In the final section, I presented remaining and new issues which are challenges for future pragmatic studies of metaphor. These broadly fall into three dimensions: (1) Some issues pertain to internal dimensions of pragmatics, its conceptualisation of metaphor as a pragmatic phenomenon ([1], [2]), and how its insights are to be corroborated through empirical research ([5]); (2) Some issues are to be dealt with in any theory of metaphor ([4] and [6]); (3) Some issues ask for an opening up of the pragmatic study of metaphor to engage with insights from other frameworks ([1], [2], [4], [5]). Some of these insights can be regarded as part of a broader picture of the pragmatics of metaphor, as the study of the knowledge we need in order to use (create and understand) metaphors in texts.This knowledge is not only psychological knowledge of the rules of ostensive communication (i.e. narrow pragmatic knowledge), but also conceptual knowledge about the world, and social-semiotic knowledge (linguistic-semiotic knowledge about the language of metaphors and sociocultural knowledge about the use and function of metaphor in registers/text types and genres). Thus this chapter, which started off as an overview of metaphor in pragmatics, ends with pointers for incorporating other dimensions of metaphor and thus for pursuing a pragmatics of metaphor in the broader sense of the term.

Suggestions for further reading Kövecses, Z. (2010) Metaphor: A Practical Introduction. Oxford: Oxford University Press. This book is an accessible introduction to metaphor from a cognitive-linguistic point of view, with specific attention paid to how metaphor is used in different text types. Ritchie, L. D. (2013) Metaphor: Key Topics in Semantics and Pragmatics. Cambridge: Cambridge University Press. This book offers an overview of how metaphor is used in a range of discourse contexts, with discussions of recent metaphor models. Steen, G. (2007) Finding Metaphor in Grammar and Usage: A Methodological Analysis of Theory and Research. Amsterdam: Benjamins. This is an in-depth meta-theoretical overview of models of metaphor, distinctions and methodological steps in metaphor research. Tendahl, M. and Gibbs, R. W., Jr. (2008) ‘Complementary perspectives on metaphor: Cognitive linguistics and relevance theory’, Journal of Pragmatics, 40(11): 1823–1864. This article explores cognitive linguistic and relevance theoretic accounts of metaphor as complementary perspectives on metaphor, which should both be taken into account in a comprehensive theory of metaphor.

Notes 1 A well-known criticism of this characterisation of the sentence meaning in a metaphor is that it cannot be used as a general criterion for all metaphors, because some metaphors, for instance negative statements as in (i), are literally true (i.e. true in their literal sense). (i)  Life is not a pigsty. 2 I will stick to the common convention in pragmatics to refer to the speaker by she and the hearer by he. 3 More specifically, it is Grice’s particularised conversational implicature which is kept as the last process in deriving the speaker’s intended meaning. The other types of processes which are introduced in various pragmatic theories (and which will be cursorily mentioned later) can be seen as further interpretations of Grice’s conventional implicature (closer to ‘what is said’) and generalised conversational implicature (closer to ‘what is communicated’). 4 This further implicational meaning can be seen as (less) conventional and/or (more) creative (but potentially prone to conventionalization). Compare Dickins (1998: 322), who argues that Sperber and Wilson

338

Metaphor in pragmatics (1986a) in fact treat This room is a pigsty as a creative metaphor since they analyse it in terms of that extra level of implicature, which derives the meaning ‘the room is untidy or filthy beyond the norm’. 5 Other intermediate levels of meaning which are similar to Sperber and Wilson’s (1986a) explicature are impliciture (Bach 1994), utterance-type meaning or presumptive meaning (Levinson 2000), what is said/prag (Récanati 2004).

References Austin, J. L. (1962) How to Do Things with Words. Oxford: Oxford University Press. Bach, K. (1994) Conversational impliciture. Mind & Language: 284. Bambini, V. and Resta, D. (2012) ‘Metaphor in experimental pragmatics: When theory meets empirical investigation’, HumanaMente, 23: 37–59. Special issue: Erva, Francesca & Elisabetta Gola (eds.), Philosophical perspectives on experimental pragmatics. Black, M. (1954) ‘Metaphor’, Proceedings of the Aristotelian Society, 55: 273–294. Black, M. (1979) ‘How metaphors work: A reply to Donald Davidson’, in S. Sacks (ed.) On Metaphor. Chicago: University of Chicago Press. 181–192. Carston, R. (2002) Thoughts and Utterances. Oxford: Blackwell. Carston, Robyn. 2004. Relevance theory and the saying/implicating distiction. In: Horn, Laurence R. & Gregory Ward (eds.) The Handbook of Pragmatics. Oxford: Blackwell. 633–656. Carston, R. (2005) ‘Relevance theory, Grice and the neo-Griceans: A response to Laurence Horn’s “Current issues in neo-Gricean pragmatics” ’, Intercultural Pragmatics, 2(3): 303–319. Carston, R. (2008) ‘Linguistic communication and the semantics/pragmatics distinction’, Synthese, 165: 321–345. Carston, R. (2010) ‘Lexical pragmatics, ad hoc concepts and metaphor: From a relevance theory perspective’, Italian Journal of Linguistics, 22(1): 153–180. Carston, R. and Wearing, C. (2011) ‘Metaphor, hyperbole and simile: A pragmatic approach’, Language and Cognition, 3(2): 283–312. Dickins, J. (1998) Extended Axiomatic Linguistics. Berlin/New York: Mouton De Gruyter. Gibbs, Raymond W. & Malaika Santa Cruz. 2012. Temporal unfolding of conceptual metaphoric experience. Metaphor & Symbol, 27(4): 299–311. Goatly, A. (1997) The Language of Metaphors. London: Routledge. Grice, H. P. (1975) ‘Logic and conversation’, in P. Cole and J. L. Morgan (eds.) Syntax and Semantics: Speech Acts. London: Academic Press. 41–58. Halliday, M. A. K. (1978) Language as Social Semiotic: The Social Interpretation of Language and Meaning. London: Arnold. Horn, L. R. (1984) ‘Toward a new taxonomy for pragmatic inference: Q-based and R-based implicature’, in D. Schiffrin (ed.) Meaning, Form, and Use in Context: Linguistic Applications.Washington, DC: Georgetown University Press. 11–42. Horn, L. R. (2004) ‘Implicature’, in L. Horn and G.Ward (eds.) Handbook of Pragmatics. Oxford: Blackwell. 3–28. Jucker, A. H. and Taavitsainen, I. (2013) English Historical Pragmatics. Edinburgh: Edinburgh University Press. Leezenberg, M. (2001) Contexts of Metaphor. Amsterdam: Elsevier. Levin, S. R. (1977) The Semantics of Metaphor. Baltimore, MD: John Hopkins University Press. Levinson, S. C. (2000) Presumptive Meanings:The Theory of Generalized Conversational Implicature. Cambridge, MA: Cambridge University Press. Loewenberg, I. (1975) ‘Identifying metaphors’, Foundations of Language, 11: 315–338. Mack, D. (1975) ‘Metaphoring as speech act: Some happiness conditions for implicit similes and simple metaphors’, Poetics, 4(2): 221–256. Matthews, R. (1971) ‘Concerning a “linguistic theory” of metaphor’, Foundations of Language, 7: 413–425. Mazzone, M. (2011) ‘Schemata and associative processes in pragmatics’, Journal of Pragmatics, 42: 2148–2159. Morgan, J. (1978) ‘Two types of convention in indirect speech acts’, in P. Cole (ed.) Pragmatics. New York: Academic Press. 261–280. Récanati, F. (2004) Literal Meaning. Cambridge: Cambridge University Press. Reddy, M. J. (1969) ‘A semantic approach to metaphor’, Chicago Linguistic Society, 5: 210–251. Sanders, R. E. (1973) ‘Aspects of figurative language’, Linguistics, 96: 56–100. Searle, J. R. (1975) ‘Indirect speech acts’, Syntax and Semantics, 3: 59–82. Searle, J. R. (1979) ‘Metaphor’, in A. Ortony (ed.) Metaphor and Thought. Cambridge: Cambridge University Press. 356–408.

339

Miriam Taverniers Searle, J. R. (1989) ‘Individual intentionality and social phenomena in the theory of speech acts’, in G. Deladalle (ed.) Semiotics and Pragmatics. Amsterdam: John Benjamins. 3–17. Sperber, D. and Wilson, D. (1986a) Relevance: Communication and Cognition. Oxford: Blackwell. Sperber, D. and Wilson, D. (1986b) ‘Loose talk’, Proceedings of the Aristotelian Society, 86: 153–171. Sperber, D. and Wilson, D. (2008) ‘A deflationary account of metaphor’, in R. W. Gibbs (ed.) Handbook of Metaphor and Thought. Cambridge: Cambridge University Press. 84–105. Steen, G. (2007) Finding Metaphor in Grammar and Usage: A Methodological Analysis of Theory and Research. Amsterdam: John Benjamins. Taverniers, M. (2002) ‘Metaphor’, in J. Verschueren, J. Östman, J. Blommaert and C. Bulcaen (eds.) Handbook of Pragmatics. Amsterdam: John Benjamins. 1–36. Thibault, P.J. (1997) Re-reading Saussure. The Dynamics of Signs in Social Life. London: Routledge. Thibault, P.J. (1999) ‘Communicating and interpreting relevance through discourse negotation: An alternative to relevance theory – A reply to Franken’, Journal of Pragmatics, 31: 557–594. Wilson, D. and Carston, R. (2006) ‘Metaphor, relevance and the “emergent property” issue’, Mind and Language, 21(3): 404–433. Wilson, D. and Carston, R. (2007) ‘A unitary approach to lexical pragmatics: Relevance, inference and ad hoc concepts’, in N. Burton-Roberts (ed.) Pragmatics. Basingstoke: Palgrave Macmillan. 230–260. Wilson, D. and Sperber, D. (2002) ‘Truthfulness and relevance’, Mind, 111: 583–632.

340

26 ENRICHMENT Alison Hall

Introduction It is widely accepted that there is a distinction between what a speaker ‘says’, or explicitly expresses, in making an utterance, and what she implicates, or communicates indirectly. For example, imagine (1) is uttered in response to a question about how a student called Neil is getting on in a new flatshare: (1) He hated everyone and moved out after a month. The speaker of (1) is, clearly, implicating something along the lines that Neil’s new housing arrangement was not to his liking. The hearer grasps this implicature by inferring it from what the speaker explicitly expresses plus easily accessible background knowledge about what makes for a successful flatsharing experience. This much is uncontroversial. But what of the content explicitly expressed by the utterance of (1)? That some pragmatic inference is needed for the hearer to come to grasp an utterance’s explicit content has long been recognized. Certain expressions – pronouns and demonstratives, such as ‘he’ in (1) – require the hearer to assign reference, and where the utterance contains ambiguity, either lexical (e.g., ‘bank’) or structural (‘He hit the man with an umbrella’), the hearer has to disambiguate to arrive at the explicit content. Recent work in semantics and pragmatics, however, has argued that the pragmatic processes involved in getting from the meaning encoded in the sentence uttered to the speaker’s explicit meaning go beyond disambiguation and reference assignment. The subject of this chapter is these other pragmatic processes involved in recovering explicit content, which are known as ‘enrichment’. To see why these enrichment processes are posited, consider that the linguistic meaning of (1) together with assignment of reference to ‘he’ would give (1a). I use small caps to indicate mental representations/contents, as opposed to natural-language expressions. Subscripts on names, pronouns and so forth indicate that reference has been assigned: (1) a.

neili hated everyone and neili moved out after a month.

However, there is a widely shared intuition that what the hearer would understand (and the speaker would intend) to be directly communicated – not merely implicated – by this utterance 341

Alison Hall

is more specific than the meaning given in (1a). For example, the linguistically encoded meaning of ‘everyone’ in (1) includes many more people than the speaker intends to refer to. Instead, it would be clear, given the context, that she means that Neil hated everyone else who was living in the flat. The explicit content plausibly also contains other elements that differ in some way from what was linguistically encoded or resulted from reference assignment; these are shown in (1b), where pragmatically provided meanings are highlighted in bold: (1) b. neili hated everyone else

who was living in flatj at timet and as a result neili

moved out of flatj after living therej a month

Most theorists adopt Grice’s (1967) position that the explicit content of an utterance – what he called ‘what is said’ – is the truth-conditional content of the utterance: that is, the content on the basis of which speaker-hearers would judge the utterance true or false. Implicatures, on the other hand, are separate communicated propositions that are not part of the utterance’s truthconditional content. So which of (1a) or (1b) forms the basis for our truth-value judgments? Here, many would depart from Grice and say that it is (1b): we agree/disagree, judge the utterance true/false, based not on whether Neil hated everyone in the world, but on whether he hated his flatmates; not on whether he moved out of somewhere or other but on whether he moved out of that particular flat, and so on. Since it was recognized that there may be a bigger gap between linguistic meaning and truthconditional content than Grice envisaged, many examples have been identified where pragmatic contributions to explicit content go beyond disambiguation and the assignment of values to pronouns and demonstratives. Consider for now just these two other frequently cited kinds of case: (2) It’s snowing. (3) Typical. In both (2) and (3), it seems that what the speaker would intend to explicitly communicate is something more specific than what results simply from recovering the encoded linguistic meaning, disambiguating and assigning reference to any indexical elements. If (2) is uttered by a speaker in Paris, at midday on 18 December 2013, then intuitively she is expressing the proposition it is snowing in Parisx at midday on 18 december 2013. This looks to be confirmed by the truthconditional content criterion:The truth conditions of the utterance of (2) are not that it is snowing just anywhere, or ‘snowing punkt’, as it is often put, with the specific location featuring only at the level of implicature. To see this, consider that, despite the fact that it is no doubt true that it will be snowing somewhere or other in the world, you would take the speaker of (2) to have spoken falsely unless it is snowing at some specific location that is salient and relevant in the context – in most cases, the location of utterance. The inferred location, then, is part of the explicit content (see Carston and Hall 2012: section 5 for discussion of truth conditions as a criterion for drawing the explicit-implicit distinction). Example (3), meanwhile, appears not to be a full sentence: it does not encode a propositional form, but the speaker would succeed, it seems, in explicitly expressing a proposition, presumably quite obvious in context, such as that behaviour i is typical of johnj. Taking appearances at face value, then, the examples discussed so far all communicate elements of meaning that are not the value of any element of the linguistically encoded material but that are intuitively part of what is directly, explicitly communicated rather than being conversationally implicated.What this suggests is that the explicit content (truth-conditional content) is recovered not only through linguistic decoding plus the linguistically mandated processes

342

Enrichment

of reference assignment and disambiguation, but that pragmatically motivated processes are also involved. These pragmatic processes are known as ‘enrichment’ (or ‘free enrichment’, ‘pragmatic enrichment’,‘free pragmatic enrichment’). Like conversational implicatures, they are linguistically optional – that is, ‘free’ – occurring when triggered by pragmatic requirements (the search for an interpretation that is relevant, informative, coherent and so on), but unlike implicatures, enrichment processes contribute to the explicit content of an utterance. According to many linguists and philosophers, including Bach (1994), Bezuidenhout (2002), Carston (2002), Neale (2004), Recanati (2004, 2010) and Sperber and Wilson (1986/1995), practically every utterance involves enrichment to arrive at the explicit content. That enrichment is pervasive is to be expected on the assumption, shared by the authors listed earlier and many others, that the meaning encoded in the linguistic expression used generally vastly underdetermines any communicated proposition. Given our ability to infer what the speaker means, usually quickly and with little effort, full encoding of a speaker’s message is unnecessary (even if it is possible), and generally undesirable: the linguistic meaning that forms the input to pragmatic inference need not even come close to encoding the thoughts we communicate, as is especially obvious from subsentential examples such as (3). The role of the linguistic meaning, on this picture, rather than to encode a thought or even a thought-shaped logical form, is to provide enough evidence to set the hearer on the right track; our inferential pragmatic ability, drawing on various clues of which the linguistic meaning is just one, can be relied on to do the rest (see especially Sperber and Wilson 1986/1995 for development of this approach, including support for the idea that it is impossible to fully encode our thoughts). Setting aside for now the subsentential cases such as (3), and concentrating just on cases where the linguistic input is a syntactically complete sentence, discussion of three kinds of free pragmatic contribution to explicit content can be found in the recent literature. These are illustrated in (4)–(6), (7)–(9) and (10): (4) a. Mary left John and he started drinking. b. maryx left johny & as a result hey started frequently drinking excessive quantities of alcohol

(5) a. Said by a mother to her child, who is crying about a scratched knee: You’re not going to die b. youx are not going to die from thatj scratch (bach 1994) (6) a. No one goes there anymore. b. no one of any importance/interest goes therex anymore (7) To buy a house in this neighbourhood you need money. (8) I’m tired: let’s stay in and watch a movie tonight. (9) The empty bottles can go in the trash. (10) I’m parked out back. In (4)–(6), the sentence uttered is given in (a); in (b) is the explicit content of each utterance (when appropriately contextualized). The explicit content in each example seems to include a constituent that is not the value of any element of the linguistic meaning, as was also the case with the inferred location in (2) and the domain restriction in (1). Such extra elements are known as ‘unarticulated constituents’, where what is meant by ‘unarticulated’ is not just that these elements are not phonologically realized, but that they are not present at any level of linguistic representation (that is, they are not part of any representation in the language faculty). As discussed in the next section, unarticulated constituents have been the subject of much debate in recent years.

343

Alison Hall

What happens with the expressions in italics in (7)–(9), on the other hand, seems better construed as the pragmatic modification of a simple concept in the encoded logical form so that the concept that forms a constituent of the speaker’s meaning replaces the encoded one. The proposition that buying a house requires some (any) amount of money would, in most contexts, be uninformative and irrelevant, so the denotation of the lexically encoded concept money would be narrowed to exclude sums under a certain amount. In (8), the idea is that the encoded concept tired is likely to be very general, covering a range of states, including a level of exhaustion that would preclude any activity other than sleeping.The communicated concept would be narrower, denoting a far more specific degree of tiredness that excludes those states where one would, say, want to go clubbing or for a run or engage in any more energetic activity, but also excludes those states of exhaustion at the other end of the scale. (9) is a case where the communicated concept would have a broader denotation than the encoded concept: Imagine (9) is uttered while clearing up after a party; in this situation, the denotation of ‘empty’ is likely to include, for example, bottles with small amounts of beer left in them, which are, therefore, not literally empty. Finally, (10) illustrates the phenomenon of ‘predicate transfer’: since the predicate ‘parked out back’ cannot literally apply to the subject, ‘I’, it must be interpreted non-literally, as conveying the property of having a car parked out back (for evidence that it is the predicate that is interpreted non-literally, as opposed to ‘I’, see Nunberg 1995). These modulations of encoded word meaning – narrowing and loosening/broadening – and metonymic transfers are generally agreed to contribute to the explicit content of the utterance. If they didn’t – if the narrower or broader concepts entered the picture only at the level of implicature – then the utterance would be judged trivially true in the cases of narrowing such as (7) and (8), and false in many cases of loosening. But this is not the case: a speaker who included bottles with a few millilitres of beer in them in the denotation of ‘empty bottles’ would, in the context described earlier, not be judged to have said or explicitly expressed something false and only implicated that the bottles in question are close enough to empty to go in the trash. To conclude this subsection, then, enrichment is a free pragmatic process that contributes to the explicit content of utterances, either by adjusting the encoded meaning of individual words used, as in the narrowing/loosening examples (7)–(9) and in (10), or by composing extra concepts into the explicit content, as in (1)–(6). What is meant by saying that a pragmatic effect is ‘free’ is that it is not mandated by any element of the linguistic meaning (indexicality or ambiguous words/structures), but is motivated entirely by pragmatic requirements, such as the requirement for an interpretation that is sufficiently relevant and informative.

Definitions of enrichment It is fairly standard to use ‘enrichment’ to cover whatever free pragmatic processes contribute to explicit content, that is, it would include all the examples used so far. However, there are different uses in the literature. ‘Enrichment’ is sometimes used to refer only to the process of making the interpretation of a particular sub-propositional constituent more specific: for Recanati (2010: 22–24) and Pagin (2014), examples (7) and (8) would count as enrichment; so would the interpretation of ‘drinking’ in (4), and the domain restriction of ‘everyone’ and ‘no one’ in (1) and (6). The broadening in (9) would not be enrichment, nor would the development of the subsentential fragment ‘Typical’ in (3) into a whole proposition, as these do not involve making any constituent more specific. It is unclear whether all of the other free pragmatic effects mentioned so far would count as enrichment for Recanati and Pagin: with the inferred location in (1), ‘It’s snowing’, it is not obvious that it is the interpretation of a constituent, as opposed to the interpretation of the whole sentence, that is made more specific. Despite these differences 344

Enrichment

between different groups of examples, some considerations justify the more inclusive use of ‘enrichment’. First, the broadening illustrated in (9) and the narrowing in (7)–(8) are, arguably, not different processes (see Carston 2002: 334), but rather, different outcomes of the same process of ad hoc concept construction (where an ‘ad hoc concept’ is understood as a non-lexicalized concept, though it need not be a genuinely ad hoc, i.e. novel, construction). The role of the linguistic meaning encoded by an expression is to give access to the material for inferring the intended meaning given the available contextual information. For example, in a context similar to (8), with speaker and hearer deliberating what to do that evening, for ‘I’m tired’ to warrant the intended inferences that the speaker does not want to go out, might like to watch a movie at home and so forth, a concept of tiredness that would warrant such implicatures will be retrieved or constructed. For the utterance in (9), about putting the empty bottles in the trash, to be relevant in a context where there are no literally empty bottles but several with dregs of beer or wine in them, a concept that applies to such bottles, paraphraseable as ‘almost empty of liquid’, will be retrieved/constructed and composed into the explicit content. Whether the communicated concept happens to be narrower or broader than the encoded concept, the process can be seen as a single one of ad hoc concept construction. Support for this position comes from the fact that the ad hoc concept is frequently both narrower and broader than the encoded concept, in that it excludes some objects or properties that fell under the denotation of the encoded concepts, but includes some that did not. Examples of this are (11) and (12), from Carston (2002: 344): (11) There’s a rectangle of lawn at the back of the house. (12) Ken’s a real bachelor [where Ken is technically married] As Carston says, ‘rectangle’ in (11) is used somewhat loosely, in that the lawn in question is unlikely to be strictly rectangular, but given the context of describing the shape of a lawn, the ad hoc concept might also exclude some shapes that are, strictly, rectangles, such as those with two parallel sides a mile long and the other two sides a few millimetres long. In (12), ‘bachelor’ would be understood as denoting a set of human male adults with a certain kind of attitude/ behaviour typical of some unmarried men.This set might include individuals who would not be denoted by the literal meaning of ‘bachelor’, such as Ken, but it would also exclude a substantial proportion of those who do fall in the literal denotation of ‘bachelor’ but do not exhibit the relevant behaviour. If it is correct, then, that the broadening in (9), (11) and (12) and the narrowing in (7)–(8) and (11)–(12) are essentially the same process of modulating encoded word meaning, we are justified in treating them together as a single phenomenon of enrichment. Besides, as Recanati (2010) acknowledges, using ‘enrichment’ as a generic term for free pragmatic contributions to the proposition expressed makes sense on the conception of these pragmatic processes as supplementing or developing the linguistically decoded form into a further mental representation.

Current work Enrichment has been one of the most contentious issues in semantics and pragmatics for the past decade, with many objecting to the idea that pragmatic processes have such a large role in deriving the explicit content. The debate has focused largely on cases of so-called unarticulated constituents such as the inferred location when weather predicates are used (example [2]), domain restriction ([1] and [6]), and various causal relations ([4]–[5]). A closely related, though 345

Alison Hall

separate, debate concerns the apparent subsentential fragments illustrated in (3). I will discuss these in turn.

Pragmatic enrichment versus hidden indexicals Certain theorists, while sharing the view of the advocates of enrichment that the explicit content corresponds to the utterance’s intuitive truth-conditional content, deny that free pragmatic processes contribute to this level of content. They take the view that any pragmatically contributed element of the proposition expressed is linguistically articulated – that is, is represented in the logical form that is the output of linguistic decoding and input to pragmatic inference. The only free pragmatic processes, on this view, are those that result in conversational implicatures. I will label this approach ‘indexicalism’, for reasons which will emerge very soon. As Stanley states this view, ‘all truth-conditional effects of extra-linguistic context can be traced to logical form’ (2000: 391). To account for the elements of the proposition expressed that do not appear to be the values of anything in the decoded sentence meaning (logical form), indexicalists posit silent (hidden, covert) indexicals attached to certain lexical items. For example, weather predicates (‘rain’, ‘snow’, ‘be sunny’ etc.) might encode a location variable, and on Stanley and Szabó’s (2000) proposal, every nominal encodes a pair of domain restriction indices. (13)–(15) roughly indicate the kinds of structures envisaged. Imagine (14) uttered by someone checking all the bottles in a house in search of water, and (15) uttered in the context of reporting on an ice-skating competition where the competitors and judges each include a Russian: (13) It is raining (14) Every is empty (15) The voted for the Whenever a weather predicate or a quantified noun phrase is uttered, these variables/indices require the provision of a specific value: a location in (13); a domain in (14)–(15). In (15), for instance, the value assigned to the first occurrence of the domain restriction indices would be, roughly, ‘members of the judging panel’; the value assigned to the second would be ‘competitors’. What motivates the rejection of the enrichment approach, and this development of an alternative, is a set of related concerns about how the enrichment process is supposed to work and, important, what constrains it. The traditional account of how we grasp the explicit content of an utterance is that we work out the semantic values of the expressions used (assign referents to indexicals, demonstratives, tense indicators, etc.) and combine these values according to rules of semantic composition, which form part of our linguistic knowledge. If non-linguistically mandated pragmatic processes intrude into explicit content, this compositional account no longer works so straightforwardly, and many authors have objected that an approach that posits enrichment will not be able to provide a systematic, principled account of how explicit content is derived. One version of this objection is provided by Stanley (2002), who argues that enrichment theories massively over-generate interpretations. For instance, he asks, given that enrichment can allegedly add to nominals such things as quantifier domains, what stops it modifying nominals in other ways, such as by supplying the element in italics in (16b), given an utterance of (16a): (16) a. John likes Sally. b. john likes sally and his mother. 346

Enrichment

Both sides in the debate agree that (16b) could not be the explicit content of (16a), but that given the right context, say if there is a salient contextual assumption that everyone who likes Sally also likes his or her own mother, a speaker uttering (16a) could implicate (16b), as Stanley points out.The indexicalist can easily explain why (16b) could not be the explicit content: (16a) contains no indexical of which the extra conjunct and his mother could be the value, and, with there being no such thing as pragmatic enrichment, there is no other way for this conjunct to get into the explicit content of (16a). A full response to this over-generation argument remains to be provided by those who support the enrichment approach, though there are attempts to address it in Elbourne (2008), Hall (2008) and Recanati (2012). However, the indexicalist alternative faces problems of its own: a lack of evidence for the extensive hidden linguistic structure that it posits, and the question of how many of the apparent free pragmatic effects on explicit content it can really account for. The main attempt to provide evidence for the hidden indexicals illustrated in (13)–(15) has been the ‘binding argument’ (Stanley 2000: 412–420). This appeals to the fact that some of the apparent enrichments, including locations and quantifier domains, behave like the values of overt indexicals in that they can be either bound or free. (17a) illustrates this for overt indexicals: ‘he’ can co-vary with ‘everyone’, or can refer to a salient individual in the context, such as John. These two interpretations are represented in (17b)–(17c): (17) a. Every man loves his mother. b. FOR EVERY MAN X, X LOVES X’S MOTHER. c. FOR EVERY MAN X, X LOVES JOHNi’S MOTHER. Similarly, according to Stanley, (18a) has, as one of its possible interpretations, (18b), where the location of the rain co-varies with the location of the cigarette-lighting; also possible is the free reading illustrated in (18c). The domain of the quantifier ‘every bottle’ in (19) can also be bound (every bottle in that room), or free (a bizarre reading where, say, John carries his collection of bottles from room to room and places them on the top shelf in whichever room he is in at the time): (18) a. Every time John lights a cigarette, it rains. b. FOR EVERY TIME T AT WHICH JOHNx LIGHTS A CIGARETTE, IT RAINS AT T AT THE LOCATION L AT WHICH JOHNx LIGHTS A CIGARETTE c. FOR EVERY TIME T AT WHICH JOHNx LIGHTS A CIGARETTE, IT RAINS AT T IN PARIS (19) In every room in John’s house, he keeps every bottle on the top shelf. While this similarity between the values of overt variable elements and these apparent enrichments lends some plausibility to the idea that there could be, in the logical form of (18a) and (19), covert variables/indexicals for location/domain, the indexicalist has not shown that this behaviour must be explained by the presence of hidden indexicals (see Carston 2002: 199–200 and Neale 2004: 128 for assessment of Stanley’s binding argument). Moreover, there are obvious dissimilarities between most possible cases of enrichment and overt indexicals. Overt indexicals are obligatorily assigned a contextual value (or they are bound). This is the case also for uncontroversial cases of covert indexicals, such as the third-person pronoun in pro-drop languages such as Spanish, which is generally not made overt unless it is needed for clarification or emphasis. Enrichment, by contrast, is linguistically optional by definition, being a ‘free’ pragmatic process (though of course it may be required, in context, in order to derive the intended interpretation). This optionality is especially clear in some cases, such as the causal and temporal enrichment 347

Alison Hall

of ‘and’-conjunctions, illustrated in (4) (‘Mary left John and he started drinking’); not all ‘and’conjunctions undergo enrichment. Various nouns whose meaning is adjusted in the examples (‘money’, ‘drink’, and so on) can be used, it seems, entirely literally, to express their encoded meanings as opposed to the more specific meanings that can be paraphrased as ‘a lot of money’ and ‘alcoholic drink’. And even with weather predicates, it has been argued (see especially Recanati 2004), the location does not always need to be inferred, as it is not always relevant (for example ‘Once, in antiquity, it rained blood. Since then, it has never happened again’). Recanati (2004) proposes that optionality should replace bindability as the criterion for deciding whether an apparent enrichment is actually the result of assigning a value to a hidden indexical, and many others agree that it is, at least, a good test for deciding whether hidden indexicality is plausible in any given case. (It does not apply to subsentential fragments. As I discuss in the next subsection, the debate about these is not whether there is hidden indexicality, but elided syntactic structure.) Assignment of values to indexicals is obligatory, so, if a pragmatic process appears obligatory (must occur in every context where the relevant expression is uttered), it is plausible that it is indeed mandated by the presence of an indexical – a covert one – in the sentence uttered. This optionality criterion suggests some potential cases of covert linguistic context-sensitivity: most authors agree expressions such as ‘enough’ and ‘ready’ are inherently underdetermined, requiring a specific value (ready for/to x, enough of y) to be provided on every occasion of use, as do relational expressions such as ‘local’, ‘friend’ and ‘enemy’. However, most of the effects discussed previously do not share the obligatoriness characteristic of the assignment of values to indexicals. An additional difference is that, unlike the context-specific values of indexicals, the contextspecific interpretations in many of these apparent enrichment cases can differ from each other in arbitrarily many ways, including novel ways, and these parameters of variation could not all be written into the linguistic meaning in the form of indexicals. The point can be clearly illustrated with colour adjectives, which have been widely discussed in work on how to account for context-sensitivity: (20) These leaves are green. (Travis 1997: 89) Consider (20) as uttered in two contexts. In context 1, some botanists are looking for green leaves for a study of green-leaf chemistry. Pia has some leaves that are naturally red but have been painted green, and offers them to the botanists, uttering (20).The intuition is that she has spoken falsely. In context 2, green leaves are required for decorating a picture. Pia offers some more of her painted green leaves and utters (20). The intuition is that her utterance this time is true, and this must be because ‘green’ is making a different truth-conditional contribution in this second context than it did in the first. Szabó (2001) proposes that colour adjectives encode a hidden variable for the part of the object that is claimed to be that colour. That is supposed to account for the different interpretations of (20) as follows. In context 2, the relevant part of the leaves is their surface; their surface is green, so Pia’s utterance is true. In context 1, the relevant part is the whole leaf, and since not the whole leaf but only part of it is green, the utterance is false. The relevant part of the coloured object is not the only dimension along which the interpretation of ‘green’ can vary: an object could be truly identified as ‘The green one’ if a larger proportion of the relevant surface is green, or perhaps if it is a more intense shade of green than any other salient objects. So Szabó posits a second hidden variable, for comparison class. Two variables quickly prove inadequate, however. Kennedy and McNally (2010: 83) give a scenario that shows that the part-variable does not even capture Travis’s original example: if the naturally red leaves had been dipped in dye that turned them green through and through, the contextually relevant part of the leaves is still the whole leaves, so Szabó’s account wrongly predicts that the utterance 348

Enrichment

in context 1 should be judged true. This scenario requires an extra variable in the meaning of ‘green’, to accommodate the contrast between being naturally versus artificially green. With a little imagination, one can come up with examples that also require variables for viewing conditions (consider Austin’s 1962: 65–66 case of a fish that looks luminescent when seen in an ocean trench, but dull grey when examined out of the water), and several other parameters of variation, as seen in this quote from Lahav (1989: 264): For a bird to be red (in the normal case), it should have most of the surface of its body red, though not its beak, legs, eyes, and of course its inner organs. Furthermore, the red color should be the bird’s natural color, since we normally regard a bird as being ‘really’ red even if it is painted white all over. . . . . A red star only needs to appear red from the earth, a red glaze needs to be red only after it is fired, and a red mist or a red powder are red not simply inside or outside. A red pen need not even have any red part (the ink may turn red only when in contact with the paper). Even if all the possible parameters could be specified, furthermore, we would end up with massive redundancy, as the majority would not be saturated on any given occasion of use of the adjective. All this supports the conclusion that, when a hearer is required to work out, say, which part of an object ‘green’ is being predicated of, this requirement arises not from a ‘part’-variable in the linguistic meaning of ‘green’ requiring saturation, but simply from the fact that ‘green’ is getting predicated of an object with different parts and it is relevant to understanding the speaker to work out which one(s). It is a phenomenon that arises from use rather than in the semantics. Many of the other candidates for enrichment analyses can be used to make this point about the redundancy of linguistic parameters/variables and the improbability of them capturing the productive modification illustrated by the Austin and Lahav cases: see Bosch (2007) for application of the argument to several examples including weather predicates. The exact semantics of many of the context-sensitive expressions discussed earlier is still far from settled, and ongoing investigation might well uncover evidence for more cases of hidden indexicality or other analyses compatible with the indexicalist approach (for example, Kennedy and McNally (2010) argue that colour adjectives are lexically ambiguous). However, the conclusion of this subsection is that, while there may well be a limited amount of indexicality, the indexicalists’ strategy looks highly unlikely to cover the full range of cases, from the provision of unarticulated constituents through ad hoc concept construction to predicate transfer. Many of these are, therefore, better analyzed as pragmatic enrichment, and the complaints about lack of systematicity are no reason to abandon the approach, but highlight the need for more work on how enrichment operates and what constrains it.

Subsentential utterances: enrichment or ellipsis? Apparently subsentential utterances such as (21) and (22) are also central to the debate about enrichment, though are generally discussed separately because the evidence appealed to, and the alternative accounts, are different than for the sentential utterances that were the subject of the previous subsection: (21) [Uttered by someone trying to pick a robber out of a police line-up] The second man from the right. (22) [Holding up a bottle of wine] From Italy. 349

Alison Hall

We can utter what appear to be isolated noun phrases, prepositional phrases, adjective phrases and so on, and thereby communicate propositional contents – for (21) and (22), these could be something like the second man from the right of the line-up robbed me, and this wine is from italy. That many such utterances are used to explicitly express complete propositions (and not only to convey implicatures) is strongly suggested by the fact that they are evaluable as true/ false: consider (22) uttered by a speaker who knows that the wine in question is Romanian. Such utterances display important differences from those constructions that are usually agreed to be syntactically elliptical, such as gapping and sluicing, illustrated in (23) and (24): (23) Linda speaks French and Jane German. (24) Sally’s left and we have no idea why. In syntactically elliptical utterances, the fragment is syntactically a full sentence, and it is assumed that the elided content is recovered from the antecedent by processes internal to the language faculty, without appeal to pragmatic inference. One difference between the fragments in (21)– (22) and the elliptical sentences (23)–(24) is that the latter cannot occur discourse-initially – that is, without the right kind of linguistic antecedent – whereas the former can, suggesting that pragmatic inference, rather than language-internal processes, is responsible for developing (21)– (22)’s logical forms into the content expressed. A second difference is that, in the latter pair of cases, the elided material is uniquely identifiable, as is to be expected if the source is the overt linguistic antecedent. In the former pair, in contrast, we cannot uniquely identify the allegedly elided linguistic material. The discourse-initial fragments look, then, like cases where pragmatic enrichment supplies the unarticulated predicate or argument required for a complete proposition. Much more detail on the enrichment account, and criticisms of alternatives that invoke ellipsis, can be found in Stainton (2006). Few theorists now would argue that the discourse-initial fragments involve regular syntactic ellipsis; however, there is still a question about whether they really involve enrichment, or whether there is more linguistic structure there than meets the eye. Current work on this question is prompted by a set of effects, largely invisible in English fragments, that are not expected on the assumption that what is uttered is an isolated noun phrase/adjective phrase/etc., and which can be demonstrated by case-marking. In German, to order a coffee, you can utter the apparent subsentence ‘Einen Kaffee’ (a.ACC coffee). Accusative case on the determiner is obligatory, as it would be if the fragment were embedded in a full overt sentence: ‘Ein Kaffee’ (a.NOM coffee) is ungrammatical. This fact could be easily explained on the assumption that there is elided syntax here, containing a transitive verb such as ‘gerne haben’ (‘would like’) which ‘einen Kaffee’ could be the direct object of. Merchant (2010) presents a range of case connectivity effects in a wide variety of languages, which points to the conclusion that all apparently isolated noun phrases, preposition phrases, etc., might turn out to display connectivity effects of some kind – case agreement, gender agreement, and others. Those who defend a nonsententialist account have begun to address case agreement, and suggested that case markers are constraints on interpretation – that is, on the kinds of conceptual structures to be built, rather than on natural-language syntactic structures, so they constrain how the fragmentary logical form is to be enriched. For example, Stainton’s (2006: 108) idea is that case-marking plays a semantic role in the coffee example: accusative case indicates that the denotation of the NP is the patient. In the same spirit, Kempson and colleagues (forthcoming) analyze accusative case-marking (on an NP such as ‘einen Kaffee’) as indicating that the immediate combination of this NP with another conceptual structure must result in a predicate. It remains to be seen whether this idea, worked out in detail, can apply to the whole range of connectivity effects displayed by 350

Enrichment

apparently subsentential utterances and thus enable an account of their interpretation in terms of enrichment.

Critical issues and future directions A major challenge for theories of enrichment is to fully address the criticism that enrichment is hopelessly unconstrained, levelled by indexicalists such as Stanley (2002) and discussed in the last section. However, other very recent strands of work also raise the possibility of alternative analyses of some or all of the pragmatic effects that have been discussed under the label ‘enrichment’. One such trend is the move to adopt some form or other of ‘semantic relativism’, one variety of which I will briefly discuss below. A second challenge comes from work on word meaning, much of it from authors who have strongly defended enrichment theories. Recanati (2007, 2010) advocates a moderate form of semantic relativism, where the truth of an utterance depends not just on the utterance content and the world, but further parameters of evaluation. He suggests that unarticulated constituents such as inferred locations (as in (2) above, ‘It’s snowing’) should be treated not as part of the content of an utterance but as part of the situation, or ‘circumstance of evaluation’, with respect to which the content is evaluated for truth or falsity. The idea that the location is not represented as part of the utterance content – or at all – is appealing for many uses of weather predicates, namely those where there is no contrast with another location. For example, if a speaker and hearer are in the same room, the speaker looks out the window and reports ‘It’s raining’, and no other location apart from their current one is salient, one might question whether speaker or hearer really entertain a thought containing the location constituent, such as it’s raining in paris or it’s raining here? It would certainly be odd to actually utter ‘It’s raining here/in Paris’ in such a context. But, even if the location does not need to be inferred in such cases, it is still a factor on which the truth of the utterance depends, and needs to be accommodated somewhere in the analysis, and the circumstances of evaluation approach could serve this purpose. The idea of accounting for some possible enrichment phenomena in terms of circumstances of evaluation is one that still needs to be fully worked out and assessed in detail, but if Recanati is correct, enrichment would not involve composing extra concepts (unarticulated constituents) into explicit content, but would be limited to ad hoc concept construction, at least in sentential utterances, and this might extend to subsentential utterances too. Recanati says nothing about these, but Hinzen (2015) presents an account that has similarities with Recanati’s relativism, and denies that subsentential fragments explicitly express complete (truth-evaluable) contents. Related to Recanati’s relativist proposal, there is a more general question about what varieties of enrichment there are: are ad hoc concepts and unarticulated constituents, at least as I have used the terms in this chapter, really two different kinds of enrichment processes, or could unarticulated constituent cases be reanalyzed as ad hoc concept construction? One issue that is relevant to this question is that of what words encode (restricting our attention just to openclass words: nouns, adjectives, adverbs, verbs). The traditional view has been that word meanings are concepts, that is, constituents of thoughts (Fodor 1975, 1998), but recently such authors as Recanati (2004), Pietroski (2010) and Carston (2012) have begun questioning this. The very pervasiveness of ad hoc concept construction – the fact that words rarely – if ever – seem to be used to convey the concept they encode, but virtually always need some modification – is a good reason for questioning the idea that they do encode stable meanings that could be a constituent of the thought explicitly expressed. Other considerations that point in this direction are the impossibility of identifying a meaning for many words that is general enough to 351

Alison Hall

be compatible with all their literal uses and could serve as the concept communicated on at least some occasions. There is increasing consensus that what words encode are not the kind of thing that could be a constituent of thought, and the authors mentioned earlier are among those developing ideas of what to replace concepts with: sets of constraints that the communicated meaning must meet, instructions for constructing a concept, grab-bags of encyclopaedic information or collections of memory traces of past uses of the word in context. This work will undoubtedly have consequences for pragmatic enrichment accounts and for the debate with indexicalism. On one hand, depending on the nature of the proposals for word meanings, these advocates of enrichment may end up with something close to the indexicalist view, on which linguistically encoded context-sensitivity is generalized: it will be the linguistic meaning that mandates the provision of pragmatically inferred concepts. On the other hand, more work will be left to pragmatics in even arriving at a minimal propositional content. As discussed in the last section, using the example of colour adjectives, it is not plausible that this pragmatic work would be constrained by a myriad of specific parameters written into the linguistic meaning of words, as in the case of overt pronouns and demonstratives such as ‘he’ or ‘tomorrow’, so the main commitments of the enrichment approach would be preserved.

Conclusion Recent work in semantics and pragmatics posits a range of enrichment processes affecting the interpretation of virtually all utterances. While their existence remains the subject of much debate, and a few examples may ultimately prove better analyzed as saturation or disambiguation, it seems that the main current alternative approach, indexicalism, cannot cover the full range of data. As a result, some amount of enrichment will have to be admitted. Among the major questions to be considered in future work are whether some varieties of enrichment are better treated in terms of circumstances of evaluation, as relativists suggest, and the implications for enrichment accounts of recent ideas about the nature of word meaning.

Suggestions for further reading Carston, R. (2002) Thoughts and Utterances:The Pragmatics of Explicit Communication. Oxford: Blackwell. See especially chapter 1 for a survey of linguistic underdeterminacy, and chapter 5 on lexical pragmatics. Recanati, F. (2012) ‘Pragmatic enrichment’, in G. Russell and D. Graff Fara (eds.) The Routledge Companion to the Philosophy of Language. New York: Routledge. 67–78. Discussion of different conceptions of pragmatic enrichment, distinguishing it from semantic underspecification and saturation, and addressing the issue of systematicity. Stainton, R. J. (2006) Words and Thoughts: Subsentences, Ellipsis, and the Philosophy of Language. Oxford: Oxford University Press. An account of subsentential utterances, and defence of enrichment against ellipsis accounts. Stanley, J. (2002) ‘Making it articulated’, Mind and Language, 17(1–2): 149–168. Criticizes enrichment on the grounds of unsystematicity and over-generation.

References Austin, J. L. (1962) Sense and Sensibilia. Oxford: Oxford University Press. Bach, K. (1994) ‘Conversational impliciture’, Mind and Language, 9(2): 124–162. Bezuidenhout, A. (2002) ‘Truth-conditional pragmatics’, Philosophical Perspectives, 16: 105–134. Bosch, P. (2007) ‘Productivity, polysemy, and predicate indexicality’, in B. ten Cate and H. Zeevat (eds.) Proceedings of the Sixth International Tbilisi Symposium on Language, Logic and Computation. Berlin: Springer. 58–71. Carston, R. (2002) Thoughts and Utterances:The Pragmatics of Explicit Communication. Oxford: Blackwell.

352

Enrichment Carston, R. (2012) ‘Word meaning and concept expressed’, The Linguistic Review, 29(4): 607–623. Carston, R. and Hall, A. (2012) ‘Implicature and explicature’, in H. J. Schmid (ed.) Handbooks of Pragmatics: Cognitive Pragmatics. Berlin: Mouton de Gruyter. 47–84. Elbourne, P. (2008) ‘The argument from binding’, Philosophical Perspectives, 22(1): 89–110. Fodor, J. (1975) The Language of Thought. New York: T.Y. Crowell. Fodor, J. (1998) Concepts:Where Cognitive Science Went Wrong. New York: Oxford University Press. Grice, H. P. (1967) ‘Logic and conversation’, William James lectures delivered at Harvard University. Reprinted in H. P. Grice (1989) Studies in the Way of Words. Cambridge, MA: Harvard University Press. 1–143. Hall, A. (2008) ‘Free enrichment or hidden indexicals?’, Mind and Language, 23(4): 426–456. Hinzen, W. (2015) ‘Nothing is hidden: Contextualism and the grammar-meaning interface’, Mind and Language, 30(3): 259–291. Kempson, R., Gregoromichelaki, E., Eshghi, A. and Hough, J. (forthcoming) ‘Ellipsis in dynamic syntax’, in J. Van Craenenbroeck and T. Temmerman (eds.) The Oxford Handbook of Ellipsis. Oxford: Oxford University Press. Kennedy, C., and McNally, L. (2010) ‘Color, context, and compositionality’, Synthese, 174(1): 79–98. Lahav, R. (1989) ‘Against compositionality: The case of adjectives’, Philosophical Studies, 57: 261–279. Merchant, J. (2010) ‘Three kinds of ellipsis’, in F. Recanati, I. Stojanovic and N. Villanueva (eds.) ContextDependence, Perspective, and Relativity. Berlin: Walter de Gruyter. 141–192. Neale, S. (2004) ‘This, that, and the other’, in A. Bezuidenhout and M. Reimer (eds.) Descriptions and Beyond. Oxford: Oxford University Press. 68–182. Nunberg, G. (1995) ‘Transfers of meaning’, Journal of Semantics, 12: 109–132. Pagin, P. (2014) ‘Pragmatic enrichment as coherence raising’, Philosophical Studies, 168: 59–100. Pietroski, P. (2010) ‘Concepts, meanings, and truth: First nature, second nature, and hard work’, Mind and Language, 25: 247–278. Recanati, F. (2004) Literal Meaning. Cambridge: Cambridge University Press. Recanati, F. (2007) Perspectival Thought. Oxford: Oxford University Press. Recanati, F. (2010) Truth Conditional Pragmatics. Oxford: Oxford University Press. Recanati, F. (2012) ‘Pragmatic enrichment’, in G. Russell and D. Graff Fara (eds.) The Routledge Companion to the Philosophy of Language. New York: Routledge. 67–78. Sperber, D. and Wilson D. (1986/1995) Relevance: Communication and Cognition. Oxford: Blackwell. Stanley, J. (2000) ‘Context and logical form’, Linguistics and Philosophy, 23: 391–434. Stanley, J. (2002) ‘Making it articulated’, Mind and Language, 17(1–2): 149–168. Stanley, J. and Szabó, Z. G. (2000) ‘On quantifier domain restriction’, Mind and Language, 15: 219–261. Stainton, R. J. (2006) Words and Thoughts: Subsentences, Ellipsis, and the Philosophy of Language. Oxford: Oxford University Press. Szabó, Z. G. (2001) ‘Adjectives in context’, in I. Kenesei and R. M. Harnish (eds.) Perspectives on Semantics, Pragmatics, and Discourse. Amsterdam: John Benjamins. Travis, C. (1997) ‘Pragmatics’, in B. Hale and C. Wright (eds.) A Companion to the Philosophy of Language. Oxford: Blackwell. 87–107.

353

Interactional pragmatics

27 CONVERSATION1 Hansun Zhang Waring

Introduction In the beginning, there were speech acts. As Kasper (2006: 281) observes, ‘it is beyond dispute that speech acts have a central place in the field’ of pragmatics. Conversation, on the other hand, is a relative newcomer on the stage of pragmatics, and its rise is often considered a reaction to or entwined with certain discontent with speech act theory and research (Schegloff 1988). The official status of conversation as an integral part of pragmatics has been engrained in such foundational volumes as Levinson (1983) and Yule (1996), where the study of conversation is first and foremost linked to the theory and method of conversation analysis (CA). To date, CA studies of naturally occurring conversation have become a staple at the international pragmatics conferences hosted by the International Pragmatics Association (IPrA), as well as in journals such as Pragmatics and Journal of Pragmatics. Conversation is clearly here to stay.The purpose of this chapter is to charter its journey into the field of pragmatics, offer a snapshot of our current understandings of conversation, identify critical issues and topics, and discuss future directions. Before we proceed, let me emphasize that social interaction is an object of inquiry for a variety of discourse analytic approaches including, but not limited to, interactional sociolinguistics (IS) and critical discourse analysis (CDA). None, however, treats conversation as a focus of analysis in and of itself as does CA. The trademarks of other approaches may privilege, for example, questions such as ‘How do culturally specific inferential processes lead to miscommunications?’ or ‘How are ideologies negotiated in discourse?’ CA, on the other hand, treats conversation as ‘the primordial site of sociality’ (Schegloff 1986: 112) and wrestles with such fundamental questions as ‘What are the structures of conversation?’ and ‘How do these structures become resources for producing actions?’ That said, let me also acknowledge that by highlighting CA as a major approach to studying conversation, my account is necessarily partial, or one might even say, biased. Admittedly, recent developments in interactional sociolinguistics have moved away from its earlier focus on miscommunication. Indeed, the goal of IS to show how speakers signal and interpret meaning in everyday communicative practice (Gordon 2009) is not unlike CA’s aim at discovering and describing the methods of producing and understanding actions (Heritage and Clayman 2010). It is, however, in most of the CA work on ordinary conversation that we gain crucial insights into what constitutes the foundational practices of conversation (e.g., turn-taking, sequencing, repair) and how such basic social actions as compliments or requests are produced and recognized. 357

Hansun Zhang Waring

Yule (1996: 71), for instance, introduces CA as the approach to studying ‘the fundamental kind of interaction we acquire first and use most often.’ Levinson (1983: 285) calls CA an ‘outstanding empirical tradition in pragmatics,’ one that ‘has the most to offer in the way of substantial insight into the nature of conversation’ (294). In the remainder of this chapter, I maintain this focus on ordinary conversation noted by both Yule and Levinson – the kind that we acquire first and use most often. Readers interested in CA findings on institutional talk should consult such edited volumes as Drew and Heritage (1992), Sarangi and Roberts (1999), and Antaki (2011), as well as monographs, such as Silverman (1997), Clayman and Heritage (2002), and Maynard (2003). In what follows, I first outline the historical dissatisfaction with speech act theory within pragmatics and how CA emerged as a solution to that dissatisfaction. Next, I briefly explicate and illustrate CA findings on such fundamental conversational practices as turn-taking, sequencing, overall structuring, and repair. Critical issues and topics related to children, embodied practices, and the teaching of pragmatics are then identified. This chapter ends with some overall reflections on the study of conversation and its future directions.

Historical perspective Dissatisfaction with speech act theory In a series of papers written in the 1980s, Schegloff repeatedly demonstrated the inadequacy of the speech act approach to studying conversation. In perhaps his most extensive treatment of speech act research vs. conversation analysis, Schegloff (1988) offers a compelling illustration of how understandings of questions such as Do you know who/what /when/where? can be obtained only by considering these questions in their sequential contexts of naturally occurring conversation. A speech act analysis of Do you know who’s going to that meeting? would yield two interpretations: a direct speech act where the knowledge state of the addressee is being queried and an indirect speech act where the answer to the embedded wh-question is being sought, the latter being the more conventional understanding. Natural conversations tell quite a different story. In the following family dinner conversation, Mother asks just this question: (1) [from Schegloff 1988] 01 Mother: Do you know who’s going to that meeting? 02 Russ: Who. 03 Mother: I don’t know. 04 Russ: Oh::. Prob’ly Mrs. McOwen. . .and some of the teachers In response to the question, Russ asks Who, neither displaying his knowledge state nor answering the embedded wh-question. In other words, neither interpretation from the speech act analysis would apply. He does, instead, treat Mother’s question first and foremost as a pre-announcement – an understanding outside the purview of the speech act analysis. On the other hand, Mother’s question was indeed designed to seek the answer to the embedded wh-question (i.e., produced as an indirect speech act), which Russ then provides in line 04. Others have expressed similar sentiments. Bilmes (1986: 108) posits that CA has the most to offer in the study of utterance meaning – a central preoccupation of pragmatics. As perhaps the only proponent of CA whose earlier work was situated in linguistic pragmatics, Kasper (2006) argues for a discursive approach to speech act pragmatics – one that applies CA to speech act research. Her critique of the latter is concisely articulated in her remark that in order to determine the illocutionary force of a particular utterance, ‘the analyst must make assumptions about 358

Conversation

the speaker’s intentions. But the intentions reside in the speaker’s head and are thus hidden from sight’ (Kasper 2006: 286). She goes on to specify the distinction between the two: ‘speech act pragmatics explains how an utterance was responded to according to what the utterance meant. Conversation analysis explains what the utterance meant according to how it was responded to’ (Kasper 2006: 289).

Conversation analysis as a theory and a methodology Although typically applauded for its methodological advantage in critiques of speech act theory and research, CA is not simply a methodology, but a theory as well (Heritage 2008). As a theoretical framework, CA is informed by Harold Garfinkel’s (1967) ethnomethodology and Erving Goffman’s (1967) study of face-to-face interaction, and as such, comprises the following set of assumptions: (1) social interaction is orderly at all points, that is, no detail can be dismissed a priori; (2) participants orient to that order themselves, that is, order is not a result of the analyst’s conceptions or of any pre-formulated theoretical categories; (3) such order can be discovered and described through examining the details of interaction. By engaging in analytic induction, conversation analysts discover in each specific instance of interaction a theory that undergirds the practices of social interaction. Concepts such as turn-constructional unit (TCU), transitionrelevance place (TRP), adjacency pair (AP), and preference constitute theoretical tools with which social interaction may be viewed and understood. As Clayman and Gill (2004: 590) write, CA is ‘guided by a well-developed conceptual foundation grounded in empirical findings from past research.’ CA insists on using data collected from naturally occurring interaction as opposed to interviews, field notes, native intuitions, and experimental methodologies (Heritage 1984: 236). Analysts work with audio or video recordings along with the transcripts of these recordings, using transcription notations originally developed by Gail Jefferson (2004). A CA transcript captures a full range of interactional details such as volume, pitch, pace, intonation, overlap, inbreath, smiley voice, and the length of silence, as well as nonverbal conduct (see Appendix). It is in these minute details that evidence is located for how social actions such as requesting or complaining are accomplished. Analysis is guided by a simple question: why that now?, that is, why is a particular bit of talk produced in that particular way at that particular moment?

CA understandings of conversation CA understandings of conversation are usefully articulated in Wong and Waring’s (2010) model of interactional practices, which synthesizes decades of accumulative and robust conversation analytic findings on the nature of ordinary conversations with regard to turn-taking, sequencing, overall structuring, and repair (e.g., Sacks et al. 1974; Schegloff et al. 1977; Schegloff 2007). Briefly, in order to participate in any conversation, one must know when and how to construct a turn (i.e., a unit of talk that has conveyed a particular action such as agreement, request, etc.), take a turn, and yield a turn. But people don’t take turns for the sake of taking turns. They take turns to get things done, or in CA terms, to perform actions, such as gloating, fishing, offering, advising, or story-telling (cf. speech acts).With that, we enter the realm of sequence. A sequence is a course of action, and sequencing practices are the methods through which actions are implemented. The action of responding to a compliment in American English, for example, can be implemented through such methods as ‘praise downgrade’ (e.g., Not very solid through. in response to Good shot!) or ‘reference shift’ (e.g., So are you in response to You’re looking good.) (Pomerantz 1978). Knowing how to get things done with the turns in hand is, therefore, a major 359

Hansun Zhang Waring

component of one’s interactional competence, but there is more. Imagine beginning a conversation without an opening or ending a conversation without a closing. Such behaviour would typically be deemed bizarre, if not downright rude. Such overall structuring practices as opening or closing a conversation are also integral to one’s interactional competence (see Taguchi, Chapter 14, this volume). Finally, any self-organizing system will have a repair mechanism – one designed to manage the breakdown of operations within the system. In conversations, we engage in repair practices during moments when we have trouble speaking, hearing, or understanding. Each repair practice can involve ‘self ’ or ‘other’ vis a vis whose talk is being repaired, and each practice contains such elements as trouble source, repair initiation, and repair solution. Space precludes any in-depth treatment of the full array of intricacies, but suffice it to say that doing repair is a profoundly interactional and sequential endeavor – one that cannot be reduced to a list of ‘useful expressions’ such as Pardon? or Could you please say that again? For the purpose of illustrating these practices, I now offer a line-by-line analysis of a short excerpt from a telephone conversation between two friends, where one calls the other at work. I hope to demonstrate, with this brief analysis, the array of pragmatic insights that may be gained from applying a conversation analytic approach. In particular, I illustrate how the practices of turn-taking, sequencing, overall structuring, and repair are manifested in this brief extract. Debbie is calling Shelley at work. As becomes clearer later in the call (not shown), the two are close friends who have planned on a trip together, but it looks like Shelley is dropping out of that trip – allegedly because of a boyfriend. (2) [Shelley and Debbie] ((ring))-(5.0) 01 02 Shelley: District Attorney’s office. 03 Debbie: Shelley:¿ 04 Shelley: Debbie:¿= =what is the dea::l. 05 Debbie: 06 Shelley: whadayou ↑mean. yuh not gonna go::? 07 Debbie: 08 (0.2) well. hh now: my boss wants me to go: an: uhm finish 09 Shelley: 10 this >stupid< trial thing, u[hm 11 Debbie:            [