Cognitive Pragmatics 9783110214215, 9783110214208

Citation preview

Cognitive Pragmatics HoPs 4

Handbooks of Pragmatics

Editors

Wolfram Bublitz Andreas H. Jucker Klaus P. Schneider Volume 4

De Gruyter Mouton

Cognitive Pragmatics

Edited by

Hans-Jörg Schmid

De Gruyter Mouton

ISBN 978-3-11-021420-8 e-ISBN 978-3-11-021421-5 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.dnb.de. ” 2012 Walter de Gruyter GmbH & Co. KG, Berlin/Boston Cover image: Lisa Thornberg/iStockphoto Typesetting: Dörlemann Satz GmbH & Co. KG, Lemförde Printing: Hubert & Co. GmbH & Co. KG, Göttingen 앝 Printed on acid-free paper 앪 Printed in Germany www.degruyter.com

Preface to the handbook series Wolfram Bublitz, Andreas H. Jucker and Klaus P. Schneider The series Handbooks of Pragmatics, which comprises nine self-contained volumes, provides a comprehensive overview of the entire field of pragmatics. It is meant to reflect the substantial and wide-ranging significance of pragmatics as a genuinely multi- and transdisciplinary field for nearly all areas of language description, and also to account for its remarkable and continuously rising popularity in linguistics and adjoining disciplines. All nine handbooks share the same wide understanding of pragmatics as the scientific study of all aspects of linguistic behaviour. Its purview includes patterns of linguistic actions, language functions, types of inferences, principles of communication, frames of knowledge, attitude and belief, as well as organizational principles of text and discourse. Pragmatics deals with meaning-in-context, which for analytical purposes can be viewed from different perspectives (that of the speaker, the recipient, the analyst, etc.). It bridges the gap between the system side of language and the use side, and relates both of them at the same time. Unlike syntax, semantics, sociolinguistics and other linguistic disciplines, pragmatics is defined by its point of view more than by its objects of investigation. The former precedes (actually creates) the latter. Researchers in pragmatics work in all areas of linguistics (and beyond), but from a distinctive perspective that makes their work pragmatic and leads to new findings and to reinterpretations of old findings. The focal point of pragmatics (from the Greek prãgma ‘act’) is linguistic action (and inter-action): it is the hub around which all accounts in these handbooks revolve. Despite its roots in philosophy, classical rhetorical tradition and stylistics, pragmatics is a relatively recent discipline within linguistics. C.S. Peirce and C. Morris introduced pragmatics into semiotics early in the twentieth century. But it was not until the late 1960s and early 1970s that linguists took note of the term and began referring to performance phenomena and, subsequently, to ideas developed and advanced by Wittgenstein, Ryle, Austin and other ordinary language philosophers. Since the ensuing pragmatic turn, pragmatics has developed more rapidly and diversely than any other linguistic discipline. The series is characterized by two general objectives. Firstly, it sets out to reflect the field by presenting in-depth articles covering the central and multifarious theories and methodological approaches as well as core concepts and topics characteristic of pragmatics as the analysis of language use in social contexts. All articles are both state of the art reviews and critical evaluations of their topic in the light of recent developments. Secondly, while we accept its extraordinary complexity and diversity (which we consider a decided asset), we suggest a definite structure, which gives coherence to the entire field of pragmatics and provides

vi

Wolfram Bublitz, Andreas H. Jucker and Klaus P. Schneider

orientation to the user of these handbooks. The series specifically pursues the following aims: – it operates with a wide conception of pragmatics, dealing with approaches that are traditional and contemporary, linguistic and philosophical, social and cultural, text- and context-based, as well as diachronic and synchronic; – it views pragmatics from both theoretical and applied perspectives; – it reflects the state of the art in a comprehensive and coherent way, providing a systematic overview of past, present and possible future developments; – it describes theoretical paradigms, methodological accounts and a large number and variety of topical areas comprehensively yet concisely; – it is organized in a principled fashion reflecting our understanding of the structure of the field, with entries appearing in conceptually related groups; – it serves as a comprehensive, reliable, authoritative guide to the central issues in pragmatics; – it is internationally oriented, meeting the needs of the international pragmatic community; – it is interdisciplinary, including pragmatically relevant entries from adjacent fields such as philosophy, anthropology and sociology, neuroscience and psychology, semantics, grammar and discourse analysis; – it provides reliable orientational overviews useful both to students and more advanced scholars and teachers. The nine volumes are arranged according to the following principles. The first three volumes are dedicated to the foundations of pragmatics with a focus on micro and macro units: Foundations must be at the beginning (volume 1), followed by the core concepts in pragmatics, speech actions (micro level in volume 2) and discourse (macro level in volume 3). The following three volumes provide cognitive (volume 4), societal (volume 5) and interactional (volume 6) perspectives. The remaining three volumes discuss variability from a cultural and contrastive (volume 7), a diachronic (volume 8) and a medial perspective (volume 9): 1. Foundations of pragmatics Wolfram Bublitz and Neal R. Norrick 2. Pragmatics of speech actions Marina Sbisà and Ken Turner 3. Pragmatics of discourse Klaus P. Schneider and Anne Barron 4. Cognitive pragmatics Hans-Jörg Schmid 5. Pragmatics of society Gisle Andersen and Karin Aijmer

Preface to the handbook series

6. Interpersonal pragmatics Miriam A. Locher and Sage L. Graham 7. Pragmatics across languages and cultures Anna Trosborg 8. Historical pragmatics Andreas H. Jucker and Irma Taavitsainen 9. Pragmatics of computer-mediated communication Susan Herring, Dieter Stein and Tuija Virtanen

vii

Acknowledgements This volume is the outcome of a collaborative effort involving a large number of people. First and foremost, I would like to thank the contributors for producing such excellent papers and for their professional discipline in meeting deadlines and sticking to the niceties of stylesheet regulations. Wolfram Bublitz, who acted as supervising series editor, gave me all the guidance and support that I needed by reviewing papers, sharing his experience and giving me a free hand to make all the right and wrong decisions that I did. Thank you, Wolfram! Dirk Geeraerts, who started out as co-editor of this volume but unfortunately had to withdraw to see to other, more urgent matters, has provided invaluable input which is still visible in the overall conception and structure of the Handbook. I am indebted to the large number of anonymous reviewers whose feedback contributed a great deal to the quality of the contributions. I very much regret that they must remain what they agreed to be, i.e. anonymous. Closer to home, my heartfelt thanks go to Sue Bollinger, Maura BresnanEnders, Claudia Höger and Johannes Mürter, who shared the burden of the nittygritty with me and did a marvellous job diligently correcting, making suggestions for stylistic improvements, proofreading, desk editing, as well as double and triple checking the contributions. Finally, I want to express my gratitude to Barbara Karlson and Wolfgang Konwitschny of De Gruyter, who were as pleasant as they were professional in making sure that this volume eventually materialized.

Table of contents

Preface to the handbook series . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v ix

Part I: Introduction 1.

Generalizing the apparently ungeneralizable. Basic ingredients of a cognitive-pragmatic approach to the construal of meaningin-context Hans-Jörg Schmid . . . . . . . . . . . . . . . . . . . . . . . . . . .

3

Part II: Cognitive principles of pragmatic competence “Pragmatic” principles 2.

3.

4.

Relevance and neo-Gricean pragmatic principles Yan Huang . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

25

Implicature and explicature Robyn Carston and Alison Hall . . . . . . . . . . . . . . . . . . . .

47

Inference and reasoning in discourse comprehension Murray Singer and R. Brooke Lea . . . . . . . . . . . . . . . . . . .

85

“Semantic” principles 5.

6.

7.

Conceptual principles and relations Małgorzata Fabiszak . . . . . . . . . . . . . . . . . . . . . . . . . .

123

Contextual salience, domains, and active zones John R. Taylor . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

151

Encyclopaedic knowledge and cultural models Istvan Kecskes . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

175

xii

Table of contents

Part III: The psychology of pragmatics Processing and acquisition 8.

9.

10.

The processing of pragmatic information in discourse Ted J. M. Sanders and Anneloes R. Canestrelli . . . . . . . . . . . .

201

Happy New War – The role of salient meanings and salience-based interpretations in processing utterances Rachel Giora . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

233

Components of pragmatic ability and children’s pragmatic language development Daniela O’Neill . . . . . . . . . . . . . . . . . . . . . . . . . . . .

261

Disorders 11.

12.

13.

Pragmatic disorders Louise Cummings . . . . . . . . . . . . . . . . . . . . . . . . . . .

291

Autism from a cognitive-pragmatic perspective Anne Reboul, Sabine Manificat and Nadège Foudon . . . . . . . . .

317

Aphasia: The pragmatics of everyday conversation Suzanne Beeke . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

345

Part IV: The construal of non-explicit and non-literal meaning-in-context The construal of non-explicit meaning-in-context 14.

15.

Shared knowledge, mutual understanding and meaning negotiation William S. Horton . . . . . . . . . . . . . . . . . . . . . . . . . . .

375

Conversational and conventional implicatures Jacques Moeschler . . . . . . . . . . . . . . . . . . . . . . . . . . .

405

Table of contents

xiii

Figurative language in discourse Alice Deignan . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

437

Humour and irony in cognitive pragmatics Geert Brône . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

463

The construal of non-literal meaning-in-context 16.

17.

Part V: The emergence of linguistic structures from meaning-in-context 18.

Emergent and usage-based models of grammar Peter Harder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

507

Grammaticalization, lexicalization and constructionalization from a cognitive-pragmatic perspective Graeme Trousdale . . . . . . . . . . . . . . . . . . . . . . . . . . .

533

Sociopragmatics of language change Terttu Nevalainen . . . . . . . . . . . . . . . . . . . . . . . . . . .

559

The semantics of pragmatic expressions Maj-Britt Mosegaard Hansen . . . . . . . . . . . . . . . . . . . . .

587

About the authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

613

Subject index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Name index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

623 634

19.

20.

21.

xiv

Table of contents

Part I: Introduction

1.

Generalizing the apparently ungeneralizable. Basic ingredients of a cognitive-pragmatic approach to the construal of meaning-in-context Hans-Jörg Schmid

1.

The term Cognitive Pragmatics

At the time of writing, on 15 March 2012, the query “cognitive pragmatics” (inserted between inverted commas in a Google search) harvested a mere 714 real (37,400 estimated) websites from the Internet. The majority of these pages related to one of three sources: a book by Bruno Bara (2010) entitled Cognitive Pragmatics, a research initiative referred to by that name by Asa Kasher or, indeed, to advance announcements of this handbook and individual contributions to it. At present, then, there is little evidence that the term Cognitive Pragmatics is well established, and this provokes the following questions: What is Cognitive Pragmatics? What is the niche it is supposed to fill in the already highly diversified landscape of approaches to the study of language? And why (on earth, the reader may well be inclined to add) should a voluminous handbook be devoted to this so far apparently rather marginal field of inquiry? Cognitive Pragmatics can initially be defined as dealing with the reciprocal relationship between pragmatics and cognition. Considering that pragmatics is concerned with “meaning-in-context” (Bublitz and Norrick 2011: 4), it follows that Cognitive Pragmatics focuses on the cognitive aspects of the construal of meaning-in-context. This pertains to both language production and comprehension, and it specifically concerns one of the key questions that pragmatics has set out to answer: What are the cognitive abilities and processes required to be able to arrive at “what can or must be said” in order to get across “what is meant” and to arrive at “what is meant” on the basis of “what is said”? This conception of Cognitive Pragmatics is, to a large extent, compatible with that proposed by Bara (2010: 1), who defines it as “the study of the mental states of people who are engaged in communication”. However, the present conception of Cognitive Pragmatics is, on the one hand, more specific than Bara’s in that it focuses on the “construal of meaning” rather than on “communication” as such, and, on the other hand, more general in that it does not talk about “mental states”, but “cognitive aspects” in general. The present purview of Cognitive Pragmatics is also much more restricted and focused than the one demarcated by the journal Pragmatics & Cognition, which

4

Hans-Jörg Schmid seeks to explore relations of all sorts between semiotic systems as used by humans, animals and machines, in connection with mental activities: logical and causal dependence; condition of acquisition, development of loss; modeling, simulation of formalization, shared or separate biological and neurological bases; social and cultural variation; aesthetic expression; historical development. (quoted from http://benjamins.com/#catalog/ journals/pc, accessed 15 March 2012)

2.

The concept of Cognitive Pragmatics

For some linguists – especially those who study what is called the “core” of grammar with the aim of producing formal representations of its structure – the idea of there actually being a linguistic discipline that goes by the name of Cognitive Pragmatics may well be a rather hair-raising thought. The interbreeding of two approaches to the study of language, the cognitive-linguistic one and the pragmatic one, each of which is notorious for defying all attempts to formulate hard and fast rules and generalizations, can only result in a hybrid that epitomizes adhocness, slipperiness and vagueness. This understandable reaction precisely pinpoints the challenge that Cognitive Pragmatics and the current Handbook are facing – the challenge, as it were, of generalizing what appears to be ungeneralizable. While cognitive processes are, by definition, carried out in individual minds, which renders them to a considerable extent idiosyncratic, and while pragmatic processes are, again more or less by definition, context-dependent and thus largely unpredictable, the aim of this handbook is to identify the general cognitive-pragmatic principles and processes that underlie and determine the construal of meaning-in-context. A second group of linguists – those with a “pragmatic” bent – are maybe not unlikely to observe that, in a sense at least, the expression Cognitive Pragmatics is a tautology. And, indeed, reading some of the classics in the pragmatic literature such as Grice’s (1975) account of implicatures and how they are worked out, or Searle’s (1975) description of the ten steps which hearers have to go through in order to arrive at the interpretation of indirect speech acts, the impression that pragmatics has been cognitive all along is clearly substantiated. The title of Sperber and Wilson’s seminal book Relevance: Communication and Cognition (1986, 2nd edition 1995) and their formulation of a cognitive principle of relevance alongside a communicative one provide further support. Nevertheless, there can be no doubt that neither the “narrow”, “Anglo-American”, nor the “broad”, “Continental [European]” strand of pragmatics (Huang 2007: xi; cf. Bublitz and Norrick 2011: 3) is rooted in psychological or cognitive-science approaches but rather in philosophical, action-theoretical and sociological ones. The major markers of what a given scientific approach is like, i.e. its research questions and topics, methods and argumentation patterns, indicate very clearly that scholars and researchers who work in the field of pragmatics traditionally do not target psychologically plau-

Generalizing the apparently ungeneralizable

5

sible, let alone “realistic” models of the construal of meaning-in-context, but prioritize criteria such as the parsimoniousness, elegance and descriptive and explanatory power of a theory. To be sure, the classic pragmatic theories have spawned a range of approaches that are firmly placed in cognitive-science and cognitive-linguistic frameworks. These can indeed be considered as being distinctly cognitive-pragmatic, even though this term has not been applied to them so far. The following survey does not even try to do justice to this body of work, since this mission will be accomplished by the chapters of the Handbook (which also include references to relevant publications): –

–

–

–

–

Firstly, the rich body of experimental and theoretical publications by Herbert Clark and his collaborators on a wide range of cognitive-pragmatic topics such as common ground, shared knowledge, reference tracking, conversational collaboration and many others has to be mentioned here (cf. Chapter 13). Secondly, originally inspired by the claims made by Searle, Grice and others, Seana Coulson, Raymond Gibbs, Rachel Giora, Sam Glucksberg, Anthony Sanford and their collaborators have contributed substantially to our understanding of the construal of meaning-in-context through their work on the processing of figurative, idiomatic, ironic, humorous and other non-literal uses of language (cf. Chapters 9 and 17). This is complemented by the work by Lynn Cameron, Alice Deignan and others on the discursive and pragmatic dimensions of metaphor (cf. Chapter 16). Thirdly, the work by Suzanne Beeke, Dorothy Bishop, Louise Cummings, Daniela O’Neill, Ann Reboul and others in the field of developmental and clinical pragmatics has been invaluable, not only for developing therapeutic strategies for the treatment of cognitive-pragmatic deficits and developmental disorders, but also for obtaining theoretical insights into pragmatic competence (cf. Chapters 9, 10, 11 and 12). Fourthly, the substantial body of research into discourse processing, reference tracking and inferencing associated with such names as Mira Ariel, Simon Garrod, Morton Ann Gernsbacher, Art Graesser, Walter Kintsch, Ted Sanders, Anthony Sanford and Rolf Zwaan covers an important dimension of pragmatic processing, viz. the cognitive underpinnings of the way in which semantic and pragmatic content are incrementally “put together” during the construal of meaning-in-context (cf. Chapters 3 and 8). Fifthly, the term experimental pragmatics has entered the scene rather recently (cf. Noveck and Sperber 2004), subsuming attempts to apply established experimental psycholinguistic and psychological methods to test theoretical claims. A survey of such approaches by Breheny (2011) can be found in the first volume of this Handbook series edited by Wolfram Bublitz and Neal Norrick. Experimental pragmatics is, to a large extent, a spin-off from Relevance Theory

6

Hans-Jörg Schmid

(Sperber and Wilson 1986/1995), aiming to complement the dominantly theoretical work typically based on fabricated examples by experimental studies. – Sixthly, an increasing number of cognitive linguists are becoming acutely aware of the need to complement the cognitive approach with pragmatic and socio-cultural dimensions of inquiry. Besides Gilles Fauconnier and Mark Turner, whose conceptual blending theory has integrated a context-dependent component right from the start, key representatives include René Dirven, Dirk Geeraerts, Peter Harder, Istvan Keczkes, Gitte Kristiansen and John Taylor (cf. Chapters 5, 6 and 7). An important corollary of this development is the insight that the role played by cognitive-pragmatic principles and processes in the emergence, constitution and change of the “linguistic system” has been grossly underestimated in the past and should be taken much more seriously (cf. Chapters 18, 19, 20 and 21). – Finally, of course, mention must be made of Bara’s monograph Cognitive Pragmatics (2010), which is an updated and translated version of an Italian book published in 1999. Bara presents an extremely wide-ranging account of communicative competence. As one reviewer remarks, “Bara’s Cognitive Pragmatics is a unique exploration of human mental processes in communication with many insightful connections to areas beyond cognitive science” (Wang 2011: no page numbers). And, indeed, the book contains, among other things, discussions of and references to cybernetics, animal communication, paleolithic graffiti, game theory, ethiology, evolutionary anthropology and theories of language origin and evolution. On a more critical note, the same reviewer comments that “it seems that Bara’s Cognitive Pragmatics framework is more descriptive than interpretive”, and that “[t]he author pays more attention to a static description of human mental processes of communication, leaving the individual’s dynamic mental process in real communicative interaction by the wayside” (Wang 2011: no page numbers). As a result, the book is less relevant for the study of the construal of meaning-in-context than its title seems to suggest.

3.

The demands on Cognitive Pragmatics

What, then, are the basic demands that a viable cognitive-pragmatic theory of language has to meet? Or, in other words, how do we model a human mind that is equipped to construe meaning-in-context? To flesh out the bare bones of these questions, a short example taken from the movie Last Chance Harvey featuring Emma Thompson and Dustin Hoffman in the roles of Kate and Harvey will certainly be helpful. Harvey is an American commercial composer who travels to London looking forward to giving away his daughter at her wedding but is ousted from this role by his daughter’s stepfather.

Generalizing the apparently ungeneralizable

7

Having left the wedding in disappointment to fly back home, he learns on the phone that he has been fired and finds a bar in the airport to recover from this shock. Kate, whose job is to collect statistical information from travellers at Heathrow airport, is spending her lunch break in this bar and reading a book. The two had shared a rather fleeting and unpleasant moment the day before when Harvey somewhat harshly declined to answer Kate’s questions while making his way out of the terminal after having arrived in London. Harvey has found a place at the bar, Kate is sitting at a table; the only other person present is the barman. After a short conversation initiated by Harvey picking up on their previous encounter, Kate, who has watched Harvey down his third whisky, initiates the following brief exchange: (1) Kate: Harvey: Kate: Harvey: Kate: Harvey:

That’ll help. Sorry? I said, that’ll help. Believe me, it will. Right. I reckon it’ll help as much as that trashy novel and a glass of Chardonnay.

Let us focus on Kate’s first utterance and discuss what is needed to construe a plausible meaning-in-context of the type ‘drinking large of amounts of alcohol will not solve your problems’. Beginning with general cognitive prerequisites, Kate and Harvey must of course have the motor ability to produce spoken utterance and the sensory ability to perceive them. Secondly, both must have acquired, at some earlier point in their lives, what could be called “linguistic competence”, i.e. lexical and grammatical knowledge (however that is to be modelled) of English, which enables them to associate meanings with the individual lexical items, grammatical elements and grammatical structures. Note that this kind of knowledge is, of course, far from sufficient to arrive at the meaning-in-context sketched above. Thirdly, Kate and Harvey have to be willing to engage in a conversation in the first place, which must not be taken for granted given the type of situation they are in. This includes the willingness to cooperate communicatively; indeed, the way this willingness is gradually building up, especially on Kate’s side, is a major part of the appeal of the scene for the viewer. Fourthly, both must have acquired “pragmatic competence”, that is, the general ability and willingness to interpret other interlocutors’ communicative intentions. Fifthly, Kate and Harvey must have acquired a certain degree of “social competence” allowing them to make informed guesses as to the nature of this social situation: a meeting of strangers in a public place; the social norms governing such situations in Western culture; a man trying to “chat up” a woman; the social role of the barman; etc. And sixthly, Kate and Harvey must have at their disposal general world knowledge and cultural knowledge pertaining to a vast range of issues such as airports, bars, alcoholic drinks and their effects, and so on. Without taking recourse to this kind of knowledge, there is defi-

8

Hans-Jörg Schmid

nitely no way of proceeding from that’ll help to ‘drinking large quantities of alcohol will not solve your problems’. Moving next to necessary cognitive abilities, both interlocutors must first of all be able to keep track of the situational context, the linguistic cotext and the ways in which both constantly change. For example, the fact that Harvey is progressing towards his fourth shot is clearly relevant for the way in which Kate phrases her utterance and Harvey eventually understands it. What is equally important to keep track of are the hypothetical mental states of the other interlocutors: Kate and Harvey, like indeed all competent speakers and hearers, seem to be acutely aware of each others’ current mental states. What does the other person know on the basis of what has happened before, of what has been said before and of what can be garnered from the perception of the situation? Significantly, Kate is not only aware of Harvey’s attempt to get drunk, she is also aware of the fact that he is aware of her knowledge and that she is aware of the fact that he is aware of it. In addition, Harvey has to have the ability to connect the individual elements of the sequentially aligned linguistic input he is confronted with (i.e. that with ’ll and help), keeping in mind the situational input and currently activated and stored knowledge. In the course of this, he has to compute conventionally and conversationally underdetermined and implicit meanings, including, among other things, the target of the deictic that and the contextually appropriate sense of help. While it would be tempting to claim that at this point Harvey has “understood” the literal meaning that ‘that will help you solve your problems’ and is then able to proceed to figuring out the ironic meaning of Kate’s utterance, this presumably misses the point: both the background knowledge already activated and the potential familiarity with situations where that’ll help is used with an ironic meaning render it rather unlikely that Harvey will process the utterance in a sequential manner by first construing a literal meaning-in-context and then construing the non-literal, ironic one on the basis of additional cognitive principles. Be that as it may, Harvey does have to compute the conventionally and conversationally nonliteral meanings, taking into account, as before, the utterance itself, as well as cotext, context, pragmatic, social, cultural and world knowledge, in order to arrive at a contextually appropriate ironic meaning along the lines suggested above. Abstracting and abducting from this innocuous but sufficiently complex example, it can be stated that a realistic cognitive-pragmatic model of the construal of meaning-in-context has to accommodate at least the following cognitive prerequisites and abilities (cf. Table 1). The table also includes key terms and research fields traditionally associated with these prerequisites and abilities, which will not be detailed any further here, however, as they are dealt with in the subsequent chapters of this Handbook and are included in the subject index. While providing such a general list of the major cognitive foundations of the construal of meaning-in-context does not seem to be a particularly daunting task, the proof of the pudding is, as observed above, in the generalizing. It is the mission

Generalizing the apparently ungeneralizable Table 1.

9

Survey of demands on a cognitive-pragmatic theory

Cognitive prerequisites:

Key terms:

– motor and sensory ability to produce and perceive utterances

articulation and auditory perception

– linguistic competence

grammatical and lexical knowledge

– willingness to engage in communication

cooperation, cooperative principle

– pragmatic competence

joint attention, intention-reading

– social competence, cultural knowledge and world knowledge

social norms, context of culture, frames, scripts, cognitive and cultural models

Cognitive abilities:

Key terms:

– keep track of situational context and linguistic cotext

deixis, anaphora, cohesion and coherence

– keep track of mental states of other interlocutors

common ground, shared knowledge, mutual knowledge, audience design, given – new, accessibility, topicality

– connect linguistic and situational input and construe meanings of elements and chunks in the input

sense disambiguation (polysemy), reference tracking, anaphora resolution, pragmatic enrichment, explicature

– construe conventionally implicit meaning (taking into account cotext, context and pragmatic, social and cultural knowledge)

inferencing, presupposition, conventional (and, to some extent, generalized conversational) implicature

– construe contextually implicit meaning (dto.)

inferencing, reasoning, conversational (esp. particularized) implicature

– construe conventionally and contextually non-literal meaning (dto.)

implicature, irony, banter, humour, figurative language, metaphor, metonymy

of this Handbook to make some progress towards delivering key components of such an account.

4.

The structure of this Handbook

4.1.

The rationale behind the structure of this Handbook1

To translate the conception of Cognitive Pragmatics introduced so far into a coherent arrangement of chapters in this Handbook, a distinction between four types of input factors and three major targets of construal is introduced:

10

Hans-Jörg Schmid

Types of input factors used in the construal of meaning-in-context: 1. The utterance/text/discourse (including preceding cotext). 2. The non-linguistic situational, social and cultural environment. 3. The linguistic system (somewhat simplistically hypostatized as a cognitively given structure). 4. Cognitive principles (i.e. general, entrenched routines) and processes (i.e. online applications of routines and novel cognitive processes). Major targets of the construal of meaning-in-context: A. Underdetermined meaning (e.g. sense disambiguation, anaphora resolution, pragmatic enrichment). B. Non-explicit meaning (e.g. inferencing, reasoning, implicature, world knowledge). C. Non-literal meaning (e.g. humour, irony, figurative language). While traditional, i.e. non-cognitive, pragmatics has mainly been concerned with the effects of 1 and 2 on A, and particularly on B and C, the picture of cognitive pragmatics drawn so far indicates that this field of inquiry focuses on the effects of the interplay of 4, on the one hand, and of 1 and 2, on the other, on A, B and C. In addition, as pointed out above, Cognitive Pragmatics should be interested in “feedback” effects of the interaction of 1, 2 and 4 with A, B and C on 3, i.e. the emergence of linguistic structure from actual language usage and processing, especially recurrent processing routines. A very loose cross-tabulation of these two sets of categories, complemented by the “feedback loop”, yields four broad domains, which are covered by the four parts of this Handbook that follow the present introduction (which constitutes Part I). Part II: The cognitive principles of pragmatic competence This part deals with entrenched cognitive routines of pragmatic interpretation, i.e. the influence of 4 on A, B and C, mainly from an off-line perspective. Key issues addressed in this part are: – relevance as a fundamental communicative and cognitive principle (Chapter 2) – implicature and explicature as basic cognitive-pragmatic macro-processes (Chapter 3) – inferencing and reasoning as cognitive processes in the construal of meaningin-context (Chapter 4) – basic conceptual principles and relations (Chapter 5) – salience phenomena in cognitive domains and conceptual networks (Chapter 6) – the role of encyclopaedic knowledge and cultural models (Chapter 7)

Generalizing the apparently ungeneralizable

11

Part III: The psychology of pragmatics This part brings the online perspective to the fore and surveys key aspects related to pragmatic processing, the acquisition of pragmatic competence and pragmatic disorders, i.e. the influence of 4 on A, B and C from an online perspective: – – – – – –

the processing of pragmatic information in discourse (Chapter 8) the role of salience-based interpretations in the processing of utterances (Chapter 9) components of pragmatic ability and children’s pragmatic language development (Chapter 10) pragmatic disorders in general (Chapter 11) autistic spectrum disorders (Chapter 12) aphasia from a cognitive-pragmatic and conversation-analytical perspective (Chapter 13)

Part IV: The construal of non-explicit and non-literal meaning-in-context This part focuses on the cognitive principles and online processes involved in the construal of non-explicit and non-literal meaning, i.e. the joint influence of 1, 2 and 4 on B and C: – – – –

shared knowledge, mutual understanding and meaning negotiation (Chapter 14) conversational and conventional implicatures in the construal of non-explicit meaning-in-context (Chapter 15) figurative language in discourse (Chapter 16) the cognitive pragmatics of humour and irony (Chapter 17)

Part V: The emergence of linguistic structure from the construal of meaningin-context This part deals with the effects of the interaction of 1, 2 and 4 with A, B and C on 3: – – – –

a survey of emergentist and usage-based models of grammar (Chapter 18) the cognitive pragmatics of language change (Chapter 19) the sociopragmatics of language change (Chapter 20) the semantics of pragmatic expressions (Chapter 21)

Since the cognitive, psycholinguistic and psychological perspectives are to dominate this Handbook, it is considered appropriate to start out from input factor 4 and move on later to aspects closer to utterances and contexts traditionally covered by pragmatics. In the remainder of this introduction a survey of the chapters will be given.

12 4.2.

Hans-Jörg Schmid

Survey of the chapters

Part II: Cognitive principles of pragmatic competence Part II on “Cognitive principles of pragmatic competence” is subdivided into those cognitive principles and processes which would traditionally be considered as “pragmatic” insofar as they pertain to online processing and context-dependent content, on the one hand, and the ones that are closer to “semantic”, i.e., represented, conceptual and less context-dependent, aspects of meaning construal, on the other. Topics such as inferencing, reasoning, implicature and explicature as well as the relevance principle are at the heart of the former section (“Pragmatic” Principles), while the latter (“Semantic” Principles) focuses on fundamental conceptual principles underlying meaning construction, including conceptual networks and domains, contextual salience, metaphor, metonymy and conceptual blending, and the role and representation of encyclopaedic knowledge. “Pragmatic” principles Chapter 2 by Yan Huang on “Relevance and neo-Gricean pragmatic principles” gives a survey of the development and theoretical ramifications of the fundamental pragmatic principle of relevance. This principle has been a cornerstone of a range of pragmatic theories – most prominently those proposed by Sperber and Wilson (1986/1995), Horn (1984, 2009) and Levinson (2000) – which are inspired by Grice’s (1975) maxim of relation and his notion of implicature (see also Carston and Hall, Chapter 3, and Moeschler, Chapter 15). Huang summarizes these approaches and compares them with regard to how they explain the way in which language users construe “what is meant” on the basis of “what is said” by the application of a very limited number of pragmatic principles. Squarely set within the relevance-theoretical framework (Sperber and Wilson 1986/1995), Chapter 3 by Robyn Carston and Alison Hall details this model’s account of the cognitive-communicative processes involved in the online, relevance-driven construal of meaning in context. The processes of implicature and explicature take centre stage and provide the title of this chapter. While explicature subsumes pragmatic enrichment processes that are, metaphorically speaking, close to the actually communicated propositions, including sense disambiguation (see also Taylor, Chapter 6) and anaphora resolution (see also Sanders and Canestrelli, Chapter 8), implicatures are more distant contextual implications intended to be communicated by the speaker. In the chapter, the two concepts are explained and analyzed with regard to their explanatory potential and compared to related notions in competing frameworks. Chapter 4 by Murray Singer and R. Brooke Lea takes a psychological, rather than linguistic, vantage point and surveys the empirical evidence available on “Inferencing and reasoning in discourse comprehension”. Partly basing their

Generalizing the apparently ungeneralizable

13

account on Kintsch’s (1988) influential construction-integration model of discourse comprehension, the authors provide an in-depth discussion of the role of bridging inferences, which are necessary for comprehension, and elaborative inferences, which additionally enrich the construed situation model. While inference processes are regarded as tapping into stored memory representations, allowing the language user, for example, to connect information about objects to knowledge of their parts, or about events to their causes, deductive reasoning processes are based on conclusions arrived at by means of syllogisms and other forms of basic propositional logic. “Semantic” principles Chapter 5 by Małgorzata Fabiszak entitled “Conceptual principles and relations” opens the section on pragmatically relevant conceptual principles underlying the construal of meaning-in-context. Stressing the fluid boundaries between semantics and pragmatics, the author presents the current state of mainly cognitive-linguistic research on fundamental conceptual principles and relations. Key issues of the chapter include the prototypical structure of categories, conceptual networks, image schemata, idealized cognitive models, frames, metonymy and metaphor and conceptual blending (see Ungerer and Schmid 2006 for a general introduction to these notions). As befits a contribution to a handbook of cognitive pragmatics, the chapter focuses on the online processes of meaning construction integrating those conceptual structures in the context of communicative acts in social settings and also touches upon the role of individuals’ embodied experience and socialization history as well as the shared knowledge of the members of a speech community. Chapter 6 by John Taylor on “Contextual salience, domains, and active zones” follows up on issues discussed by Fabiszak, including the semantics-pragmatics continuum, and deals from a semanticist’s perspective with the following long-standing questions (approached from a pragmatic perspective by Carston and Hall in Chapter 3 and from a psycholinguistic perspective by Giora in Chapter 9): How are hypothetically stored, represented “semantic” meanings of lexical items – the large majority of which are of course highly polysemous – instantiated in given contexts? What cognitive processes are involved in this? And how can the representation of lexical meanings be modelled in such a way that it does justice to their amazing flexibility and adaptability to variable contexts? Taking up usage-based cognitive-linguistic approaches (e.g. Lakoff 1987; Langacker 1993; Evans 2009), Taylor concludes his discussion by suggesting that the semantics-pragmatics distinction may presumably turn out to be artificial, if it can indeed be shown that representations of word meanings are essentially routinized activation patterns in conceptual networks which are extracted from actual uses in social settings (see also Harder, Chapter 18).

14

Hans-Jörg Schmid

Chapter 7 by Istvan Kecskes details another conceptual foundation of the construal of meaning-in-context which is already mentioned in Fabiszak’s contribution, viz. the role played by “Encyclopaedic knowledge and cultural models”. While Fabiszak mainly dwells on the cognitive aspects of the representation of encyclopaedic knowledge, Kecskes systematically widens the scope to include the social and cultural dimensions of linguistically relevant extra-linguistic knowledge. The paper sketches the major pillars of a dynamic socio-cognitive model of encyclopaedic knowledge (cf. Kecskes 2008) and explains its implications for the study of meaning-in-context, thus spinning further the thread begun in the two preceding chapters. Kecskes’s account of intersubjectively shared cultural models complements Fabiszak’s survey of idealized cognitive models and other cognitive knowledge structures such as frames, scripts and their role in the construction of mental spaces. Part III: The psychology of pragmatics Part III moves the spotlight to the psychology of pragmatics and elucidates the processing of contextual information, the acquisition of pragmatic competence and pragmatic impairment and disorders. Processing and acquisition Chapter 8 by Ted Sanders and Anneloes Canestrelli surveys psycholinguistic research which investigates the mental operations involved in “The processing of pragmatic information in discourse”. As indicated by the title, the authors’ focus is on the level of discourse and on the cognitive processes behind traditional notions such as coherence and cohesion. The two major areas of investigation are the cognitive processes involved in establishing referential coherence, i.e. the ways in which language processors keep track of multiple references to representations of the same referent, and relational coherence, i.e. connecting discourse by means of logical, thematic and argumentative links. The chapter discusses variables that affect the choice and interpretation of anaphoric items, including accessibility (Ariel 1990), topicality (Givón 1995), recency of mention and others (Walker, Joshi, and Prince 1998), and probes the question whether or not the overt signalling of coherence relations by means of connectors such as but, yet or because is invariably conducive to ease of reading and recall of text content. The authors conclude with some reservations concerning the generalizability of experimental findings and plead for an integration of text-linguistic and psycholinguistic approaches. Chapter 9 by Rachel Giora entitled “Happy New War – The role of salient meanings and salience-based interpretations in processing utterances” constitutes a psycholinguistic and pragmatic counterpart to the chapter by John Taylor, who looks at ambiguity and polysemy resolution from a cognitive-linguistic stance.

Generalizing the apparently ungeneralizable

15

Like Taylor, Giora deals with the cognitive principles and processes determining the context-dependent access to the meanings of lexical items and larger meaningcarrying chunks, but her focus is on experimental work in the field. According to Giora, the notion of graded salience is the key to answering the question of how interlocutors swiftly and unconsciously arrive at contextually appropriate lexical meanings. Reporting on a wealth of empirical studies and illustrating their findings with original material, Giora highlights the potential of her approach in comparison with other models, among them the so-called direct-access-view (Gibbs 1994) and the relevance-theoretical approach (Sperber and Wilson 1986/1995). Chapter 10 by Daniela O’Neill on “Components of pragmatic ability and children’s pragmatic language development” closes the first group of chapters in Part III and links it with the second one. It starts out from a systematic description of pragmatic competence, which is seen to rely on three distinguishable types of knowledge: social knowledge, social-cognitive knowledge and cognitive knowledge. These are related to social pragmatics, mindful pragmatics and cognitive pragmatics respectively. Exploiting this distinction, the author then proceeds to a rich and detailed account of empirical findings on children’s pragmatic language development in these three areas. O’Neill’s account of ‘normal’ paths for the acquisition of pragmatic competence provides the backdrop for the following three chapters on pragmatic disorders. Disorders Chapter 11 by Louise Cummings opens the next section by giving a bird’s eye survey of “Pragmatic disorders”. The author summarizes the current state of our knowledge of pragmatic impairment by examining the features of specific language impairments (SLI), autistic spectrum disorders (ASD), types of right-hemisphere damage and other disorders in children and adults (cf. Cummings 2009). Specifically, Cummings considers how breakdown in the pragmatics of language adversely affects the comprehension and expression of classic pragmatic phenomena including speech acts, the processing of implicatures, the use and understanding of deictic expressions and presuppositional phenomena, the utilization of context during utterance interpretation, and the processing of non-literal language. Impairments in pragmatic aspects of non-verbal communication are also touched upon. Chapter 12 by Anne Reboul, Sabine Manificat and Nadège Foudon (“Autism from a cognitive-pragmatic perspective”) zooms in on a group of developmental deficits subsumed under the label autistic spectrum disorders (ASD). They discuss the impact of social-pragmatic deficits on language acquisition in autistic people, distinguishing various syndromes including autism (also known as Kanner’s autism) and Asperger Syndrome. Moving to more narrowly “pragmatic” deficits, the authors explain the intention-reading ability commonly referred to as

16

Hans-Jörg Schmid

Theory of mind (Premack and Woodruff 1978) and demonstrate its role as a key prerequisite for successful pragmatic processing. Pragmatic deficits typical of autistic adults, which relate to problems in interpreting indirect speech acts and other non-explicit and non-literal utterances, are also discussed. The focus of Chapter 13 by Suzanne Beeke entitled “Aphasia: The pragmatics of everyday conversation” is a second classic field of language impairment. From a cognitive-pragmatic point of view, it is particularly remarkable that patients who show the lexical and/or syntactic deficits commonly associated with different types of aphasia, as Beeke puts it, “communicate better than they speak”. The author first summarizes the findings of early pragmatic approaches to aphasic spoken language, which were taken to indicate that people with aphasia seemed to have relatively intact pragmatic abilities. Following this, the bulk of Beeke’s chapter demonstrates the potential of a conversation-analytical approach to researching and treating aphasia and presents key findings corroborating the value of this perspective in relation to aphasic conversation. Richly illustrated, the chapter includes an analysis of extracts from conversations involving one speaker who suffers from agrammatic aphasia. Part IV: The construal of non-explicit and non-literal meaning-in-context Part IV takes the Handbook into the next level of cognitive-pragmatic inquiry: the cognitive processes involved in the construal of non-explicit and non-literal meaning-in-context. The part is divided into two sections focussing on these two aspects respectively. The construal of non-explicit meaning-in-context Chapter 14 on “Shared knowledge, mutual understanding and meaning negotiation” by William Horton follows up on Chapter 3 on inferencing and reasoning, on the one hand, and Chapter 7 on encyclopaedic knowledge, on the other, and reviews the historical development of psycholinguistic studies into common ground, shared knowledge, mutual understanding and meaning negotiation. Particular attention is devoted to controversies over the amount of mutual knowledge actually required for successful language processing and the time course of its activation. Optimal audience design models (e.g. Clark 1996), which regard common ground as a prerequisite for successful communication, are shown to compete with approaches that attribute a much less pervasive role to mutual knowledge (e.g. Keysar et al. 2000) or see it as only exerting a probabilistic influence on language use (e.g. Brown-Schmidt, Gunlogson, and Tanenhaus 2008). Picking up themes from chapters 3 and 4, Chapter 15 by Jacques Moeschler focuses on “Conversational and conventional implicatures”. This chapter presents a historical survey of the concept of implicature. Criteria for distinguishing be-

Generalizing the apparently ungeneralizable

17

tween different types of implicatures are discussed. Moeschler then surveys implications, ramifications and modifications of the notion of implicature in the work of Levinson (2000), Horn (e.g. 2004), Sperber and Wilson (1986/1995) and others, concluding with an assessment of the role of implicatures in comprehension and communication in general. The construal of non-literal meaning-in-context Chapter 16 on “Figurative language in discourse” by Alice Deignan is the first of two contributions which deal with the use and comprehension of conventional and novel non-literal language. The topic of Deignan’s chapter is the role of metaphors and other types of figurative language, mainly metonymy, in discourse. The author demonstrates the potential of conceptual metaphor theory (Lakoff and Johnson 2003) for the analysis of figurative language in actual spoken and written discourse, but indeed also highlights its limitations and shortcomings (cf. Cameron 2008; Semino 2008; Deignan, Littlemore, and Semino, forthcoming). The chapter reviews empirical research into the evaluative, interpersonal and textual functions of metaphor in discourse and briefly introduces the critical discourse approach to metaphor analysis. The forms that figurative language takes in discourse, the functionally motivated, non-random distribution of manifestations of figurative language in conversation and text, as well as the implications of the discourse-related research on our understanding of figurative language are also discussed. In Chapter 17, Geert Brône tackles “Humour and irony in cognitive pragmatics”. Like figurative language, humorous and ironic language presents a special challenge to any pragmatic theory, since “what is meant” is particularly distant from “what is said”. Brône provides a comprehensive state-of-the-art review of theoretical and empirical research into the cognitive aspects of humour and irony. The section on humorous language takes Attardo and Raskin’s (1991) General Theory of Verbal Humour as a reference-point for a discussion of the ways in which canned but also conversational jokes are processed. The section on irony presents the work by Giora (2003), Coulson (2001), Gibbs (1986) and others mentioned in Section 2 above (cf. also Chapter 9) and probes the potential of cognitivelinguistic theories for the study of irony and sarcasm. Part V: The emergence of linguistic structures from meaning-in-context Parts II to IV of the Handbook follow common practice in pragmatics insofar as they take for granted that the online construal of “pragmatic” meaning-in-context relies on stored and entrenched “semantic” knowledge representations. Part V reverses this perspective and highlights the ways in which cognitive-pragmatic principles and processes contribute to the emergence and change of individually entrenched and collectively shared knowledge that is grammar. Among the best-

18

Hans-Jörg Schmid

known approaches emphasizing this contribution are Traugott’s (cf. Traugott and Dasher 2004) invited-inference theory of semantic change, Hopper’s (1987) concept of emergent grammar and an increasingly wide range of usage-based approaches which regard grammar as the product of language use (cf., e.g., Langacker 1988; Tomasello 2003; Bybee 2010). Chapter 18 by Peter Harder opens this part by providing a survey of “Emergent and usage-based models of grammar”. Harder’s account focuses not on the role of specific cognitive-pragmatic principles and processes but on the theoretical background assumptions and claims that motivate such models and the methodological challenges they have to face. Major running themes in Harder’s contribution are the tension, as yet unresolved, between system and usage in linguistic theorizing and the related controversies over how to model the interplay of lexical, grammatical and pragmatic aspects of language adequately. Chapter 19 by Graeme Trousdale entitled “Grammaticalization, lexicalization and constructionalization from a cognitive-pragmatic perspective” explores the potential of cognitive pragmatics for explaining various aspects of language change. Picking up threads from Harder, the chapter looks at attempts to explain the co-evolution of meanings and forms in grammaticalization studies (e.g. Bybee, Perkins, and Pagliuca 1994). The role of cognitive construal operations such as viewpoint and subjectivization (Langacker 1987) on pragmatic processes such as invited inferencing (Traugott and Dasher 2004) and particular aspects of language change are discussed. In the later section of the chapter, the focus shifts to constructionalization, a recent term referring to changes affecting conventionalized or novel pairings of form and meaning. Chapter 20 by Terttu Nevalainen widens the scope of investigation to include consideration of the “Sociopragmatics of language change”. Recognizing the fact that a pragmatic perspective on the emergence of meaning must not be restricted to the interplay between the language system and language use in specific situations, the chapter emphasizes the need to describe how changes and innovations spread across a linguistic community. Providing a wealth of illustrations and empirical findings, Nevalainen approaches this task by looking at the main sources and loci of innovation, discussing local and long-term effects of sociopragmatic processes such as accommodation (Giles, Coupland, and Coupland 1991; Auer and Hinskens 2005) and investigating the role of social networks and acts of social identity. Chapter 21 by Maj-Britt Mosegaard Hansen aptly closes Part V and the whole Handbook with an investigation of “The semantics of pragmatic expressions”. This final chapter demonstrates the relevance of a number of issues raised earlier in Part V and elsewhere in the Handbook for a notoriously difficult area of linguistic inquiry: the linguistic description of the meanings of pragmatic expressions such as connectives, discourse markers, pragmatically motivated multi-word expressions and other “context-level expressions”, as Hansen calls them. Among these issues are: the contested boundary between semantics and

Generalizing the apparently ungeneralizable

19

pragmatics, which is supplemented by the psycholinguistic question as to whether the meanings of context-level expressions are referential in nature (“semantic”) or procedural (“pragmatic”); the vexing question of how to make sense of the polysemy/polyfunctionality of context-level expressions; and the need to construct adequate models of generalizable cognitive-pragmatic factors behind the semantic change of these expressions.

5.

Conclusion

Does the collection of chapters in this Handbook fill a “much-needed gap”? Ignoring the fact that the editor should of course leave it to others to answer this legitimate question, I would like to venture one or two concluding observations. It should be fairly uncontroversial that the development of a coherent and comprehensive cognitive-pragmatic account of how language users construe meanings-in-context falls within the remit of the language sciences. So far, however, I would argue, in spite of the existence of the inspiring and illuminating work referred to in Section 2 above, that none of the existing linguistic “disciplines” has managed to provide a natural breeding-ground for such an account. Traditional “philosophical” pragmatics has failed or, better, not even tried to supply solid empirical evidence demonstrating the cognitive plausibility and validity of its largely theoretical claims, however appealing and convincing they may be. Psycholinguistics and the psychology of language, true to the nature of experimental approaches, tend to lose sight of the bigger picture behind the controlled observation of highly specific processes and phenomena related to the construal of meaning-incontext. Cognitive linguistics, with its focus on stored representations and their cognitive foundations, has so far not even attempted, let alone managed, to produce systematic models of how stored knowledge and online processing interact in the construal of linguistically underdetermined, non-explicit and non-literal meanings-in-context. In short, a linguistic discipline that feels responsible for targeting the seemingly simple general question as to how interlocutors connect “what is said” to “what is meant” and vice versa has not yet been established. It goes without saying that this Handbook, partly due to the fact that it is a handbook rather than a monograph, is not a suitable medium for offering a comprehensive and coherent answer to this question either. What it might be able to do, however, is give a fresh impetus and provide some assistance for future efforts in the field by providing a rich survey of the questions to be asked, of avenues that promise to lead to answers to these questions and of ways to approach the task of systematically generalizing the apparently ungeneralizable.

20

Hans-Jörg Schmid

Notes 1. I would like to thank Dirk Geeraerts for his contribution to designing the overall structure of this Handbook, which is, in spite of some changes, still reflected in Section 4.1.

References Ariel, Mira 1990 Accessing Noun-phrase Antecedents. London: Routledge. Attardo, Salvatore and Victor Raskin 1991 Script theory revis(it)ed: joke similarity and joke representational model. Humor. International Journal of Humor Research 4(3): 293–347. Auer, Peter and Frans Hinskens 2005 The role of interpersonal accommodation in a theory of language change. In: Peter Auer, Frans Hinskens and Paul Kerswill (eds.), Dialect Change, 335–357. Cambridge: Cambridge University Press. Bara, Bruno 2010 Cognitive Pragmatics. The Mental Processes of Communication, translated from Italian by John Douthwaite. Cambridge, MA: MIT Press. Breheny, Richard 2011 Experimental pragmatics. In: Wolfram Bublitz and Neal Norrick (eds.), Foundations of Pragmatics. [= Handbooks of Pragmatics, Vol. 1]. 561–585. Berlin/ New York: De Gruyter Mouton. Brown-Schmidt, Sarah, Christine Gunlogson and Michael K. Tanenhaus 2008 Addressees distinguish shared from private information when interpreting questions during interactive conversation. Cognition 107: 1122–1134. Bublitz, Wolfram and Neal Norrick (eds.) 2011 Foundations of Pragmatics. [= Handbooks of Pragmatics, Vol. 1]. Berlin/New York: De Gruyter Mouton. Bybee, Joan 2010 Language, Usage and Cognition. Cambridge: Cambridge University Press. Bybee, Joan, Revere Perkins and William Pagliuca 1994 The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago: University of Chicago Press. Cameron, Lynne 2008 Metaphor and talk. In: Raymond Gibbs (ed.), The Cambridge Handbook of Metaphor and Thought, 197–211. Cambridge: Cambridge University Press. Clark, Herbert H. 1996 Using Language. Cambridge: Cambridge University Press. Clark, Herbert H. and Catherine R. Marshall 1981 Definite reference and mutual knowledge. In: Arvind K. Joshi, Bonnie L Webber and Ivan A. Sag (eds.), Elements of Discourse Understanding, 10–63. Cambridge: Cambridge University Press. Coulson, Seana 2001 Semantic Leaps. Frame-Shifting and Conceptual Blending in Meaning Construction. Cambridge: Cambridge University Press. Cummings, Louise 2009 Clinical Pragmatics. Cambridge: Cambridge University Press.

Generalizing the apparently ungeneralizable

21

Deignan, Alice, Jeannette Littlemore and Elena Semino forthcoming Figurative Language, Genre and Register. Cambridge: Cambridge University Press. Evans, Vyvyan 2009 How Words Mean: Lexical Concepts, Cognitive Models, and Meaning Construction. Oxford: Oxford University Press. Gibbs, Raymond W. Jr. 1986 On the psycholinguistics of sarcasm. Journal of Experimental Psychology: General 115: 3–15. Gibbs, Raymond W. Jr. 1994 The Poetics of Mind. Cambridge: Cambridge University Press. Giles, Howard, Nikolas Coupland and Justine Coupland (eds.) 1991 Contexts of Accommodation: Developments in Applied Sociolinguistics. Cambridge: Cambridge University Press. Giora, Rachel 2003 On our Mind: Salience, Context and Figurative Language. New York: Oxford University Press. Givón, Talmy 1995 Coherence in text vs. coherence in mind. In: Morton Ann Gernsbacher and Talmy Givón (eds.), Coherence in Spontaneous Text, 59–115. Amsterdam: John Benjamins. Grice, H. P. 1975 Logic and conversation. In: Peter Cole and Jerry Morgan (eds.), Syntax and Semantics 3: Speech Acts, 41–58. London: Academic Press. Hopper, Paul 1987 Emergent Grammar. In: Jon Aske, Natasha Beery, Laura Michaelis and Hana Filip (eds.), Proceedings of the Thirteenth Annual Meeting of the Berkeley Linguistics Society, 139–157. Berkeley: Berkeley Linguistics Society. Horn, Laurence R. 1984 Toward a new taxonomy for pragmatic inference: Q-based and R-based implicature. In: Deborah Schiffrin (ed.), Meaning, Form, and Use in Context: Linguistic Applications, 11–42. Washington, DC: Georgetown University Press. Horn, Laurence R. 2004 Implicature. In: Laurence R. Horn and George Ward (eds.), The Handbook of Pragmatics, 3–28. Oxford: Basil Blackwell. Horn, Laurence R. 2009 WJ-40: Implicature, truth, and meaning. International Review of Pragmatics 1: 3–34. Horton, William S. and Boaz Keysar 1996 When do speakers take into account common ground? Cognition 59: 91–117. Huang, Yan 2007 Pragmatics. Oxford: Oxford University Press. Kecskes, Istvan 2008 Dueling contexts: A dynamic model of meaning. Journal of Pragmatics 40(3): 385–406. Keysar, Boaz, Dale J. Barr, Jennifer A. Balin and Jason S. Brauner 2000 Taking perspective in conversation: The role of mutual knowledge in comprehension. Psychological Science 11: 32–38.

22

Hans-Jörg Schmid

Kintsch, Walter 1988 The role of knowledge in discourse comprehension: A construction-integration model. Psychological Review 95: 163–182. Lakoff, George 1987 Women, Fire and Dangerous Things. What Categories Reveal about the Mind. Chicago: University of Chicago Press. Lakoff, George and Mark Johnson 2003 Metaphors We Live By. 2nd edn. Chicago: University of Chicago Press. Langacker, Ronald W. 1987 Foundations of Cognitive Grammar. Vol. I: Theoretical Prerequisites. Stanford: Stanford University Press. Langacker, Ronald W. 1988 A usage-based model. In: Brygida Rudzka-Ostyn (ed.), Topics in Cognitive Linguistics, 127–161. Amsterdam/Philadelphia: Benjamins. Langacker, Ronald W. 1993 Reference point constructions. Cognitive Linguistics 4, 1–38. [Reprinted in: Grammar and Conceptualization (1999), 171–202, Berlin: Mouton de Gruyter.] Levinson, Stephen C. 2000 Presumptive Meanings. The Theory of Generalized Conversational Implicature. Cambridge, MA: The MIT Press. Noveck, Ira A. and Dan Sperber (eds.) 2004 Experimental Pragmatics. New York: Palgrave. Premack, David and Guy Woodruff 1978 Does the chimpanzee have a theory of mind? Behavioral and Brain Sciences 1(4): 515–526 Searle, John R. 1975 Indirect speech acts. In: Peter Cole and Jerry L. Morgan (eds.), Syntax and Semantics 3. Speech Acts, 59–82. New York: Academic Press. Semino, Elena 2008 Metaphor and Discourse. Cambridge: Cambridge University Press. Sperber, Dan and Deirdre Wilson 1985/96 Relevance: Communication and Cognition. Oxford: Blackwell. [Second edition with Postface 1995] Tomasello, Michael 2003 Constructing a Language. A Usage-Based Theory of Language Acquisition. Cambridge, MA/London: Harvard University Press. Traugott, Elisabeth Closs and Richard B. Dasher 2004 Regularity in Semantic Change. Cambridge: Cambridge University Press. Ungerer, Friedrich and Hans-Jörg Schmid 2006 An Introduction to Cognitive Linguistics. 2nd edn. Harlow etc.: Pearson Longman. Walker, Marilyn, Aravind Joshi and Ellen Prince (eds.) 1998 Centering Theory in Discourse. Oxford: Clarendon Press. Wang, Jinrui 2011 Review of Bruno Bara, Cognitive Pragmatics. The Mental Process of Communication, Cambridge, MA: MIT Press. LinguistList, posted 13 July 2011. [http://linguistlist.org/issues/22/22–2872.html; accessed 15 March 2012)

Part II: Cognitive principles of pragmatic competence “Pragmatic” principles

2.

Relevance and neo-Gricean pragmatic principles Yan Huang

This chapter is dedicated to Professor Anna Morpurgo Davies, my mentor at Oxford, on the occasion of her seventy-fifth birthday.

1.

Introduction

The maxim of Relation is one of a set of pragmatic sub-principles originally postulated by Grice (1975, 1989). However, in Sperber and Wilson’s (1986/1995) relevance theory, the whole centre of gravity of pragmatics is reduced to a single, technical concept of relevance, which is embodied in two principles of relevance: the principle of cognitive relevance and the principle of communicative relevance. This contrasts with the neo-Gricean pragmatic models advanced by both Horn (1984, 2009) and Levinson (1987, 2000), where the original Gricean maxim of Relation is not given any prominent place. This chapter endeavours to examine relevance and neo-Gricean pragmatic principles. I shall start by discussing the role played by the maxim of Relation in the original Gricean programme in section 2. Next, in section 3, I shall navigate around the notion of relevance and the cognitive and communicative principles of relevance. I shall then explore both Horn’s (1984, 2009) and Levinson’s (1987, 2000) neo-Gricean pragmatic principles in section 4. Finally, in section 5, a comparison between relevance and neo-Gricean pragmatic principles and their related areas will be offered.

2.

Relation in classical Gricean pragmatics

In classical Gricean pragmatics, the maxim of Relation is just one of a set of pragmatic sub-principles articulated by Grice (1975, 1989) to provide an account of how language is used maximally efficiently and effectively to achieve rational interaction in communication, as can be seen in the simplified version of classical Gricean theory of conversational implicature in (1) (see e.g. Huang 2000a: 206, 2007: 26, 2010a).

26

Yan Huang

(1) Classical Gricean theory of conversational implicature: a. The co-operative principle Be co-operative. b. The maxims of conversation Quality: Be truthful. (i) Don’t say what is false. (ii) Don’t say what lacks evidence. Quantity: (i) Don’t say less than is required. (ii) Don’t say more than is required. Relation: Be relevant. Manner: Be perspicuous. (i) Avoid obscurity. (ii) Avoid ambiguity. (iii) Be brief. (iv) Be orderly. Assuming that the co-operative principle and its associated maxims of conversation are normally adhered to by both the speaker and the addressee in a verbal interaction, Grice suggested that a conversational implicature – roughly, any meaning implied or expressed by, and inferred or understood from, the utterance of a sentence which is meant without being part of what is strictly said1 – can then arise from either strictly observing or ostentatiously flouting the maxims. In Huang (2003, 2007), I called conversational implicatures that are engendered by way of directly observing the maxims conversational implicaturesO, and conversational implicatures that are generated by way of the speaker’s deliberately flouting the maxims conversational implicaturesF. Given the Gricean maxim of Relation, while the Relation-implicature in (2) is a conversational implicatureO, that in (3) is a conversational implicatureF. (I use ‘+>’ to stand for “conversationally implicate”.) (2) A: I am out of petrol. B: There is a garage round the corner. (Grice 1989: 32) +> The garage is open and sells patrol. (3) John: Susan can be such a cow sometimes! Mary: Oh, what a lovely day today! (Huang 2007: 30) +> e.g. One shouldn’t speak ill of people behind their back.

3.

Relevance in relevance theory

In Sperber and Wilson’s (1986/1995) relevance theory, however, the whole centre of gravity of pragmatic theory is reduced to a single, technical concept of relevance.2 This notion of relevance is embodied in the two principles of relevance

Relevance and neo-Gricean pragmatic principles

27

put forward by Sperber and Wilson (1986/1995): a first or cognitive principle of relevance (4), and a second or communicative principle of relevance (5).3 (4) The cognitive principle of relevance Human cognition tends to be geared to the maximization of relevance. (5) The communicative principle of relevance Every utterance conveys a presumption of its own optimal relevance. What, then, is relevance? According to Sperber and Wilson (1986/1995), relevance is a function or measure of two factors: (i) cognitive or contextual effects, and (ii) processing effort. The first factor is the fruitful outcome of an interaction between a newly impinging stimulus and a subset of the assumptions that are already established in a cognitive system. The second factor is the effort a cognitive system has to expend in order to yield a satisfactory interpretation of any incoming information processed. Defined thus, relevance is a matter of degree. The degree of relevance of an input to an individual is a balance struck between cognitive effects (i.e. the reward) and processing effort (i.e. the cost). The greater the positive cognitive effects achieved, and the smaller the processing effort required, the greater the relevance of the input to the individual at the time. Furthermore, Sperber and Wilson isolated three main types of positive cognitive effects to which the processing of new information in a context may give rise: (i) a generating of a conclusion derivable from new and old information together, but from neither new nor old information separately, which is called a contextual implication, (ii) a warranted strengthening of an existing assumption, and (iii) a warranted revision (including contradicting and cancelling) of an existing assumption. As an illustration, consider the following scenario, taken from Blakemore (2002: 60–61). Suppose that a bus driver is about to leave from a bus stop. He sees in his rear mirror the reflection of an anxious-looking woman carrying a bus pass, trying to cross the road behind him. In this situation, the bus driver’s overall representation of the world can be improved in the following three ways corresponding to the three types of cognitive effects mentioned above. In the first place, given the assumption that if a person is holding a bus pass, then he or she intends to travel on a bus, the bus driver will derive the new assumption or the contextual implication that the woman in question has the intention of travelling on his bus. Secondly, the bus driver’s existing assumption that the woman is trying to cross the road to catch his bus may be supported and strengthened by the assumption that she is carrying a bus pass. Thirdly and finally, the bus driver’s existing assumption that the woman intends to take his bus is contradicted and eliminated when he sees the woman walk off in the opposite direction after handing the bus pass to someone who is standing by the bus stop. Next, I move to processing effort. Consider another scenario: Suppose that Xiaoming, a Chinese student who has just arrived in England, wants to rent a room

28

Yan Huang

from the Smiths, his potential landlords. He wants to know whether the Smiths keep any cats. His question in (6) could receive any of the replies in (7). (6) Do the Smiths have any cats? (7) a. They have three cats. b. They have three pets. c. Either they have three cats or the capital of China is not Beijing. Given (4) and (5), all the three replies in (7) are relevant to Xiaoming’s question in (6). However, according to the notion of relevance defined above, (7a) is more relevant than either (7b) or (7c). On the one hand, (7a) is more relevant than (7b), because the former requires less processing effort than the latter. The reason is because (7a) entails (7b), and hence generates all the conclusions, among other things, deducible from (7b), together with the context. On the other hand, (7a) is also more relevant than (7c). This is because, although (7a) and (7c), being logically equivalent, yield exactly the same amount of cognitive effect, the effect is much more easily obtained from (7a) than (7c), since the latter has a second, false disjunct to be computed. In other words, the processing effort required to compute (7a) is smaller than that required to compute (7c). Given that relevance stated in this way is a trade-off of effect and effort, it is comparative rather than quantitative. One feature of such a comparative notion is that, as Sperber and Wilson were aware, it provides clear comparisons only in some but not all cases. This gives rise to one of the most common criticisms levelled against relevance theory, namely that it fails to provide an explanation of how to measure contextual effects and processing effort in an objective way, how to make them commensurate with each other, and why there is always a unique way to meet the cognitive and communicative principles of relevance (see e.g. Levinson 1989, and Bach 1999, but see Wilson and Sperber 2004 for a spirited defence). By way of summary, what the cognitive principle of relevance basically says is this: ‘as a result of constant selection pressures, the human cognitive system has developed a variety of dedicated (innate or acquired) mental mechanisms or biases which tend to allocate attention to inputs with the greatest expected relevance, and process them in the most relevance-enhancing way’ (Wilson 2010: 394). In other words, in human cognition, there is a tendency for communicators to achieve as many cognitive effects as possible for as little processing effort as possible. This means that relevance plays an essential role not only for utterance interpretation, but for all external stimuli or internal mental representations like sights, smell and thoughts as well (Wilson 2010). Now, given the cognitive principle of relevance, it follows that a speaker, by his or her act of uttering a sentence, indicates that his or her utterance should be seen as relevant enough to be worth processing by the addressee. This is in fact what the communicative principle of relevance in (5) above actually says. Next, the pre-

Relevance and neo-Gricean pragmatic principles

29

sumption of optimal relevance included in the principle entitles the addressee to presume not only that the utterance is relevant enough to be worth his or her processing effort, but also that it is the most relevant utterance compatible with the speaker’s abilities and preferences. Finally, on the basis of the communicative principle of relevance and the presumption of optimal relevance included in it, there is a specific relevance-theoretic procedure employed by the comprehension system, which is given in (8) (Wilson 2010). (8) The relevance-guided comprehension heuristic a. Follow a path of least effort in constructing an interpretation of the utterance (and in particular in resolving ambiguities and referential indeterminacies, adjusting lexical meaning, supplying contextual assumptions, deriving implicatures, etc.). b. Stop when your expectations of relevance are satisfied. The communicative principle of relevance is responsible for the recovery of both the explicit content (i.e. explicature) and the implicit content (i.e. what I called “r-implicature” in Huang 2007) of an utterance. An explicature is an inferential development of one of the linguistically-given incomplete conceptual representations or logical forms of the sentence uttered (but see Burton-Roberts 2007 for the comments that ‘development’ in the definition is not defined in relevance theory and Bach’s 2004 comments on the term ‘explicature’). Defined thus, it is a pragmatically inferred component of the Gricean notion of what is said (though what is said is abandoned in relevance theory). It corresponds roughly to Bach’s (2004) notion of “impliciture” and Recanati’s (2004) notion of “pragmatically enriched said” (see also Huang 2010d on impliciture). Depending on the context, the explicature of (9a) may be the bracketed part of (9b). (9) a. Everyone enjoys classical music. b. Everyone [in John’s class] enjoys classical music. I turn next to what is taken as implicit content, or r-implicatures in relevance theory. An r-implicature is defined as a communicated assumption derivable solely via pragmatic inference. Thus, the recovery of it differs from that of an explicature in that while the latter involves both decoding and inference, the former involves only inference. Two kinds of r-implicatures can then be identified: (i) implicated premises, and (ii) implicated conclusions. The former is a contextual assumption intended by the speaker and supplied by the addressee, and the latter is a contextual implication communicated by the speaker. By way of illustration, take (10). (10) Car salesman: Are you interested in test-driving a Rolls Royce? John: I’m afraid I’m not interested in test-driving any expensive car. (Huang 2007: 195) John’s reply may yield the following two r-implicatures:

30

Yan Huang

(11) a. A Rolls Royce is an expensive car b. John isn’t interested in test-driving a Rolls Royce Here, (11a) is an implicated premise and (11b) is an implicated conclusion of John’s reply. (11b) follows deductively from (11a) combined with (10). However, this analysis is not unequivocally accepted. In the first place, in Recanati’s (2004: 48) view, an implicated premise is not an implicature, because it is not part of what the speaker means. Rather, it is part of what he or she takes for granted or presupposes and expects the addressee to take for granted. Secondly, as pointed out in Huang (2007: 195), the treatment of (11a) as an implicated premise is limited to the class of examples like (10) in the relevance theory literature, and more generally, is applicable only to what is treated as a particularized conversational implicature (PCI) in the Gricean and neo-Gricean framework. How it can be applied to a generalized conversational implicature (GCI) in the Gricean sense like (12) is, to say the least, not clear. (12) John and Mary watched the Berlin Wall crumble in 1989. +> John and Mary watched the Berlin Wall crumble in 1989 together4 To sum up, in relevance theory, the recovery of both explicatures and r-implicatures is guided by the communicative principle of relevance. More specifically, in an overall, relevance-theoretic comprehension process, three sub-tasks are involved, as listed in (13) (Wilson and Sperber 2004: 615). (13) Sub-tasks in the overall comprehension process a. Constructing an appropriate hypothesis about explicit content (explicatures) via decoding, disambiguation, reference resolution, and other pragmatic enrichment processes. b. Constructing an appropriate hypothesis about the intended contextual assumptions (implicated premises). c. Constructing an appropriate hypothesis about the intended contextual implications (implicated conclusions). In other words, in relevance theory, utterance interpretation involves (i) the completion or expansion of incomplete logical forms, and the derivation of explicatures via the completion or expansion of the logical forms, and (ii) the generation of r-implicatures including the recovery of r-implicated premises and r-implicated conclusions. Furthermore, the first satisfactory interpretation worked out by the use of the relevance-theoretic comprehension procedure is the only satisfactory reading. This interpretation is the one that the addressee should select. Put in a different way, the relevance-guided comprehension mechanism has a clear stopping point. Once the addressee discovers a reading that crosses the threshold set up by the relevance-theoretic comprehension heuristic, he or she will take it to be what the speaker intends to communicate (Wilson 2010).

Relevance and neo-Gricean pragmatic principles

31

This relevance-theoretic analysis is, however, not without problems. For instance, it is incapable of accounting for examples of the following type, discussed in Horn (2006). (14) If it’s warm, we’ll lie out in the sun. But if it’s hot, we’ll go inside and sit in front of the air-conditioner. As pointed out by Horn (2006), in (14), the recovery of the stronger, pragmatically enriched reading of what is said, namely, the explicature ‘warm but not hot’ arising from the use of warm in the first clause can be obtained only when the stronger scalar expression hot in the third clause is reached. In other words, the explicature under consideration is not the first satisfactory interpretation picked up by the addressee. Rather, it is the outcome of a retroactive adjustment of the earlier, weaker reading ‘at least warm’. Clearly, in order to derive the correct explicature of (14), the addressee cannot stop where he or she is predicated to stop by the relevancetheoretic comprehension procedure (see also Borg’s 2009 discussion of various relevance-theoretic possible but untenable manoeuvres to tackle counterexamples of this kind).

4.

Neo-Gricean pragmatic principles

In the last section, I discussed the notion of relevance and two principles of relevance in relevance theory. In this section, I turn to neo-Gricean pragmatic principles. 4.1.

The Hornian bipartite model

Horn (1984, 2009) developed a bipartite, neo-Gricean pragmatic model, namely what he (2007) called “a Manichaean manifesto”.5 In Horn’s opinion, all of Grice’s maxims (except the maxim of Quality) can be replaced with two fundamental and counterpoising principles: the Q[uantity]- and R[elation]-principles. (15) Horn’s Q- and R-principles a. The Q-principle Make your contribution sufficient; Say as much as you can (given the R-principle). b. The R-principle Make your contribution necessary; Say no more than you must (given the Q-principle). In terms of information structure, Horn’s Q-principle, which collects Grice’s first sub-maxim of Quantity and his first two sub-maxims of Manner, is a lower-bounding pragmatic principle which may be (and characteristically is) exploited to en-

32

Yan Huang

gender upper-bounding conversational implicatures: a speaker, in saying ‘ … p …’, ceteris paribus conversationally implicates that (for all he or she knows) ‘ … at most p …’. The locus classicus here is those conversational implicatures that arise from a prototype Horn-scale. Prototype Horn-scales are defined in (16) (see e.g. Atlas and Levinson 1981, Levinson 2000: 76, Huang, 1991, 2007: 38), with exemplification given in (17). (16) Prototype Horn-scales For to form a Horn-scale, (i) A(S) entails A(W) for some arbitrary sentence frame A; (ii) S and W are equally lexicalized, of the same word class, and from the same register; and (iii) S and W are ‘about’ the same semantic relation, or from the same semantic field where S stands for ‘semantically strong expression’ and W stands for ‘semantically weak expression’. (17) a. b. c. An instance of Q-implicatures is given in (18). (18) Some of the children hate broccoli. +> Not many/most/all of the children hate broccoli6 In (18), the speaker used the semantically weaker some, where a semantically stronger many, most or all could have been employed. From a semantic point of view, some has the ‘at least’ reading, namely ‘at least some’, and this is its lowerbounded interpretation. By the Q-principle, the use of some then gives rise to the ‘at most’ or ‘not more than’ Q-scalar implicature, namely ‘at most some’ or ‘not more than some’. This reading is the upper-bounded interpretation (Huang 2009, 2010b, 2010c). On the other hand, the counterbalancing R-principle, which subsumes Grice’s second sub-maxim of Quantity, his maxim of Relation, and his last two submaxims of Manner, and which is based on Atlas and Levinson’s (1981) principle of informativeness, is an upper-bounding pragmatic law which may be (and systematically is) exploited to invite lower-bounding conversational implicatures: a speaker, in saying ‘ … p …’, conversationally implicates that (for all he or she knows) ‘ … more than p …’. This is illustrated in (19) below. However, more recently Horn (2007) has been of the opinion that the R-principle is not in itself subsumable under Grice’s co-operative principle, but under rationality.

Relevance and neo-Gricean pragmatic principles

33

(19) The Berlin Wall came down and the Wolfgang family moved to Munich. +> The Berlin Wall came down and then the Wolfgang family moved to Munich. Viewing the Q- and R-principles as a mere instantiation of Zipfian economy, Horn (1984, 2009) explicitly equated the Q-principle (“a hearer-oriented economy for the maximization of informational content”) with Zipf’s “Auditor’s Economy” (the Force of Diversification, which tends towards a vocabulary of m different words with one distinct meaning for each word) and the R-principle (“a speakeroriented economy for the minimization of linguistic form”) with Zipf’s “Speaker’s Economy” (the Force of Unification, which tends towards a vocabulary of one word which will refer to all the m distinct meanings). The notion of Speaker’s Economy is further distinguishable between mental inertia or paradigmatic economy (économie mémorielle) and articulatory or physical inertia or syntagmatic economy (économie discursive), hence internally dialectic in its operation. The former is concerned with the reduction in the inventory of mental lexicon; the latter, with the reduction in the number of linguistic units (Horn 2007: 173–174). While the Auditor’s Economy places a lower bound on the informational content of the message, the Speaker’s Economy places an upper bound on its form. Furthermore, Horn argued that the whole Gricean mechanism for pragmatically contributed meaning can be derived from the dialectic interaction (in the classical Hegelian sense) between the two mutually constraining mirror-image forces in the following way. (20) Horn’s division of pragmatic labour The use of a marked (relatively complex and/or prolix) expression when a corresponding unmarked (simpler, less “effortful”) alternate expression is available tends to be interpreted as conveying a marked message (one which the unmarked alternative would not or could not have conveyed). In effect, what the communicative equilibrium in (20) basically says is this: the R-principle generally takes precedence until the use of a contrastive linguistic form induces a Q-implicature to the non-applicability of the pertinent R-implicature (see also e.g. Huang 1994/2007, 2000, 2003, 2004, 2007 for further discussion). 4.2.

The Levinsonian tripartite model

Horn’s proposal to reduce Grice’s maxims to the Q- and R-principles was called into question by Levinson (1987, 2000). In Levinson’s opinion, Horn failed to draw a distinction between what Levinson called “semantic minimization” (“semantically general expressions are preferred to semantically specific ones.”) and “expression minimization” (“shorter” expressions are preferred to “longer” ones). Consequently, inconsistency arises with Horn’s use of the Q- and R-principles. For

34

Yan Huang

example, in Horn’s division of pragmatic labour, the Q-principle operates primarily in terms of units of speech production whereas elsewhere, in Horn-scales, for instance, it operates primarily in terms of semantic informativeness. Considerations along these lines led Levinson to argue for a clear separation between pragmatic principles governing an utterance’s surface form and pragmatic principles governing its informational content (but see Horn 2007 for a spirited defence of his bipartite model). He proposed that the original Gricean program (the maxim of Quality apart) be reduced to three neo-Gricean pragmatic principles: what he dubbed the Q[uantity]-, I[nformativeness]- and M[anner]-principles. Each of the three principles has two sides: a speaker’s maxim, which specifies what the principle enjoins the speaker to say and a recipient’s corollary, which dictates what it allows the addressee to infer. Let me take them one by one. (21) The Q-principle (simplified) (e.g. Huang 2007: 41, see also Levinson 2000: 31) Speaker: Do not say less than is required (bearing the I-principle in mind). Addressee: What is not said is not the case. The basic idea of the metalinguistic Q-principle is that the use of an expression (especially a semantically weaker one) in a set of contrastive semantic alternates (such as a Horn-scale) Q-implicates the negation of the interpretation associated with the use of another expression (especially a semantically stronger one) in the same set. Seen the other way around, from the absence of an informationally stronger expression, one infers that the interpretation associated with the use of that expression does not hold. Hence, the Q-principle is essentially negative in nature. Three types of Q-implicature can then be identified: (i) Q-scalar implicatures, as in (18) above; (ii) Q-clausal implicatures, and (iii) what I dubbed Q-alternate implicatures in Huang (2007: 42–44). As mentioned above, Q-scalar implicatures are derived from Horn-scales. Next, Q-clausal implicatures are pragmatically enriched meanings of epistemic uncertainty. Like Q-scalar implicatures, Q-clausal implicatures also rest on a set of contrastive semantic alternates, but, in this case, of a constructional kind (see (22)). Finally, we have Q-alternate implicatures, which come from a non-entailment semantic (contrast) set. Roughly, we have two subtypes here. In the first, the lexical expressions in the set are informationally ranked, as in (23). Following Huang (2007: 42–44), let me call Q-implicatures deriving from such a set Q-ordered alternate implicatures. By contrast, in the second subtype, the lexical expressions in the set are of equal semantic strength, as in (24). Let me term Q-implicatures thus induced Q-unordered alternate implicatures. (22) Mary believed that the cat was hiding behind the sofa. +> The cat might or might not be hiding behind the sofa – Mary didn’t know which.

Relevance and neo-Gricean pragmatic principles

35

(23) John tried to give up smoking. +> John did not succeed in giving up smoking. (24) The language school offers a three-week immersion course in English, French, and German. +> The language school doesn’t offer a three-week immersion course in e.g. Russian. Furthermore, based in part on what he called rank orders, in (25), where S unilaterally entails ~ W), Horn (2007: 168–170) distinguished two kinds of pragmatic strengthening: informative and rhetorical. While R- or I-implicature (to be discussed below) increases both informative and rhetorical strength, Q-implicature is informatively but not rhetorically stronger than the sentence uttered without the implicature. (25) Next, there is the I-principle. (26) The I-principle (simplified) (e.g. Huang 2007: 46, see also Levinson 2000: 32) Speaker: Do not say more than is required (bearing the Q-principle in mind). Addressee: What is generally said is stereotypically and specifically exemplified. Mirroring the effects of the Q-principle, the central tenet of the I-principle is that the use of a semantically general expression I-implicates a semantically specific interpretation. More accurately, the conversational implicature engendered by the I-principle is one that accords best with the most stereotypical and explanatory expectation given our knowledge about the world. (27) If you clean the kitchen, I’ll give you twenty euros. +>If, and only if, you clean the kitchen will I give you twenty euros. Finally, we come to the M-principle. (28) The M-principle (simplified) (e.g. Huang 2007: 50, Levinson 2000: 33) Speaker: Do not use a marked expression without reason. Addressee: What is said in a marked way conveys a marked message. Unlike the Q- and I-principles, which operate primarily in terms of semantic informativeness, the metalinguistic M-principle operates primarily in terms of a set of alternates that contrast in form. The fundamental axiom upon which this principle rests is that the use of a marked expression M-implicates the negation of the interpretation associated with the use of an alternative, unmarked expression in the same set. To put it another way, from the use of a marked linguistic expression, one infers that the stereotypical interpretation associated with the use of an alternative,

36

Yan Huang

unmarked linguistic expression does not obtain. An M-implicature is illustrated by (29b) below, which is marked. (29) a. This government is efficient. I +> This government is efficient in the stereotypical sense. b. This government is not inefficient. M +> This government is less efficient than the utterance of (29a) suggests. Given the above tripartite classification of neo-Gricean pragmatic principles, the question that arises next is how inconsistencies arising from these potentially conflicting implicature apparatuses can be resolved. According to Levinson (2000), they can be resolved by an ordered set of precedence, incorporating Horn’s (1984) division of pragmatic labour. (30) Levinson’s resolution schema for the interaction of the Q-, I- and M-principles (e.g. Huang 2007: 51–54): a. Level of genus: Q > M > I b. Level of species: e.g. Q-clausal > Q-scalar This is tantamount to saying that genuine Q-implicatures (where Q-clausal cancels rival Q-scalar) precede inconsistent I-implicatures, but otherwise I-implicatures take precedence until the use of a marked linguistic expression triggers a complementary M-implicature to the negation of the applicability of the pertinent I-implicature (see also Huang 1991, 2000, 2003, 2004, 2007, 2011 for further discussion). In summary, while both Horn’s and Levinson’s neo-Gricean endeavours have put the classical Gricean theory on a much more elegant and rigorous basis, neither of them, however, has given the original Gricean maxim of Relation a more prominent place.

5.

A comparison of relevance and neo-Gricean pragmatic principles and related areas

What, then, are the main similarities and differences between relevance and neoGricean pragmatic principles and their related areas? There are at least three main similarities. In the first place, both relevance and neo-Gricean pragmatic theories share the Gricean assumption that sentence meaning is a vehicle for conveying meaning-nn or speaker meaning, and the essence of meaning-nn is communication which is intended to be recognized as having been intended. In other words, speaker meaning is a matter of expressing and recognizing intention (Huang 2007). A second similarity is that both theories maintain with Grice that speaker meaning cannot be simply decoded, but has to be pragmatically enriched. This indicates that both theories belong to the camp of “contextualism” in the current divide between contextualism and semantic minimalism or literalism in the philosophy of lan-

Relevance and neo-Gricean pragmatic principles

37

guage. Thirdly and finally, both theories assume that in computing speaker meaning, an addressee is guided by the expectation that an utterance should meet certain standards (see also Wilson 2010). On the other hand, there are major differences between relevance and neoGricean pragmatic principles and their related areas. The first difference is concerned with whether a pragmatic theory of human communication should be based on the study of usage principles or cognitive principles. One of the central issues of any pragmatic theory is to explain how the addressee works out the speaker’s intended meaning on the basis of the evidence available (but see note 10 below). The answer provided by Grice is that utterances automatically raise certain expectations, and these expectations guide the addressee towards what the speaker intends. More specifically, Grice put forward an account of these expectations in terms of a co-operative principle and a set of attendant conversational maxims. The Gricean co-operative principle and its associated maxims of conversation are essentially usage principles based on the rational nature of human communication, and indeed of any goal-oriented (human) activity (Grice 1989: 28). In other words, they are general communicative norms recognized jointly, though tacitly, by the speaker and the addressee in order to communicate effectively and efficiently. Where do the co-operative principle and its attendant conversational maxims come from, and how do the speaker and the addressee come to know them? Whereas Grice was non-committed on the source of these pragmatic principles and their place in our overall cognitive architecture, one possible answer provided by him is that they are likely to be learned. To quote him: “it is just a well-recognized empirical fact that people do behave in these ways; they learned to do so in childhood and have not lost the habit of doing so; and, indeed, it would involve a good deal of effort to make a radical departure from the habit” (Grice 1989: 29). In the relevance-theoretic framework, on the other hand, pragmatics is reduced to a single notion of relevance, which is realized in two principles of relevance. But unlike Grice’s co-operative principle and its attendant maxims of conversation, the principles of relevance are not a maxim addressed to the speaker, known by the addressee, and obeyed or exploited in communication. Rather, grounded in a general view of human cognition, they are an automatic reflex of the human mental capacity that works without the communicators having any overt knowledge of it. How do the speaker and the addressee follow the principles of relevance? They do not. According to Sperber and Wilson (1995: 162), “[c]ommunicators and audience need no more know the principle of relevance to communicate than they need to know the principles of genetics to reproduce. Communicators do not ‘follow’ the principle of relevance; and they could not violate it even if they wanted to. The principle of relevance applies without exception: every act of ostensive communication communicates a presumption of relevance”. Relevance is thus a form of unconscious inference. In other words, the principles of relevance are governing cognitive principles that are not themselves an object of processing.7 This raises the

38

Yan Huang

larger issue of whether relevance theory can be falsified or not. Given that relevance is an exceptionless generalization, it is likely to be immune from any possible counterexamples (see e.g. Levinson 1989, Huang 1994/2007, 2000 and 2007, but see Wilson and Sperber 2004, Wilson 2009 for counterarguments).8 In the second place, relevance theory and classical/neo-Gricean pragmatic theory differ in whether or not separate notions of explicature and implicature are needed in the recovery of so-called explicit and implicit content, respectively. As is well-known, in Grice’s opinion, the propositional content of what is said is not fully worked out until reference is identified, deixis is interpreted and ambiguity is resolved. How can all this be done? Grice (1989: 25) seemed to take the recovery of the explicit truth-conditional, propositional content as largely the outcome of linguistic and contextual decoding. This led Sperber and Wilson to their criticism that, in the Gricean paradigm, only the recovery of implicit context (r-implicature) is taken to be properly pragmatic, and that as a result, the key part pragmatics plays in computing explicit content is overlooked. Consequently, Sperber and Wilson posited their notion of explicature. Explicature plays an extremely important role in relevance theory, and as a consequence, many types of implicature in the Gricean and neo-Gricean sense are reduced to explicature. This, for example, is the case for “scalar implicature”, “conjunction buttressing”, and many others. By contrast, in the neo-Gricean pragmatic framework, the relevance-theoretic explicature-implicature distinction is rejected. In Levinson’s (2000: 195–196) opinion, pragmatic intrusion into what is said, that is, into determining the truth-conditional, propositional content is neither an explicature, nor the pragmatically enriched said, as argued by Recanati (2004); nor an impliciture, as urged by Bach (2004). Rather, it is the same beast as a neo-Gricean conversational implicature. In my opinion, the reason why it is a conversational implicature is threefold. First, so-called explicature is engendered largely by the same Gricean pragmatic mechanism that yields a conversational implicature. Secondly, Recanati (1993) put forward two tests, i.e. the availability principle and the scope principle, to differentiate explicature, the pragmatically enriched said, and impliciture from conversational implicature. But as I argued in Huang (2007), neither of Recanati’s tests really works from a theoretical point of view. This is also the case with work carried out in experimental pragmatics. I do not think that there is any experiment that can differentiate explicature from conversational implicature. Therefore, currently there is no failsafe test (either conceptual or experimental) that can be employed to distinguish alleged explicature or the pragmatically enriched said or impliciture from conversational implicature on a principled basis. Thirdly, other things being equal, given the metatheoretical principle known as “Occam’s razor” (“theoretical entities are not to be multiplied beyond necessity”), the implicature analysis is theoretically and methodologically preferable, because it postulates less representational levels in the interpretation of an utterance than the relevance-theoretic explicature/implicature analysis. If neo-Gricean conversational implicature

Relevance and neo-Gricean pragmatic principles

39

can intrude into the truth-conditional content of an utterance, then a problem known as Grice’s circle arises, namely, how what is conversationally implicated can be defined in contrast to, and calculated on the basis of what is said, given that what is said seems to both determine and to be determined by what is conversationally implicated (e.g. Levinson 2000, Huang 2007).9 Levinson’s proposal was that one should reject the “received” view of the pragmatics-semantics interface, namely the view that the output of semantics is the input of pragmatics, and to allow implicatures to play a systematic role in “pre”-semantics, that is, to help determine the truth-conditional, propositional content of an utterance (Levinson 2000, see also Huang 2003, 2007). To put it slightly differently, in order to avoid Grice’s circle, one needs both “pre”-semantic pragmatics and “post”-semantic pragmatics or what Korta and Perry (2008) called “near-side pragmatics” and “farside pragmatics”. Thirdly, a heated debate has been going on for the last two decades or so, focusing on the nature of Gricean and neo-Gricean generalised conversational implicatures (GCIs) in general and Q-scalar implicatures in particular. One view is that GCIs in general and Q-scalar implicatures in particular convey default meaning, that is, their meaning is automatically worked out by an addressee on encountering a scalar implicature trigger like some. In other words, according to this default inference theory, scalar implicatures are inferences sans a conscious inferential process and irrespective of a particular context (see e.g. Levinson 2000 for strong defaultism and Horn 2009 for weak defaultism). A second, relevance-theoretic position, known as the contextual inference theory, maintains that the derivation of scalar implicatures depends heavily on contextual factors. Another way of saying this is that scalar implicatures can only be derived if the context warrants it. In addition, a third contender has recently entered the debating chamber. According to this position, scalar implicatures arise in an upward entailing environment but are quite weak and even blocked in a downward entailing environment. On the basis of this claim, it is argued that the derivation of Q-scalar implicatures relies heavily on structural or grammatical factors. Consequently, scalar implicatures must be computed compositionally, and therefore fall under compositional semantics, hence part of grammar or innate linguistic mechanism. This position is associated with the work of Chierchia (2004) and is called the structural inference theory. Furthermore, all the three views have recently been subject to studies in experimental pragmatics. While much of the relevance-theoretically oriented experimental work naturally favours the contextual inference approach (e.g. Noveck and Sperber 2007, Noveck and Reboul 2008), there is also experimental evidence in support of the default inference theory (e.g. Grodner et al. 2007) and the structural inference view (e.g. Panizza and Chierchia 2008, see also Garrett and Harnish 2009). In the fourth place, there is the debate relating to whether a pragmatic theory of human communication should contain two levels or three levels. Building on the Gricean generalized versus particularized implicature dichotomy, Levinson (2000)

40

Yan Huang

developed a theory of presumptive meaning. From a traditional, standard point of view, there are only two levels of meaning to a theory of communication: a level of sentence-meaning versus a level of speaker-meaning, or, to make use of the typetoken distinction, a level of sentence-type-meaning versus a level of utterancetoken-meaning. But Levinson (2000: 23) argued that such a view “is surely inadequate, indeed potentially pernicious, because it underestimates the regularity, recurrence, and systematicity of many kinds of pragmatic inferences”. He proposed to add a third level – utterance-type-meaning – to the two generally accepted levels of sentence-type-meaning and utterance-token-meaning. This third layer is the level of generalized, preferred or default interpretation, which is not dependent upon direct computations about speaker-intentions but rather upon expectations about how language is characteristically used. Generalized conversational implicatures, Levinson argued, should be included in this layer, as these pragmatic inferences have an expectable, stable and even standardized or conventionalized interpretation. Stated in this way, a neo-Gricean pragmatic theory of conversational implicature, which is largely concerned with generalized rather than particularized implicature, is essentially a theory of utterance-type-meaning on a level intermediate between sentence-type-meaning on the one hand and utterance-token-meaning on the other. In other words, it is ultimately a theory of presumptive meaning – pragmatic inference that is generalized, default and presumed. However, as pointed out by Levinson (2000), this middle-layer of utterance-type-meaning has been constantly subject to attempts to reduce it on the one hand to the upper-layer of sentence-type-meaning, as in e.g. Chierchia’s (2004) structural inference theory, and on the other to the lower-layer of utterance-token-meaning, as in, e.g. relevance theory. In my opinion, such reductionist efforts, though highly desirable given “Occam’s razor”, cannot be successful. The reason they will fail is this: on the one hand, generalized conversational implicatures are defeasible, thus not code-like, as claimed by Sperber and Wilson. This will make it difficult for them to be semanticized. On the other hand, other things being equal, a theory about types is in principle better than a theory about tokens in that the former enjoys more predictive and explanatory power. Therefore, any attempts to reduce generalized conversational implicatures to a kind of context-induced, “nonce” inferences should be resisted. If these arguments are correct, a three-tiered theory of communication with a layer of default but defeasible interpretation sitting midway is in principle to be preferred over a two-levelled one without such an intermediate layer (Huang 1994, 2000, 2003, 2004, 2007). Finally, somewhat related to the difference concerning communication theory levels above is the difference with respect to the distinction in conversational implicature type or context type. In the Gricean framework, there are two types of conversational implicature, i.e. generalized and particularized. Alternatively, one can argue that there is only one type of conversational implicature but two types of context, i.e. default and specific. In relevance theory, by contrast, there is neither

Relevance and neo-Gricean pragmatic principles

41

any distinction in conversational implicature type nor any distinction in context type. As noted already, all conversational implicatures are reduced to a kind of context-induced, “nonce” inference, namely what I have called r-implicatures (see also Huang 2007).10

Acknowledgements I would like to thank the anonymous referees, Hans-Jörg Schmid, the editor of this volume, and Wolfram Bublitz, one of the editors of the series for their helpful comments on an earlier version of this chapter. The research reported on here was partially supported by a grant from the University of Auckland Faculty Research Development Funds, which is hereby gratefully acknowledged.

Notes 1. Currently there is a split on the issue of whether a conversational implicature is a component of speaker-meaning or a pragmatic inference. Saul (2002) was of the opinion that Grice’s main goal is to develop a theory of speaker-meaning, but not a theory of psychological processing. This view is echoed in Bach (2006). Following Saul (2002) and Bach (2006) and biting the bullet, Horn (2009) now holds that a conversational implicature is a component of speaker-meaning rather than a pragmatic inference. By contrast, Sperber and Wilson (1986/1995), Levinson (2000) and Atlas (2005), for example, are still treating a conversational implicature as a pragmatic inference. My definition is applicable to both sides. 2. In fact, a similar notion of relevance can be traced back to the work of the logicians Noel Belnap (1969) and A. R. Anderson and the social phenomenologist Alfred Schutz (1970). Anderson and Belnap (1975/1992) developed a logic of relevance to account for inferences that are relevant but not quite valid. Schutz (1970) argued that relevance is a principle according to which an individual organizes his or her cognitive structures into “provinces of meaning”. See also Allwood (2000). 3. Notice that in Sperber and Wilson (1986), there was only one principle of relevance, namely the communicative principle of relevance, which goes thus: every ostensive stimulus conveys a presumption of its own optimal relevance. On Sperber and Wilson’s (2005: 261) view, “[t]he change is, of course, expository and not substantive”. 4. Strictly speaking, ‘together’ is what is implicated. ‘John and Mary watched the Berlin Wall crumble in 1989 together’ is what is communicated, that is, the sum of what is said and what is implicated. In what follows, however, I shall not make such a difference. 5. Manichaeus or Manes was an ancient Persian teacher who advocated the doctrine known as Manichaeanism or Manicheism, namely, the world is governed by a balance of the forces of good and evil. 6. An issue arises concerning the epistemic commitment or strength of Q-implicatures, especially Q-scalar implicatures. The issue has to do with the question of what it is a speaker Q-scalar implicates against. Two main neo-Gricean pragmatic positions can be identified:

42

7.

8.

9.

10.

Yan Huang the weak epistemic force one represented by e.g. Hirschberg (1991), Sauerland (2004), Geurts (2009), and Horn (2009), and the strong epistemic force one advocated by e.g. Gazdar (1979) and Levinson (2000: 77–79). In addition, there is a third neo-Gricean position. In Atlas’s (2005) view, the whole issue is derived from the confusion between speaker meaning on the one hand and a speaker’s psychological states on the other. See also Cummings (2005) for a philosophical critique of three critical features of relevance theory, namely (i) the elimination rules of the logical entries of concepts, (ii) the role of deduction in utterance interpretation, and (iii) the functional concept of confirmation. In fact, Cummings’ criticism is in part based on Putnam’s (1981) similar charge of self-refutation against logical positivism. In Cummings’ opinion, the relevance-theoretic claims display a striking similarity to those of logical positivism. One of the minimal Popperian criteria for a scientific theory is falsifiability, which dictates that empirically based theories (under which linguistics falls) can only be refuted, but not proven true. Grice’s circle is labelled the “pragmatic circle” by Korta and Perry (2008). According to Korta and Perry, the circle runs like this: Gricean reasoning requires what is said to get started and is needed to get to what is said. Cf. Saul (2002), who was of the opinion that classical/neo-Gricean pragmatic theory and relevance theory are not as incompatible as they may appear. This is because while Grice’s main goal is to develop a theory of speaker-meaning or utterance production, Sperber and Wilson (1986/1995) are primarily concerned with the construction of a cognitive psychological theory of utterance comprehension. Their theory does not attend to the question of how and why the speaker, given what he or she wants to convey, utters what he or she utters (Horn 2006). Put slightly differently, in Mate and Tirassa’s (2010: 240) words, “however satisfactory as theory of language comprehension, relevance theory has never been able to become a theory of language generation or of dialogue, nor has it ever even tried to”. When there are meaning discrepancies between utterance production and comprehension, one has what Saul called an “audience-implicature”. This opinion was also shared by Bach (1999), who argued that utterance interpretation is not a problem for pragmatics, but for cognitive and social psychology. Both Saul and Bach thought that relevance theory has misunderstood this main goal of Grice’s thinking.

References Allwood, Jens 2000 An activity-based approach to pragmatics. In: Harry Bunt and William Black (eds.), Abduction, Belief and Context in Dialogue, 47–80. Amsterdam: John Benjamins. Anderson, Alan Ross and Noel Belnap 1975/1992 Entailments. 2 vols. Princeton: Princeton University Press. Atlas, Jay D. 2005 Logic, Meaning, and Conversation. Oxford: Oxford University Press. Atlas, Jay D. and Stephen C. Levinson 1981 It-clefts, informativeness and logical form: radical pragmatics. In: Peter Cole (ed.), Radical Pragmatics, 1–61. London: Academic Press.

Relevance and neo-Gricean pragmatic principles

43

Bach, Kent 1999 The semantics-pragmatics distinction: what it is and why it matters. In: Ken Turner (ed.), The Semantics-Pragmatics Interface from Different Points of View, 65–84. New York: Elsevier Science. Bach, Kent 2004 Pragmatics and the philosophy of language. In: Laurence R. Horn and Gregory Ward (eds.), The Handbook of Pragmatics, 463–87. Oxford: Blackwell. Bach, Kent 2006 The top ten misconceptions about implicature. In: Betty Birner & Gregory Ward (eds.), Drawing the Boundaries of Meaning, 21–30. Amsterdam: John Benjamins. Belnap, Noel 1969 Questions: their presuppositions, and how they can fail to arise. In: Karel Lambert (ed.), The Logical Way of Doing Things. New Haven: Yale University Press. Blakemore, Diane 2002 Relevance and Linguistic Meaning. Cambridge: Cambridge University Press. Borg, Emma 2009 On three theories of implicature: default theory, relevance theory and minimalism. International Review of Pragmatics 1: 63–83. Burton-Roberts, Noel 2007 Varieties of semantics and encoding: negation, narrowing/loosening and numericals. In: N. Burton-Roberts (ed.), Pragmatics, 90–114. London: Palgrave Macmillan. Burton-Roberts, Noel (ed.) 2007 Pragmatics. London: Palgrave Macmillan. Chierchia, Gennaro 2004 Scalar implicatures, polarity phenomena, and the syntax/pragmatics interface. In: Adriana Belletti (ed.), Structures and Beyond, 39–103 Oxford: Oxford University Press. Cummings, Louise 2005 Pragmatics. Edinburgh: Edinburgh University Press. Cummings, Louise (ed.) 2010 The Pragmatics Encyclopaedia. London: Routledge. Garrett, Merrill and Robert Harnish 2009 Q-phenomena, I-phenomena and impliciture: some experimental pragmatics. International Review of Pragmatics 1: 84–117. Gazdar, Gerald 1979 Pragmatics: Implicature, Presupposition and Logical Form. London: Academic Press. Geurts, Bart 2009 Scalar implicature and local pragmatics. Mind and Language 51–79. Grice, H. P. 1975 Logic and conversation. In: Peter Cole and Jerry Morgan (eds.), Syntax and Semantics 3: Speech Acts, 41–58. London: Academic Press. Grice, H. P. 1989 Studies in the Way of Words, Cambridge, MA: Harvard University Press.

44

Yan Huang

Grodner, Daniel, Natalie Klein, Kathleen Carbary and Michael Tanenhaus 2007 Experimental evidence for rapid interpretation of pragmatic some. Paper presented at XPRAG, Berlin. Hirschberg, Julia 1991 A Theory of Scalar Implicature. New York: Garland. Horn, Laurence R. 1984 Toward a new taxonomy for pragmatic inference: Q-based and R-based implicature. In: Deborah Schiffrin (ed.), Meaning, Form, and Use in Context: Linguistic Applications, 11–42. Washington DC: Georgetown University Press. Horn, Laurence R. 2004 Implicature. In: Laurence R. Horn and Gregory Ward (eds.), The Handbook of Pragmatics, 3–28. Oxford: Blackwell. Horn, Laurence R. 2006 The border wars: A neo-Gricean perspective. In: K. von Heusinger and K. Turner (eds.), Where Semantics Meets Pragmatics, 21–48. Oxford: Elsevier. Horn, Laurence R. 2009 WJ-40: Implicature, truth, and meaning. International Review of Pragmatics 1: 3–34. Horn, Laurence R. and Gregory Ward (eds.) 2004 The Handbook of Pragmatics. Oxford: Blackwell. Huang, Yan 1991 A neo-Gricean pragmatic theory of anaphora. Journal of Linguistics 27: 301–335. Huang, Yan 1994/2007 The Syntax and Pragmatics of Anaphora. Cambridge: Cambridge University Press. Huang, Yan 2000 Anaphora: A Cross-linguistic Study. Oxford: Oxford University Press. Huang, Yan 2003 On neo-Gricean pragmatics. International Journal of Pragmatics 13: 87–110. Huang, Yan 2004 Anaphora and the pragmatics-syntax interface. In: Laurence R. Horn and Gregory Ward (eds.), The Handbook of Pragmatics, 288–314. Oxford: Blackwell. Huang, Yan 2007 Pragmatics. Oxford: Oxford University Press. Huang, Yan 2009 Neo-Gricean pragmatics and the lexicon. International Review of Pragmatics 1: 118–153. Huang, Yan 2010a Neo-Gricean pragmatic theory of conversational implicature. In: Bernd Heine and Heiko Narrog (eds.), The Oxford Handbook of Linguistic Analysis, 669–696. Oxford: Oxford University Press. Huang, Yan 2010b Implicature. In: Louise Cummings (ed.), The Pragmatics Encyclopaedia, 234–38. London: Routledge. Huang, Yan 2010c Scalar implicature. In: Louise Cummings (ed.), The Pragmatics Encyclopaedia, 441–444. London: Routledge.

Relevance and neo-Gricean pragmatic principles

45

Huang, Yan 2010d Impliciture. In: Louise Cummings (ed.), The Pragmatics Encyclopaedia, 238–40. London: Routledge. Huang, Yan 2011 Types of inference: entailment, presupposition, and implicature. In: Wolfram Bublitz and Neil Norrick (eds.), Foundations of Pragmatics, 397–421. Berlin: Mouton de Gruyter. Korta, Kepa and John Perry 2008 The pragmatic circle. Syntheses 165: 347–357. Levinson, Stephen C. 1987 Pragmatics and the grammar of anaphora. Journal of Linguistics 23: 379–434. Levinson, Stephen C. 1989 A review of Relevance. Journal of Linguistics 25: 455–472. Levinson, Stephen C. 2000 Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge, MA: The MIT Press. Mate, David and Maurizio Tirassa 2010 Knowledge. In: Louise Cummings (ed.), The Pragmatics Encyclopaedia, 239–242. London: Routledge. Noveck, Ira and Anne Reboul 2008 Experimental pragmatics: A Gricean turn in the study of language. Trends in Cognitive Sciences 12: 425–31. Noveck, Ira and Dan Sperber 2007 The why and how of experimental pragmatics: The case of ‘scalar inference’. In: Noel Burton-Roberts (ed.), Pragmatics, 184–212. London: Palgrave Macmillan. Panizza, D. and Chierchia Gennaro 2008 Two experiments on the interpretation of numerals. Paper presented at Workshop on Experimental Pragmatics/Semantics, the 30th Annual Convention of the German Society of Linguistics. Putnam, Hilary 1981 Reason, Truth and History. Cambridge: Cambridge University Press. Recanati, François 1993 Direct Reference. Oxford. Blackwell. Recanati, François 2004 Literal Meaning. Cambridge: Cambridge University Press. Sauerland, Uli 2004 Scalar implicatures in complex sentences. Linguistics and Philosophy 27: 367–391. Saul, Jennifer 2002 What is said and psychological reality: Grice’s project and relevance theorists’ criticisms. Linguistics and Philosophy 25: 347–372. Schutz, Alfred 1970 Reflections on the Problems of Relevance. New Haven: Yale University Press. Sperber, Dan and Deidre Wilson 1986/95 Relevance: Communication and Cognition, 2nd edition, Oxford: Blackwell.

46

Yan Huang

Wilson, Deidre 2010 Relevance theory. In: Louise Cummings (ed.), The Pragmatics Encyclopaedia, 394–398. London: Routledge. Wilson, Deirdre and Dan Sperber 2004 Relevance theory. In: Laurence R. Horn and Gregory Ward (eds.), The Handbook of Pragmatics, 607–632. Oxford: Blackwell.

3.

Implicature and explicature Robyn Carston and Alison Hall1

1.

Introduction

The explicature/implicature distinction is one manifestation of the distinction between the explicit content of an utterance and its implicit import. On certain ‘minimalist’ approaches, the explicit/implicit distinction is equated with the semantics/ pragmatics distinction or with Paul Grice’s saying/implicating distinction. However, the concept of ‘explicature’, which belongs to the relevance-theoretic pragmatic framework, has closer affinities with the wider ‘contextualist’ perspective, according to which context-sensitive pragmatic processes make a much greater contribution to the proposition explicitly communicated than merely resolving ambiguities and providing referents for indexicals. Crucially, there are pragmatic processes of meaning enrichment and adjustment which have no linguistic mandate but are wholly motivated by considerations of communicative relevance. Two consequences of this are that (a) explicit utterance content can include constituents which are not articulated in the linguistic form of the utterance, and (b) certain Gricean implicatures are reanalysed as components of the explicitly communicated truth-conditional content. In this chapter, we outline the explicature/implicature distinction and highlight some of the issues it raises for semantic/pragmatic theorizing. To get a preliminary idea of how the distinction between implicature and explicature works, consider what Amy communicates in the following exchange: (1) Max: Amy: (2) T HE

How was the party? Did it go well? There wasn’t enough drink and everyone left early.

PARTY DID NOT GO WELL 2

It seems fairly clear that she is communicating (2). This is not something she says explicitly; rather, it is an indirect or implied answer to Max’s question – a conversational implicature, as such implicitly communicated propositions are known. The hearer (Max) derives this implicated meaning by inferring it from the proposition which is more directly and more explicitly communicated by Amy, together with his readily available assumptions about the characteristics of successful versus unsuccessful parties (e.g. people tend to leave early when a party isn’t very good). What is the more directly communicated proposition, the ‘explicature’ of the utterance? One possibility is that it is simply the linguistically encoded meaning of the sentence that she uttered, so it is the conjunction of the context-free meaning of the two simple sentence types:

48

Robyn Carston and Alison Hall

(3) (i) T HERE

WASN ’ T ENOUGH DRINK

(ii) E VERYONE

LEFT EARLY

This is, undeniably, the most explicit component of the meaning of Amy’s utterance, but it is very unclear what exactly it amounts to and it seems to lack the specificity of the understood content. For instance, the extension of the noun drink includes green tea, tap water, and medicines in liquid form, to mention but a few of the many drinks which are unlikely to be relevant in the context of Amy’s utterance. Similarly, the linguistically encoded meaning of the bare quantifier everyone includes many people whom Amy has no intention of denoting. In the context of the dialogue above, it is clear that she intends to convey that everyone who came to the particular party that Max asked her about left that party early. So, although the linguistic expressions employed by Amy, the words she actually uttered, have a meaning and that meaning is, arguably, the most explicit meaning that her utterance provides, it seems to be quite remote from the proposition that Max is likely to take her to have directly communicated (to have said, stated, or asserted). That seems to be more like the content in (4) (where the elements highlighted in bold all go beyond the encoded meaning of the linguistic expressions uttered): (4) T HERE

WASN ’ T ENOUGH ALCOHOLIC DRINK TO SATISFY THE PEOPLE AT THE

PARTY I AND SO EVERYONE WHO CAME TO THE PARTY I LEFT IT I EARLY.

This is the proposition on the basis of which Amy’s utterance would be judged as true or false, would be agreed or disagreed with (‘Yes, there was so little alcohol that we all had to go off to the pub’, or ‘No, not everyone left the party early and those who did had an exam the next morning’). Notice also that it is this proposition (and not the very general encoded linguistic meaning) which plays the crucial role of premise in the reasoning process which leads to the implicated conclusion that the party didn’t go well. We take it that this (or something very similar to it) is the explicature of Amy’s utterance. The distinction between two kinds of communicated propositions, explicatures and implicatures, has been developed within the relevance-theoretic account of communication and utterance interpretation (Sperber and Wilson 1986/1995; Carston 2002; Wilson and Sperber 2004). This framework is resolutely cognitivescientific in orientation: the central notion of informational ‘relevance’ which drives the account is defined in terms of the positive cognitive implications that a new input has in a cognitive system weighted against the costs (in such resources as attention, inferential effort, etc.) that it imposes on that system. The greater the range of cognitive implications and the lower their cost, the more the relevance of the input. Verbal utterances and other kinds of ostensive communicative acts are special inputs in that there is an inevitable presumption that they will be ‘optimally relevant’, that is, they will provide at least a sufficient array of cognitive implications and other positive cognitive effects to offset the processing effort they

Implicature and explicature

49

require.3 It is in the (non-demonstrative) inferential process of looking for an interpretation consistent with this presumption that a hearer derives an explicature, by enriching and modulating the conceptual schema provided by decoded linguistic meaning. This occurs in parallel with the derivation of implicatures (cognitive implications manifestly intended by the speaker), and the two kinds of propositional meanings are mutually constraining. The ultimate interpretation should be one in which the explicature together with intended contextual assumptions provides an inferentially sound basis for the implications derived. In this paper, we look in detail at some of the particular micro-processes involved in the online, relevance-driven derivation of explicature and implicatures. But before that, we set out some background intellectual history tracing the development of the concept of implicature. It did not arise within linguistics or cognitive science but within the philosophy of language, where its main purpose was to help in the delineation of a favoured notion of ‘semantic content’. Its subsequent adoption into the cognitively-based relevance-theoretic account of communication where it is placed in opposition to explicature rather than semantics has naturally led to its being somewhat altered with regard to the domain of utterance meaning it encompasses and the role it plays.

2.

How it all began: semantic content and implicature

In his ‘logic of conversation’, Paul Grice (1967/89) sought to separate out the semantic content of an uttered sentence, i.e. what it says, from any other thoughts and ideas that a speaker might mean or communicate by her action of uttering the sentence. His collective term for all those extra or secondary meanings that might be conveyed was ‘implicature’, where implicatures are intended propositional components of the utterance’s overall significance but are not the basis on which the utterance is judged as true or false. Implicatures can arise in two ways: via presumptions concerning rational communicative behaviour or via certain linguistic conventions. The implicature of (5) is an example of the first sort, a ‘conversational’ implicature, and the implicature of (6) is an example of the second sort, a ‘conventional’ implicature (see Huang, this volume, and Moeschler, this volume, for more details): (5) That material looks red to me. Implicature: T HERE IS SOME DOUBT NOT .

ABOUT WHETHER THE MATERIAL IS RED OR

(6) Mary is a housewife but she is very intelligent. Implicature: T HERE IS A CONTRAST OF SOME SORT BETWEEN BEING A HOUSEWIFE AND BEING VERY INTELLIGENT.

50

Robyn Carston and Alison Hall

With (5), the idea is that if the speaker was completely certain of the redness of the material she should have made the more informative statement that the material is red; since she did not and since, other things being equal, speakers are expected to be as informative as they can relevantly be, she must be implicating that there is some doubt about the redness of the material. Thus this conversational implicature follows from one of the several conversational maxims that Grice sets out as regulating rational communication (Grice 1975). Note that the proposition in (2) above, implicated by Amy in her response to Max in (1), is also a conversational implicature, one that would be dependent on Grice’s conversational maxim of relevance. He drew a further distinction among conversational implicatures between ‘generalized’ ones, such as (5), which arise across a great many contexts of use and ‘particularized’ ones, such as (2), which are dependent on the properties of specific, often one-off, contexts, in this case the conversation about a particular party. Whether Grice intended this generalized/particularized distinction to carry any theoretical weight is unclear, but, as we will see when we move to the explicature/implicature distinction, it turns out that the status of many of the generalized cases is quite controversial. With (6), on the other hand, the implicature does not depend on any conversational presumptions and occurs across all contexts because it is generated on the basis of the conventional linguistic meaning of the connective but. In all these cases, the meaning allegedly implicated is separated off from the primary meaning, that is, the semantic content of the uttered sentence, which is the propositional content on the basis of which the utterance is judged as true or false. In the case of (5), this propositional content is that the material in question appears to the speaker to be red, so if the speaker is, in fact, in no doubt that the material is red, her utterance is somewhat infelicitous or inappropriate, but she has not spoken falsely on that basis. In the case of (6), the propositional content is that Mary is a housewife and she is very intelligent, and, again, if there is, in fact, no contrast between these two properties, the utterance is not thereby made false, though it is certainly inappropriate and very misleading. In Grice’s account, what implicatures of any stripe are set apart from is ‘what is said’ by a speaker, which Grice took to be the truth-conditional content of the utterance. What is said is determinable from the conventional linguistic meaning of the sentence uttered together with some minor context-dependent considerations, specifically, selection of the occasion-specific sense of any ambiguous words or structures and fixing of referents of indexical elements. The centrality of this semantic ‘what is said’, the proposition expressed by a sentence, goes back some way in the history of the philosophy of language, to at least Frege, Russell and Carnap. They were first and foremost logicians, interested in the syntactic and, especially, the semantic properties of formal languages, such as the predicate calculus. However, they extrapolated from these artificial languages to human (natural) languages, which they assumed would be found to have the same fundamental properties, abstracting away from such imperfections as ambiguity and vagueness. So,

Implicature and explicature

51

just as the semantics of logical formulae was taken to be a matter of how the external world must be for them to be true (that is, their truth conditions) and the semantics of logical connectives such as ‘&’, ‘v’, ‘f’ was fully captured by truth tables, it was assumed that natural language sentences also have truth conditions and natural language connectives such as and, or, if … then are truth-functional. The presence within natural languages of such elements as indexicals which depend on a context of use for their ‘semantic value’ was seen as an interesting extra issue to be dealt with but not a threat to the overall picture. Adherence to truth-conditional semantics and to explicit logical formalism continues today in contemporary formal semantic work on natural languages. This ‘ideal language’ approach was challenged by Austin, Strawson and the later Wittgenstein, who developed the ‘ordinary language’ approach, aimed at describing natural language phenomena rather than forcing them into the logical mould. They rejected the equation of sentence meaning with truth conditions, maintaining that although a sentence abstracted from use has a meaning, it is only in the context of a speech act (an utterance) that it expresses a proposition and so has truth conditions; it is the statement thus made that has truth conditions, the sentence per se does not. This aspect of ordinary language philosophy is very much reflected in the relevance-theoretic framework according to which sentence meaning is not truth-conditional but provides merely a template or schema which is contextually enriched on an occasion of utterance into a complete proposition, a proposition which the speaker has explicitly communicated (the explicature of the utterance). A second aspect of the descriptive investigative work of the ‘ordinary language’ philosophers was the close observation of the meaning of words as used in ordinary communication. This included attention to a range of linguistic expressions that lay beyond the reach of the truth-conditional paradigm, including such connectives as but, yet, despite, after all, whereas, moreover, so and anyway, sentence adverbials like frankly, seriously, evidently, unfortunately, and incidentally, and such discourse elements as alas, indeed, oh, for goodness sake, and well. Quite a few of these seemed explainable in speech act terms; for instance, but in (6) above could be characterized as introducing a second-order speech act of contrasting the two first-order speech acts of stating (‘Mary is a housewife’ and ‘Mary is very intelligent’); seriously could be characterized as modifying the speech act a speaker is performing, e.g. an act of asserting (‘Seriously, you will regret this’) or an act of requesting information (‘Seriously, where is my key?’) (for discussion, see Ifantidou 2001). This focus on language in use also led to a reappraisal of analyses of words whose semantics was of central importance within the formal truth-conditional paradigm. Strawson (1952) pointed out that the natural language counterparts of logical connectives are often used with much richer (non-truth-functional) meanings than their counterparts in logical languages, such as the temporal and causal

52

Robyn Carston and Alison Hall

connotations of many instances of sentences conjoined with and (e.g. She insulted him and he resigned), the implication often carried by conversational disjunctions that the speaker’s grounds for uttering a sentence of the form ‘P or Q’ are not that she knows P to be true or that she knows Q to be true, but that she has some other (non-truth-functional) basis for her utterance. Close attention to a range of predicates led to strong views on their precise meanings and thus on how they ought to be used: the word know should not be used about certainties (e.g. I know this is my hand is a misuse), the word try only applies if there is some difficulty, the phrase looks to me (as in (5) above) is used only when there is some doubt about the reality, and so on (for discussion, see Travis 1991). Initially, the two approaches, the truth-theoretic and the use-theoretic, were seen as diametrically opposed and exclusive: either you analyse natural language meaning in the logical truth-conditional way or you describe it as it occurs and is understood in everyday use. One of Grice’s great contributions was to bring the two traditions together into a complementary rather than rivalrous relation. His own formative philosophical development lay within the ordinary language camp, and its emphasis on speaking as an action (rationally-based and with consequences) is evident throughout his work on implicature. However, his saying/implicating distinction and his analyses of particular natural language words, including the connectives and quantifiers, are informed by insights from both traditions. He insisted – against the central tenet of most ordinary language theorists – that it is important not to equate meaning and use (Grice 1967/89: 4); in other words, he distinguished semantics and pragmatics. So, in the case of (5) above, someone who utters it looks red to me when she has absolute certainty about the redness of the item in question may, in many contexts, be somewhat misleading, saying something weaker than she could have said, given the state of her knowledge; nevertheless, the proposition that comprises the semantic content of her utterance, what she has said (as opposed to what she merely implicated), namely, that the item in question looks red to her, is true. Most important for Grice’s case here is that the implication of doubt or uncertainty is cancellable: it is possible to conceive of a context in which what is at issue is people’s perceptions, how things look or sound or feel to them, in which case the utterances X feels hot to me and Y looks red to me would not carry any implications of doubt or uncertainty. Thus, those implications, prevalent though they may be, are not part of the meaning (the semantics) of those words, not part of what is said by them; when they do arise they are a product of speaker-hearer assumptions about normal conversational use. The explanatory power of an account that combines logico-semantic analysis with considerations of language use is particularly well-exemplified in Grice’s treatment of the connectives and and or. Their semantics, he argues, is identical with that of their logical counterparts, hence truth-functional, and the stronger implications that they seem to have in many contexts can be explained by the ‘logic of conversation’:

Implicature and explicature

53

(7) Amy insulted Max and he resigned. Propositions meant: a. A MY INSULTED M AX & M AX RESIGNED b. M AX RESIGNED AFTER A MY INSULTED HIM c. M AX RESIGNED BECAUSE A MY INSULTED HIM (8) Max is working on his lecture or he is watching TV. Propositions meant: a. M AX IS WORKING ON HIS LECTURE V M AX IS WATCHING TV b. M AX IS NOT BOTH WORKING ON HIS LECTURE & WATCHING TV c. T HE SPEAKER DOESN ’ T KNOW THAT M AX IS WORKING ON HIS LECTURE d. T HE SPEAKER DOESN ’ T KNOW THAT M AX IS WATCHING TV In each case, on Grice’s account, the proposition that constitutes the semantics of the utterance (what is said) is the first one and all the others are conversational implicatures. Various conversational maxims play their part in the (non-demonstrative) inference process by which these implicatures are derived: the maxim of orderliness for (7b), probably the maxim of relevance for (7c), the maxim of informativeness for (8b)–(8d).4 One of the great strengths, in Grice’s opinion, of an account which distinguishes the statement made by an utterance (hence its semantics) from its implicatures is that it allows for the very general patterns of valid inference formulated within the logical semantic tradition to be carried over into the semantics of natural language (for discussion, see Grice 1975: 41–43). Thus, in Grice’s conception of ‘what is said’ we see the preservation of a notion of semantic content much akin to that of Frege and Russell, that is, closely tied to the context-free semantics of the words in the uttered sentence with only a very minimal context-dependent component, restricted to choosing between the senses of ambiguous words and supplying values for indexicals, both apparently achieved on the basis of best contextual fit (Grice 1975: 44). However, Grice’s ‘what is said’ has another important property that distinguishes it from truth-conditional sentence meaning. His interest in language in use, in actions performed by speaking, such as asserting something or implicating something, required that, for him, ‘what is said’ by an utterance must be a component of speaker meaning (also referred to as meaning-intended, or m-intended, content), that is, it is overtly endorsed by the speaker.5 Hence what is said and what is implicated together constitute what the speaker meant by her utterance (for discussion, see Neale 1992; Recanati 2004: chapter 1).

3.

From ‘what is said’ to explicature

The two-fold nature of Grice’s ‘what is said’ is reflected in the fact that Griceans sometimes talk of what the speaker says and sometimes of what the utterance or the words themselves say. Unfortunately, this very combination of features – speaker

54

Robyn Carston and Alison Hall

meantness and a minimalist, albeit truth-conditional, semantics – in a single notion leads to a problem. It seems that quite often what a speaker says and means is something more specific than that which results from the minimal identification processes that Grice allowed for in arriving at what is said. Consider (9): (9) It’s snowing. What sentence semantics (linguistic conventions, in Grice’s terms) plus values for indexical elements delivers is, at most, the proposition that, say, IT IS SNOWING AT MIDDAY ON 31 M ARCH 2010. But is this the proposition that the speaker asserts (explicitly communicates)? Consider an utterance of (9) as a response to an enquiry about why the speaker, who is in London, has abandoned plans to go for a run in the local park. It does not seem that the speaker intends to ‘say’/assert/express the proposition that it is snowing tout court (that is, just anywhere): If it is snowing in Oslo, that fact would not make this particular utterance of (9) true. What makes it true is that it is snowing in London, which suggests that a specific location-ofsnowing constituent, despite apparently not being the value of anything in the linguistic form, is part of what is said, rather than being merely implicated. Here is another example, well-known in the literature (see Bach 1994), which illustrates the same problem: (10) Mother (to child crying over a cut on his knee): You’re not going to die. a. Y OU (B ILLY ) ARE NOT GOING TO DIE FROM THAT CUT b. Y OU (B ILLY ) SHOULD STOP MAKING SUCH A FUSS ABOUT IT What the mother means (what she intends to communicate to the child) is given in (a) and (b), where (b) is clearly an implicature and (a) seems to be what she has said (explicitly communicated). But the proposition delivered by conventional linguistic meaning and the assignment of a referent to the pronoun you (hence the Gricean ‘what is said’) is B ILLY IS NOT GOING TO DIE , which seems to entail that Billy is immortal, something that the mother has no intention of conveying. What these examples indicate is that it is just not generally right that what a speaker says (and means) is as close to the semantics of the sentence uttered as Grice’s definition of ‘what is said’ requires (Grice 1975: 44). In other words, it is not the case that a single level of meaning can do double duty as both sentence semantics and speaker-meant primary meaning (explicitly communicated content). Faced with this dilemma, there are essentially three moves that different theorists have made. The first is to maintain the traditional minimal truth-conditional semantics and drop the requirement that it be speaker-meant (communicated); the second is to make the opposite move, that is, maintain a level of meant/communicated content that is distinct from merely implicated content and accept that this may require much more elaboration of the linguistic meaning than Grice allowed; the third is to separate the two Gricean requirements on ‘what is said’ and have both a truth-conditional semantic content (which is not a part of speaker meaning)

Implicature and explicature

55

and a fully pragmatic (speaker-meant) notion of ‘what is said’. For discussion of these options and their takers, see Carston (2004a) and Recanati (2004: chapters 2 and 4). The relevance-theoretic approach that we support takes the second option: we see the explicit-implicit distinction as a communicative distinction – a distinction between two types of communicated assumptions or thoughts6. As our discussion of examples (1), (9) and (10) indicates, the explicitly communicated content may go well beyond Grice’s what is said. Before looking at the kinds of pragmatic processes that can contribute to the content of an explicature we need to say something about our view of linguistic meaning or sentence semantics. The label ‘semantic minimalism’ has been appropriated by philosophers who take a resolutely truth-conditional stance on sentence meaning (Borg 2004; Cappelen and Lepore 2005), but, in fact, our relevance-theoretic view of the linguistic semantics on which pragmatics builds in constructing an explicature is even more minimalist. The view is that, typically, context-free sentence meaning is subpropositional (does not determine a fully truth-conditional content) and provides a schema or set of clues that constrains the occasion-specific pragmatic process of determining the proposition the speaker explicitly communicated. This view marks a strong shift away from the traditional ideal language philosophy of Frege and Russell to a much more cognitive, empirically-informed position. The cognitive turn in linguistics spearheaded by Chomsky (1980, 2000), experimental work on language processing (Swinney 1979; Altmann and Steedman 1988, among hundreds of others), and hypotheses about cognitive architecture emerging from both philosophy of mind (Fodor 1983, 1987) and evolutionary psychology (Cosmides and Tooby 1994; Sperber 1994, 2002) have converged on a view of the mind as consisting, in large part, of modular systems each dedicated to dealing with a particular problem that human minds confront. These include an array of perceptual modules, systems for spatial reasoning, a language faculty (dedicated to decoding linguistic forms), and a cluster of systems dedicated to social cognition, including a mental state interpreter (a system or systems that attribute intentions, beliefs, motives, etc. to others) and, perhaps as a submodule of this, a system that interprets ostensive stimuli, that is, a pragmatics module, which recovers the content of speakers’ communicative intentions. Within this cognitive landscape, sentences (linguistic entities) and acts of ostensive communication (whether verbal or non-verbal) fall in the domain of distinct, albeit interfacing, systems. The language faculty operates in accordance with its own parsing procedures, which are encapsulated from general world knowledge, and delivers a logical form, that is, a non-propositional conceptual representation, generated entirely from the context-free meaning of lexical items and the syntax of the sentence uttered. This is a vital input to the pragmatic processor, which uses it, together with a circumscribed set of contextual assumptions, to construct the explicatures and implicatures of the utterance. For further discussion, see Sperber and Wilson (2002) and Wilson and Sperber (2004).

56

Robyn Carston and Alison Hall

The emphasis in cognitive science is on the nature of the faculties of the human mind/brain that house our linguistic and interpretive abilities and on the mechanisms that do the work of comprehension, a quite distinct set of concerns from those of the anti-psychological ‘ideal language’ philosophers. However, alongside these trends in cognitively-based research, there have been some roughly parallel developments in the philosophy of language. Kaplan’s (1977/89) distinction between the character and content of linguistic expressions parallels the cognitivist distinction between context-invariant, encoded (or standing) linguistic meaning and the content inferred in specific contexts.7 And the ‘contextualist’ semantics stance, that word meanings quite generally (not just indexicals) are highly context-sensitive (Searle 1978; Travis 1985; Recanati 1993, 2004) and so may make different contributions to truth-conditional content on different occasions of their use, is very much at one with the relevance-theoretic view on explicitly communicated utterance content, as will be evident in the next section.

4.

The pragmatics of explicature

On the cognitive pragmatic approach that we endorse, the explicit content of an utterance (its explicature) is taken to be that content which ordinary speaker-hearer intuitions would identify as having been said or asserted by the speaker. In section 5, we discuss the role of intuitions as a criterion for drawing the distinction between explicature and implicature. In this section, we look at a range of examples which, together with (1), (9) and (10) above, indicate the extent and kinds of contributions that pragmatics may make to the explicature of an utterance. Clearly, disambiguation and the assignment of context-specific values to indexicals and demonstratives are essential pragmatic processes in ascertaining the explicitly communicated content. On this, we agree with Grice and virtually everybody who has discussed the proposition explicitly expressed or communicated by a speaker. So, in the following subsections, we focus on pragmatic contributions to explicature which go well beyond the effects of those two obvious processes. 4.1.

Linguistically unarticulated constituents of content

In the following examples, the sentence uttered is given in (a) and a likely explicature of the utterance (dependent on context, of course) is given in (b): (11) a. No-one goes there anymore. b. H ARDLY ANYONE OF ANY WORTH / TASTE (12) a. There’s milk in the fridge. b. T HERE ’ S MILK OF SUFFICIENT IN THE FRIDGE Y

GOES TO LOCATION X ANY MORE

QUANTITY / QUALITY FOR ADDING TO COFFEE

Implicature and explicature

57

(13) a. Max: Would you like to stay for supper. Amy: No thanks, I’ve already eaten. b. A MY HAS ALREADY EATEN SUPPER THIS EVENING Intuitively, the speaker of (11a) does not assert that no-one (that is, not a single person in existence) goes to the place in question, which is what seems to be predicted by a strictly Gricean account; rather, she is directly communicating that very few people of a particular sort (important, fashionable, etc.) go to the place in question these days. This is the content with which other people may express agreement or disagreement, on which the truth or falsity of the utterance would be judged, and which would ground a possible implicature that the speaker doesn’t want to go there. In the case of (12), the presence of just a few stale drips of milk on the bottom shelf of the fridge would be compatible with the conventional/encoded linguistic meaning of the utterance, but in a context in which the participants are intent on having a cup of coffee a more specific enriched conception of ‘milk in the fridge’ is inferred and contributes to the proposition explicitly communicated. These examples, like several of the earlier ones (see (1), (9) and (10) above), suggest that there are explicatures which include constituents of content that do not appear to be the value of any element in the linguistic form of the utterance; for instance, the causal component in (1), the location in (9), the object of eating in (13). Such constituents have been the subject of extensive debate in recent years, concerning their source and the processes that are responsible for their recovery. One way of accounting for these elements is to assume that there is a lot more linguistic structure in the utterances than meets the eye (or ear). The proposed structure takes the form of covert indexicals attached to the relevant (overt) linguistic items, or parameters in the lexical semantics of the linguistic expression, requiring a particular type of value to be assigned. For example, the domain restriction on the quantifier in (11) (and perhaps (12)) would arise as the result of saturating an indexical that is encoded in the noun phrase,8 while the location constituent in (9) would be the value assigned to a location variable encoded with meteorological verbs or to an event variable that accompanies predicates quite generally. According to our cognitive pragmatic stance, it is quite possible that these elements of content are supplied by a process of ‘free’ pragmatic enrichment. Such a process is wholly pragmatic in that, not only is the specific content recovered pragmatically, but the motivation for its recovery is entirely pragmatic, that is, it is driven by the search for an interpretation that is consistent with the presumption of optimal relevance. If this is right, then there are cases of genuinely ‘unarticulated constituents’, that is, constituents of explicature content that are not just unpronounced but are not present at any level of linguistic representation, where, by ‘linguistic’, we are referring to the processes or representations of the language faculty, free from any pragmatic influence.

58

Robyn Carston and Alison Hall

There is a vigorous ongoing discussion about whether it is extensive covert linguistic structure (plus pragmatic saturation) or a process of free pragmatic enrichment that is responsible for the recovery of these non-phonologically-realized constituents of content. Both sides of the debate have developed various criteria and tests for deciding in particular instances whether the content is linguistically mandated or pragmatically motivated. We cannot review this extensive literature here, but it seems clear that the decision will have to be made on a case by case basis. As things stand, it looks as if some cases will be best explained in terms of covert linguistic structure (the domain restriction on quantifiers, for instance), while in other cases there is just no linguistic evidence for a covert linguistic element underpinning the content in question (for instance, the causal component in the explicature of certain cases of and-conjunction, such as example (1)). For strong representatives of the covert linguistic structure view, see Stanley (2000, 2002), King and Stanley (2005) and Martí (2006); for the opposing pragmatic view, see Carston (2000), Recanati (2002), Wilson and Sperber (2002), Hall (2008a, 2008b). A case that has received considerable attention is the location constituent in utterances about the weather (or other atmospheric conditions), such as (9), repeated here: (9) It’s snowing. It seems at first blush that this is a good candidate for the presence in linguistic logical form of a covert location indexical which has to be given a value in context. Its apparent obligatoriness makes it like cases of overt indexicals that have to be pragmatically saturated in context: (14) She put it there. Arguably, for a full understanding of what a speaker of this sentence has explicitly communicated, a hearer needs to be able to supply a value for all the overt indexical elements: she, it and there. If he can’t do this, if, for instance, the proposition that he derives is that a particular individual, Mary, put something on a particular table, he won’t have recovered the proposition expressed; he needs to find the intended referent of it. However, several authors (e.g. Recanati 2002, 2007a; Cappelen and Hawthorne 2007) have argued that the provision of a location with weather verbs is optional. For instance, it doesn’t seem to be required in the following cases: (15) Once, in the middle ages, it rained constantly for 100 years. (16) Why does it rain? It rains because water vapour in the air condenses and […] While it is true that the majority of uses of weather predicates are understood as communicating a location, (15) and (16) demonstrate that this is not mandatory. It is, then, highly unlikely that the logical form of weather predicates contains a covert indexical demanding saturation on every occasion of utterance.9 That the lo-

Implicature and explicature

59

cation is so frequently communicated is because the relevance of an episode of snowing or raining most often turns on its location: the motivation for the hearer to infer it comes not from the linguistic system, but from the fact that the implications he is most likely to draw from weather information concern its effects on the particular place and its inhabitants.10 Furthermore, there is a range of other cases where a communicated element which is part of the explicature (the intuitive truth-conditional content) of the utterance cannot be accounted for by positing some element of hidden linguistic structure. This includes cases of deferred reference or metonymy, such as (17a), where the definite description is taken to refer to the customer who ordered the ham sandwich, and (17b), where the predicate parked out back cannot literally apply to the subject I, and referential uses of definite descriptions quite generally, as in (17c), which expresses a singular (non-descriptive) proposition: a IS OUR NEW EDITOR . There are also a lot of cases where the minimal (unintuitive) proposition, derivable on the basis of the encoded linguistic content and reference assignment alone, is either trivially true or patently false, as in (17d) and (17e): (17) a. b. c. d. e.

The ham sandwich wants his bill. I’m parked out back. The woman standing in the doorway is our new editor. That guy has a brain. She isn’t a human being.

So even if some phenomena, such as quantifier domain restriction, turn out to be better analyzed as involving saturation of hidden indexicals, examples such as these strongly suggest that, in the end, free pragmatic enrichment cannot be entirely avoided and everyone who recognizes this speaker-meant level of explicit content will need an account of it. Nevertheless, those of us who advocate free pragmatic enrichment still have some work to do to shore up our position against criticism from the hidden indexicalists. The key objection, made most explicitly by Stanley (2002), is that the process of free enrichment is not sufficiently constrained, allowing for enrichments that clearly do not occur. Two examples of the kind of overgeneration he presents are given in (18) and (19). According to the pragmaticist, an utterance of the sentence in (18a) could communicate the explicature in (18b) in an appropriate context, the quantifier domain having been supplied by a process of free pragmatic enrichment. Stanley’s question, then, is what prevents that same pragmatic process from supplying the constituent [OR D UTCHMAN ] so that the utterance is predicted (wrongly) to communicate the proposition in (18c) in a context in which it would be relevant. Similarly, free enrichment should enable, he says, an utterance of the sentence in (19a), in a context in which it is common ground that John likes his mother, to communicate the explicature in (19b), which, however, it clearly does not:

60

Robyn Carston and Alison Hall

(18) a. Every Frenchman is seated. b. E VERY F RENCHMAN IN THE CLASS IS SEATED c. E VERY F RENCHMAN OR D UTCHMAN IN THE CLASS (19) a. John likes Sally. b. J OHN I LIKES S ALLY

IS SEATED

AND HIS I MOTHER

This overgeneration objection is potentially serious and it has been addressed in some detail by Hall (2008a, 2008b, 2009), who sets out to show that a relevancetheoretic account of free pragmatic enrichment would not make the alleged predictions. We cannot review the arguments here (though the distinction we make in section 5.2. between local and global processes is relevant to the issue), but what emerges clearly from Hall’s discussion is that when proper attention is given to the nature of the principles and processes at work in pragmatics, ‘free’ enrichment is in fact quite tightly constrained. 4.2.

Pragmatic adjustment of word meaning

The controversy about the existence of a pragmatic process of free enrichment has focussed mainly on unarticulated constituents – cases where there is a component of the explicature which is not the value of any element of the linguistic form. However, it has been claimed that there is another pragmatic process contributing to explicature that is not linguistically mandated and therefore can also be considered a kind of free enrichment. This is the process of adjusting or ‘modulating’ lexically encoded meanings, so that the concept understood as communicated by the use of a word differs from the concept encoded – it may be narrower or looser, or some combination of the two. Consider some examples: (20) To buy a house in London you need money. (21) Max: Amy:

Would you like to go to a club tonight? I’m tired. Let’s go to a movie.

(22) I’ll put the empty bottles in the garbage. (23) The water is boiling. Grasping the proposition explicitly communicated by these examples is very likely to involve an optional process of meaning modulation (of the concept encoded by the underlined word in each case). In most contexts, the proposition that buying something requires (some/any amount of) money will be trivial and uninformative, so the lexically encoded concept MONEY is likely to be narrowed to a concept, represented as MONEY *, that denotes just those quantities that would count as sufficiently large amounts of money in the context of house-buying. Similar comments apply to Amy’s response in (21); the concept she communicates with the word tired

Implicature and explicature

61

is more specific than the very general encoded concept TIRED , which includes states of exhaustion which would preclude doing anything other than lying down to sleep. The examples in (22) and (23) appear to require a different sort of adjustment: imagine (22) uttered in a context where people are clearing up the morning after a party, so that many of the bottles to be put in the garbage may contain the dregs of wine or beer and various other items of debris, so are not strictly and literally empty. Here the communicated concept EMPTY * has a broader denotation than the encoded concept EMPTY. A similar point can be made about (23), which might be used to communicate any of a range of broader concepts than the encoded one: it might be a rough approximation (the water is near enough to boiling for the difference not to matter for current purposes) or a hyperbolic use as when someone utters it upon stepping into bath water that is a bit too hot for an entirely comfortable immersion. The general consensus is that these meaning modulations affect the explicature, that is, the proposition that the speaker communicates directly and that grounds any implicatures of the utterance. If this were not the case – if these pragmatically derived concepts were themselves registered only at the level of implicature – then we would expect that many of them, particularly the examples involving concept broadening, would be judged false. Yet, intuitively, this is not the case: someone who, in the context described just above, includes in the denotation of empty bottles, some that contain a few millilitres of flat beer, would not be judged to have spoken falsely and only implicated that the bottle is near enough to being empty that it can be thrown out. Recently, it has been argued by both relevance theorists and some philosophers of language that many instances of metaphorically used words can be explained in the same way, that is, as cases whose comprehension is achieved by an adjustment of the literal encoded concept, resulting in a radical broadening of denotation with often some degree of narrowing as well (Bezuidenhout 2001; Carston 2002; Wilson and Carston 2006, 2007). Attempts to give an account of these lexical modulation processes are recent and have raised many issues that remain to be investigated in detail (see the discussion in Carston 2010). One interesting question is the nature of the input to the process – that is, the nature of word meanings themselves. In the discussion so far, we have gone along with the view that most open-class words (nouns, verbs, adjectives) encode concepts, but the idea that words might not encode full-fledged concepts but something more like constraints, instructions for building concepts, or rules for use, has been gaining popularity recently (for discussion, see Carston 2002: chapter 5.4.; Pietroski 2005; Recanati 2004: chapter 9; Bosch 2007). What motivates this view is chiefly the observation that understanding the intended meaning of a word quite typically requires some degree of modulation of its encoded meaning. In many cases, words appear to encode something too abstract to serve as a constituent of a thought, hence of a communicated proposition (see the discussion in Searle (1983) of the verbs open and cut, and in Carston (2002) of adjectives like happy). Given the ubiquity of modulation, an appealing view is that

62

Robyn Carston and Alison Hall

word meanings are far more schematic and abstract than full-fledged concepts, and, therefore, on every occasion of understanding a word a pragmatic process of concept retrieval or construction is required. Of course, if this view of the nature of word meaning is correct, then modulation would not be a ‘free’ pragmatic process because the underspecified lexical encoding would make the process obligatory. Another development enabled by the recognition of lexical pragmatic adjustment as an important contributor to explicature is the reanalysis of several of the cases previously taken to be instances of unarticulated constituents. For example, cases such as (17d), which might previously have been analysed as expressing T HAT GUY HAS A BRILLIANT BRAIN , or (20) as Y OU NEED A LOT OF MONEY, have been recast as cases of modulation of the lexical concepts BRAIN and MONEY. This raises the question of how many more of them it is possible to reanalyze in this way, and thus, whether there will turn out to be any truly ‘free’ (hence optional) pragmatic processes at the level of explicature at all. If all cases previously thought of as unarticulated constituents could be reanalyzed as cases of lexical concept adjustment, this would enable an isomorphic structural relationship between linguistic logical form and explicature to be maintained. Such an outcome should find favour with the semanticists mentioned above (Stanley, Martí) who are concerned about the violation of semantic compositionality that the existence of unarticulated constituents would entail (see note 10). At the same time, though, such a move would seem to result in pragmatic inference being even more pervasive in determining explicit content, since it makes what Carston (2010) calls ‘pragmatic susceptibility’ (to distinguish it from the linguistically indicated context-sensitivity of indexicals) a property of virtually every expression in the language. However, while this is in many ways an appealing way to go, there are some cases for which it does not look plausible. Consider again the case of weather predicates, discussed earlier, where a particular utterance of (24a) has the explicature in (24b): (24) a. It’s snowing. b. I T ’ S SNOWING

IN

L ONDON

AT TIME

t

It is at least very counter-intuitive (and no-one has tried to argue for it) that SNOWING - IN - LONDON is a single concept that results from pragmatic modulation (narrowing) of the lexical concept SNOWING . Similar remarks apply to the explicature of certain and-conjunction utterances, such as the cause-consequence interpretation in (1). So it seems that some cases are likely to remain best analyzed as instances of unarticulated constituents. 4.3.

Reconstrual of some cases of Gricean implicature

What should be clear from the discussion in the two preceding subsections is that explicatures can contain a great deal more pragmatically inferred material than Grice allowed for in his ‘what is said’; as well as processes of disambiguation and

Implicature and explicature

63

indexical saturation, there is a range of pragmatic processes that enrich or otherwise modify linguistically encoded content in the course of recovering the proposition understood to have been explicitly communicated.11 As we will discuss in more detail in the next section, this view of explicit utterance content has implications for some of Grice’s key cases of conversational implicature. Consider the following: (25) a. I’ve read some of Deirdre Wilson’s papers. b. I HAVEN ’ T READ ALL OF D EIRDRE W ILSON ’ S

PAPERS

(26) a. Hannah reported Joe for misconduct and he was fired. b. J OE WAS FIRED AS A RESULT OF H ANNAH REPORTING HIM

FOR MISCONDUCT

In many contexts, an utterance of (25a) will communicate the proposition in (25b), and, in many contexts, an utterance of (26a) will communicate (26b). Neither of these appears to be the result of a linguistically mandated process of variable saturation and Grice treated them as cases of conversational implicature. Note that both fall into his category of ‘generalized’ implicatures: although they are not inevitable and can be cancelled without contradiction (e.g. I’ve read some of Deirdre Wilson’s papers – in fact, I’ve read all of them is a consistent utterance), they would occur across the majority of contexts of utterance. This is reflected in the fact that we can present the sentences uttered in (25a) and (26a) without indicating any previous discourse or other contextual specifics and assume readers will agree that the propositions in (25b) and (26b) are likely to be inferred in each case. This distinguishes them from particularized implicatures, as in (1), (2) and others to be seen shortly. Once it is recognized that there is far more linguistic underdeterminacy of content than was envisaged by Grice, the status of these elements of meaning becomes much less clear, since it is quite possible that they are in fact constituents of explicature. The price we pay, then, for such a pragmatic conception of explicit utterance content is that it is no longer always a straightforward matter to distinguish which elements of speaker meaning are implicatures and which belong to explicature. In the next section, we consider some of the various tests and criteria that have been proposed for drawing the explicit-implicit distinction.

5.

Explicature/implicature – How to draw the distinction

5.1.

Intuitions and tests

Explicatures and implicatures are characterized as follows within relevance theory: An assumption communicated by an utterance U is explicit (hence is an explicature) if and only if it is a development of a logical form12 encoded by U, where explicitness is a matter of degree: the greater the contribution of encoded meaning the more explicit the

64

Robyn Carston and Alison Hall explicature is and the greater the contribution of pragmatically inferred content the less explicit it is. Any assumption communicated, but not explicitly so, is implicitly communicated: it is an implicature. (Sperber and Wilson 1986/1995: 182)

These definitions are fine as far as they go, but they do not provide us with a clear criterion for deciding, in any particular case of pragmatically derived meaning, whether it constitutes a distinct implicitly communicated proposition (an implicature) or a pragmatic contribution to explicature. For instance, how can we tell whether the ‘not all’ meaning in (25) above and the inferred cause-consequence relation in (26) are developments of the logical form of the sentence uttered or are distinct implicated propositions? Like relevance theorists, Recanati (1989, 1993, 2004) is a long-term advocate of a pragmatically enriched level of communicated content, which he often refers to as the ‘intuitive (or enriched) what is said’ or the ‘intuitive truth-conditional content’ of the utterance, distinguishing it from Grice’s minimalist semantic notion. He has, in fact, elevated the status of native speaker intuitions about this to the level of a criterion for distinguishing explicature (enriched what is said) from implicature: “Availability Principle: In deciding whether a pragmatically determined aspect of utterance meaning is part of what is said, that is, in making a decision concerning what is said, we should always try to preserve our pre-theoretic intuitions on the matter” (Recanati 1989: 309–10; 1993: 248). The principle yields clear results for many cases of particularized implicatures, such as Amy’s indirect answer given in (2) above, for which there is a consensus of intuitions that it falls outside what is said. But since there are no conflicting predictions among different theories for cases like this, the criterion isn’t really doing any work here. Consider a slightly more contentious case: (27) a. b. c. d.

Robert broke a finger last night. R OBERT BROKE A FINGER , EITHER HIS OWN OR SOMEONE ELSE ’ S , ON NIGHT n. R OBERT BROKE HIS OWN FINGER ON NIGHT n. R OBERT CAN ’ T PLAY IN THE MATCH TODAY. (Carston 2002: 167)

On a Gricean account, what is said by an utterance of (27a) would be as given (roughly) in (27b). But (27b) is not available to the conscious awareness of the speaker and hearer, so the Availability Principle denies that it is what is said. Intuitively, the inference that the finger broken was Robert’s – which Grice treated as a generalized conversational implicature – is part of what is ‘said’, or explicitly communicated, and is what provides the basis for the hearer’s inference to the (particularized) implicature in (27d). In this case, there is strong consensus that (27c), rather than (27b), is what was said or explicitly communicated. But there are many cases where intuitions are less consistent:

Implicature and explicature

(28) a. b. c. d.

65

Some lecturers are insecure. The ham sandwich left without paying. When Joe saw Hanna approaching he crossed the street. The man over there drinking champagne is a famous philosopher.

Arguably, a pragmatic component of content contributes to the explicature in each of these cases: ‘NOT ALL LECTURERS ’ in (28a), ‘THE PERSON WHO ORDERED THE HAM SANDWICH ’ in (28b), a cause-consequence relation in (28c), an individual concept denoting a particular person (say, John Perry) in (28d). But, in all these cases, a more minimal content, closer to the linguistically encoded meaning, is for many people ‘available’ to intuitions. Although, as most theorists agree, converging intuitions are an important source of data, in many cases we do not have consistent intuitions that are directly about ‘what is said’ or explicit content, so we need to get at them more indirectly. We said above that the explicature is taken to be the ‘intuitive truth-conditional content’ of the utterance – that is, that content on the basis of which ordinary speaker-hearers would judge the utterance true or false. The rationale for this seems to be that the speaker has more responsibility for her explicatures than her implicatures: an explicature is based on decoded meaning, which is algorithmically derived, whereas implicatures are entirely inferred, so are more the hearer’s responsibility. If, as seems to be the case, our judgments about the utterance (true/ false; agree/disagree) and our behavioural responses to it are focused on this level of content, as opposed to the more minimal what is said, then there is good reason to believe that the explicature, rather than the Gricean what is said, is a psychologically real level of representation. In accordance with this, Recanati (2004: 14) assumes that “whoever fully understands a declarative utterance knows which state of affairs would possibly constitute a truth-maker for that utterance, that is, knows in what sort of circumstance it would be true”. He therefore suggests that the way to elicit intuitions concerning the truth-conditional content is “to present subjects with scenarios describing situations, or, even better, with – possibly animated – pictures of situations, and to ask them to evaluate the target utterance as true or false with respect to the situations in question” (2004: 15). This method of eliciting intuitions has been employed by Ira Noveck and colleagues in a series of experiments to test people’s understanding of scalar terms such as some and or (for a summary of findings, see Noveck 2004; Noveck and Sperber 2007). As mentioned above, these terms often give rise to scalar pragmatic inferences: from some of the F, hearers often infer NOT ALL OF THE F; from P or Q, hearers often infer NOT BOTH P & Q. In an experiment requiring participants to evaluate the truth/falsity of scalar utterances in particular scenarios, one of the test utterances was Some of the turtles are in the boxes and it was presented together with one of several possible accompanying scenarios. For instance, in one case some of the turtles were in the boxes and some were lying outside the boxes, and in

66

Robyn Carston and Alison Hall

another one, the crucial case, all of the turtles were in the boxes. Presented with an utterance-scenario pair, adult subjects were asked to judge the utterance as ‘true’ or ‘false’ in the scenario, where the response ‘false’ in the test case (where all the turtles were in the boxes) should, according to Recanati’s principle, indicate that the scalar inference (NOT ALL OF THE TURTLES ) is contributing to truth-conditional content (‘what is said’). The problem for the Availability Principle is that adult responses were very far from univocal. In this particular experiment, 53 % gave the response ‘true’, 47 % ‘false’ (Noveck 2004: 308) and this result is consistent with findings across a wide number of other experiments carried out on such scalar terms. In short, the pre-theoretic intuitions of ordinary speaker-hearers concerning the truth-conditional content (hence ‘what is said’) of utterances involving scalar terms (presented with scenarios which strongly encourage the pragmatic inference) are highly divergent. The same inconsistency arises for a number of other contentious cases, including metonymies (such as (28b) above), metaphorical uses (e.g. My lawyer is a shark), and some of the quantifier cases (e.g. when the TV guide lists continuous gardening and antiques programmes and the speaker reports There’s nothing on TV tonight, a lot of people judge the utterance as (strictly) false). Another well-known example is the case of misuses of definite descriptions, as discussed by Donnellan (1966). Consider again (28d) above and suppose that as a matter of fact the man in question is not drinking champagne (say, he has sparkling water in a champagne flute), but he is indeed a famous philosopher. Is the utterance of (28d) true or false? Unfortunately, the full range of possible responses has been attested (‘true’, ‘false’, ‘neither’, and, remarkably, ‘both’). Should we, then, abandon the idea that there is a propositional content that can be considered ‘the’ intuitive truth-conditional content of the utterance (which is stable across different interpreters)? The difficulties with this method of eliciting truth conditions do not force us to that conclusion, since such tests can be argued to be not eliciting intuitions about the right thing. What we want is the proposition that speaker-hearers intuitively take to be explicitly expressed, but what Recanati’s method taps seems to be something slightly different because it introduces an element of reflection which makes salient various possible truth-conditional contents, including some of a more strictly literal nature than the intended explicature. A better guide than such judgments of truth and falsity may be hearers’ responses to utterances in online comprehension. Consider B’s responses in (29) – (32): (29) A: It’ll take time for that cut to heal. B: No it won’t; I’m having the stitches out tomorrow. (30) A: John double-parked and Mary ran into him. B: No, Mary ran into him and that’s why he stopped where he did.

Implicature and explicature

67

(31) A: Bob’s a bit of a bulldozer. B: God, yes, he completely ignored the majority opinion on the council. (32) A: No one goes to that club any more. B: I know: there were only first-year students there last Saturday. Clearly, what B is agreeing or disagreeing with here is not any ‘minimal’ proposition of A’s utterances corresponding to the Gricean ‘what is said’. Rather, in each case the proposition responded to includes some element that is pragmatically inferred: A FAIRLY LONG TIME in (29); a cause-consequence connection in (30); the metaphorical meaning in (31), and the domain, something like NO ONE IN OUR GROUP OF FRIENDS , in (32).13 It seems clear that we would not want to abandon the Availability Principle (we do want to preserve pre-theoretic intuitions regarding what a speaker says/implicates as much as we can). The problem is that applying the principle (eliciting the desired intuitions) in contentious cases is quite difficult. One way of sharpening intuitions involves embedding the utterance/sentence at issue in the scope of an operator such as negation, disjunction, the conditional, or the comparative. Consider the following cases of and-conjunction: (33) a. The old king has died of a heart attack and a republic has been declared. b. A republic has been declared and the old king has died of a heart attack. Recall that Grice held that the meaning of and is simply the truth-functional logical conjunction ‘&’ and, given his minimalist view of what is said (which is the semantic content of the utterance), it follows that, for him, what is said by utterances of (33a) and (33b) is the same (they have identical truth-conditional content). Any difference in what they communicate, specifically in their temporal sequence and cause-consequence relations, is a matter of conversational implicatures, based on Grice’s maxims of orderliness and of relevance. As Cohen (1971) first pointed out, a problem emerges when these conjunctions are embedded in the antecedent of a conditional, as in (34): (34) a. If the old king has died of a heart attack and a republic has been declared then Tom will happy. b. If a republic has been declared and the old king has died of a heart attack, then Tom will be unhappy. The Gricean account predicts that the antecedents of the two conditionals must be truth-conditionally equivalent and, assuming the truth of this single antecedent proposition, it must be that one or other of the conditionals is false (Tom can’t be both happy and unhappy). However, the intuitive consensus is that these could both be true and that this is because the antecedents differ in the temporal-consequential relations they express about the two events. It follows that these non-truth-func-

68

Robyn Carston and Alison Hall

tional aspects of the conjunctions contribute to the explicatures of the conjunctive utterances in (33). This ‘scope test’, as it became known, can be used to demonstrate that other pragmatically inferred constituents are components of the explicature rather than distinct implicated propositions: (35) a. If it rains tomorrow, we’ll stay at home. b. Sam doesn’t care if some of his students fail, but if they all do, he’ll be worried. c. We don’t have any milk; there are just a few rancid drops on the tray. In (35a), it’s clear that the basis for staying home is rain falling in the specific location of the speaker; in (35b), Sam won’t be upset if just some (not all) of his students fail;14 in (35c), where the test employs a negation operator rather than a conditional, we understand that the milk that is lacking is of the sort that would be usable for a specific purpose (such as drinking with coffee). This embedding procedure played an important role in the early days of relevance theory as a source of evidence indicating that certain pragmatic elements of meaning, which Grice had taken to be implicatures, can contribute to explicature instead (see, for instance, Carston 1988). So it was a useful tool for establishing the extensive role of pragmatic inference in determining explicature, in particular, the role of processes of free pragmatic enrichment.15 However, it cannot predict or decide in advance whether some such element does contribute to explicature or not on any particular occasion of utterance. Since the provision of these elements is optional, one has to look at the utterance token – and particularly the occasion-specific context in which it was tokened – in order to decide whether there is enrichment of content or not. As the embedding procedure essentially tests utterance types, it does not necessarily allow one to draw the explicit-implicit distinction for any particular occurrence.16 Given the existence of free (optional) pragmatic enrichment, the line between explicature and implicature can vary from occasion to occasion of utterance of the same sentence type, and will be determined by how the utterance is processed. 5.2.

Relevance-based derivation of explicatures and implicatures: local versus global processes

Let us consider now how the pragmatic system, as envisaged on the relevance-theoretic account, derives explicatures and implicatures. The claim is that they are derived in parallel by a unitary, online inferential process, which takes the decoded linguistic meaning and accessible contextual assumptions as evidence for the interpretation, which must ultimately be both inferentially sound and consistent with the presumption of optimal relevance. As most theorists (across different approaches) would agree, a hearer does not first decode the entire utterance, then

Implicature and explicature

69

saturate, disambiguate, enrich, and modulate the decoded meaning in order to arrive at the explicature (what is said), and only then use the explicature, together with contextual assumptions, to form hypotheses about implicatures. Instead, the explicatures, implicatures, and contextual assumptions are mutually adjusted in parallel until they form an inferentially sound relation, with premises (explicature, contextual assumptions) warranting conclusions (implicatures). It follows that a hypothesis about an implicature can both precede and shape a hypothesis about an explicature. Here is a brief example involving the adjustment of explicit content in response to hypothesized implicatures and where the outcome is a narrowing of a lexically encoded meaning: (36) A (to B): Be careful. The path is uneven. Given that the first part of A’s utterance warns B to take care, B is very likely to expect the second part of the utterance to achieve relevance by explaining or elaborating on why, or in which way, he should take care. Now, virtually every path is, strictly speaking, uneven to some degree or other (i.e. not perfectly plane), but given that B is looking for a particular kind of implication, he will enrich the very general encoded concept UNEVEN so that the proposition explicitly communicated provides appropriate inferential warrant for such implications of the utterance as: B might trip over, B should take small steps, B should keep his eye on the path, etc. The result is a concept, which we can label UNEVEN *, whose denotation is a proper subset of the denotation of the lexical concept UNEVEN . Provided the interpretation as a whole satisfies the hearer’s context-specific expectations of relevance (licensed by the general presumption of optimal relevance carried by all utterances), it is accepted as the intended meaning. For much more detailed exemplifications of the relevance-theoretic account of lexical meaning modulation, see Wilson and Sperber (2002), and for demonstrations of how the mutual adjustment mechanism brings about other pragmatic developments, including disambiguation and saturation, in deriving explicatures, see Wilson and Sperber (2004). So are we any closer to characterising the notion of a pragmatic ‘development’ of logical form, which plays such a key role in the definition of explicature as distinct from implicature? We believe there is an interesting difference in both the domain and the output of the processes involved in deriving the two kinds of speaker meaning: whereas implicatures are derived by global inference from fully propositional premises, what is common to the various pragmatic processes that develop the logical form into an explicature is that they are local. What we mean by a pragmatic process being local is that it applies to a subpropositional constituent, or, more specifically, a subpart of the linguistically encoded logical form and it is the output of this pragmatic process that goes into the meaning composition process.17 The most obvious processes exemplifying the localness of ‘development’ are disambiguation and reference assignment: in the online comprehension of (37), for in-

70

Robyn Carston and Alison Hall

stance, the indexical is assigned a referent and it is this referent that is composed together with the semantic/pragmatic values of the other expressions to give a fully propositional content: (37) He is swimming. This is widely agreed to be a far more plausible story about processing than that the hearer first composes the standing linguistic meanings of the lexical items, deriving something like ‘some male is swimming’, and then uses that as the starting point for explicature development. The lexical modulations (narrowings and broadenings) discussed in section 4.2 are also, clearly, local in this sense: (38) a. France is hexagonal. b. This steak is raw. Here the encoded concepts HEXAGONAL and RAW are each replaced by a pragmatically inferred concept (HEXAGONAL * and RAW *, respectively), which shares relevant implications with the encoded concept, and it is these latter concepts that enter into the composition process. A similar point can be made about (possible) cases of unarticulated constituents: (39) a. b. c. d.

The ham sandwich [ORDERER ] wants his bill. It will take [A LONG ] time. Every boy [IN THE CLASS ] was there. I’ve got nothing [SUITABLE FOR A WEDDING ] to wear.

When appropriately contextualized, the comprehension of a metonymy like (39a), does not involve first computing the absurd ‘literal’ meaning on which a culinary item wants the bill, then, once the absurdity is recognized, inferring that the speaker was referring to the person who ordered it. Instead, the deferred meaning is computed at the local level on a first processing pass and it is what goes into the composition process (for discussion, see Recanati 1993, 2004; Sag 1981; Nunberg 1995). Similarly, the enrichment of (39b) involves a local modification of the decoded concept TIME , rather than first recovering the trivially true proposition (that the activity in question will take place over a period of time) and only then inferring the relevant communicated proposition. Being trivially true, the unenriched proposition would, of course, have no relevant implications of its own, so the enrichment is necessary if the explicit content is to play any role in further inference. Likewise, domain restriction, as in utterances of (39c) and (39d), can also be seen to be local; for instance, Recanati (1993: 262–263) treats (39c) as enrichment of the predicate BOY to BOY IN THE CLASS . Inferring implicatures, in contrast, is not a case of just modifying a subpart of the linguistically encoded meaning; rather, it is a global process in that it operates on fully propositional forms.18,19

Implicature and explicature

71

Assuming it is correct that the pragmatically inferred elements of explicatures are derived locally, the question is: how does this line up with the requirement that the explicature be the ‘intuitive truth-conditional content’ of the utterance? As we have seen, ordinary speaker-hearers’ intuitions are sensitive not only to the truthconditional content of one particular proposition communicated by the utterance, but also to those of other related propositions and people vary with regard to how strict (how literal) they are in judging utterances as true/false. The distinction based on the locality criterion will often coincide with one based on judgments about the truth-conditional content of the utterance since the explicature is generally going to be highly salient to our judgments and, given that a substantial element of its content is linguistically encoded meaning, it is content that the speaker can be held largely responsible for. However, when intuitions diverge, it is the derivational distinction between local and global pragmatic inference that is the ultimate arbiter about what constitutes an explicature and what an implicature. (See Hall 2008b for further discussion of the local/global processing distinction.)

6.

Conclusion: implicature, explicature and semantics

It seems that, in the move from Grice’s philosophically based saying/implicating distinction to the cognitive communicative distinction between explicature and implicature in relevance theory, the class of implicatures has become much reduced. Most, if not all, cases of generalized implicatures have turned out to be local adjustments to subparts of the decoded logical form rather than propositions derived by global inference. Thus their treatment as constituents of the proposition that is explicitly communicated (explicature) is supported both intuitively and theoretically. However, the shift from implicature to explicature is not confined to the generalized cases. Several kinds of non-literal language use, including hyperbole, metaphor and metonymy, whose communicated (speaker-meant) content was analysed as a matter of particularized implicature by Grice (and many after him), have been reanalyzed as cases of local adjustments of encoded meaning, at the lexical or phrasal level. Thus they also contribute elements of content to the proposition most directly communicated by the speaker, that is, the explicature (see Carston 2002; Wilson and Sperber 2002; Wilson and Carston 2007). So what is left of implicature within this cognitive communicative picture?20 Implicatures are contextual implications that are speaker-meant (i.e. which fall within the speaker’s communicative intention) and, as discussed in the previous section, they are derived globally. Consider the following exchange: (40) A: Who’s eaten my last chocolate egg? B: Not me. Mary mentioned needing a chocolate fix.

72

Robyn Carston and Alison Hall

B’s utterance contextually implies that Mary is probably the person who ate A’s chocolate egg and this implication is, in fact, strongly implicated by B since it would be very difficult for A to satisfy his expectations regarding the relevance of B’s utterance without inferring this conclusion. This implicature is inferentially warranted by the explicature of the utterance together with highly accessible contextual assumptions including: people who are in need of a chocolate fix are likely to eat whatever chocolate they can lay their hands on; A’s chocolate egg was within Mary’s reach.21 This implicature is clearly entirely dependent on the specific context, in particular the preceding question, and this is typical of implicatures on the relevance-theoretic account; there are no default implicatures (compare Levinson 2000) nor implicatures that are highly generalized across contexts. In this respect, the class of implicatures is a more unified communicative phenomenon than it was in Grice’s account (and in current neo-Gricean frameworks). According to the relevance-theoretic account, implicatures are more or less strongly communicated, depending on the extent to which they can be taken to have been specifically intended by the speaker. The example just discussed is a case of a strongly communicated implicature since establishing the relevance of the utterance depends heavily on the recovery of that particular implication. In other cases, the speaker may have in mind a wider range of possible implications, various subsets of which would contribute adequately and equally to the relevance of her utterance, and none of which she intends more specifically than any others. A simple example of this is the following: (41) A has been devoting her time and energy for many weeks to helping B with his dissertation. Finally, she says: It’s up to you now. Among the possible implications of her utterance are: I have given you enough help with your dissertation; I cannot give you any more help; you need to take responsibility for your own work; you should not continue to ask me for advice; you have the ability to complete the dissertation; and so on. It is not clear that she intends any particular one (or two) of these and provided B derives several of them as implicatures of the utterance he will have grasped the relevance of A’s utterance. There are, of course, cases of even more weakly communicated implicatures as in some highly creative and evocative uses of language for which hearers/readers may derive implications that are so weakly implicated that it is not clear whether they are intended by the speaker/writer at all. For further discussion of degrees of strength of implicature, see Sperber and Wilson (1986/1995: 197–202; 2008). Another respect in which the way implicatures are construed in relevance theory differs from the Gricean conception is that entailments can, on occasion, be implicated. Consider the following exchange:

Implicature and explicature

73

(42) X: Does John like cats? Y: He doesn’t like any animals. a. C ATS ARE ANIMALS . b. J OHN DOESN ’ T LIKE CATS . c. D OGS ARE ANIMALS . d. J OHN DOESN ’ T LIKE DOGS . According to the relevance-theoretic account, all of (a)–(d) are (potential) implicatures of Y’s utterance in the context of X’s question, with (a) and (c) as implicated premises and (b) and (d) as implicated conclusions. The (a)/(b) pair are strongly communicated in that X must recover them in order to understand the utterance. The (c)/(d) pair are communicated less strongly since assumptions with this exact content need not be derived, though some assumptions of this sort are likely to be recovered, given Y’s general and indirect response to X. Both (42b) and (42d) are entailed by Y’s utterance of He doesn’t like any animals and, on this basis, Griceans tend to assume that they cannot be implicatures and instead treat them as part of what is said (explicitly communicated). However, according to the relevance-theoretic view, since they are communicated by the utterance, they are either explicatures or implicatures, and they cannot be explicatures because the utterance does not encode a logical form from which they could be developed. This prediction is backed up by the fact that the way the example works is essentially parallel to the following one, where there is no dispute about (a)–(d) being (potential) implicatures, rather than explicatures, of Y’s utterance: (43) X: Have you read Susan’s book? Y: I don’t read crime fiction. a. S USAN ’ S BOOK IS CRIME FICTION . b. Y HASN ’ T READ S USAN ’ S BOOK . c. A GATHA C HRISTIE ’ S BOOKS ARE CRIME FICTION d. Y HASN ’ T READ A GATHA C HRISTIE ’ S BOOKS . The only difference between the two cases is that there happens to be an entailment relation between the proposition expressed and the (alleged) implicatures in (42b) and (42d), but no such entailment relation in (43). The derivation process in both cases is the same: in order to establish the relevance of Y’s utterance as an answer to his question, X has to access the premise in (a) in each case, from which the conclusion in (b), which answers his question, follows. There is not even, necessarily, any difference in the accessibility of the premises in the two cases, since X may or may not already have them stored as part of his general knowledge. If he does, he can retrieve them ready-made; if he does not, he has to construct them in accordance with a standard inferential procedure based on the speaker’s indirect response to his question. In the (c)/(d) pairs in each case, there is only one possible processing route: in the first example, the hearer looks into his encyclopaedic entry for ani-

74

Robyn Carston and Alison Hall

mals and pulls out his assumption that dogs are animals, from which, given the explicature, the conclusion in (42d) follows; in the second case, he consults his knowledge of crime fiction and retrieves the assumption about Agatha Christie’s books, from which, given the explicature, the conclusion in (43d) follows. We would extend this analysis to the following examples, although they are perhaps a little more controversial: (44) A: Have you invited any men to the dinner? B: I’ve invited my father. Implicature: B HAS INVITED AT LEAST ONE MAN . (45) A: I can’t face lentil bake again tonight; I’m desperate for some meat. B: Good. I’ve just bought some pork. Implicature: B HAS JUST BOUGHT SOME MEAT . On the relevance-theoretic account, word meanings are atomic (i.e. unstructured) rather than decompositional: father encodes FATHER, pork encodes PORK (for strong supporting arguments, see Fodor 1998; Sperber and Wilson 1986/1995: 90–93). Thus, the derivation of the communicated assumptions that B has invited a man, in (44), and that B has bought some meat, in (45), is an entirely inferential process, in fact a straightforward logical inference, so the mechanism involved is essentially the same as that for any implicated conclusion.22 For Grice, entailments and implicatures were mutually exclusive, a view which remains widespread among neo-Griceans and which is a natural consequence of an account in which a notion of ‘what is said’ is doing double duty as both semantics and explicitly communicated proposition (as discussed in section 2). In our view, the concept of entailment and the concept of implicature belong to different explanatory levels, in fact different sorts of theory, the one a static description of our native speaker knowledge of linguistic meaning, the other an account of the cognitive processes and representations involved in understanding utterances, so there is no reason at all why one and the same element of meaning should not fall into both categories. Finally, let us return to the issue of whether there is a minimal proposition semantically expressed by a sentence. Unlike Grice, current semantic minimalists acknowledge that ‘what is said’ by a speaker (explicature) is very much a matter of pragmatics and thus they make a clear distinction between it and their favoured notion of semantic content (Borg 2004: 127–131; Cappelen and Lepore 2005: 204). For Cappelen and Lepore, the minimal semantic content of an utterance is that proposition that results from the conventional linguistic meaning of the sentence (its lexical items and syntax) and the fixing of the referents of any overt indexicals/ demonstratives and the disambiguation of any ambiguous terms. They consider this notion of semantic content to be of central importance in an account of human verbal communication for two (related) reasons: it provides a reliable fallback con-

Implicature and explicature

75

tent when something goes awry in the speaker-hearer communicative enterprise and it constitutes a content that can be shared across divergent contexts, ensuring a reasonable degree of stability in the propositions we attribute to each other, agree or disagree with, debate about, and act upon, across time and place. Cappelen and Lepore’s reasoning is that the amount of pragmatic work (non-demonstrative inference) that goes into explicature undermines the possibility of its being a shareable content as it is very unlikely that all the necessary details of the original context and speaker’s intentions are carried across subsequent contexts enabling that content to be reliably inferred time after time. So explicature cannot serve as shared content, but nor can the kind of encoded linguistic meaning envisaged in relevance theory, since this is generally a subpropositional logical form while the kind of content looked for is propositional (truth-evaluable). Cappelen and Lepore’s minimal propositions appear to fulfil the requirements of ‘shared content’, being truth-evaluable, yet determined largely semantically, requiring relatively little pragmatic work (just disambiguation and saturation of overt indexicals).23 However, several authors have argued (e.g. Carston 2008b; Wedgwood 2007) that Cappelen and Lepore’s minimal proposition is not really well suited to play the role of shared content and this is for two quite distinct reasons. First, it just is not reliably shared across contexts because it must incorporate the results of variable saturation, a process which frequently requires substantial knowledge of the original context/intentions: consider the specific information required for working out the referents of this and that on different occasions of use, or of deciding whether here means IN E NGLAND , IN L ONDON , IN THIS BUILDING , IN THIS ROOM , or something even more specific. All that a contextually disoriented addressee (or reporter) can turn to for some degree of semantic anchoring is the linguistically encoded meaning, which is typically nonpropositional. Second, even if the minimal proposition were common across all contexts in which the original utterance is reported and its content evaluated, it would not provide the right kind of shared content for most of the purposes to which we would want to put such a notion. As Cappelen and Lepore seem to concede (2005: 13–14, 58; 2007: 210), most instances of minimal semantic content are not the communicated content of speech acts, from which it follows that they are not the kind of content that enters into our debates and negotiations with each other, the kind of content that we hold people responsible for and on which we base our actions – rather, that content is the explicature (pragmatically enriched ‘what is said’) of our utterances. Thus, on the one hand, the minimal proposition that is claimed to be semantically expressed is not minimal enough and, on the other hand, it is too minimal. We conclude that it is not a psychologically real construct.24 On our account of linguistic communication, there are two quite distinct kinds of semantics. First, there is linguistic semantics, which concerns the meaning encoded by linguistic expression types. This consists of concepts (and procedures – see note 20), that is, constituents of the language of thought (Mentalese), and

76

Robyn Carston and Alison Hall

grasping linguistic meaning is a matter of mapping or translating linguistic forms onto conceptual constituents of thought. This kind of semantic representation is usually subpropositional, hence not truth-evaluable. Second, there is truth-conditional semantics, which captures the content of thoughts. It is thoughts (realized as sentences in Mentalese) that are the primary bearers of truth conditions, from which it follows that explicatures and implicatures (communicated thoughts) have a truth-conditional semantics.

Notes 1. We would like to acknowledge funding to Robyn Carston from the Centre for the Study of Mind in Nature at the University of Oslo, and to Alison Hall from a British Academy Postdoctoral Fellowship. 2. Throughout the chapter, we use small caps to represent propositions, thoughts or conceptual representations, in order to distinguish them from sentences and other natural language expressions. 3. For a particularly clear explanation of why communicative acts carry a presumption of optimal relevance, and of the comprehension strategy it warrants and its nonapplicability to both noncommunicative intentional acts and unintentional phenomena observed in the world, see Sperber and Wilson (2002). 4. Grice’s attempt to extend this treatment to natural language indicative conditionals and so maintain that they are semantically identical to material implication, hence wholly truthfunctional, foundered on the basis of his observation that for natural language conditionals that carry an implicature (concerning, say, a cause-consequence relation between antecedent and consequent), the negation of the conditional has to be interpreted as denying not the semantic content but the implicature (Grice: 1967/89: 83). However, given the possibility of extensive pragmatic enrichment at the level of explicit content, which we discuss in section 4, it may be that Grice’s preferred truth-functional treatment can be maintained for if as well as for and and or while accommodating the richer non-truthfunctional understanding of the utterances containing them. 5. For this reason, Grice was driven to his talk of ‘making as if to say’ in the case of metaphorical and other non-literal uses of language. Although a propositional semantic content is expressed by such uses, it is not part of the speaker’s meaning, so it is not ‘said’, in his favoured sense. See Grice (1975). 6. Most contemporary semantic minimalists would agree about the existence of this distinction between explicitly and implicitly communicated contents: Borg (2004) distinguishes what is asserted from what is implicated; Salmon (1991) and Soames (2002) recognize a ‘pragmatic what is said’ in contrast to both implicatures and a ‘strict, semantic’ notion of what is said; Bach (1994) has ‘impliciture’ and implicature, which correspond roughly to the explicature-implicature distinction we discuss here. Our difference with the minimalist approach is that we reject the idea of an extra level of content, intermediate between standing linguistic meaning and explicitly communicated meaning. 7. Kaplan’s distinction is much more limited in scope, though, than the cognitive distinction between encoded meaning and explicit content – it’s only in the case of pronouns and other indexical expressions that the distinction really does some work. Strawson (1950)

Implicature and explicature

8.

9.

10.

11.

12.

13.

14.

77

had already made a more thorough-going distinction between rules for use and expressed content, a precursor of some current views on encoded word meaning (e.g. Pietroski 2005). Different versions of this proposal have the hidden indexical attached to the nominal (Stanley and Szabó 2000), the determiner (von Fintel 1994; Martí 2003; Elbourne 2008), or adjoinable in different positions (Stanley 2005). Martí (2006) suggests that such optionality can be accounted for without appeal to free pragmatic enrichment by making the linguistic variables themselves optional, so that, on the occasions when they do appear, saturation operates mandatorily, just as it does for overt indexicals and uncontroversial cases of covert ones, such as pro-drop in Spanish and other languages. We consider (and reject) this idea in Carston and Hall (forthcoming). A possible concern raised by such alleged unarticulated constituents is that they would violate a basic semantic principle, the principle of compositionality, according to which the semantic (truth-conditional) content of the whole utterance is derived from the semantic value of each of its parts and the way in which they are syntactically combined. However, this strict version of the principle has been questioned in recent years and a pragmatics-sensitive compositional process has been proposed (Recanati 2004, 2012; Pagin 2005). In this paper, we do not discuss the notion of ‘higher-level’ explicatures concerning the speaker’s attitude to the basic explicature content or the speech act she performs in expressing this basic explicature. Sentence adverbials such as frankly, seriously, happily, and unfortunately, discussed in section 2 above, find a natural place in the relevancetheoretic account as modifiers of such higher-level explicitly communicated propositions (for discussion, see Wilson and Sperber 1993 and Ifantidou 2001). The wording is ‘“a” logical form’ (as opposed to “the”) in order to cover the case of ambiguity, where a surface form encodes more than one logical form. It would also include utterances which could be analyzed as encoding multiple logical forms, such as those with parentheticals and non-restrictive relative clauses. For discussion of this definition and a suggested amendment to it, see Carston (2002: 116–124). A reasonable question here is: how can we be sure that B is not responding to an implicature of A’s utterance? We probably cannot be completely sure, but it does seem to be quite pragmatically infelicitous to immediately and directly contradict the implicature of someone’s utterance. Consider C’s response to B in the following conversation: A: Does Max have a girlfriend these days? B: He visits New York every weekend. C: No, he hasn’t. He goes to New York to see his ill mother. C’s direct denial of B’s implicature that Max probably has a girlfriend seems odd (not fully coherent), which is not at all the case for the examples in (29) – (32). For more discussion of this point, see Carston (2004b). In a discussion of similar cases of embedding of scalar terms like some, two and or, King and Stanley (2005: 147–150) maintain that there is standardly a focal stress on the element that is to be pragmatically modified (some in this case) and this makes salient the requirement to find a contextually appropriate contrast set (here ). The upshot of their analysis is that, while the enrichment of some does contribute to the explicature, it is not a case of ‘free’ pragmatic enrichment but is a linguistically mandated pragmatic process.

78

Robyn Carston and Alison Hall

15. For discussion of this and other criteria for distinguishing pragmatic contributions to explicature from implicatures, see Carston (2002: 183–197). Recanati (1993: 269–274) elevates the embedding test to what he labels the ‘Scope Principle’ and discusses some of the pitfalls in its application. For related discussion, see also Levinson (2000) on what he calls ‘intrusive constructions’ and King and Stanley’s (2005) response. 16. However, when the sentences are presented decontextualized (as we have done in (33) – (35)), what hearers/readers do is try to treat them as utterance tokens, that is, they tend to imagine a kind of default context – the first one that comes to mind in which it would seem appropriate to utter such a sentence. 17. Recanati (1995, 2004) makes a much stronger distinction between pragmatic processes of explicature derivation and those responsible for implicatures. He calls the former ‘primary’ pragmatic processes and characterizes them as subpersonal and non-inferential (associative), while the latter ‘secondary’ processes are person-level and properly inferential. See Carston (2007) for critical discussion of this primary/secondary process distinction and a comparison with the relevance-theoretic approach, and see Recanati (2007b) for his reply. 18. The recent analysis of scalar inferences by Noveck and Sperber (2007) provides an excellent demonstration of this local/global distinction: in their view, local lexical modulation of scalar terms like some, most, often, good, probable, etc. is general and frequent across utterances, but, occasionally, when expectations of relevance demand it, a distinct proposition, an implicature, is derived by a global inferential process (e.g. the proposition N OT ALL OF THE STUDENTS PASSED THE EXAM would very likely be implicated by a speaker who utters ‘Some of them did’ in response to the question ‘Did all of the students pass the exam?’). 19. An interesting question we have yet to consider is whether ‘higher-level’ explicatures concerning speaker attitude and speech act (see note 11) meet our ‘local development’ criterion. 20. Grice’s category of conventional implicature (turning on such non-truth-conditional words as but, moreover and therefore, discussed above in section 2) has been entirely transformed in the move to the cognitively-based relevance-theoretic framework. Blakemore (1987, 2002) has argued persuasively for these words being analysed as encoding ‘procedural’ meaning, which functions to constrain the pragmatic inferential process of deriving contextual implications. See also Wilson and Sperber (1993), who extend the notion of procedural encoding to a wider range of words, some of which constrain pragmatic processes that contribute to explicature. Coming from a different (noncognitive) perspective, Bach (1999) argues that there is no such thing as conventional implicatures and reanalyses a number of Grice’s cases in other terms. 21. In relevance theory, speaker-intended contextual assumptions such as these are also classified as implicatures (implicated premises). These are, of course, also globally derived since they are retrieved as a whole from perception or memory. Recanati (2004: 48–49) does not think that so-called ‘implicated premises’ are genuine implicatures in that they tend to be not part of what is communicated but part of what the speaker takes for granted (presupposes) and expects the hearer to take for granted. This is an interesting issue that needs further thought. 22. Many entailments of an uttered sentence are not communicated at all and when they are communicated some are implicatures (as discussed above) and some are explicatures (e.g. the conjuncts in and-conjunctions). For further discussion, see Carston (2002: 121–125, 139–141).

Implicature and explicature

79

23. It is less clear why Borg (2004) is so insistent on the propositionality of sentence meaning other than that she has an antecedent commitment to the Davidsonian truth-conditional programme for natural language semantics. For some discussion of Borg’s semantic minimalism, see Carston (2008a). 24. However, the issue of shared content that drives Cappelen and Lepore is an important one and remains to be tackled within an account like ours on which propositions expressed and statements made are very pragmatic (context-sensitive) entities.

References Altmann, Gerry and Mark Steedman 1988 Interaction with context during human sentence processing. Cognition 30: 191–238. Bach, Kent 1994 Conversational impliciture. Mind and Language 9: 124–162. Bach, Kent 1999 The myth of conventional implicature. Linguistics and Philosophy 22: 327–366. Bezuidenhout, Anne 2001 Metaphor and what is said: A defense of a direct expression view of metaphor. Midwest Studies in Philosophy 25: 156–186. Blakemore, Diane 1987 Semantic Constraints on Relevance. Oxford: Blackwell. Blakemore, Diane 2002 Relevance and Linguistic Meaning: the Semantics and Pragmatics of Discourse Markers. Cambridge: Cambridge University Press. Borg, Emma 2004 Minimal Semantics. Oxford: Oxford University Press. Bosch, Peter 2007 Productivity, polysemy, and predicate indexicality. In: Ben ten Cate and Henk Zeevat (eds.), Proceedings of the Sixth International Tbilisi Symposium on Language, Logic and Computation, 58–71. Berlin: Springer. Cappelen, Herman and John Hawthorne 2007 Locations and binding. Analysis 67: 95–105. Cappelen, Herman and Ernie Lepore 2005 Insensitive Semantics. A Defense of Semantic Minimalism and Speech Act Pluralism. Oxford: Blackwell. Cappelen, Herman and Ernie Lepore 2007 The myth of unarticulated constituents. In: Michael O’Rourke and Corey Washington (eds.), Situating Semantics: Essays on the Philosophy of John Perry, 199–214. Cambridge, MA.: MIT Press. Carston, Robyn 1988 Implicature, explicature and truth-theoretic semantics. In: Ruth Kempson (ed.), Mental Representations: the Interface between Language and Reality, 155–81. Cambridge: Cambridge University Press. Reprinted in Steven Davis (ed.) 1991 Pragmatics: A Reader, 33–51. Oxford: Oxford University Press.

80

Robyn Carston and Alison Hall

Carston, Robyn 2000 Explicature and semantics. UCL Working Papers in Linguistics 12: 1–44. Reprinted in: Steven Davis and Brendan Gillon (eds.) 2004, Semantics: A Reader, 817–845. Oxford: Oxford University Press. Carston, Robyn 2002 Thoughts and Utterances: The Pragmatics of Explicit Communication. Oxford: Blackwell. Carston, Robyn 2004a Relevance theory and the saying/implicating distinction. In: Laurence Horn and Gregory Ward (eds.), The Handbook of Pragmatics, 633–656. Oxford: Blackwell. Carston, Robyn 2004b Truth-conditional content and conversational implicature. In: Claudia Bianchi (ed.), The Semantics/Pragmatics Distinction, 65–100. Stanford University: CSLI Publications. Carston, Robyn 2007 How many pragmatic systems are there? In: Maria J. Frapolli (ed.), Saying, Meaning, Referring: Essays on François Recanati’s Philosophy of Language, 18–48. Basingstoke: Palgrave Macmillan. Carston, Robyn 2008a A review of Emma Borg, 2004, Minimal Semantics. Oxford: Clarendon Press. Mind and Language 23: 359–367. Carston, Robyn 2008b Linguistic communication and the semantics-pragmatics distinction. Synthese 165: 321–345. [Special issue on the semantics/pragmatics distinction] Carston, Robyn 2010 Explicit communication and ‘free’ pragmatic enrichment. In: Belén Soria and Esther Romero (eds.), Explicit Communication: Robyn Carston’s Pragmatics, 217–287. Basingstoke: Palgrave Macmillan. Carston, Robyn and Alison Hall forthcoming Pragmatic pragmatics and the extent of linguistic systematicity. Chomsky, Noam 1980 Rules and Representations. Oxford: Blackwell. Chomsky, Noam 2000 New Horizons in the Study of Language and Mind. Cambridge: Cambridge University Press. Cohen, L. Jonathan 1971 Some remarks on Grice’s views about the logical particles of natural language. In: Yehoshua Bar-Hillel (ed.), Pragmatics of Natural Language, 50–68. Dordrecht: Reidel. Cosmides, Leda and John Tooby 1994 Origins of domain-specificity: the evolution of functional organization. In: Laurence Hirschfeld and Susan Gelman (eds.), Mapping the Mind: Domain Specificity in Cognition and Culture, 85–116. New York: Cambridge University Press. Donnellan, Keith 1966 Reference and definite descriptions. The Philosophical Review 75: 281–304.

Implicature and explicature

81

Elbourne, Paul 2008 The argument from binding. Philosophical Perspectives 22: 89–110. von Fintel, Kai 1994 Restrictions on quantifier domains. PhD dissertation, University of Massachussetts, Amherst. Fodor, Jerry A. 1983 Modularity of Mind. Cambridge, MA: MIT Press. Fodor, Jerry A. 1987 Modules, frames, fridgeons, sleeping dogs, and the music of the spheres. In: Zenon Pylyshyn (ed.), The Robot’s Dilemma, 139–150. Norwood, New Jersey: Ablex. Fodor, Jerry A. 1998 Concepts: Where Cognitive Science Went Wrong. Oxford: Clarendon Press. Grice, H. Paul 1967/89 Logic and Conversation, 1967 William James lectures, Harvard University. Published in H. Paul Grice (1989), Studies in the Way of Words, 1–143. Cambridge, MA: Harvard University Press. Grice, H. Paul 1975 Logic and conversation. In: Peter Cole and Jerry Morgan (eds.), Syntax and Semantics 3: Speech Acts, 41–58. New York: Academic Press. [Reprinted in H. Paul Grice (1989), 22–40]. Grice, H. Paul 1989 Studies in the Way of Words. Cambridge, MA: Harvard University Press. Hall, Alison 2008a Free enrichment or hidden indexicals? Mind and Language 23: 426–456. Hall, Alison 2008b Free pragmatic processes and explicit utterance content. PhD dissertation, University of London. Hall, Alison 2009 ‘Free’ enrichment and the nature of pragmatic constraints. UCL Working Papers in Linguistics 21: 93–123. Ifantidou, Elly 2001 Evidentials and Relevance. Amsterdam: John Benjamins. Kaplan, David 1977/89 Demonstratives. In: Joseph Almog, John Perry and Howard Wettstein (eds.), 1989, Themes from Kaplan, 481–563. Oxford: Oxford University Press. King, Jeffrey C. and Jason Stanley 2005 Semantics, pragmatics, and the role of semantic content. In: Zoltan Gendler Szabó (ed.), Semantics versus Pragmatics, 111–164. Oxford: Oxford University Press. Levinson, Stephen 2000 Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge, MA: MIT Press. Martí, Luisa 2003 Contextual variables. PhD dissertation, University of Connecticut. Martí, Luisa 2006 Unarticulated constituents revisited. Linguistics and Philosophy 29: 135–166.

82

Robyn Carston and Alison Hall

Neale, Stephen 1992 Paul Grice and the philosophy of language. Linguistics and Philosophy 15: 509–559. Noveck, Ira 2004 Pragmatic inferences related to logical terms. In: Ira Noveck and Dan Sperber (eds.), Experimental Pragmatics, 301–321. Basingstoke: Palgrave Macmillan. Noveck, Ira and Dan Sperber 2007 The why and how of experimental pragmatics: The case of “scalar inferences”. In: Noel Burton-Roberts (ed.), Advances in Pragmatics, 184–212. Basingstoke: Palgrave Macmillan. Nunberg, Geoffrey 1995 Transfers of Meaning. Journal of Semantics 12: 109–32. Pagin, Peter 2005 Compositionality and context. In: Gerhard Preyer and Georg Peter (eds.), Contextualism in Philosophy: Knowledge, Meaning, and Truth, 303–348. Oxford: Oxford University Press. Pietroski, Paul 2005 Meaning before truth. In: Gerhard Preyer and Georg Peter (eds.), Contextualism in Philosophy: Knowledge, Meaning, and Truth, 255–302. Oxford: Oxford University Press. Recanati, François 1989 The pragmatics of what is said. Mind and Language 4: 295–329. Recanati, François 1993 Direct Reference: From Language to Thought. Oxford: Blackwell. Recanati, François 1995 The alleged priority of literal interpretation. Cognitive Science 19: 207–232. Recanati, François 2002 Unarticulated constituents. Linguistics and Philosophy 25: 299–345. Recanati, François 2004 Literal Meaning. Cambridge: Cambridge University Press. Recanati, François 2007a It is raining (somewhere). Linguistics and Philosophy 30: 123–146. Recanati, François 2007b Reply to Carston. In: Maria J. Frapolli (ed.), Saying, Meaning, Referring: Essays on François Recanati’s Philosophy of Language, 49–54. Basingstoke: Palgrave Macmillan. Recanati, François 2012 Compositionality, flexibility, and context-dependence. In: Markus Werning, Wolfram Hinzen and Edourd Machery (eds.), The Oxford Handbook of Compositionality, ch. 8, Oxford: Oxford University Press. Sag, Ivan 1981 Formal semantics and extra-linguistic context. In: Peter Cole (ed.), Syntax and Semantics 13. Radical Pragmatics, 273–94. New York: Academic Press. Salmon, Nathan 1991 The pragmatic fallacy. Philosophical Studies 42: 37–45. Searle, John 1978 Literal meaning. Erkenntnis 13: 207–224.

Implicature and explicature

83

Searle, John 1983 Intentionality. Cambridge: Cambridge University Press. Soames, Scott 2002 Beyond Rigidity: The Unfinished Semantic Agenda of Naming and Necessity. Oxford: Oxford University Press Sperber, Dan 1994 The modularity of thought and the epidemiology of representations. In: Laurence A. Hirschfeld and Susan A. Gelman (eds.), Mapping the Mind: Domain Specificity in Cognition and Culture, 39–67. Cambridge: Cambridge University Press. Sperber, Dan 2002 In defense of massive modularity. In: Emmanuel Dupoux (ed.), Language, Brain and Cognitive Development: Essays in Honor of Jacques Mehler, 47–57. Cambridge MA: MIT Press. Sperber, Dan and Deirdre Wilson 1986/95 Relevance: Communication and Cognition. Oxford: Blackwell. Second edition with Postface 1995. Sperber, Dan and Deirdre Wilson 2002 Pragmatics, modularity and mind-reading. Mind and Language 17(1/2): 3–23. Sperber, Dan and Deirdre Wilson 2008 A deflationary account of metaphors. In: Ray Gibbs (ed.), Cambridge Handbook of Metaphor and Thought, 84–105. New York: Cambridge University Press. Stanley, Jason 2000 Context and logical form. Linguistics and Philosophy 23: 391–434. Stanley, Jason 2002 Making it articulated. Mind and Language 17: 149–168. Stanley, Jason 2005 Semantics in context. In: Gerhard Preyer and Georg Peter (eds.), Contextualism in Philosophy: Knowledge, Meaning, and Truth, 221–54. Oxford: Oxford University Press. Stanley, Jason and Zoltan Gendler Szabó 2000 On quantifier domain restriction. Mind and Language 15: 219–261. Strawson, Peter F. 1950 On referring. Mind 59: 320–344. Strawson, Peter F. 1952 Introduction to Logical Theory. London: Methuen. Swinney, David A. 1979 Lexical access during sentence comprehension: a (re)consideration of context effects. Journal of Verbal Learning and Verbal Behavior 18: 645–659. Travis, Charles 1985 On what is strictly speaking true. Canadian Journal of Philosophy 15: 187–229. Travis, Charles 1991 Annals of analysis: Studies in the Way of Words, by H.P. Grice. Mind 100: 237–264.

84

Robyn Carston and Alison Hall

Wedgwood, Daniel 2007 Shared assumptions: Semantic minimalism and Relevance Theory. Journal of Linguistics 43: 647–681. Wilson, Deirdre and Robyn Carston 2006 Metaphor, relevance and the ‘emergent property’ issue. Mind and Language 21: 404–433. Wilson, Deirdre and Robyn Carston 2007 A unitary approach to lexical pragmatics: relevance, inference and ad hoc concepts. In: Noel Burton-Roberts (ed.), Advances in Pragmatics, 230–260. Basingstoke: Palgrave Macmillan. Wilson, Deirdre and Dan Sperber 1993 Linguistic form and relevance. Lingua 90: 1–25. Wilson, Deirdre and Dan Sperber 2002 Truthfulness and relevance. Mind 111: 583–632. Wilson, Deirdre and Dan Sperber 2004 Relevance theory. In: Laurence Horn and Gregory Ward (eds.), Handbook of Pragmatics, 607–632, Oxford: Blackwell.

4.

Inference and reasoning in discourse comprehension Murray Singer and R. Brooke Lea

1.

Introduction

A folktale begins, There once was a fisherman who had no luck at all. He and his wife and four children sold everything to their name. Full understanding requires both comprehending the explicit content and extracting innumerable inferences. Trivially, the fisherman is interpreted to be a human being. Pronouns such as he and their must be inferentially associated with their proper referents. However, many candidate inferences are deceptively complex. No luck probably implies the failure to catch fish. The reader understands that the family sold everything to provide needed money. Inferences so pervade language comprehension that Schank (1976) identified inference processing as the core of the comprehension problem. This chapter addresses these problems in the framework of empirical studies of cognitive psychology (Anderson 2005: 10). Early psychological researchers considered the possibilities that discourse inference processing favoured (a) a message’s logical implications (e.g., Tim gave one of his two apples to Jim implies that Tim has one left) and/or (b) those implications highly consistent with world knowledge (The plane landed implies the participation of a pilot). However, these hypotheses were proven to oversimplify matters. This chapter is intended to analyze forty years of progress concerning these issues. First, an outline is presented of one contemporary theoretical framework of discourse inference, along with definitions of “inference” that it affords (section 2). Then, the chapter inspects inferences necessary for comprehension (section 3), others that usefully elaborate message content (section 4), and ones that represent the deductive implications of a message (section 5). Finally, some important trends in ongoing research will be examined (section 6).

2.

Theoretical framework and terminology

2.1.

Theoretical framework of inference processing

Theories of language comprehension address the message representations constructed by the reader1 and the cognitive processes that enable that construction. Discourse comprehension is considered to result in the representation of (a) the surface form of the message, (b) a network of the explicitly stated idea units (or “propositions”), and (c) mental models of the described situations (Johnson-Laird

86

Inference and reasoning in discourse comprehension

1983b; van Dijk and Kintsch 1983). This multilevel analysis is supported by findings such as the faster decay of surface than situational representations (Jarvella 1971; Kintsch and van Dijk 1978); and the predominance, in post-comprehension tasks, of situational representations over propositional and surface ones (e.g. Potts 1972). The influential construction-integration theory (Kintsch 1988, 1998) addresses the cognition of comprehension in general and of inference in particular. It posits that, at an initial construction stage, language processes ranging from word recognition to syntactic parsing extract the idea units of the message, termed “propositions.” Upon reading Jane saw a cardinal in the nest, one extracted proposition might be denoted as JANE SAW A CARDINAL.2 This initial propositional network is an enriched one: It includes message generalizations, such as JANE WAS BIRDWATCHING; and ideas that preserve message coherence. It includes associates that are both relevant (SONGBIRD) and even irrelevant (cardinal suggests CLERGYMAN; Swinney 1979) to the message. During a subsequent integration processing stage, activation accumulates in the most highly interconnected message elements.3 As a result, irrelevant concepts, such as the CLERGYMAN interpretation of cardinal, plus other tangential text ideas are effectively eliminated from the representation. Thus, integration guides the construction of a coherent representation. The most highly activated message elements are carried over in working memory for the processing of subsequent discourse chunks (Fletcher 1981; Kintsch and van Dijk 1978). Furthermore, the integrated representation is considered to encompass the surface, propositional, and situation levels discussed earlier. Recent advances have further refined the construction-integration framework. The memory processes of retrieving relevant information during construction are addressed by the memory-based text processing analysis (Greene et al. 1994; O’Brien, Lorch, and Myers 1998). From this perspective, the current discourse chunk provides cues for the passive retrieval of prior discourse ideas and relevant world knowledge, which are said to “resonate” to the cues (Ratcliff 1978). Retrieval is regulated by factors such as the similarity and message-distance between the current chunk and its antecedents; and the degree of elaboration, typicality, and distinctiveness of those antecedents (Albrecht and Myers 1995, 1998; O’Brien and Albrecht 1991; O’Brien et al.1995; McKoon and Ratcliff 1992). As a result, comprehending the current clause interrelates the discourse and world knowledge. Likewise, there is accumulating evidence that Kintsch’s (1988) integration processing stage involves a substage of validating the accuracy of the current chunk (O’Brien and Albrecht 1992; Singer, Revlin, and Halldorson 1990; van Berkum, Hagoort, and Brown 1999). Then, depending on the outcome of validation, the discourse situation model may be updated (e.g., O’Brien et al. 1998; Radvansky and Copeland 2001; Rinck, Hahnel, and Becker 2001). The contribution of these processes to inference computation will be considered throughout.

Murray Singer and R. Brooke Lea

2.2.

87

Terminology: Psychological definitions of inference

Several features of psychological terminology will serve to clarify this chapter. First, the explicit and implicit content that the understander successfully derives from a message and stores in memory is said to be encoded. This differs from the linguistic usage, according to which the speaker “encodes” mental content acoustically or orthographically. Second, the act of implying is assigned to the communicator whereas inferring is associated with the understander. However, even indisputable implications of the speaker’s message are not considered to definitively reflect the speaker’s intention; which contrasts with the thrust of implicatures (Grice 1975; cf. Carston and Hall, this volume; Moeschler, this volume). Third, the term entailment complies with linguistic usage, according to which concepts and constructs entail their semantic elements. In this regard, bachelor entails unmarried. Psychological analyses predominantly treat the inferences that accompany comprehension as being encoded in the discourse situation model (e.g., Schmalhofer, McDaniel, and Keefe 2002). Section 3 will scrutinize the numerous dimensions of the situation model. In this regard, detecting the incongruence of Jane was waiting outside the door of her health club when the instructor came in must reflect the reader’s construction of a spatial situation model – the surface composition of the sentence does not signal the incongruence. Certain inferences are likely represented at the propositional level in the form of supplementary arguments and whole propositions. Thus, the explicit proposition underlying The patient was examined, denoted as (EXAMINE, OBJECT: PATIENT), might be inferentially augmented to (EXAMINE, OBJECT: PATIENT, AGENT: DOCTOR).2 Fully understanding The lightning struck. The hut collapsed would reasonably involve constructing a causal proposition, although none was explicit. Relevant psychological evidence appears in sections 3 and 4. This treatment largely disregards inferences stemming from low-level processing; such as deducing that The orange had no *eel, with the obscured speech sound [p], refers to a peel. Likewise deemphasized are syntactic inferences such as determining that, in Before the king rides his horse is always groomed, horse is not the direct object of ride (Frazier and Rayner 1982). Figurative inferences that enable comprehension of metaphors such as My lawyer is a shark are considered in other chapters of this volume (cf. Giora, this volume; Deignan, this volume).

3.

Bridging inferences

For comprehension to proceed smoothly, the understander must be able to relate each successive text idea to the prior discourse. The contributing processes are deceptively complicated. The connection between a fisherman having no luck and selling his possessions may seem apparent. In contrast, A fisherman had no luck at

88

Inference and reasoning in discourse comprehension

all. He vacuumed his carpets, seems disjointed. However, the luck sentence has the same degree of surface overlap with selling possessions and vacuuming carpets. The difference is that world knowledge permits the reader to infer a reason for selling possessions if one has been unlucky. Here, that reason serves to bridge the message ideas and so preserve message coherence. Discourse bridges are often signalled by the cohesive devices of syntax that distinguish suitable from inappropriate wording (Halliday and Hasan 1976). Also central to discourse coherence is the distinction between given and new information (Clark and Haviland 1977; Haviland and Clark 1974). The syntactic form of What the audience enjoyed was the music suggests that it was previously understood (“given”) that the audience enjoyed something, whereas music constitutes new information. To maintain coherence, the reader must identify the given information of the current clause, retrieve its referent from the prior discourse, and encode the relation between the referent and the new information. The bearing of the given-new analysis on inference processing was highlighted by the classic demonstration that readers take less time to report understanding The beer was warm when it is preceded by We got some beer out of the trunk than by We checked the picnic supplies (Haviland and Clark 1974). In both sequences, the definite article The designates beer as given information. However, the picnic supplies sentence mentions no beer, so the reader must draw a time-consuming bridging inference establishing beer as a sensible component of picnic supplies. Those findings enabled further investigations of the integral contribution of bridging inferences to comprehension.4 The following sections inspect this contribution at both the propositional and situational levels of representation, and with regard to numerous semantic relations. 3.1.

Anaphoric resolution

Anaphors are linguistic devices for referring to prior elements of a message. As such, they serve fundamental bridging functions. The pronoun is the prototypical anaphor. In We got some beer out of the trunk. It was warm, coherence would be disrupted if the reader did not detect that It referred to the beer. Detecting this relation bridges the second sentence to the first. Pronouns vary on numerous meaning dimensions, such as number, gender, person (she – I), and syntactic role (he – him). As a result, pronouns may or may not unambiguously denote their antecedents. In this regard, comprehension time is less for sentences with pronouns with a single possible referent (Sally rewarded Ron because HE was punctual) than multiple referents (Tom rewarded Ron because HE was punctual; Ehrlich 1980). It also takes less time to understand sentences with pronouns whose antecedents are nearby than distant (Lesgold, Roth, and Curtis 1979). These are all essentially bridging-inference phenomena: Expressions that complicate anaphoric resolution systematically inflate comprehension time.

Murray Singer and R. Brooke Lea

89

Pronoun resolution involves subtle and intricate processes. For example, pronoun use is not merely an optional shortcut. Rather, pronouns convey the continuity of discourse topic whereas definite noun phrases signify topic change (Vonk, Hustinx, and Simons 1992; although it is the indefinite noun phrase that introduces entirely new information, Halliday and Hasan 1976). As a result, in Bob thinks that Bob will win, repeating Bob instead of using a pronoun is awkward and tends to impedes comprehension (Gordon and Scearce 1995). These phenomena of pronoun resolution and bridging extend to other anaphors, such as definite noun phrases. The vehicle is anaphoric because the definite article the is a indicator of given information – it suggests that a vehicle has been previously mentioned. Resolving these anaphors, like pronoun resolution, involves complex bridging processes. To cite one example, it takes less time to understand a sentence beginning The vehicle … when its discourse antecedent is a typical vehicle (car) than an atypical one (tank; Corbett 1984; Garrod and Sanford 1977). 3.2.

Semantic bridging inferences

Sequences such as The patient was examined. The doctor made some notes pose the understander little trouble. Brief consideration, however, indicates that the first sentence includes no explicit referent for the given element of the second sentence, doctor. However, the sentences are readily bridged by detecting that doctor is a sensible agent for examining a patient. Examples like this pervade ordinary language. Early investigations inspected readers’ inferences about the likely agents, objects, instruments, locations, and “recipients” of discourse (Fillmore 1968; Johnson, Bransford, and Solomon 1973). Further study demonstrated that, to a degree, encoding inferences depends less on the particular semantic role of a concept than its necessity for coherence. Any of the mentioned roles may contribute either to necessary inferences or merely informative inferences. In one relevant study, subjects read stimuli like The patient was examined. The doctor made some notes and then verified statements that probed posited bridging inferences, such as The patient was examined by a doctor (Singer 1980; see also Kintsch 1974). Verification time was about equal to that measured in sequences that explicitly stated the crucial idea; and less than for sequences in which the inference plausibly elaborated the message (see next section) but was not necessary for coherence. This suggests that readers routinely encode bridging inferences about the agents, objects, and instruments of actions. Bridging inferences might, as suggested earlier, be represented in encoded propositions (e.g., Singer and Halldorson 1996): As discussed earlier, The patient was examined, denoted propositionally as (EXAMINE, OBJECT: PATIENT), might be elaborated to (EXAMINE, OBJECT: PATIENT, AGENT: DOCTOR), capturing the “doctor” bridging inference. Perhaps more importantly, inferences

90

Inference and reasoning in discourse comprehension

that accompany comprehension are proposed to be encoded (perhaps redundantly) in the discourse situation model (e.g., Schmalhofer, McDaniel, and Keefe 2002). Situation models encompass numerous discourse dimensions, including cause, motivation, narrative character, space, and time (Zwaan, Magliano, and Graesser 1995; Zwaan and Radvansky 1998). Readers’ construction of coherent situation models is central to comprehension (Graesser, Singer, and Trabasso 1994). It is informative to inspect the bridging inferences that foster coherence on some of these situational dimensions. 3.2.1.

Causal and motive bridging inferences

The causal structure of discourse encompasses both physical cause, such as the sun melting a snowman; and motivation, such as finding work for the purpose of earning money (e.g., Schank and Abelson 1977). According to the principle of explanation, understanders routinely seek candidate causes of both physical outcomes and people’s intentional actions (Graesser, Singer, and Trabasso 1994). Candidate causes are evaluated on their sufficiency and “necessity in the circumstances” for their outcomes (Trabasso, van den Broek, and Suh 1989). One feature of events and intentional actions that is proposed to initiate causal inference is that they signal changes in physical or motivational state (Cheng and Holyoak 1985; Trabasso et al. 1989). More broadly, identifying explanations is proposed to enhance people’s control over their psychological environment (Weiner 1985). The necessity of causal bridges for comprehension is illustrated by the sequence, The lightning struck. The hut collapsed. Coherence requires the bridging inference that the first event caused the second, because the two sentences share no content words. There is extensive behavioural evidence that such inferences accompany comprehension. Reading time increases systematically with the causal distance between text events: Reading time (interpreted to partly reflect people’s comprehension time) is greater for The ashes floated up the chimney when preceded by the causally distant The spy needed to dispose of the report than the causally close The spy threw the report in the fireplace (Keenan, Baillet, and Brown 1984; Myers, Shinjo, and Duffy 1987). This suggests that readers strive to bridge the current sentence to its antecedent. Furthermore, Mason and Just (2004) observed that functional magnetic resonance imaging (fMRI) responses in certain right-hemisphere areas of the temporal cortex were greater in the close-causal than distant-causal condition. They concluded that readers detected satisfactory causal bridges only in the former case. Congruent with the latter findings, judgment time in narrative understanding is less for test prompts that capture plausible causal bridges than for those that do not. This is observed both for single-word prompts (e.g., burn for the prior “spy”

Murray Singer and R. Brooke Lea

91

example; Potts, Keenan, and Golding 1988) and full sentences (The report burned up; Singer and Ferreira 1983). Analogously, in comprehending expository text, lexical decisions (do strings like SCARF and SNARF form English words?) take less time for a cause word like gravity when it appears in the context of the collapse of a star than in a control context (Millis and Graesser 1994; see also Graesser and Bertus 1998). In such studies, the causal bridging condition and control condition are equated for the degree to which the test item and its antecedent overlap in their explicit wording and even associations among their words (McKoon and Ratcliff 1986; Potts et al. 1988). 3.2.2.

Temporal bridging inferences

Temporal or time relations are fundamental to verbal expression and comprehension. In many languages, a tense inflection of the verb (e.g., present, past) identifies the relative time of occurrence of the event. Time markers such as adverbials likewise organize events in relative (much later) or absolute (on 1 December 2015) ways. These signals characterize the time dimension of the situation model. Understanders seem strikingly sensitive to these discourse relations (Anderson, Garrod, and Sanford 1983; Carreiras et al. 1997; Münte, Schiltz, and Kutas 1998; Zwaan 1996; Zwaan, Madden, and Whitten 2000). Much of the construction of the temporal situation depends on bridging inferences. In one typical study, subjects encountered one version of Claudia’s train arrived in Dresden on time (before/after) Markus’s train. It subsequently took longer to read Claudia was already waiting for him when he got off the train when the subject had received the inconsistent after version than the consistent before version (Rinck, Hahnel, and Becker 2001). This consistency effect depends on the reader’s construction of the temporal situation. It emerges even though the sentences are locally coherent by the repetition of Claudia. Linguistic aspect marks the extension or completion of an event but, like tense, pertains to the temporal features of situations (Lyons 1977: 485, 687). In this regard, the imperfective She was changing the flat tire describes an ongoing activity whereas the perfective She changed the flat tire describes a completed action. After reading text sequences including such statements, readers took less time to recognize the statement CHANGE FLAT with reference to the imperfective aspect than the perfective (Magliano and Schleich 2000; see also Bergen and Wheeler 2010; Morrow 1985). The investigators concluded that completed activities are less available in working memory than ongoing ones. 3.2.3.

Spatial bridging inferences

Messages of many genres organize concepts in spatial contexts. Full comprehension arguably ought to include encoding the relative position of objects and places,

92

Inference and reasoning in discourse comprehension

and even the earliest language-processes studies were consistent with this proposal (Bransford, Barclay, and Franks 1972). Detecting the spatial configurations described in discourse often depends on bridging inferences rather than explicit assertions. However, the picture seems more equivocal than for the causal and temporal dimensions. Readers are certainly aware of conspicuous spatial anomalies. Thus, it takes longer to read the second sentence of the incongruent, Jane waited outside the door of her health club. The instructor came in than of the congruent, Jane waited inside the door of her health club. The instructor came in (O’Brien and Albrecht 1992). However, complex arrangements of entities, as might be depicted in a mystery novel, do not seem to be reliably extracted (Zwaan and van Oostendorp 1993). Inferring all possible spatial relations, even ones that serve bridging functions, would likely overwhelm the understander’s cognitive resources. In a renowned paradigm for studying spatial situations, subjects first memorize the layout of a complex setting. When they subsequently read a story set in that site, characters’ positions and movement paths affect the relative cognitive availability of the site’s objects and rooms (Morrow, Greenspan, and Bower 1987). Thus, when the setting is familiar, readers seem to monitor location and space. However, these representations mainly capture ordinal locations along paths rather than Euclidean distances (Rinck et al. 1997). In addition, the tracking of protagonist movement seems curtailed when test items probe only objects and not characters (Wilson et al. 1993). Spatial inferences are also affected by text quality, such as whether a story depicts an ordinal route or compass positions (Perrig and Kintsch 1985). The encoding of spatial information also seems regulated by its relevance: For example, statements like standing UNDER THE BRIDGE were read more quickly and remembered better in the relevant context, starting to rain, than the less relevant context, blocking the moonlight (Radvansky and Copeland 2000). These findings suggest that discourse spatial information is constructed in a constrained manner. 3.2.4.

Conclusion

Discourse situation models are as complex and varied as the scope of human cognition. From this perspective, bridging inferences are likely to contribute to yet other situational dimensions; such as complex psycholinguistic models of the discourse (Sanford and Garrod 1998), the tacit speaker-hearer contracts of speech act theory (Searle 1975; Grice 1975), and the bodily reactions that cognitive embodiment theorists propose to support language comprehension (Glenberg and Robertson 2000; Zwaan, Stanfield, and Yaxley 2002).

Murray Singer and R. Brooke Lea

3.3.

93

Validation of bridging inferences

Successful reading requires the continual monitoring of discourse congruence and accuracy. As discussed throughout, readers are sensitive to inconsistencies at surface and propositional levels of representation (Klin 1995; Long and Chong 2001) and with reference to the various dimensions of the situation model (Albrecht and Myers 1995; Lea, Mulligan, and Walton 2005; O’Brien and Albrecht 1992; Rinck, Hahnel, and Becker 2001; Zwaan, Magliano, and Graesser 1995). Consistencymonitoring in comprehension enables the validation of the current discourse with reference to its antecedents and to general knowledge. Discourse consistency-effects converge with evidence that an evaluative stage is common to many domains of cognition (Kolodner 1983; Traxler et al. 1997; Tulving 1983). In the realm of discourse comprehension, it is proposed that tentative bridging inferences must be validated against world knowledge before they are accepted by the reader (Noordman, Vonk, and Kempff 1992; Singer, Revlin, and Halldorson 1990). For example, to comprehend Dorothy poured the bucket of water on the bonfire. The fire went out, one analysis states that the reader must first construct an idea which, combined with the causal antecedent, accounts for the causal outcome. Then, if the constructed idea, such as “water extinguishes fire”, is consistent with world knowledge, the inference has been validated. Consistent with this proposal, it takes less time to answer Does water extinguish fire? after reading the former causal sequence than after the control “temporal sequence”, Dorothy placed the bucket of water by the bonfire. The fire went out (Singer et al. 1992). Without validation processing, the contrasting sequence, Dorothy poured the bucket of water on the bonfire. The fire grew hotter, would seem entirely agreeable (Singer 1993). Readers apparently strive to validate bridging inferences even in challenging, expository contexts. Consider the sentence, Some fireworks burn with a crimson flame because they contain calcium salts. Full understanding depends on a construction such as Calcium salts burn with a crimson flame. Having encountered the “fireworks” sentence in a passage, subjects took no longer to answer a question about the hypothetically constructed idea than when the message had explicitly asserted this fact (Singer, Harkness, and Stewart 1997). In this circumstance, of course, most readers have no relevant knowledge against which to evaluate the constructed idea. However, they apparently accept this new idea on the basis of their belief in the cooperation of the writer (Grice 1975). It is noted that readers’ tendencies to validate discourse inferences can be defeated by sufficiently complex expository text (Noordman, Vonk, and Kempff 1992).

94 4.

Inference and reasoning in discourse comprehension

Elaborative inferences

As discussed throughout, ordinary sentences carry innumerable implications. The aunt drove downtown implies that the agent was human, the agent was female, the aunt drove a car, the aunt drove along streets, and many other ideas. These implications, some certain (“female”) but others probabilistic (“car”), stem from our world knowledge. They support “elaborative inferences” because they sensibly elaborate the explicit content of a message. Researchers originally considered that elaborative inferences might reliably accompany comprehension, but extensive evidence tended to deny this. Many studies compared elaborative inferences with corresponding bridging inferences and suitable control conditions. For example, The spy threw the report in the fireplace. He called the airline permits the elaborative inference that the report burned, but this inference does not serve to bridge the two sentences. Having read that sequence, people took longer to verify the inference statement, The spy burned the report, than when the statement probed either a bridging inference or an explicit message idea (Singer and Ferreira 1983). Likewise, judgment time for inference words (e.g., burn) is greater in an elaborative condition than a bridging condition (Potts, Keenan, and Golding 1988; see also McKoon and Ratcliff 1986). More recently, the electrophysiological N400 ERP signal suggested that elaborative inference statements are less congruent with the evolving discourse representation than bridging statements (St. George, Mannes, and Hoffman 1997; Yang, Perfetti, and Schmalhofer 2007). Theorists reasoned that encoding all reasonable elaborative inferences during comprehension would create an inference explosion (Charniak 1975) that would overwhelm people’s limited cognitive resources (Baddeley 1986). These findings, coupled with the converse conviction that elaborative inferences usefully add to the message representation, created a framework for several programs of research. 4.1.

Highly constrained elaborative inferences

Elaborative inferences are relatively likely to be encoded when they are highly constrained by the discourse context. Implied categories may be so constrained, as when juice for breakfast strongly suggests orange and woman in the drama implies actress. Experiments examining (a) inferential cued-recall (does the implied probe, orange, bring the original sentence to mind?) and (b) reading time of continuation sentences suggested that such elaborations are robustly encoded (Anderson et al. 1976; Dubois and Denis 1988; Garrod et al. 1990; McKoon and Ratcliff 1989; Whitney 1986). Some messages highlight one property or another of a concept. For example, play the piano emphasizes music whereas lift the piano focuses on weight. Readers appear to encode relevant characteristics of discourse concepts (McKoon and Rat-

Murray Singer and R. Brooke Lea

95

cliff 1988). Readers also appear to elaboratively infer the theme or topic of text. Thus, timed word judgments indicated that, upon reading The townspeople found that all the buildings had collapsed except the mint, one infers an earthquake, at the expense of inferences about the “candy” sense of mint (Till, Mross, and Kintsch 1988). The mere presence in a discourse of a concept promotes elaborative inferences about it. In the sequence, There was a broom in the closet next to the kitchen. Bill swept the floor every week on Saturday, the word broom does not bridge the second sentence back to the first. Nevertheless, people make faster decisions about broom than about suitable control words after reading the second sentence of the sequence (Lucas, Tanenhaus, and Carlson 1990). Because broom is available in the discourse context (McKoon and Ratcliff 1992), it likely resonates to swept. Therefore, it may participate in inference computations. In conclusion, the joint constraint upon and relevance of certain potential discourse elaborations appears to result in their reliable encoding. 4.2.

Predictive inferences

The continued scrutiny of elaborative inferences probably reflects researchers’ intuition, spanning the history of this field (Miller 1980; Otten and van Berkum 2008), that understanders might predict the ideas or the very words that will be presented next. Investigators have particularly inspected predictive inferences about the physical and motivational causes and the consequences of discourse events. Several investigations have examined sequences like No longer able to control his fury, the husband hurled the delicate porcelain vase against the wall, which implies that the vase broke (McKoon and Ratcliff 1986). This prediction is elaborative – it does not bridge explicit text ideas. Shortly after reading the sequence, readers’ naming times, recognition times, and lexical decisions about “broke” are faster than when “broke” follows a control sequence (Keefe and McDaniel 1993; Klin, Guzman, and Levine 1999; McKoon and Ratcliff 1986; Murray, Klin, and Myers 1993). This suggests that some degree of encoding of the prediction takes place. Chronometric studies indicate that the inference activation may require about 1.25 seconds (Calvo, Castillo, and Estevez 1999). These findings have been informatively qualified in other studies. Significantly, some encoded predictive inferences appear to be transient: If probed after just one more continuation sentence, decisions about the inference are no longer faster than in the control condition unless it acquires a bridging function (Keefe and McDaniel 1993). Furthermore, potential predictive inferences compete with one another. For example, when the husband’s throwing of the vase appeared in the discourse context indicating that, if he had another tantrum, his wife would leave him, the predictive inference about the vase breaking was delayed if not prevented (Klin, Guzman, Levine 1999; Weingartner et al. 2003).

96

Inference and reasoning in discourse comprehension

Inferences about people’s motives, such as to avoid marital breakdown, are paralleled by results about physical causation. Thus, 500 ms after reading the sequence Jimmy (’s throwing of a rock) missed, though, and he accidentally hit the door of a new car, readers named the implication “dent” faster than after reading a control passage in which a sponge ball rather than a rock was thrown (Cook, Limber, and O’Brien 2001). The authors emphasized that the implication “dent” was not a simple associate of words or word combinations of their materials. However, they noted that naming time was lower still when the outcome was explicit, which tends to deny the robust encoding of the predictive inference. 4.3.

Toward unified accounts of elaborative and bridging inference

Theorists have strived to integrate these diverse findings. Proposed analyses have noted that bridging inferences are likely encoded both in the situation model, and, by virtue of their coherence functions, the propositional textbase (Fincher-Kiefer and D’Agostino 2004; Kintsch 1988; Singer and Kintsch 2001). Elaborative inferences, in contrast, are neither mentioned in their message nor do they preserve coherence; so they might be restricted to the situation model (Schmalhofer, McDaniel, and Keefe 2002). Schmalhofer, McDaniel and Keefe advanced a unified analysis that drew on Kintsch’s (1988, 1998) construction-integration framework. They noted that elaborations, such as that throwing a vase results in its breaking, are situationally supported by other text ideas and world knowledge. Therefore, integration processes may confer moderate activation on them. However, if the consequences of the breakage are not pursued in the message, that activation quickly dissipates (Keefe and McDaniel 1993), resulting in weak or negligible encoding. The promise of unified analyses is their explanation of atypical or complex cases, such as the weak representation of bridging inferences or the robust representation of elaborations. It was discussed earlier that the failure of an antecedent to resonate may prevent it being bridged to the current clause (Albrecht and Myers 1995). In another study suggesting the curtailment of bridging, subjects read about a father learning about a broken window and then becoming angry. However, the motivational inference that might have bridged the two ideas was precluded when the text provided an earlier, highly elaborated alternative cause of the anger (“lost keys”; Rizzella and O’Brien 1996). Conversely, it was discussed earlier that highly constrained elaborative inferences may be strongly encoded. Non-encoded bridges and encoded elaborations both may be addressed with reference to the abundance of strong and weak connections that the target inference bears to elements within and between the surface, propositional, and situational levels of text representation. There is a danger of circularity in presuming that empirical evidence of inference or non-inference diagnoses rich versus sparse representational connections.

Murray Singer and R. Brooke Lea

97

However, computational models of competing inference theories can provide converging evidence that may corroborate proposals concerning text connectivity (e.g., Frank et al. 2003; Myers and O’Brien 1998).

5.

Reasoning and deduction in discourse comprehension

Many of the inferences considered thus far are based on readers’ knowledge about the world. If you do not have experience with brick walls and porcelain vases, then you will not infer the likely outcome when the two collide with force. These inferences are inductive in that they are (a) based on many past individual observations, and (b) probabilistic in nature. For example, when one reads The actress fell from the roof of a 14 story building, probable outcomes come to mind but none of them are guaranteed by the discourse circumstances (McKoon and Ratcliff 1986). By contrast, deductive inferences are certain to be true if the premises they are based on are true. Propositional-logic inferences recently have attracted considerable attention. These include deductions based on linguistic particles that convey logic, such as if, and, or, and not. Although perhaps not as common as pragmatic inferences (Franks 1996), logical inferences have been shown to play an important role in discourse comprehension. Contemporary research on logic and behaviour began in the reasoning literature. Early work on psychological models of deduction followed the Piagetian tradition by investigating children’s basic competence with logic (e.g., Braine 1959; Braine and Rumain 1981; Hiebert 1980; Lowerre and Scandura 1973; Paris 1976). Later work extended this research to adults, and sought to distinguish the relative contributions of logic and pragmatics (e.g., Johnson-Laird, Legrenzi, and Legrenzi 1972). A rich set of so-called content effects demonstrated that logically complex puzzles (10 % solution rate) could be solved easily if they were presented in a helpful, real-world context (up to 90 % solution rate). Wason’s (1966) notorious Selection Task is perhaps the best known example of the effect of content on performance: Subjects could more easily solve deductive problems expressed in realistic terms, such as the required postage on an envelope; than with reference to arbitrary relationships among numbers and letters (see also Cheng and Holyoak 1985; Cosmides 1989; Griggs and Cox 1982; Johnson-Laird et al. 1972). Dramatic content effects led researchers interested in logical competence to use stimuli devoid of meaning (e.g. ps and qs) in an effort to control pragmatic effects. These studies showed that both adults and children are proficient with a set of propositional-logic inferences (e.g. Braine 1978; Rips 1983). Abstract-content tasks were helpful in sorting out the relative roles of pragmatics and logic, but were not particularly informative about the use of logic in everyday behaviour. Eventually, psychological theories of deduction needed to be tested in the domain in which our logical competence is expressed – ordinary language comprehension.

98

Inference and reasoning in discourse comprehension

5.1

Deductive inferences in discourse

5.1.1.

Deductive bridges

Early studies of logical inferences in text comprehension scrutinized scenarios that were written around logical syllogisms and examined subjects’ memory of the passages immediately after reading them. Consider the following story about a young man dressing for an important event (Lea et al. 1990). (1) Jerry was trying to decide what to wear to meet his fiancée’s parents. (2) “I’d like to wear either my striped or my checkered shirt,” he thought. (3) “If I wear my striped shirt, I’ll have to wear my blue pants since they match the shirt,” he thought, “but I’m not sure where my blue pants are.” (4) He looked in his closet for his checkered shirt and saw that it was wrinkled, so he decided he could not wear that. (5) “So I’d better find my blue pants,” he thought. Comprehending this short passage depends in part on propositional-logic inferences. Unless the disjunction or in Sentence 2, the conditional if in Sentence 3, and the negation not in Sentence 4 are understood and the appropriate inferences are drawn from them, the story will be difficult if not impossible to comprehend. The passage is structured around the logical syllogism shown in Table 1. Table 1.

Syllogism supporting the comprehension of the Dressing Up passage A or B If A, then C Not B Therefore A (by or-elimination) Therefore C (by modus-ponens)

After reading passages like “Dressing Up,” the subjects judged whether the information expressed by test sentences had been stated explicitly or implicitly in its passage. Overwhelmingly, and more often compared to a control, subjects falsely identified, as explicit, information that they had inferred (Lea et al. 1990, Experiments 1 and 2). For example, readers indicated that the “Dressing Up” passage explicitly stated that Jerry had decided to wear his striped shirt, when in fact that outcome must be deduced via the or-elimination inference (striped or checkered shirt; not checkered; therefore striped). Lea et al. (1990) showed that readers were highly accurate at drawing deductive inferences (95 % correct) and that they were making them easily enough to be largely unaware of doing so (see Rodrigues, Dias, and Roazzi 2002, for a replication in Portuguese). This result was striking for two reasons: First, the literature abounded with evidence of inept human reasoning (e.g. Tversky and Kahneman 1973). Second, to the extent that people could reason deductively, logical infer-

Murray Singer and R. Brooke Lea

99

ences were assumed to require substantial and focused effort on the part of the reasoner. Logical reasoning was not thought to occur without intention or awareness. 5.1.2.

Deductive elaborations

The inferences investigated by Lea et al. (1990) bridged explicitly mentioned text concepts, such as “blue pants” in sentence 5 above, to earlier discourse ideas. A follow-up study used passages such as Sniffing Out The Goods, shown next, to ask whether deductive inferences are made online when their output does not fulfil a bridging function (Lea 1995). (6) Special Investigator Evans was trying to crack a Cuban cigar smuggling ring. (7) As he entered the dark warehouse he whispered into his walkie-talkie: “If the dog starts to bark, then we know there’s tobacco around.” (8) “Ok, I’ll be listening,” replied his partner Newstead, who was monitoring the situation from their unmarked patrol car. (9a) There was quiet for a few minutes, then the silence was broken by the sound of Evans’s dog barking. After reading, subjects were asked to respond to words that were semantically associated with the entailed proposition. In this passage, the reader was given the conditional premise if the dog barks, then tobacco, and the minor premise the dog barked. Together, these premises yield the inference that there is tobacco around, via modus ponens.5 For this passage, SMOKE was the probe word in either a lexical decision task or a naming task. Responses to the target were faster when the texts contained the second premise, as in (9a) above; compared to when the second premise was replaced by the similarly worded control phrase of sentence (9b) that did not satisfy the antecedent of the conditional. (9b) There was quiet for a few minutes, and Newstead wondered whether he was close enough to hear the sound of the dog barking. This result indicated that readers were making the modus ponens inference at the moment that both premises were available to them. Together with a parallel finding for or-elimination, these findings favour the drawing of certain logical inferences easily and spontaneously – even when the inference is not needed to make sense of the text or to answer a question. The latter consideration is especially relevant to identifying the conditions in which elaborative, non-bridging inferences are made online. Two reasons that elaborative inferences are not made during reading are the following: First, it is inefficient to infer what will likely be given in the text. Second, there is the risk that one may draw the wrong inference (Murray, Klin, and Myers 1993). Logical inferences are different from most pragmatic inferences in that, if the premises are true, then

100

Inference and reasoning in discourse comprehension

the conclusion (inference) must be true. Thus, the second reason for elaborative inferences not accompanying comprehension may not hold – or hold as strongly- – for logical inferences such as modus ponens and or-elimination. The measure of certainty afforded by logical inference may explain why elaborative logical inferences may be computed during comprehension more reliably than elaborative pragmatic inferences (Campion 2004, 2006; Lea 1995; Rader and Sloutsky 2001, 2002). Evidence of Campion (2004) converged with the latter principles. He proposed that deductive inferences might be represented as relatively certain whereas pragmatic inferences, such as predicted consequences, are represented as hypothetical facts. Consistent with this analysis, readers needed less time verify statements pertaining to deductive inferences when the statements were phrased as certainties than hypothetically; but vice versa for predicted-consequence inferences. 5.1.3.

Discourse deductions and negation

As logical deductions became recognized as reliably accompanying comprehension, questions arose as to how the deductive processes interact with other cognitive functions associated with reading. For example, a robust finding in the text processing literature is that readers have difficulty verifying negated propositions that appear in a text. This observation suggested that negated propositions are difficult to understand and remember. For example, MacDonald and Just (1989) found that subjects were slower and less accurate at verifying that the word bread had appeared in the sentence, Almost every weekend, Elizabeth bakes no bread but only cookies for the children, compared to a control sentence in which cookies was negated instead of bread. Do psychological models of deduction make predictions about impact of negation on the inference process? One theory very specifically predicts that inference schemas that produce negated deductions (e.g., therefore not q) are not more difficult to compute than schemas that generate affirmative ones (e.g., therefore q; Braine, Reiser, and Rumain 1984). Using abstract-content syllogisms, Braine et al. (1984) found that two very similar logical inferences produced almost identical difficulty ratings. This was despite the fact that one form, or-elimination, produces an affirmative conclusion (p or q, not p, therefore q); whereas the other, not-bothelimination, yields a negated inference (not both p and q; p, therefore not q). This in turn raises the question of whether the predictions of the model apply to discourse processing or whether the suppressing effect of negation on memory for explicit propositions extends to the production of negative information in a discourse setting. If the latter possibility is true, then readers should infer affirmative propositions (therefore q) more readily than negated ones (therefore not q). This would be consistent with the theoretical position, considered earlier, that portrays readers as relatively passive about making elaborative inferences. According to that approach, readers should be less likely to generate negative information about

Murray Singer and R. Brooke Lea

101

what did not happen in a story than to generate affirmative information about what did happen. In a series of experiments, Lea and Mulligan (2002) observed that readers readily derived both affirmative and negative logical inferences during reading, and that only the output of the inference process is affected by negation. That is, negation did not affect the generation of the logical inference. Rather, it affected the availability of the inference once it was made. 5.2.

Discourse reasoning and basic cognition

5.2.1.

Memory-based processing

How do deductive inferences operate within the context of current theories of comprehension? Some investigators have scrutinized the contribution to discourse deductions of the passive activation of backgrounded information by the resonance processes that were considered toward the outset of the chapter. In one pertinent study, logical premises, such as that a person who was shopping could not afford to buy both a scarf and a hat, appeared near the beginning of stories (Lea et al. 2005). The focus was whether, upon encountering the second premise, she decided to buy a hat, readers infer that the shopper did not buy the scarf. The results indicated that the inference was drawn when the second premise followed the first immediately but not when approximately 12 sentences intervened between the two. However, when, in the intervening-sentences condition, the second premise was accompanied by a contextual cue (e.g. leather sofa) that had co-occurred with the first, readers did make the logical inference. This suggested that the contextual cue provided sufficient contextual overlap to cause the distant premise to resonate. Thus, memory-based text processes reactivated the backgrounded premises required to draw the logical inference. This illustrates how basic cognitive functions collaborate with language processes to produce inferences during reading. Resonance is a fundamental memory process. Psychological models of deduction concern the integration of information via logical connectives. Thus, successful language comprehension involves both language-specific cognitive processes and more general ones. A related point concerns the limits of any single account of inference processing in discourse. Passive activation processes such as resonance can account for how premises enter working memory, but they cannot explain how a deduction is derived from them once they are activated. Logical inferences require the involvement of higher-level processes of deduction (e.g., Braine and O’Brien 1991; Johnson-Laird, Byrne, and Schaeken 1992; Rips 1994). Conversely, psychological models of deduction cannot explain how two non-contiguous premises co-occur in working memory, thereby enabling an inference. Any complete model of discourse inferences will need to specify how both the critical antecedent information and world knowledge are activated so as to permit their integration into appropriate inferences.

102 5.2.2.

Inference and reasoning in discourse comprehension

Commonalities of deductive and pragmatic inference processing

Here, pragmatic inferences and logical inferences have been treated separately, reflecting somewhat distinct literatures concerning the two. When people read stories, however, the two classes of inference are more difficult to distinguish (Lea 1998). Consider the following exchange: (10) Joan was asking Joy about her new car. “What did you decide about which type to get?” “Well,” replied Joy, “I live in San Francisco, so I decided not to get a stick.” In this example, Joan can infer (deductively, via or-elimination) that Joy bought a car with an automatic transmission. Unlike in other studies, the required premises were not both provided by the text. Rather, Joan must derive, from world knowledge, the major premise that there are automatic and stick-shift car transmissions. In designing her utterance, Joy can take this mutual knowledge about automobiles for granted, assume that Joan will retrieve the appropriate premise, and be assured that she has provided sufficient information to answer Joan’s question (Sperber and Wilson 1986/1995). This effortless collaboration between logic and pragmatics permits economy in discourse: Joy can disregard two important elements of their exchange (the major premise and the conclusion), and is free to focus on what motivated the decision – her desire to avoid driving a stick on the hilly streets of San Francisco. This exchange amounts to a syllogism with a missing premise, or an “enthymeme” (Singer, Revlin, and Halldorson 1990). For comprehension to succeed, Joan – and the reader – need to construct and retrieve the missing premise that car transmissions are either automatic or manual. This serves to validate the tentative deductive inference that Joy has opted for an automatic transmission. 5.2.3.

Discourse reasoning and inference validation

The latter analysis suggests a convergence of the processes of validating pragmatic inferences, as discussed earlier; and drawing deductive inferences in comprehension. Consider again the sequence Dorothy poured the bucket of water on the bonfire. The fire went out. In this case, the enthymeme requiring solution might be expressed as follows (Singer et al. 1992): (11) Major premise = Dorothy poured the bucket of water on the bonfire. Minor premise = X. Conclusion = The fire went out. The enthymeme is readily solved as X = Water extinguishes fire. Because the solution is congruent with world knowledge, the causal inference that bridges sen-

Murray Singer and R. Brooke Lea

103

tences (13) and (15) is validated. What is noteworthy is that ordinary readers are not trained in logic. They are also likely unaware of the logical processes that permit them to compute the implications of Jerry decided he could not wear his checkered shirt or to be satisfied that the bucket of water sequence is coherent. It has been emphasized that pragmatic discourse inferences depend on people’s factual world knowledge. Deductive inference likewise depends on people’s knowledge – namely, their knowledge of the deductive procedures that capture one’s intuitive logical abilities. Although pragmatic knowledge accumulates through experience, mainly, it is less clear how we come to possess our deductive abilities. A previously cited body of research supports the hypothesis that some facility with logic reveals itself rather early in life (e.g., Braine and Rumain 1981; Hiebert 1980); and recent work has even shown that children can think more logically than adults (Noveck 2001). Indeed, a basic logical ability is hypothesized by some to be a part of our natural cognitive endowment, variously called a natural, or mental logic (e.g., Braine and O’Brien 1998; Rips 1994), a syntax of thought, or mentalese (Fodor 1975; Pinker 1994), or a logic instinct (Crain and Khlentzos 2010). These hypotheses, of course, should not be misunderstood as claims that all forms of logic come naturally to humans; clearly they do not. However, just as undergraduates’ difficulty with integral calculus does not gainsay research that reveals a basic mathematical competence in young children (e.g. Wynn 1996), the depth to which people can be made to look foolish with logic should not impeach or eclipse evidence that we possess a basic logical competence. To summarize, these observations suggest that theoretical analysis has perhaps underestimated the commonalities of pragmatic and deductive inference processing. Further integrating these domains is a contemporary challenge to researchers.

6.

Current trends in inference research

Research in any domain is bound to exhibit ongoing controversies, debates, and advances. This section briefly addresses several realms of intense current scrutiny. 6.1.

Immediacy of inference

The understander both of spoken and written messages experiences rapidly encountered strings of words. This raises the issue of when, relative to its appearance, a word is analyzed. Just and Carpenter (1980) proposed that words are immediately interpreted, as far as possible, at all levels of representation. Immediate processing precludes the need to maintain, in working memory, multiple interpretations of the current discourse chunk, such as would result from encountering words or syntactic expressions that are ambiguous. Delayed processing, in contrast, would have

104

Inference and reasoning in discourse comprehension

the advantage of avoiding backtracking when immediate processing yields an erroneous interpretation. In the domain of parsing, it was observed that a sentence such as The old man the boats puzzles the reader at the second the. The reader has interpreted man as a noun and so is not expecting an article. This “garden-path” is resolved by reinterpreting man as a verb rather than a noun. This phenomenon suggests immediate parsing: Had the reader not categorized man as a noun, the garden path would not emerge (e.g. Frazier and Rayner 1982). Closer to the present concerns, consider the sequence, The guard mocked one of the prisoners in the machine shop. He had been at the prison for only one week. The pronoun He could refer to the guard or the prisoner. Carpenter and Just (1977) presented eye fixation data revealing that readers regularly looked back in the message immediately upon encountering the pronoun, frequently fixating precisely on one of the competing nouns. Immediacy has again become topical in inference research because of the power of reading times, eye fixation data, and neural measures to distinguish immediate from delayed processing. In this regard, ERP methodology has been applied to understanders’ immediate sensitivity to inconsistencies between the current discourse and the antecedent context. N400 profiles indicate that both readers (van Berkum, Hagoort, and Brown 1999) and listeners (van Berkum et al. 2003) immediately detect such inconsistencies, such as when Jane told her brother that he was exceptionally slow is encountered in the context in which he has done something quickly. In a complementary study, reading times indicated that people immediately detected situational incongruencies, such as a vegetarian ordering a cheeseburger in a restaurant (Stewart, Kidd, and Haigh 2009). Furthermore, ERP profiles within 1 sec of a critical word dissociate explicit, paraphrased, and elaborative inference discourse continuations (Yang et al. 2007). Thus, contemporary methodologies extend the immediacy principle to subtly differing text relations. 6.2.

Embodied cognition and discourse inference

The embodied cognition framework posits that meaning is based on representations of perception and action (Barsalou 1999; Glenberg 1997). Perceptual representations are specific to sensory modalities such as vision and audition. Such representations stand in contrast with the propositions discussed here, which, by abstracting information away from perceptual experience, are amodal. The embodiment hypothesis bears on discourse inference in addressing behavioural effects that can be traced to sensorimotor phenomena. Numerous experimental results that arguably pertain to elaborative inferences are consistent with the embodiment position. In one study, people took less time to recognize the picture of a vertical nail after reading He hammered the nail into the floor, which implies a vertical orientation; than after He hammered the nail into the wall, which implies a horizontal orientation. Likewise, after reading The ranger

Murray Singer and R. Brooke Lea

105

saw the eagle in the nest, people needed less time to name a picture of an eagle with its wings folded than with its wings extended in flight (Zwaan, Stanfield, and Yaxley 2002). This is despite the fact that the sentence made no reference to the eagle’s wings. With regard to action representation, Glenberg and Kaschak’s (2002) subjects had to register sensibility judgments about statements like He closed a drawer. Response switches were arranged so that subjects had to move their hands either away from or toward their bodies. Answer time was lower with the actioncompatible motion of the hand away one’s body. It is noteworthy that propositional representations of contrasting sentences pairs (e.g., nail into wall/floor) do not entail information that could explain these results. Anderson (2005), in contrast, reported that subjects discriminated sentences differing only by wrote versus forged better than those differing by words such as scared versus frightened. This was despite the fact that the perceptions afforded by wrote and forged are ostensively identical. Anderson acknowledged, however, that the sensorimotor representations of the latter verbs might differ in subtle ways. A possible reply to embodiment findings in the discourse realm is that although the cited effects cannot be addressed by propositional representations, contemporary discourse theory has long acknowledged that propositional networks must be augmented by situation models (Anderson 1993; Bransford, Barclay, and Franks 1972; van Dijk and Kintsch 1983). Furthermore, symbols originally grounded in sensorimotor experience are likely cognitively manipulable without reference to their original grounding (Fodor et al. 1980; Newell and Simon 1972). At the least, however, embodiment theory has identified complex effects that pertain to a neglected dimension of the discourse situation model. 6.3.

Comprehension and inference: Shallow or robust?

In view of the limitations of elaborative inference processing discussed here, it might not be surprising that theorists have addressed imperfections of discourse inference. Inference might be restricted because of systematic cognitive constraints and because understanders are inattentive to discourse. These influences and some factors that tend to curtail them will be briefly examined. 6.3.1.

Cognitive constraints on discourse inference

There are numerous, well-known instances of understanders’ misapprehension of ordinary language. In response to How many animals of each kind did Moses take on the ark?, many people readily respond “two” even though it was Noah that took animals on the ark (Erickson and Mattson 1981; see also Loftus 1975). Understanders overlook verbal anomalies such as No head injury is too trivial to ignore (Sanford and Graesser 2006; read it carefully) and tranquillizing stimulants (Hannon and Daneman 2004). Garden-path sentences such as The man ate up the street

106

Inference and reasoning in discourse comprehension

offer comprehenders with alternative, conflicting grammatical interpretations (Crain and Steedman 1985). Deviations of this sort enable erroneous inferences. In many of these instances, the cognitive regularities that contribute to misunderstanding have been systematically inspected. One feature of the Moses illusion question is that Moses is given rather than new. There is evidence of curtailed matching of given information with its antecedents (Reder and Kusbit 1991; Singer 1976). As a result, one may carelessly answer “two” to the Moses illusion question. Excessive demands on working memory, sometimes subtle, are likely to thwart comprehension. In this regard, No head injury is too trivial to ignore presents three negatives either explicit or implicit in nature (e.g., IGNORE = NOT(ATTEND); Just and Clark 1973). Likewise, working memory capacity may be exceeded by complex constructions, such as the “multiply-embedded” The mouse the cat the dog hated saw escaped (Miller and Isard 1963). These observations have led to systematic analyses of curtailed comprehension. From the perspective of “good-enough” language processing, challenges to comprehension ranging from the rapid decay of the surface form of sentences to competition among interpretations of ambiguous sentences yield representations that may be partial, incomplete, or inaccurate (Ferreira, Ferraro, and Bailey 2002; see also Sanford and Graesser 2006). It is noteworthy that the construction-integration analysis posits that spurious elements are constructed during the initial phase of comprehension (Kintsch 1988, 1998). Good-enough processing adds that the final, integrated representation may systematically include spurious elements. Curtailed inference has also been addressed by the memory-based text processing analysis. That analysis holds that comprehension computations are limited to (a) the current discourse chunk, (b) ideas maintained in working memory, and (c) the other discourse ideas and general knowledge that resonate to them (O’Brien et al. 1998; see also McKoon and Ratcliff 1992). As discussed earlier, memory-based processing furthermore identifies factors that regulate resonance, such the degree of elaboration in the message of a candidate antecedent, and the antecedent’s distance from and similarity to the current chunk. In the absence of relevant resonance, the understander may fail to (a) bridge a character’s action to the failure of her goal (Albrecht and Myers 1995) or (b) inferentially detect one possible cause of an outcome in the presence of a competing cause (Rizzella and O’Brien 1998). Apart from the systematic impact of discourse factors, understanders differ on many dimensions that promote or impede inference processing. These dimensions include resources of working memory and other cognitive systems, vocabulary, reading skill, and idiosyncratic knowledge (Gernsbacher, Varner, and Faust 1990; Just and Carpenter 1992; Long, Oppy, and Seely 1994; McNamara and Kintsch 1996). Furthermore, the appreciable extent to which readers become distracted from their text would likewise tend to thwart inference construction (Smallwood, McSpadden, and Schooler 2008).

Murray Singer and R. Brooke Lea

6.3.2.

107

Robust inference processes

There are many phenomena of discourse inference that contrast with generalizations about shallow processing and restricted inference. One noteworthy example, as discussed throughout, is the understander’s sensitivity to discourse inconsistencies at the level both of message detail and its multidimensional situation model (e.g., O’Brien and Albrecht 1992). Those findings are complemented by brain-imaging data indicating that even discourse-level semantic anomalies are detected within a few hundred milliseconds of their appearance (van Berkum, Hagoort, and Brown 1999). The pervasiveness of discourse inference was addressed by the constructionist analysis of Graesser et al. (1994). Comprehension was characterized as a search after meaning. From this perspective, comprehension is guided by the reader’s goal; and it typically monitors situational as well as surface coherence. Maintaining situational coherence was, in turn, posited to entail seeking the explanations of outcomes described in discourse. Thus, constructionism identified principles that might guide the inferential derivation of situation models. Furthermore, some demonstrations have indicated that situational coherence apparently can supersede the memory-based constraints of similarity and text-distance (Lutz and Radvansky 1997; Singer and Richards 2005; van den Broek and Lorch 1993). These results tend to counterbalance findings that minimizing resonance precludes certain inferences. The prevalence of veridical inference in comprehension is likewise favoured by the apparent default status of deep comprehension. In fact, behavioural results tend to comply with an inferential profile unless the understander is directed to undertake a shallow task, such as proofreading (Singer and Halldorson 1996). Thus, the current trend to inspect systematic deficiencies in comprehension and inference defines a tension between the viewpoints of deep, accurate inferencing versus curtailed inferencing. Principled analyses of curtailed inference have begun and will likely continue to characterize comprehension as a process of interpretation. This process will likely be shown to produce useful albeit distorted representations. In that event, comprehension would be comparable to the visual analysis that makes people prone to optical illusions or to the encoding of maps in systematically imperfect ways. Systematically imperfect comprehension and inference would thus characterize those processes as comparable to rather than distinct from other cognitive systems.

Notes 1. Reading and listening comprehension exhibit many similarities (e.g., Perfetti and Goldman 1976). The large volume of research on reading comprehension justifies referring to the understander as the “reader”.

108

Inference and reasoning in discourse comprehension

2. Using Kintsch’s (1974) formal scheme, this idea would be denoted as (SEE, AGENT:JANE, PATIENT:CARDINAL, LOCATION:NEST). 3. This process is proposed to be regulated by computational algorithms of constraint satisfaction (Rumelhart and McClelland 1986). 4. It should be noted that some bridging inferences are likely realized as a relation between two propositions, as in (CAUSE, STRIKE(LIGHTNING), COLLAPSE(HUT)). Others take the form of an added argument: For example, (EXAMINE, OBJECT: PATIENT, AGENT: DOCTOR) captures a bridging inference about the agent, doctor. 5. Modus ponens refers to the following syllogistic form: If A then B (major premise) A (minor premise) Then B (conclusion)

Acknowledgements This research was supported by Discovery Grant OGP9800 from the Natural Sciences and Engineering Research Council of Canada to MS and NSF grant DUE-1044006 to RBL. We are grateful to Kathy Maxwell and Anjum Fazaluddin for technical assistance in the preparation of the manuscript. Please address correspondence to Murray Singer, Department of Psychology, University of Manitoba, Winnipeg, Canada R3T 2N2, [email protected]. References Albrecht, Jason E. and Jerome L. Myers 1995 Role of context in accessing distant information during reading. Journal of Experimental Psychology: Learning, Memory, and Cognition 21: 1459–1468. Albrecht, Jason E. and Jerome L. Myers 1998 Accessing distant text information during reading: Effects of contextual cues. Discourse Processes 26: 87–10. Anderson, Anthony, Simon C. Garrod and Anthony J. Sanford 1983 The accessibility of pronominal antecedents as a function episode shifts in narrative text. Quarterly Journal Experimental Psychology 35A: 427–440. Anderson, John R. 1993 Rules of the Mind. Hillsdale, NJ: Erlbaum. Anderson, John R. 2005 Cognitive Psychology and its Implications, 6th edn. New York: Worth Publishers. Anderson, Richard C., James V. Pichert, Ernest T. Goetz, Diane L. Schallert, Kathleen V. Stevens and Stanley R. Trollip 1976 Instantiation of general terms. Journal of Verbal Learning and Verbal Behavior 15: 667–679. Baddeley, Alan D. 1986 Working Memory. Oxford Psychology Series Volume 11. Oxford: Oxford University Press (Clarendon).

Murray Singer and R. Brooke Lea

109

Barsalou, Lawrence W. 1999 Perceptual symbol systems. Behavioral and Brain Sciences 22: 577–660. Bergen, Benjamin and Kathryn Wheeler 2010 Grammatical aspect and mental simulation. Brain and Language 112: 150–158. Braine, Martin D. S. 1959 The ontogeny of certain logical operations: Piaget’s formulation examined by nonverbal methods. Psychological Monographs 73 (5). [Whole No. 475]. Braine, Martin D. S. 1978 On the relation between the natural logic of reasoning and standard logic. Psychological Review 85: 1–21. Braine, Martin D. S. and David P. O’Brien 1991 A theory of if: A lexical entry, reasoning program, and pragmatic principles. Psychological Review 98: 182–203. Braine, Martin D. S. and David P. O’Brien (eds.) 1998 Mental Logic. Mahwah, NJ: Erlbaum. Braine, Martin D. S., Brian J. Reiser and Barbara Rumain 1998 Evidence for the theory: Predicting the difficulty of propositional logic inference problems. In: Martin D. S. Braine and David O’Brien (eds.), Mental Logic, 91–144. Mahwah, NJ: Erlbaum. Braine, Martin D. S. and Barbara Rumain 1981 Development of comprehension of “or”: Evidence for a sequence of competencies. Journal of Experimental Child Psychology 31: 46–70. Bransford, John D., J. Richard Barclay and Jeffery J. Franks 1972 Semantic memory: A constructive versus interpretive approach. Cognitive Psychology 3: 193–209. Calvo, Manuel G., M. Dolores Castillo and Adelina Estevez 1999 On-line predictive inferences in reading: Processing time during versus after the priming context. Memory and Cognition 27: 834–843. Campion, Nicolas 2004 Predictive inferences are represented as hypothetical facts. Journal of Memory and Language 50: 149–164. Campion, Nicolas 2006 Hypothetical and certain inferences from conditional arguments read in texts. Journal of Experimental Psychology: Learning, Memory, and Cognition 32: 547–558. Carpenter, Patricia A. and Marcel A. Just 1977 Reading comprehension as eyes see it. In: Marcel A. Just and Patricia A. Carpenter (eds.), Cognitive Processes in Comprehension, 109–139. Hillsdale, NJ: Erlbaum. Carreiras, Manuel, Núria Carriedo, Maria A. Alonso and Angel Fernandez 1997 The role of verbal tense and verbal aspect in the foregrounding of information in reading. Memory and Cognition 25: 438–446. Charniak, Eugene 1975, Organization and inference in a frame-like system of common sense knowledge. In: Roger Schank and Bonnie Nash-Webber (eds.), Theoretical Issues in Natural Language Processing. An Interdisciplinary Workshop, 46–55. Cambridge, MA: Massachusetts Institute of Technology.

110

Inference and reasoning in discourse comprehension

Cheng, Patricia W. and Keith J. Holyoak 1985 Pragmatic reasoning schemas. Cognitive Psychology 17: 391–416. Clark, Herbert H. and Susan E. Haviland 1977 Comprehension and the given-new contract. In: Roy Freedle (ed.), Discourse Production and Comprehension, 1–40. Hillsdale, NJ: Erlbaum. Cook, Anne E., John E. Limber and Edward J. O’Brien 2001 Situation-based context and the availability of predictive inferences. Journal of Memory and Language 44: 220–234. Corbett, Albert 1984 Prenominal adjectives and the disambiguation of anaphoric nouns. Journal of Verbal Learning and Verbal Behavior 23: 683–695. Cosmides, Leda 1989 The logic of social exchange: Has natural selection shaped how humans reason? Studies with the Wason selection task. Cognition 31: 187–276. Crain, Stephen, and Drew Khlentzo 2010 The logic instinct. Mind & Language 25: 30–65. Crain, Stephen and Mark Steedman 1985 On not being led up the garden path: The use of context by the psychological syntax processor. In: David R. Dowty, Lauri Karttunen and Arnold M. Zwicky (eds.), Natural Language Parsing: Psychological, Computational and Theoretical Perspectives. Cambridge: Cambridge University Press. Dubois, Daniele and Michel Denis 1988 Knowledge organization and instantiation of general terms in sentence comprehension. Journal of Experimental Psychology: Learning, Memory, and Cognition 14: 604–611. Ehrlich, Kate 1980 The comprehension of pronouns. Quarterly Journal of Experimental Psychology 32: 247–256. Erickson, Thomas D. and Mark E. Mattson 1981 From words to meaning: A semantic illusion. Journal of Verbal Learning and Verbal Behavior 20: 540–551. Ferreira, Fernanda, Vittoria Ferraro and Karl G. D. Bailey 2002 Good-enough representations in language comprehension. Current Directions in Psychological Science 11: 11–15. Fillmore, Charles J. 1968 The case for case. In: E. Bach and T. Harms R. (eds.), Universals of Linguistic Theory, 1–88. New York: Holt, Rinehart and Winston. Fincher-Kiefer, Rebecca and Paul R. D’Agostino 2004 The role of visuospatial resources in generating predictive and bridging inferences. Discourse Processes 37: 205–224. Fletcher, Charles R. 1981 Short-term memory processes in text comprehension. Journal of Verbal Learning and Verbal Behavior 20: 564–574. Fodor, Jerry. A. 1975 The Language of thought. Cambridge, MA: Crowell Press. Fodor, Jerry A., Merrill F. Garrett, E. C. Walker and C. H. Parkes 1980 Against definition. Cognition 8: 263–367.

Murray Singer and R. Brooke Lea

111

Frank, Stefan L., Mathieu Koopen, Leo G. M. Noordman and Wietske Vonk 2003 Modeling knowledge-based inferences in story comprehension. Cognitive Science 27: 875–910. Franks, Bridget A. 1996 Deductive reasoning in narrative contexts: Developmental trends and reading skill effects. Genetic, Social, and General Psychology Monographs 122: 75–105. Frazier, Lyn and Keith Rayner 1982 Making and correcting errors during sentence comprehension: Eye movements in the analysis of structurally ambiguous sentences. Cognitive Psychology 14: 178–210. Garrod, Simon, Edward J. O’Brien, Robin K. Morris and Keith Rayner 1990 Elaborative inferences as an active or passive process. Journal of Experimental Psychology: Learning, Memory, and Cognition 16: 250–257. Garrod, Simon, and Anthony Sanford 1977 Interpreting anaphoric relations: The integration of semantic information while reading. Journal of Verbal Learning and Verbal Behavior 16: 77–90. Gernsbacher, Morton Ann, Kathleen R.Varner, and Mark E. Faust 1990 Investigating differences in general comprehension skill. Journal of Experimental Psychology: Learning, Memory, and Cognition 16: 430–445. Glenberg, Arthur M. 1997 What memory is for. Behavioral and Brain Sciences 20: 1–19. Glenberg, Arthur M. and Michael P. Kaschak 2002 Grounding language in action. Psychonomic Bulletin and Review 9: 558–565. Glenberg, Arthur M. and D. A. Robertson 2000 Symbol grounding: A comparison of high-dimensional and embodied theories of mean. Journal of Memory and Language 43: 379–401. Gordon, Peter C. and Kimberly A. Scearce 1995 Pronominalization and discourse coherence, discourse structure and pronoun interpretation. Memory and Cognition 23: 313–323. Graesser, Arthur C. and Eugenie L. Bertus 1998 The construction of causal inferences while reading expository texts on science and technology. Scientific Studies of Reading 2: 247–269. Graesser, Arthur C., Murray Singer and Tom Trabasso 1994 Constructing inferences during narrative text comprehension. Psychological Review 101: 371–395. Greene, Steven B., Richard J. Gerrig, Gail McKoon and Roger Ratcliff 1994 Unheralded pronouns and management by common ground. Journal of Memory and Language 35: 511–526. Grice, Herbert P. 1975 Logic and conversation. In: Peter Cole and Jerry. Morgan (eds.), Syntax and semantics 3: Speech acts, 41–58. New York: Academic Press. Griggs, Richard A. and James R. Cox 1982 The elusive thematic-materials effect in Wason’s selection task. British Journal of Psychology 73: 407–420. Halliday, Michael A. K. and Ruqaiya Hasan 1976 Cohesion in English. London: Longman.

112

Inference and reasoning in discourse comprehension

Hannon, Brenda and Meredyth Daneman 2004 Shallow semantic processing of text: An individual-differences account. Discourse Processes 37: 187–204. Haviland, Susan E. and Herbert H. Clark 1974 What’s new? Acquiring new information as a process in comprehension. Journal of Verbal Learning and Verbal Behavior 13: 512–521. Hiebert, Elfrieda H. 1980 The relationship of logical reasoning ability, oral language comprehension, and home experiences to preschool children’s print awareness. Journal of Reading Behavior 8: 313–324. Jarvella, Robert J. 1971 Immediate memory and discourse processing. In: Gordon H. Bower (ed.), The Psychology of Learning and Motivation, Volume 13, 379–421. New York: Academic Press. Johnson, Marcia K., John D. Bransford and Susan K. Solomon 1973 Memory for tacit implications of sentences. Journal of Experimental Psychology 98: 203–205. Johnson-Laird, Philip N. 1983a A computational analysis of consciousness. Cognition and Brain Theory 6: 499–508. Johnson-Laird, Philip N. 1983b Mental Models. Cambridge, MA: Harvard University Press. Johnson-Laird, Philip N., Ruth M. Byrne and Walter Schaeken 1992 Propositional reasoning by model. Psychological Review 99: 418–439. Johnson-Laird, Philip N., Paolo Legrenzi and Maria S. Legrenzi 1972 Reasoning and a sense of reality. British Journal of Psychology 63: 395–400. Just, Marcel A. and Patricia A. Carpenter 1980 A theory of reading: From eye fixations to comprehension. Psychological Review 87: 329–354. Just, Marcel A. and Patricia A. Carpenter 1992 A capacity theory of comprehension: Individual differences in working memory. Psychological Review 99: 122–149. Just, Marcel A. and Herbert H. Clark 1973 Drawing inferences from the presuppositions and implications of affirmative and negative sentences. Journal of Verbal Learning and Verbal Behavior 12: 21–31. Keefe, Dennis E. and Mark McDaniel 1993 The time course and durability of predictive inferences. Journal of Memory and Language 32: 446–463. Keenan, Janice M., Susan D. Baillet and Polly Brown 1984 The effects of causal cohesion on comprehension and memory. Journal of Verbal Learning and Verbal Behavior 23: 115–126. Kintsch, Walter 1974 The Representation of Meaning in Memory. Hillsdale, NJ: Erlbaum. Kintsch, Walter 1988 The role of knowledge in discourse comprehension: A construction-integration model. Psychological Review 95: 163–182.

Murray Singer and R. Brooke Lea

113

Kintsch, Walter 1998 Comprehension. New York: Cambridge University Press. Kintsch, Walter and Teun A. van Dijk 1978 Toward a model of text comprehension and production. Psychological Review 85: 363–394. Klin, Celia M. 1995 Causal inferences in reading: From immediate activation to long term memory. Journal of Experimental Psychology: Learning, Memory, and Cognition 21: 1483–1494. Klin, Celia M., Alexandria E. Guzman and William H. Levine 1999 Prevalence and persistence of predictive inferences. Journal of Memory and Language 40: 593–604. Kolodner, Janet L. 1983 Maintaining organization in a dynamic long-term memory. Cognitive Science 7: 243–280. Lea, R. Brooke 1995 On-line evidence for elaborative logical inferences in text. Journal of Experimental Psychology: Learning, Memory, and Cognition 21: 1469–1482. Lea, R. Brooke 1998 Logical inferences and comprehension: How mental-logic and text processing theories need each other. In: Martin D. S. Braine, and David P. O’Brien (eds.), Mental logic, 63–78. Mahwah, NJ: Erlbaum. Lea, R. Brooke and Elizabeth J. Mulligan 2002 The effect of negation on deductive inferences. Journal of Experimental Psychology: Learning, Memory, and Cognition 28: 303–317. Lea, R. Brooke, Elizabeth J. Mulligan and Jennifer L. Walton 2005 Accessing distant premise information: How memory feeds reasoning. Journal of Experimental Psychology: Learning, Memory, and Cognition 31: 387–395. Lea, R. Brooke, David P. O’Brien, Shalom M. Fisch, Ira A. Noveck and Martin D. S. Braine 1990 Predicting propositional logic inferences in text comprehension. Journal of Memory and Language 29: 361–387. Lesgold, Alan M., Steven F. Roth and Mary E. Curtis 1979 Foregrounding effects in discourse comprehension. Journal of Verbal Learning and Verbal Behavior 18: 291–308. Loftus, Elizabeth F. 1975 Leading questions and the eyewitness report. Cognitive Psychology 7: 560–572. Long, Debra L. and Jennifer L. Chong 2001 Comprehension skill and global coherence: A paradoxical picture of poor comprehenders’ abilities. Journal of Experimental Psychology: Learning, Memory, and Cognition 27: 1424–1429. Long, Debra L., Brian J. Oppy, and Mark R. Seely 1994 Individual differences in the time course of inferential processing. Journal of Experimental Psychology: Learning, Memory, and Cognition 20: 1456–1470. Lowerre, George F. and Joseph M. Scandura 1973 Conceptually based development and evaluation of individualized materials for critical reading based on logical inferences. Reading Research Quarterly 9: 186–205.

114

Inference and reasoning in discourse comprehension

Lucas, Margery M., Michael K. Tanenhaus, and Greg N. Carlson 1990 Levels of representation in the interpretation of anaphoric reference and instrument inference. Memory and Cognition 18: 611–631. Lutz, Mark F. and Gabriel A. Radvansky 1997 The fate of completed goal information. Journal of Memory and Language 36: 293–310. Lyons, John 1977 Semantics, Volume 2. Cambridge: Cambridge University Press. MacDonald, Maryellen C. and Marcel A. Just 1989 Changes in activation levels with negation. Journal of Experimental Psychology: Learning, Memory, and Cognition 15: 633–642. Magliano, Joseph P. and Michelle C. Schleich 2000 Verb aspect and situation models. Discourse Processes 29: 83–112. Mason, Robert A. and Marcel A. Just 2004 How the brain processes causal inferences in text. Psychological Science 15: 1–7. McKoon, Gail and Roger Ratcliff 1986 Inferences about predictable events. Journal of Experimental Psychology: Learning, Memory, and Cognition 12: 82–91. McKoon, Gail and Roger Ratcliff 1988 Contextually relevant aspects of meaning. Journal of Experimental Psychology: Learning, Memory, and Cognition 14: 331–343. McKoon, Gail and Roger Ratcliff 1989 Inferences about contextually defined categories. Journal of Experimental Psychology: Learning, Memory, and Cognition 15: 1134–1146. McKoon, Gail and Roger Ratcliff 1992 Inference during reading. Psychological Review 99: 440–466. McNamara, Danielle S. and Walter Kintsch 1996 Learning from texts: Effects of prior knowledge and text coherence. Discourse Processes 22: 247–288. Miller, George A. and Stephen Isard 1963 Some perceptual consequences of linguistic rules. Journal of Verbal Learning and Verbal Behavior 2: 217–228. Miller, James R. 1980 The role of knowledge and text structure in prose comprehension. Unpublished manuscript, Department of Psychology, University of Colorado, Boulder. Millis, Keith K. and Arthur C. Graesser 1994 The time-course of constructing knowledge-based inferences for scientific texts. Journal of Memory and Language 33: 583–599. Morrow, Daniel G. 1985 Prepositions and verb aspect in narrative understanding. Journal of Memory and Language 24: 390–404. Morrow, Daniel G., Steven L. Greenspan and Gordon H. Bower 1987 Accessibility and situation models in narrative comprehension. Journal of Memory and Language 2: 165–187. Münte, Thomas F., Kolja Schiltz and Marta Kutas 1998 When temporal terms belie conceptual order. Nature 395: 71–73.

Murray Singer and R. Brooke Lea

115

Murray, John D., Celia M. Klin and Jerome L. Myers 1993 Forward inferences in narrative text. Journal of Memory and Language 32: 464–473. Myers, Jerome L. and Edward J. O’Brien 1998 Accessing the discourse representation during reading. Discourse Processes 26: 131–158. Myers, Jerome L., Makiko Shinjo and Susan A. Duffy 1987 Degree of causal relatedness and memory. Journal of Memory and Language 26: 453–465. Newell, Allen and Herbert A. Simon 1972 Human Problem Solving. Englewood Cliffs, NJ: Prentice Hall. Noordman, Leo G. M., Wietske Vonk and Henk J. Kempff 1992 Causal inferences during the reading of expository texts. Journal of Memory and Language 31: 573–590. Noveck, Ira A. 2001 When children are more logical than adults: Experimental investigations of scalar implicature. Cognition 78: 165–188. O’Brien, Edward J. and Jason E. Albrecht 1991 The role of context in accessing antecedents in text. Journal of Experimental Psychology: Learning, Memory, and Cognition 17: 94–102. O’Brien, Edward J. and Jason E. Albrecht 1992 Comprehension strategies in the development of a mental model. Journal of Experimental Psychology: Learning, Memory, and cognition 18: 777–784. O’Brien, Edward J., Jason E. Albrecht, Christopher M. Hakala and Michelle L. Rizzella 1995 Activation and suppression of antecedents during reinstatement. Journal of Experimental Psychology: Learning, Memory, and Cognition 21: 626–634. O’Brien, Edward J., Robert F. Lorch and Jerome L. Myers 1998 Memory-based text processing [Special issue]. Discourse Processes 26 (2–3). O’Brien, Edward J., Michelle L. Rizzella, Jason E. Albrecht and Jennifer G. Halleran 1998 Updating a situation model: A memory based test processing review. Journal of Experimental Psychology: Learning, Memory, and Cognition 24: 1200–1210. Otten, Marte and Jos J. A. van Berkum 2008 Discourse-based word anticipation during language processing: Prediction of priming? Discourse Processes 45: 464–496. Paris, Scott G. 1976 Propositional Logical Thinking and Comprehension of Language Connectives: A Developmental Analysis. Oxford: Mouton. Perfetti, Charles A. and Susan R. Goldman 1976 Discourse memory and reading comprehension skill. Journal of Verbal Learning and Verbal Behavior 14: 33–42. Perrig, Walter and Walter Kintsch 1985 Propositional and situational representations of text. Journal of Memory and Language 24: 503–518. Pinker, Steven 1994 The Language Instinct. New York: William Morrow and Co.

116

Inference and reasoning in discourse comprehension

Potts, George R. 1972 Information processing strategies used in the encoding of linear orderings. Journal of Verbal Learning and Verbal Behavior 11: 727–740. Potts, George R., Janice M. Keenan and Jonathan M. Golding 1988 Assessing the occurrence of elaborative inferences: Lexical decision versus naming. Journal of Memory and Language 27: 399–415. Rader, Aaron W. and Vladimir Sloutsky 2001 Conjunctive bias in memory representations of logical connectives. Memory and Cognition 29: 838–849. Rader, Aaron W. and Vladimir Sloutsky 2002 Processing of logically valid and logically invalid conditional inferences in discourse comprehension. Journal of Experimental Psychology: Learning, Memory, and Cognition 28: 59–68. Radvansky, Gabriel A. and David E. Copeland 2000 Functionality and spatial relations in memory and language. Memory and Cognition 28: 987–992. Radvansky, Gabriel A. and David E. Copeland 2001 Working memory and situation model updating. Memory and Cognition 29: 1073–1080. Ratcliff, Roger 1978 A theory of memory retrieval. Psychological Review 85: 59–108. Reder, Lynne M. and Gail W. Kusbit 1991 Locus of the Moses Illusion: Imperfect encoding, retrieval, or match. Journal of Memory and Language 30: 385–407. Rinck, Mike, Andrea Hahnel and Guido Becker 2001 Using temporal information to construct, update, and retrieve situation models of narratives. Journal of Experimental Psychology: Learning, Memory, and Cognition 27: 67–80. Rinck, Mike, Andrea Hahnel, Gordon H. Bower and Ulrich Glowalla 1997 The metrics of spatial situation models. Journal of Experimental Psychology: Learning, Memory, and Cognition 23: 622–637. Rips, Lance J. 1983 Cognitive processes in propositional reasoning. Psychological Review 90: 38–71. Rips, Lance J. 1994 The psychology of proof: Deductive reasoning in human thinking. Cambridge, MA: The MIT Press. Rizzella, Michelle L. and Edward J. O’Brien 1996 Accessing global causes during reading. Journal of Experimental Psychology: Learning, Memory, and Cognition 22: 1208–1218. Rodrigues, Amariles A., Maria da-G. B. B. Dias and Antonio Roazzi 2002 Raciocínio lógico na compreensão de texto / logical reasoning in text comprehension. Estudos De Psicologia 7: 117–132. Rumelhart, David E. and James L. McClelland 1986 Parallel distributed processing: Explorations in the microstructure of cognition, Volume 1. Cambridge, MA: MIT Press.

Murray Singer and R. Brooke Lea

117

Sanford, Anthony J. and Simon C. Garrod 1998 The role of scenario mapping in text comprehension. Discourse Processes 26: 159–190. Schank, Roger C. 1976 The role of memory in language processing. In: Charles Cofer (ed.), The structure of human memory. San Francisco: W. H. Freeman Press. Schank, Roger C. and Robert P. Abelson 1977 Scripts, Plans, Goals, and Understanding. Hillsdale, NJ: Erlbaum. Schmalhofer, Franz, Mark A. McDaniel and Dennis E. Keefe 2002 A unified model for predictive and bridging inferences. Discourse Processes 33: 105–132. Searle, John R. 1975 Indirect speech acts. In: Peter Cole and Jerry Morgan. (eds.), Syntax and Semantics 3: Speech acts, 59–82. New York: Academic Press. Singer, Murray 1976 Thematic structure and the integration of linguistic information. Journal of Verbal Learning and Verbal Behavior 15: 549–558. Singer, Murray 1980 The role of case-filling inferences in the coherence of brief passages. Discourse Processes 3: 185–201. Singer, Murray 1993 Causal bridging inferences: Validating consistent and inconsistent sequences. Canadian Journal of Experimental Psychology 47: 340–359. Singer, Murray and Fernanda Ferreira 1983 Inferring consequences in story comprehension. Journal of Verbal Learning and Verbal Behavior 22: 437–448. Singer, Murray and Michael Halldorson 1996 Constructing and validating motive bridging inferences. Cognitive Psychology 30: 1–38. Singer, Murray, Michael Halldorson, Jeffrey C. Lear and Peter Andrusiak 1992 Validation of causal bridging inferences. Journal of Memory and Language 31: 507–524. Singer, Murray, Dana Harkness and Susan T. Stewart 1997 Constructing inferences in expository text comprehension. Discourse Processes 24: 199–228. Singer, Murray and Walter Kintsch 2001 Text retrieval: A theoretical exploration. Discourse Processes 31: 27–59. Singer, Murray, Russell Revlin and Michael Halldorson 1990 Bridging-inferences and enthymeme. In: Arthur C. Graesser and Gordon H. Bower (eds.), The Psychology of Learning and Motivation, Volume 25, 35–51. New York: Academic Press. Smallwood, Jonathan, Merrill McSpadden and Jonathan W. Schooler 2008 When attention matters: the curious incident of the wandering mind. Memory and Cognition 36: 1144–1150. Sperber, Dan and Deirdre Wilson 1986/95 Relevance: Communication and Cognition. Cambridge, MA: Harvard University Press. [Second edition with Postface 1995]

118

Inference and reasoning in discourse comprehension

St. George, Marie, Suzanne Mannes and James E. Hoffman 1997 Individual differences in inference processing. Journal of Cognitive Neuroscience 9: 776–787. Stewart, Andrew J., Evan Kidd and Matthew Haigh 2009 Early sensitivity to discourse-level anomalies: Evidence from self-paced reading. Discourse Processes 46: 46–69. Swinney, David A. 1979 Lexical access during sentence comprehension: (Re)consideration of context effects. Journal of Verbal Learning and Verbal Behavior 18: 645–659. Till, Robert E., Ernest F. Mross and Walter Kintsch 1988 Time course of priming for associate and inference words in a discourse context. Memory and Cognition 16: 283–298. Trabasso, Thomas, Paul van den Broek and S. Y. Suh 1989 Logical necessity and transitivity of causal relations in stories. Discourse Processes 12: 1–25. Traxler, Matthew J., Anthony J. Sanford, Joy P. Aked and Linda M. Moxey 1997 Processing causal and diagnostic statements in discourse. Journal of Experimental Psychology: Learning, Memory, and Cognition 23: 88–101. Tulving, Endel 1983 Elements of Episodic Memory. New York: Oxford University Press. Tversky, Amos and Daniel Kahneman 1973 Judgment under Uncertainty: Heuristics and Biases, Volume 13. Oxford, England: Oregon Research Institute. van Berkum, Jos J. A., Peter Hagoort and Colin M. Brown 1999 Semantic integration of sentences and discourse: evidence from the N400. Journal of Cognitive Neuroscience 11: 657–671. van Berkum, Jos J. A., Pienie Zwitserlood, Peter Hagoort and Colin M. Brown 2003 When and how do listeners relate a sentence to the wider context? Evidence from the N400 effect. Cognitive Brain Research 17: 701–718. van den Broek, Paul and Robert F. Lorch Jr.1993 Network representations of causal relations in memory for narrative texts: Evidence from primed recognition. Discourse Processes 16: 75–98. van Dijk, Teun A. and Walter Kintsch 1983 Strategies of Discourse Comprehension. New York: Academic Press. Vonk, Wietske, Lettica G. M. M. Hustinx and Wim H. G. Simons 1992 The use of referential expressions in structuring discourse. Language and Cognitive Processes 7: 301–333. Wason, Peter C. 1966 Reasoning. In: B. Foss (ed.), New Horizons in Psychology, Volume 1, 135–151. Harmondsworth, Middlesex, England: Penguin. Weiner, Bernard 1985 An attributional theory of achievement motivation and emotion. Psychological Review 92: 548–573. Weingartner, Kristin M., Alexandria E. Guzman, William H. Levine and Celia M. Klin 2003 When throwing a vase has multiple conseqences: Minimal encoding of predictive inferences. Discourse Processes 36: 131–146.

Murray Singer and R. Brooke Lea

119

Whitney, Paul 1986 Processing category terms in context: Instantiations as inferences. Memory and Cognition 14: 39–48. Wilson, S. G., Mike Rinck, Timothy P. McNamara, Gordon H. Bower and Daniel G. Morrow 1993 Mental models and narrative comprehension: Some qualifications. Journal of Memory and Language 32: 141–154. Yang, Chin Lung, Charles A. Perfetti and Franz Schmalhofer 2007 Event-related potential indicators of text integration across sentence boundaries. Journal of Experimental Psychology: Learning, Memory, and Cognition 33: 55–89. Zwaan, Rolf A. 1996 Processing narrative time shifts. Journal of Experimental Psychology: Learning, Memory, and Cognition 22: 1196–1207. Zwaan, Rolf A., Carol J. Madden and Shannon N. Whitten 2000 The presence of an event in the narrated situation affects its activation. Memory and Cognition 28: 1022–1028. Zwaan, Rolf A., Joseph P. Magliano and Arthur C. Graesser 1995 Dimensions of situation-model construction in narrative comprehension. Journal of Experimental Psychology: Learning, Memory, and Cognition 21: 386–397. Zwaan, Rolf A. and Gabriel A. Radvansky 1998 Situation models in language comprehension and memory. Psychological Bulletin 123: 162–185. Zwaan, Rolf A., Robert A. Stanfield and Richard H. Yaxley 2002 Language comprehenders mentally represent the shapes of objects. Psychological Science 13: 168–171. Zwaan, Rolf A. and Herre van Oostendorp 1993 Do readers construct spatial representations in naturalistic story comprehension. Discourse Processes 16: 125–143.

“Semantic” principles

5.

Conceptual principles and relations Małgorzata Fabiszak

1.

Introduction

A cognitive approach to language, similarly to a pragmatic perspective on language, rejects the notion of the ideal native speaker/hearer and, while it attempts to uncover the conceptual processes underlying language use, it stresses the usagebased nature of language and the embodied, interactional, culture- and history-embedded character of language evolution and development. The present chapter focuses on the organizational principles and relations transparent in the systems of knowledge motivating linguistic behaviour. In particular, it presents the current state of research on such issues as prototypical structure of categories, conceptual networks, idealized cognitive models, frames, metonymy and metaphor. It also looks at the online processes of meaning construction integrating those conceptual structures (understood both as developed in an individual organism’s embodied experience and socialization history, and as shared knowledge of the speech community) with the context of a communicative act. This approach views semantics and pragmatics as a continuum, rather than as discrete modules (see Leech 1983; Cruse 2000; Croft and Cruse 2004; Evans and Green 2006).

2.

Categorization and prototypes

One of the basic cognitive processes is categorization. The cognitive linguistic approach to the categorization of semantic and lexical concepts is inspired by psychological investigations into the organization of memory, which reject the necessary and sufficient lists of features as a means of defining concepts (the classical view – on the history of lexical semantics see Geeraerts 2009). Categories are viewed as possessing a gradable structure. Numerous studies (Markman 1989; Mervis and Rosch 1981; Rosch 1973, 1978; Lloyd and Rosch 1978) into language learning, speed of processing, expectation, association and similarity judgements show that the members of a given category do not enjoy the same status. Central members of a category are learnt faster than the peripheral members. For example, the theory predicts that apples, oranges and bananas are the first types of FRUIT learnt by Western European or American children. It is important to emphasize that the structure of the category is culture-dependent, so that in North Africa dates and figs may occupy the central position (Kövecses 2006). Good members of a category are recognised faster than poor members (speed of processing). Also, nam-

124

Małgorzata Fabiszak

ing the superordinate category creates expectations, which in priming studies facilitate the recognition of central members more than the peripheral members. For instance, the prime bird will facilitate the recognition of robin as a real word more than penguin, i.e. the subjects in an experimental situation will take longer to recognize the word penguin than the word robin, testifying to their unequal status as members of the category bird. Good members of a category are listed more frequently and earlier than poor members. Finally, subjects claim that a penguin is more similar to a robin than vice versa, falsifying the classical claim in logic that similarity is reciprocal. 2.1.

Prototypical structure of linguistic categories

These psycholinguistic findings have spurred attempts to create psychologically plausible theories of meaning. Lakoff’s notion of idealized cognitive models discussed in the next section is one such attempt. A major contribution of cognitive linguistics to linguistic theory is an insistence on treating theoretical linguistic concepts in the same way as other cognitive categories, i.e. they undergo the same prototypicality effects such as members’ centrality, category overlap and fuzzy borders. Instead of struggling to define clear-cut boundaries between linguistic modules and then positing interfaces between them, cognitive linguistics views linguistic phenomena as conceptual continua. The distinction between semantics and pragmatics is a case in point. Cognitive linguistics rejects the approach within which these are two separate disciplines and attempts at integrating both semantic and pragmatic methodological tools to ensure an adequate description of language use. Saeed (2003), for example, stresses that deictic expressions like bring/take or today/tomorrow cannot be effectively defined in terms of only the coded semantic meaning, because their meaning in use is context-dependent. An application of prototype structure to linguistic categories at various levels of language (phonology, morphology, syntax, semantics) can be found in Taylor (2004). A systematic account of prototype effects in semantic change is offered in Geeraerts (1997). All of these findings strongly suggest that when investigating semantic and pragmatic phenomena it is necessary to consider the prototypicality effects within language structure.

3.

Radial structure of idealized cognitive models

The discovery of the prototypical structure of categories and the unequal status of category members led to the development of the theories of the radial structure of categories, represented in cognitive semantics by Lakoff’s (1990 [1987]: 74–76, 79–84) analysis of the category MOTHER. Lakoff calls the knowledge structures into which we organise our biological and socio-cultural experience, and which are

Conceptual principles and relations

125

stored in a form of relatively stable mental representations, “idealized cognitive models” (ICMs). An ICM for the socio-biological category MOTHER is a complex structure consisting of a number of cognitive models, such as the birth model, the genetic model, the nurturance model and the marital model, all of which constitute the ontology of the category. The evidence for such a range of models comes from the linguistic expressions that can be used to access each of them: surrogate mother, biological mother, donor mother, adoptive mother and stepmother. These models can give rise to metaphoric extensions such as in examples (1) and (2) (taken from Lakoff 1990 [1987]: 76): (1) Necessity is the mother of invention. (2) He wants his girlfriend to mother him. In (1), the birth model is the source of the metaphoric mapping and highlights the cause-effect relation between necessity and invention, constructing them in analogy to the mother- child relation. In (2), it is the nurturance model which serves as the source for the understanding of the expectations that “his girlfriend” has to meet. In accord with the prototype model of categories, not all of the models enumerated above have an equal status in the category. The central member of a socially constructed category, such as MOTHER, is called a stereotype. In American culture, as represented by Lakoff (1990 [1987]), such a social stereotype of a mother is a housewife-mother. Its strength is testified by the naturalness of sentence (3) below and semantic anomaly of sentence (4) (Lakoff 1990 [1987]: 81): (3) She is a mother, but she isn’t a housewife. (4) She is a mother, but she’s a housewife. Clearly, being a housewife is a part of the meaning of being a mother. To sum up our discussion of the MOTHER ICM, its propositional structure consists of a range of converging cognitive models with a central model of a mother who gives birth to a child, is genetically related to her/him, and rears the child together with her husband – the child’s father. This part of the ICM shows a radial structure with the conventional variations on the central model, which have to be learnt and cannot be predicted by semantic rules. The second model of MOTHER is the social stereotype of the “housewife-mother”, which often stands metonymically for the whole category, as evidenced in sentences (3) and (4). The propositional content of an ICM can also recruit our knowledge of image schemata (see Section 4 below) and serve as a basis for metaphorical and metonymic mappings (Section 8).

126 3.1.

Małgorzata Fabiszak

Idealized cognitive models and deixis

The theory of categorization in terms of idealized cognitive models has been employed by Marmaridou (2000: 65–116) to explain the intricacies of deixis1. She starts by reviewing earlier approaches, which usually draw a strict dividing line between the semantics of deictic expressions and their pragmatic use. She stresses that such an account has little explanatory value as it unnecessarily multiplies the number of deictic categories. To solve these problems she suggests looking into the cognitive processes responsible for the understanding of deixis. She argues that deixis is based on our bodily, sensory-motor experience of pointing and proposes a “pointing out ICM” as the representation underlying the interpretation of deictic expressions. As Marmaridou (2000: 99) claims, “deixis is a grammatical category which reveals our conceptualization of human beings as objects in space and of human language as an object in time.” The ontology of the “pointing out ICM” consists, then, of discourse participants, sharing of attention between them, their location in time and the utterance location in space as well as the participants’ social roles. The understanding of time, as well as that of the social structure, is metaphorically motivated by space. Marmaridou (2000: 100) expresses her position in the following words: The ICM of deixis that is proposed here involves the linguistic act of pointing to an entity in space, performed by an authorised speaker and directed to an unfocussed addressee. Accordingly, a deictic expression is one that builds a mental space in which the speaker and the addressee are co-present at a given point in time. In short, the mental space evoked by a deictic expression involves the conceptualization of a deictic centre.

Marmaridou’s proposal thus gives a unified account of the possible cognitive representation of deixis. The internal structure of the idealized cognitive model of deixis is organized by a number of image schemata, that of CENTRE - PERIPHERY, NEAR FAR and UP - DOWN . The notion of image schemata and the nature of their experiential grounding are explained in Section 4 below.

4.

Image schemata

Image schemata are considered to be abstractions from our sensory-motor experience as biological organisms (Johnson 1987; Lakoff 1987: 271–273; Lakoff and Johnson 1999: 30–36). In discussing the embodied nature of the CONTAINER schema, Johnson (1987: 30) gives the following account of our everyday experience motivating it: Consider, for example, only a few of the in-out orientations that might occur in the first few minutes of ordinary day. You wake out of the deep sleep and peer out from beneath the covers into your room. You gradually emerge out of your stupor, pull yourself out of the covers, climb into your robe, stretch out your limbs, and walk in a daze out of the bedroom and into the bathroom.

Conceptual principles and relations

127

In this fragment Johnson refers to both our spatial experience of living in containers (covers in the bed, the robe, the bedroom and the bathroom) as well as metaphorically motivated containers (sleep, stupor, daze), which are based on the conceptual metaphor STATES ARE CONTAINERS . The evidence for the embodied nature of human cognition comes not only from the linguistic expressions, but also from psycholinguistic experiments. For example, Richardson et al. (2003) demonstrate that the comprehension of verbs such as push and respect involves image schemata for spatial representation such as horizontal movement for push and vertical UP - DOWN arrangement for respect. Their results support the claim of the sensory-motor, i.e. embodied nature of linguistic representation. Moreover, image schemata have also proved useful in training a computer programme to discriminate from a number of different schematic spatial relations and evaluating them in terms of their prototypicality for the use of such linguistic expressions as English through, Russian iz-pod, and Mixtec sini (Regier 1996). Gibbs and Colston (1995) revise psychological developmental literature and show that even in infancy children are already able to identify various schematic relationships, which are the bases for the image schemata. Rohrer (2005) discusses evidence from fMRI studies showing that the same areas of motor cortex light up when subjects experience tactile stimulation of their hands and when they process sentences with the word hand used either literally or metaphorically. The activated area decreases in size, but clearly a part of the motor cortex participates in the comprehension of language. All of these studies present converging evidence on the validity of image schemata as part of the human cognitive system. 4.1.

Image schemata and value systems

Preconceptual image schemata perform not only an ontological function by providing the experiential bases of cognition but also carry an inherent axiological value. Krzeszowski (1997), in his attempt to develop axiological semantics, shows that, for example, orientational schemata, such as UP - DOWN , FRONT- BACK , FAR NEAR, CENTRE - PERIPHERY are particularly predisposed to intrinsic valuation. HAPPY is UP and SAD is DOWN exactly because UP is GOOD and DOWN is BAD . These positive/negative evaluations attached to the opposing poles of scalar schemata are not surprising, but it is also true of many other schemata including the CENTRE - PERIPHERY, PART- WHOLE , LINK and CONTAINER schemata. The evaluation of the CONTAINER schema is particularly interesting, because of its ambivalent nature. It arises from the experience of growing inside the mother’s womb and of birth understood as leaving the original CONTAINER. Krzeszowski (1997: 42) explains the ambivalence in the following words: On the one hand, we experience getting out of the container as being born and as gaining freedom. On the other hand, getting out of the original container may be experienced as

128

Małgorzata Fabiszak

leaving the security of the protective confines of a shelter and as being exposed to various external dangers. Therefore, we often long to return to some protective container. Consequently the PLUS - MINUS polarity is not consistently associated with the IN - OUT orientation of this version of the CONTAINER schema.

As a result, CONTAINERS , such as houses, hospitals, schools and the domains of experience understood in terms of CONTAINERS , such as emotions, relationships, states inherit the ambivalent axiology of the schema and can be viewed alternatively as shelters providing safety and protection or traps constraining the freedom of movement. In this way, the values inherent in image schemata influence the categorizations we make. If something is UP it will be positively evaluated, if something is DOWN it will be negatively evaluated, and if an experience is categorized in terms of the CONTAINER schema it may turn out to be either positive or negative. Axiology is inherent in our categorizations, and in this way influences our systems of values, beliefs, and social interactions. Combinations of value-laden image schemata may result in axiological clashes. Krzeszowski (1997: 140) explains this notion with example (5) below: (5) The robbers are near the bank. While the preposition near is inherently positive (in contrast to far from), the subject of the sentence – the robbers – is evaluated negatively, which results in an axiological clash, so that the actual value of the sentence (as opposed to the inherent value of its elements as stored in the mental representation) is negative. The value of the participant of the relation overrides that of the relation itself. 4.2.

Image schemata and speech acts

Axiological clashes may occur not only at the sentential level, as has been shown in the previous section, but also at the discourse level. In such a case, discourse participants often attempt to resolve the clash and restore the axiological balance of a discourse situation. Krzeszowski (1997: 248) demonstrates this using an example which he adapted from Vlassova and Vlassova (1971: 50) quoted here as (6): (6) A: I hope I haven’t interrupted … B: Of course not. Have a drink? A: Thanks. Beer? B: Sorry. We haven’t a fridge – we send out for ice. What about a Scotch? A: A small one, if you don’t mind. I’m not very keen on hard liquor. B: On the rocks? A: Plenty of soda – if you aren’t short. B: Here you are. A: Thanks a lot.

Conceptual principles and relations

129

The illocutionary intention of A in this speech act is to get a drink, in particular a beer. B’s admitting to not having one constitutes an axiological clash at the discourse level, which is partially resolved by suggesting another drink. A’s reluctant acceptance of the offer restores the axiological balance of the discourse situation to a certain degree, i.e. the conversation may go on unimpeded. This example shows how stored representations of the image schema of BALANCE and its inherent value can motivate the linguistic behaviour of speech act participants. An analysis in terms of axiologically laden image schemata explains what type of representation at the conceptual level is responsible for a specific linguistic behaviour. Grice’s cooperative principle, which could be invoked here, operates at the level of social interaction, describes linguistic behaviour of conversation participants. Image schemata and axiology attempt to explain the processes at the cognitive level. The importance of the basic image schema of FORCE has been elaborated on by Talmy (1988) in terms of force dynamics. The propositional structure of a forcedynamic idealized cognitive model includes the AGONIST , the ANTAGONIST and the FORCE operating between them. The tendency for rest and tendency for action attributed to the two participant entities are in conflict and, depending on the balance of strengths between them, may bring about a change. Force dynamics originating from our experience with physical force can be metaphorically projected into the social realm. Sociodynamics explains how social interaction can be understood in terms of pressure exerted by one discourse participant on another. These observations are further developed by Marmaridou (2000) to account for the cognitive motivation of speech acts. 4.3.

Image schemata and modality

Talmy (1988) also proposes that force dynamics is linguistically expressed by the modal system. Modality thus conceived is a conceptualization of an interaction of the sentient being with the surrounding social and biological environment. Sweetser (1990: 50) presents historical evidence for the development of the modal meanings of verbs from the physical force sense. Old English magan had several senses, among them ‘to be able’ and ‘to be strong’. The deontic meaning expressing obligation, permission or ability developed first and was then metaphorically extended to epistemic uses expressing possibility and necessity. Thus, Sweetser argues for the polysemy- and not homonymy-based understanding of English modal verbs. Her proposal is criticized by Papafragou (2000), who opts for a monosemic account of modality and a division of labour between the semantics and the pragmatics. The studies of Johnson (1987), Krzeszowski (1997), Talmy (1988) and Sweetser (1990) show that image schemata are relevant not only in the processes of categorization, but also at the level of discourse. In the following section we will demonstrate the pragmatic relevance of yet another cognitive approach to meaning representation, i.e. frame semantics.

130 5.

Małgorzata Fabiszak

Frames

Frame theory, suggested by Fillmore (1982, 1985), stems from his dissatisfaction with the description of verb semantics within the generative account of language. The first attempts at producing a more adequate theory of the meaning of verbs is Case Grammar (Fillmore 1968), which later developed into frame semantics (Fillmore 1982, 1985). The early studies were devoted to groups of semantically related verbs, such as verbs of judging or the commercial transaction frame. In his attempt to provide a satisfactory definition of verbs such as blame, accuse, criticize and buy, sell, spend and cost, Fillmore decided that the necessary and sufficient conditions proposed within the so-called truth semantics were not very successful. Instead, the meaning of these words should be understood in terms of typical situations, with the participants, such as the judge and the defender for the judging frame and the buyer, the seller and the commodity for the transaction frame. In this way, Fillmore rejected the dictionary definition of word meaning as inadequate and opted for the encyclopaedic approach, showing that more than purely linguistic knowledge is recruited in the understanding of words. 5.1.

Frames and presupposition

Frame semantics offers an elegant solution to problems of pragmatic presupposition. In Fillmore (1985: 249–251), an interpretation of presuppositional phenomena resulting from the particular word meaning are presented using the example of the verb regret: (7) John regretted signing the letter. (8) John didn’t regret signing the letter. Sentence (7) presupposes that John has signed the letter. However, if the proposition of utterance (7) is negated, as in (8), two interpretations emerge. On default interpretation, the regret is negated. John has signed the letter, but he does not regret it. But it is also possible to interpret sentence (8) in terms of presupposition negation, i.e. John does not regret signing the letter, because he has not signed it. The frame semantic analysis of these examples resolves the conflict in the following way. The frame for REGRET consists of the mental state linked to an accomplished action, which caused the feeling. When negation is used, its scope operates inside the frame and refers to the concept of regretting only – this is the first interpretation, when the mental state does not take place, but the action potentially leading to it has been accomplished. On the second account, the scope of negation is extended to the entire frame and both the mental state and the action are negated. In this way, frame semantics contributes to the explanation of the cognitive processes behind the interaction between negation and presupposition.2

Conceptual principles and relations

5.2.

131

Frames and speech acts

Frame theory can also operate at higher levels of linguistic structure, for example discourse. The frames account of speech acts presupposes that the interaction between the speaker and the hearer or between text producer and text reader is regulated in accord with the expectations embedded in the frame for particular types of discourse. They determine what forms of linguistic behaviour are considered appropriate for a given situation (compare the force-dynamic account of speech acts in Section 4.2 above). Another application of Fillmore’s theory focuses on discourse framing. It emphasizes that a certain choice of words creates particular expectations in the hearer, which may be exploited by the speaker. For instance, frame-shifting in one-line jokes (Coulson 2001: 49–69) operates through the violation of the hearer’s expectation concerning typical situations of word use: (9) By the time Mary had her fourteenth child, she’d finally run out of names to call her husband. (Coulson 2001: 49) The typical situation evoked by the first part of the sentence, that of naming a child after his or her birth has to be back-processed and reinterpreted in view of a new frame offered by the sentence ending to call her husband, i.e. the frame of marital arguments. In moral discourse, a specific framing of the situation may result in its diverging evaluation. Fillmore (1982) shows how a choice of a predicating adjective from a series STINGY- GENEROUS vs. THRIFTY- WASTEFUL can construe a given situation differently. For example, a person who is economical with money can be described from a perspective of their relationship with other people as stingy or from a perspective of using resources wisely and carefully as thrifty. Also, in discourse on abortion the pro-lifers choose to speak about the UNBORN BABY, while the pro-choice group opt for the term FETUS . The selected noun determines alternative framings of moral choices. Apart from word choice, evaluative framing effects can also be achieved through the use of conceptual metaphor, which is discussed in the following section.

6.

Metaphor

Relations between idealized cognitive models or frames can take the form of metaphorical projections or mappings (Lakoff and Johnson 1980, 1999). Cognitive linguistics postulates that some of these relations may be relatively stable in nature and may be the means through which we categorize and understand our socio-cultural experience. The basic type of conceptual metaphors is ontological, which allows us to understand abstract concepts in terms of things3.

132

Małgorzata Fabiszak

Conceptual metaphors understood as projections from one domain of experience to another are often formulated on the basis of an analysis of linguistic expressions, which demonstrate a systematic recurrence of similar patterns of meaning creation. For example, the metaphor NATION /STATE IS A FAMILY (Lakoff 1996: 154) correlates the experience of a relatively abstract concept of NATION with that of a FAMILY. Elements of the NATION domain are linked to those closer to our experience, i.e. the domain of FAMILY. It links the government with a father/parent, country citizens with children and internal policy with parental care. These correlations in experience facilitate the creation of inferences, which provide us with ways of conceptualizing abstract concepts. It is important to stress, however, that only some parts of the structure of the “source domain” are actually mapped onto the “target domain”. The restrictions on the mappings are expressed in terms of the unidirectional hypothesis (Lakoff 1993), which states that the mappings are directed from the concrete domains of experience into the abstract domains. This tendency, however, does not apply so much to complex structural metaphors like the one presented above, but rather to primary metaphors (Grady 1997), that means those whose source domains originate in our sensory motor experience represented by image schemata, which is then correlated with our subjective experience or judgement, see example (10): (10)

HELP IS SUPPORT

Subjective Judgement: Assistance Sensorimotor Domain: Physical support Example: “Support your local charities.” Primary Experience: Observing that some entities and people require physical support in order to continue functioning. (Lakoff and Johnson 1999: 52; on the basis of Grady 1997) Several primary metaphors can be combined to form complex metaphors underlying our subjective experience. 6.1.

Metaphor and belief systems

Such regular metaphorically motivated processes of understanding underlie our value and belief systems. The abstract notion of morality is metaphorically grounded in our experience of force dynamics and orientation in space. Lakoff and Johnson (1999: 300) propose the following representation for this conceptual metaphor: (11)

THE MORAL STRENGTH METAPHOR

Being Upright Being Low

f Being Good f Being Bad

Conceptual principles and relations

Falling A Destabilizing Force Strength (To Resist)

133

f Doing Evil f Evil (Internal or External) f Moral Virtue

The linguistic evidence for such mappings comes from the following examples (Lakoff and Johnson 1999: 299): (12) He’s an upstanding citizen. She is on the up and up. She’s as upright as they come. That was a low thing to do. He’s underhanded. I would never stoop to such a thing. The moral strength metaphor can be combined with a number of other conceptual metaphors and result in various ways of understanding morality, such as moral accounting or morality as nurturance. They all contribute to our understanding of moral discourse and guide our behaviour and thinking. In combination with the NATION / STATE IS A FAMILY metaphor, it gives rise to what Lakoff (1996) called “moral politics”, which will be discussed in the section below. 6.2.

Metaphor and discourse analysis

In a democratic society the importance of political discourse cannot be underestimated. What and, in particular, how politicians and political commentators speak about social, economic and international problems influences the way these problems are understood and what solutions are offered and implemented. Schön (1993) has shown how phrasing the discourse about urban renewal and development sets social policy. The metaphor applied by the government experts was that of SLUM IS A DISEASE . Such framing of social discourse sets the slum community in the role of a patient, if not a germ, on the healthy body of a city, while the government officials are set in the role of doctors. Constructing urban development in such terms creates a clear division between the slum community and state administration. It forestalls cooperation between these two social participants by setting them in opposition to each other. An alternative framing could describe the slum as a natural community, which can solve its problems from within, with a minimal incentive and aid from outside. Naming and framing of social issues is evaluative through the implicit use of such generative metaphors. A combination of NATION / STATE IS A FAMILY metaphor with the moral accounting metaphor motivates the entire political discourse in the United States (Lakoff 1996). The two parties frame their discourse in terms of a specific perspective on the role of the government. The Conservatives base their policies on the STRICT FATHER model, which enhances such values as self-discipline and self-reliance.

134

Małgorzata Fabiszak

When translated into political programmes, STRICT FATHER morality translates into financing the law enforcement and the military, as political strength is understood as military power. The Liberals adhere to the NURTURANT PARENT model, which stresses the importance of such values as protection of the weak and empathy. The liberal political programmes motivated by this metaphor focus on increasing social security and the welfare of all citizens. Discourse analytical studies originating from the various strands of critical discourse analysis (e.g. Fairclough 1989; Wodak and Krzy˙zanowski 2008; Chilton 2004; van Dijk 1998) are often inspired by the findings from the cognitive metaphor theory. This cross-breeding results in an increasing body of research represented by, among others, Cameron (2008), Charteris-Black (2004, 2005) Fabiszak (2007), Musolff (2004), Nerlich, Hamilton and Rowe (2002) and Zinken (2004). They analyze systematic patterns of metaphoric expressions in discourse and the role they play in achieving a certain framing of the social events and processes (see Deignan, this volume). Cameron (2008), for example, shows how discourse participants can negotiate, undermine or accept each others’ metaphors in classroom and reconciliation talk, and how these processes of metaphor shifting are indicative of the consensus between the speakers on the representation of certain events. Charteris-Black (2004, 2005) identifies sets of metaphors characteristic of certain politicians as well as political discourses in Britain and the US. Fabiszak (2007) explains how certain stable metaphorical patterns of enemy vilification and the glorious war myth pervade war reporting and at a certain level of generality do not change despite the change in military alliances. Musolff (2004) analyzes how various cultural models inform seemingly common metaphors of the United Europe. He shows that phrases considered translational equivalents and motivated by the same conceptual metaphors may in different language communities lead to different inference structures resulting from a choice of culture specific metaphorical scenarios. Nerlich, Hamilton and Rowe (2002) posit an unsettling question: To what extent has the policy adopted by Great Britain in dealing with foot-and-mouth disease been an effect of framing it in terms of war on the disease? Zinken (2004) shows how quantitative corpus-based research into metaphorical expressions can inform our understanding of cross-cultural differences in metaphor use.4 All of these studies emphasize how cognitive representation in the form of conceptual metaphors and cultural models (idealized cognitive models) structure the way we reason about social and political issues. They demonstrate how the choice of metaphors and framing in discourse influences our perception of the sociopolitical situation and how language, society and cognition interact dynamically to create the public sphere.

Conceptual principles and relations

7.

135

Conceptual blending: mental spaces and conceptual integration

Conceptual metaphor theory, as presented in the previous section, focuses on stable mental representations emerging from our embodied experience in the socio-cultural world. They consist of the source and the target domain and a series of regular mappings which help us understand the more abstract in terms of the more concrete. Mental space theory (Fauconnier 1997, 2007) purports to describe not the stable mental representations, but to capture the very dynamic process of online meaning formation.5 Unlike conceptual metaphor theory, which is considered a two-space model, the mental space theory allows for the simultaneous rise and interaction among more than two spaces, hence it is referred to as a multispace model. The generic space contains all the stable mental representations stored in long-term memory such as image schemas, idealized cognitive models, conceptual metaphors and frames. More specifically constructed input spaces are built on the bases of the generic space knowledge, linguistic and contextual stimuli, but contain only partial representations. Then the online representations from the input spaces are projected into the blended space, where they can undergo a serious of processes, such as composition, completion and elaboration of the elements from the input spaces6. As a result the emergent structure is created in real time for the understanding of a particular situation or utterance. The blending or conceptual integration theory, which developed from the mental spaces theory has been employed to explain a number of linguistic and non-linguistic questions including such pragmatic phenomena as deixis, the strategic use of rhetorical questions in courtroom procedures and children’s play. We will look at these illustrations in turn. 7.1.

Conceptual blending and deixis

In this section we will have a look at how blending theory can facilitate our understanding of how deixis operates in the epistolary genre and in sign language discourse. Herman (1999), in his analysis of the use of deictic expressions in letter writing, relies on the mental spaces theory’s ability to account for the dynamic continuous and changing meaning construction online. He claims that in the epistolary genre, in which the absence of the interlocutor is transformed into the imagined presence, the deictic centre needs to be projected and transposed so that a solitary act of writing or reading a letter can change into a semblance of face-to-face interaction. Liddell (2000) reinterprets some of the Sign Language gestures previously considered phonological gestures as pointing gestures and uses the concepts of the grounded mental spaces (Liddell 1995) and blended spaces (Fauconnier and Turner 2002) to propose a comprehensive account of Sign Language interaction. To argue for the distinction between deictic gestures and articulatory gestures, Lid-

136

Małgorzata Fabiszak

dell shows that the former are directed towards real things in the communication context and not at pre-established loci of signing space predicted by the language system. He also explains how deictic gestures and linguistic signs can be combined without blurring the meaning of the sign, despite the fact that, unlike in spoken languages, they use the same modality. The mental-spaces account is used to demonstrate how the real space and the signing space can be integrated to facilitate the production and construction of meaning. 7.2.

Conceptual blending in court discourse and symbolic play

Pascual (2002, 2006, 2008) proposes that, similarly to Fauconnier and Turner’s (2002) “Debate with Kant” account, Turner’s (2002) “The Dream of the Rood” analysis and Coulson and Oakley’s (2006) “Voting as Speaking”, the enactment of the American courtroom procedure requires the creation of mental spaces of, what she calls, fictive interaction. The fictive interaction emerges from the blend of faceto-face interaction with the courtroom procedure, in which many of the communicative exchanges constitute display talk in Goffman’s sense (1981). For example, during the witness testimony, the questions asked by the attorney are factively7 addressed at the witness, but fictitiously directed at the opposing party with the aim of undermining their event scenario and at the jury in order to convince them to the presented point of view. Such complex mental processing is easily accounted for within the blending theory by positing an online construction of multiple mental spaces, which blend and facilitate the emergence of a uniform discourse space. One of the discursive strategies identified by Pascual (2006, 2008) is the use of rhetorical questions, which serve the three goals at once: (i) to give a brief summary of the opponent’s position, (ii) to raise a question that the jury may potentially want to ask, and (iii) to create the discursive and cognitive space for counterargument to the opponent’s position. These three goals can only be simultaneously achieved if integrated within one blend. As Pascual shows, mental spaces theory is an effective tool for explaining meaning construction in verbal interaction. Sinha (2005) takes it one step further and shows how verbal exchanges and the use of artefacts can be integrated in meaning making. In his study of children’s symbolic play, Sinha (2005) uses mental space theory to explain how fictitious identity is created through the blending of verbal display and the use of props. When the fictitious world is constructed in an act of virtual cognition, the physical world is not replaced by an imaginary one, but the affordances of the real life artefacts are backgrounded and their new role is negotiated verbally. As Fauconnier and Turner (2002) suggest that blending can operate on various levels of generalization and on different types of representation, their theory is probably best suited to account for such complex cases of meaning making.

Conceptual principles and relations

8.

137

Metonymy

Metonymy, similarly to metaphor, is another vital conceptual process organizing the way we think and interact. In its early formulation it is defined as a within-domain mapping (Lakoff and Johnson 1980) and its referential function comes then to the fore. A more sophisticated understanding of the metonymic processes is offered in Croft (2002 [1993]). It is based on Langacker’s (1987, 1991) theory of semantic domains within which meaning of a particular linguistic expression is constructed as the result of profiling against a base in a matrix of domains. When several linguistic expressions are composed together they influence each others’ meaning so that only part of their coded meaning, called the active zone (cf. Taylor, this volume), participates in the creation of meaning. Let us take an example from an accessible presentation of Langackerian Cognitive Grammar by Taylor (2002: 111): (13) a. I had my car serviced. b. I had my car washed and waxed. c. My car got scratched (in the accident). In example (13a) above, the active zone consists of the engine, brakes and other mechanical parts of the car, in (13b) it is the entire surface of the bodywork, while in (13c) it is only a tiny part of the surface that is involved in the action expressed by the verb. Croft (2002 [1993]: 178–179) defines metonymy as domain highlighting within a domain matrix and explains it on the following example: (14) a. Time magazine is pretty vapid. b. Time took over Sunset magazine, and it’s gone downhill ever since. In the encyclopaedic view of meaning the concept activated by the expression Time is our entire knowledge about the Time magazine, together with our knowledge of its printed and online versions and the related knowledge about journalism, editing, and ownership. Thus, the use of the word Time in (14a) and the metonymic use in (14b) both access the same coded meaning structure. However, in Croft’s (2002 [1993]: 179) words: … a secondary domain for magazines is that of the process of publication, in which a publishing company is a salient entity. The metonymy that shifts reference from the magazine to the company also shifts domains from the magazine as an object with semantic content to the domain of publication. We will call this conceptual effect domain highlighting (cf. Cruse 1986: 53), since the metonymy makes primary a domain that is secondary in the literal meaning.

Croft (2002 [1993]) stresses that domain highlighting is not definitional for metonymy, as it also involves a referential function, while domain highlighting is also present in other types of lexical ambiguity.

138

Małgorzata Fabiszak

Taking Croft’s definition as a starting point, Kövecses and Radden (1998) propose a classification of metonymic access paths to the primary domain in the domain matrix. They enumerate several types, for example WHOLE THING FOR PART OF A THING , PART OF A THING FOR THE WHOLE THING , A CATEGORY FOR THE MEMBER OF CATEGORY, A MEMBER OF A CATEGORY FOR THE CATEGORY, INSTRUMENT FOR ACTION , AGENT FOR ACTION , ACTION FOR AGENT, RESULT FOR ACTION , ACTION FOR RESULT and others. The choice of a specific metonymy is governed by two main conceptual principles: HUMAN OVER NON - HUMAN and CONCRETE OVER ABSTRACT . They both predict that human and concrete entities are easier to access than non-human or abstract entities. This claim is in accord with the achieve-human-scale principle of the integrative networks (blending) theory of Fauconnier and Turner (2002: 346–347). The following sections illustrate how the conceptual process of metonymy as discussed above can contribute to our investigation of pragmatic competence. 8.1.

Metonymic motivation of anaphora

Metonymy can constitute the motivation behind the use of certain types of discourse deixis, in particular anaphora. Ruiz de Mendoza and Pérez-Hernández (2001) explain how the meaning of a referential chain can be inferred via two types of metonymy, GENERIC FOR SPECIFIC and EFFECT FOR CAUSE . To illustrate their point, they use two examples (Ruiz de Mendoza and Pérez-Hernández 2001: 349): (15) A: What’s that bird? B: That bird is a robin. (16) A: What is that noise? B: ?That noise is a burglar. The anaphoric relationship in (15) between the generic noun phrase that bird and its specific counterpart robin is motivated by the GENERIC FOR SPECIFIC metonymy characteristic of class-inclusion constructions. In (16), on the other hand, B’s utterance is anomalous, because a burglar is not a type of noise. Here the referential chain is motivated by the EFFECT FOR CAUSE metonymy, which is not a classification metonymy, hence the anomaly. However, in both cases the noun phrases from the [What’s NP] construction can be replaced by a pro-form it, as shown in examples (17) and (18) below, quoted after Ruiz de Mendoza and Pérez-Hernández (2001: 350): (17) A: What’s that bird? B: It’s a robin. (18) A: What’s that noise? B: It’s a burglar.

Conceptual principles and relations

139

The metonymic motivation of the reference chain that bird – it – robin in (17) is clear. What happens in (18) in case of the that noise – it – a burglar chain is more complex. It in (18) does not refer to the antecedent that noise, but metonymically to ‘the cause of that noise’. Only through such a metonymic complex can the meaning of (18) be constructed. The question arises why the use of a pro-form avoids the anomaly typical of the use of the repeated noun phrase in (16). Ruiz de Mendoza and Pérez-Hernández (2001: 351) explain this distinction through a reference to the nature of the metonymic mapping obtaining in the case of the two metonymies. GENERIC FOR SPECIFIC is a target-in-source, while EFFECT FOR CAUSE is a source-intarget metonymy (see Ruiz de Mendoza 1997 for the details about this classification). In both cases, the meaning of the pro-form it is interpreted within the matrix domain of the metonymy, in example (17) it is “the bird”, while in example (18) it is “the cause of that noise”. This conceptual relation between the anaphoric pro-form and the metonymic noun has been formulated as the “domain availability principle”. 8.2.

Metonymic motivation of pragmatic inferencing

Metonymy is also a process that can shed some light on pragmatic inferencing. The Gricean approach to conversation, in particular the concept of conversational implicature, can benefit from conceptual metonymy. The interpretation of utterances requires inferencing on the part of the hearer so that they can understand both the speaker’s intention and the speaker’s meaning. However, Grice gives only very general information on how such inferencing takes place. It is guided by the wellknown conversational maxims. If the literal meaning of an expression violates any of the maxims, then the inferences are made to decode the speaker’s intention (Grice 1975; Searle 1975; cf. Moeschler, this volume). Such an approach to nonliteral meaning has been questioned in cognitive semantics (Gibbs 1994) on the grounds that psycholinguistic experiments do not show any time lag in the processing of non-literal meanings. In fact, these can be recognized faster than the literal meanings depending on context. To explain these phenomena Giora (1997, 1999, 2002, 2003, this volume) has proposed the “graded salience hypothesis”, which claims that in the initial stages of language processing, lexical and contextual processes run in parallel, but the lexical access takes precedence. That means that salient (lexical) meanings of expressions are accessed automatically, whether they are literal or figurative and regardless of context. Contextually appropriate meanings can be activated as quickly as the salient lexical meanings but they cannot suppress the salient meanings. Salience depends on frequency, conventionality, familiarity and prototypicality. Overall, the violation of conversational maxims is not a sufficient explanation of the speed of inferencing or of its nature. To account for the inference patterns that the hearer employs, Panther and Thornburg (1998) suggest that these can be explained through conceptual meto-

140

Małgorzata Fabiszak

nymies operating within cognitive domains, frames or ICMs, which they call scenarios. A speech act scenario is an action scenario consisting of the knowledge about the number and functions of action participants, the pre-conditions for an action, the action itself and the action result. Panther and Thornburgh (1998: 759)8 exemplify a speech act scenario with a simplified scenario for requests quoted here as (19): (19) Simplified scenario for requests (i) The BEFORE : H can do A. S wants H to do A. (ii) The CORE : S puts H under a (more or less strong) obligation to do A. The RESULT : H is under an obligation to do A (H must/should/ought to do A). (iii) The AFTER : H will do A. In example (19), all the four basic parts of the scenario are enumerated: The BEFORE , the CORE , the RESULT , and the AFTER as well as its participants: S – speaker, H – hearer, and A – action. These elements of the speech act scenario correspond to Searle’s (1975) felicity conditions. The conceptual relations obtaining between different parts of the scenario, as well as between its parts and the whole, are metonymic in nature. Metonymy is understood here not only as performing a referential function, but also a predicative function. Panther and Thornburgh (1998: 757) explain it on the basis of the following two examples: (20) a. The Chicago Bulls were able to nail down their fifth NBA championship … [CNN Sportscaster, June 1997] b. The Chicago Bulls nailed down their fifth NBA championship. While the utterance in (20a) refers to the ability of Chicago Bulls to win the championships, (20b) refers to the past actuality of Chicago Bulls winning the championship. (20b) is clearly an implicature of (20a) based on the POTENTIALITY FOR ACTUALITY metonymy discussed in detail in Thornburgh and Panther (1997) and Panther and Thornburgh (1998). Within the scenarios of speech acts, another type of metonymy can operate: speech act or illocutionary metonymy. In this case, one illocutionary act stands for another. For example, utterance (21a), often repeated in pragmatic textbooks, is interpreted to mean (21b): (21) a. It’s hot in here. b. Please open the window. (21a), which at its face value seems to be an assertion about the speaker’s perception of temperature, has an indirect illocutionary force of a request for opening the window as expressed in (21b). The inferencing performed by the Hearer is

Conceptual principles and relations

141

motivated by the specific level metonymy ASSERTION FOR REQUEST , which is an instantiation of a general metonymy ONE ILLOCUTIONARY ACT FOR ANOTHER ILLOCUTIONARY ACT , called illocutionary metonymy (Panther and Thornburgh 1998: 757–758). The examples above illustrate how predicational metonymies and illocutionary metonymies facilitate the understanding of utterance meaning accounting for the speed of interpretation and explaining the conceptual relations underlying the process. In this way, the cognitive approach to speech acts adds a new dimension to the inferential theory of utterance meaning by positing the existence of metonymybased natural inference schemata. This work on illocutionary scenarios has led Panther (2005) to conclude that metonymies are not so much substitution relations or reference point phenomena (see Langacker 1993), but inference schemas. This model is further developed in Ruiz de Mendoza (2007) and Ruiz de Mendoza and Baicchi (2007) through expanding the number of pragmatic variables represented in the speech act scenario by incorporating power and solidarity relations as well as a cost-benefit scale (Leech 1983) to the proposed idealized cognitive model called the COST- BENEFIT cognitive model. In their proposal, Ruiz de Mendoza and Baicchi claim that sets of semantic conditions specified in a highlevel cognitive model can adequately describe the pragmatic part of the realizational potential of the model. Ruiz de Mendoza (2007) elaborates the theory of ICMs viewed as stored representations of cultural and embodied experience of an individual. He proposes to introduce a distinction between “operational” and “non-operational” ICMs. Propositional representations, such as Fillmore’s (1982) frames and image-schemata (Johnson 1987; Lakoff and Johnson 1999) are non-operational, relatively stable coded meaning representations, while metaphor and metonymy are operational in that they participate in the online construction of novel meanings from the stored representations and contextual clues. He further divides propositional ICMs into “situational”, such as the restaurant script (see Schank and Abelson’s 1977 original formulation) and “non-situational”, such as MOTHER (Lakoff 1987: 74–84). The final refinement lies in differentiating between “non-generic (low-level)” and “generic (high-level)” ICMs. The low-level ICMs constitute our encyclopaedic knowledge and stem directly from our social, cultural and biological experience, while the high-level ICMs are abstractions, generalizations from the low-level cognitive models, such as process, cause-effect, action. Following this typology, Ruiz de Mendoza and Baicchi (2007: 101) suggest that illocutionary scenarios are high-level situational ICMs. In so far as ontology is concerned, they consist of different values of the variables (discussed below) that can be viewed as conventions; as for structure, it is determined by the interaction between the variables. As we have seen, Thornburg and Panther (1997), Panther and Thornburg (1998) propose a 4 element illocutionary scenario consisting of BEFORE , CORE , RESULT, AFTER , corresponding to Searle’s (1975) felicity conditions. It is enriched in

142

Małgorzata Fabiszak

Ruiz de Mendoza and Baicchi (2007: 102–103) by a number of elements (variables) placing an illocutionary act in a wider interactive context: – – – – – –

The power relationship between the interlocutors The degree of optionality conveyed by the illocutionary act The degree of politeness The degree of prototypicality of certain utterances over others The degree of cost-benefit of the requested action The semantic motivation for other types of indirect speech acts as expressed by an oblique modal (could, would) or a negative modal (can’t, won’t) in the case of requests – The cognitive grounding of speech acts in experiential gestalts. These features interact with one another and may undergo prototypicality effects, which can be noticed in an analysis of a series of linguistic forms a request may take: (22) (23) (24) (25) (26)

Can you open the window? Will you open the window? Can you open the window, please? Could you open the window? Could you open the window, please?

Example (22) suggests a high degree of power on the part of the Speaker, limited optionality and a small degree of politeness. As the mitigating strategies are increasingly used in (23) to (26), the degrees of optionality and politeness increase, while the degree of power of the Speaker is downplayed. Ruiz de Mendoza and Baicchi (2007: 106) suggest that the more elements of the scenario are used (modality as opposed to ability, the adverb please) the more prototypical is the utterance for a given illocutionary scenario. Different low-level illocutionary scenarios (of ordering, requesting, begging etc.) may be generalized to a higher-level CostBenefit ICM, which states the conditions for human interaction in terms of cost to the Hearer and Benefit of the Speaker. This generalization is in accord with Sperber and Wilson’s (1986/1995) claim that different categories of speech acts are not stored individually, but rather their meaning is derived from an interaction of more general knowledge and situational context. While Ruiz de Mendoza and Baicchi’s proposal may not be too original from a pragmatic point of view, it nicely manages to integrate longstanding pragmatic insights into a cognitive-linguistic framework. The COST- BENEFIT model is now being tested empirically using examples from a number of linguistic corpora, for example from the Corpus of Contemporary American English (COCA). Del Campo (2009), in her paper on constructional polysemy, tries to resolve the ambiguity of interpretation of begging vs. requesting, warning vs. threatening and offering vs. inviting vs. suggesting. The same

Conceptual principles and relations

143

construction can be interpreted as either a threat or a warning depending on the active zone of the metonymically motivated profile of the expression within a CostBenefit ICM.

9.

Conclusion

The present chapter has been an attempt to present the way in which coded semantic representations are utilized in the process of meaning construction in a communicative situation. It has shown how mental categories are structured by prototypes and how they emerge from sensory-motor experience to form image schemata, which, in turn, serve as the basis for creating both basic and abstract meanings. These mental representations can be structured in the form of idealized cognitive models and frames, whose role is to provide the propositional structures on which such mental processes as conceptual metaphor, conceptual blending and conceptual metonymy operate. Some of these processes may become entrenched or conventionalized through frequent usage and form a stable representation (metaphor and metonymy), others may operate in an online meaning construction (blending). When meanings are created in context, the coded, long-term representations are recruited into the generic space and input spaces. Only those parts of the representations which are context relevant are activated and projected into the blended space, where meanings emerge in real time. It has been shown how idealized cognitive models contribute to our understanding of deixis; how image schemata and metaphor motivate value systems; how image schemata underlie modality; how image schemata, semantic frames, idealized cognitive models and metonymy form the basis for performing and interpreting speech acts; how metaphor structures discourse, how blending is necessary for constructing fictive interpretations of deixis, rhetorical questions or real world artefacts and how metonymy supplies the mental patterns for anaphora and pragmatic inferencing. All of these are examples of conceptual principles and relations, which are necessary to explain the cognitive basis of pragmatics.

Acknowledgements I would like to thank Anna Hebda, Iwona Kokorniak, Karolina Krawczak, HansJörg Schmid and an anonymous reviewer for their helpful comments on the draft version of the chapter.

144

Małgorzata Fabiszak

Notes 1. A comprehensive account of deixis can be found in Hanks (2011). 2. Another cognitive account of presupposition is offered by Marmaridou (2000). She views presupposition as a figure-ground relationship in the sense of Langacker (1991). In this account, presuppositions form the ground for the figure constituted by the proposition actually uttered. 3. Cf. – among others – Kotarbin´ski’s reism (1961 [1929]), a philosophical approach to meaning further developed within cognitive linguistics by Szwedek (2007, 2008) into a Theory of Objectification. 4. Zinken (2004) analyzes the transformation discourse of the German and Russian journalists in the late 1980s and early 1990s. He stresses that the German Wende (‘change’) discourse presented a passive perspective on the social and political changes of the period. 5. One of the first attempts to compare these two approaches is Grady, Oakley and Coulson (1999) and a succinct introduction to the theory can be found in Coulson and Oakley (2000). 6. These processes are described in detail in Fauconnier and Turner (2002). 7. On the distinction between factive and fictive see Talmy (2000 [1996]). 8. This is a corpus-based study conducted on the basis of The Bergen Corpus of London Teenage Language (COLT).

References Bach, Emmon and Robert Harms (eds.) 1968 Universals in Linguistic Theory. New York: Holt, Rinehart and Wilson. Cameron, Lynne 2008 Metaphor shifting in the dynamics of talk. In: Mara-Sophia Zanotto, Lynne Cameron and Marilda C. Cavalcanti (eds.), Confronting Metaphor in Use: An Applied Linguistic Approach, 45–62. Amsterdam/Philadelphia: Benjamins. Charteris-Black, Jonathan 2004 Corpus Approaches to Critical Discourse Analysis. London/New York: Palgrave Macmillan. Charteris-Black, Jonathan 2005 Politicians and Rhetoric: The Persuasive Power of Metaphor. London/New York: Palgrave Macmillan. Chilton, Paul 2004 Analysing Political Discourse: Theory and Practice. London: Routledge. Cole, Peter and Jerry Morgan (eds.) 1975 Speech Acts. (Syntax and semantics 3.) New York: Academic Press. Coulson, Seana 2001 Semantic Leaps. Frame-Shifting and Conceptual Blending in Meaning Construction. Cambridge: Cambridge University Press. Coulson, Seana and Todd Oakley 2000 Blending basics. Cognitive Linguistics 11 (3/4): 175–196.

Conceptual principles and relations

145

Coulson, Seana and Todd Oakley 2006 Purple persuasion: Conceptual Blending and deliberative rhetoric. In: Jean Luchjenbroers (ed.), Cognitive Linguistics: Investigations across Languages, Fields and Philosophical Boundaries, 47–65. Amsterdam/Philadelphia: John Benjamins. Croft, William and Alan D. Cruse 2004 Cognitive Linguistics. Cambridge: Cambridge University Press. Croft, William 2002 [1993] The role of domains in the interpretation of metaphors and metonymies. In: René Dirven and Ralf Pörings (eds.), Metaphor and Metonymy in Comparison and Contrast, 161–205. Berlin/New York: Mouton de Gruyter. [First published in Cognitive Linguistics 4: 335–370, 1993] Cruse, Alan D. 1986 Lexical Semantics. Cambridge: Cambridge University Press. Cruse, Alan D. 2000 Meaning in Language. An Introduction to Semantics and Pragmatics. Oxford: Oxford University Press. Del Campo-Martínez, Nuria 2009 Active zone/profile discrepancy and illocutionary meaning. A paper delivered at the International Conference on Figurative Language Learning and Figurative Language Use: Theory and Applications (CRAL 2009), 29th-31st of October 2009, Logroño, Spain. Dirven, René and Ralf Pörings (eds.) 2002 Metaphor and Metonymy in Comparison and Contrast. Berlin/New York: Mouton de Gruyter. Emmorey, Karen and Judy Reilly (eds.) 1995 Language, Gesture and Space. Hillsdale, NJ: Erlbaum. Evans, Vyvyan and Melanie Green 2006 Cognitive Linguistics. An Introduction. Edinburgh: Edinburgh University Press. Fabiszak, Małgorzata 2007 A Conceptual Metaphor Approach to War Discourse and its Applications. Poznan´: Adam Mickiewicz University Press. Fairclough, Norman 1989 Language and Power. London: Longman. Fauconnier, Gilles 1997 Mappings in Thought and Language. Cambridge: Cambridge University Press. Fauconnier, Gilles 2007 Mental spaces. In: Dirk Geeraerts and Hubert Cuyckens (eds.), Oxford Handbook of Cognitive Linguistics, 351–376. Oxford: Oxford University Press. Fauconnier, Gilles and Mark Turner 2002 The Way We Think. Conceptual Blending and the Mind’s Hidden Complexities. New York: Basic Books. Fillmore, Charles 1968 The case for case. In: Emmon Bach and Robert Harms (eds.), Universals in Linguistic Theory, 1–90. New York: Holt, Rinehart and Wilson.

146

Małgorzata Fabiszak

Fillmore, Charles 1982 Frame semantics. In: Linguistic Society of Korea (ed.), Linguistics in the Morning Calm, 111–138. Seoul: Hanshin. Fillmore, Charles 1985 Frames and the semantic of understanding. Quaderni di Semantica 6/2: 222–253. Geeraerts, Dirk 1997 Diachronic Prototype Semantics. A Contribution to Historical Lexicology. Oxford: Clarendon. Geeraerts, Dirk 2009 Lexical Semantics. Oxford: Oxford University Press. Geeraerts, Dirk and Hubert Cuyckens (eds.) 2007 Oxford Handbook of Cognitive Linguistics. Oxford: Oxford University Press. Gibbs, Raymond W. 1994 The Poetics of Mind. Figurative Thought, Language and Understanding. Cambridge: Cambridge University Press. Gibbs, Raymond. W. and Herbert L. Colston 1995 The cognitive psychological reality of image schemas and their transformations. Cognitive Linguistics 6: 347–378. Gibbs, Raymond W. and Gerard J. Steen (eds.) 1999 Metaphor in Cognitive Linguistics. 113–136. Amsterdam/Philadelphia: John Benjamins. Giora, Rachel 1997 Understanding figurative and literal language: The graded salience hypothesis. Cognitive Linguistics 7: 183–206. Giora, Rachel 1999 On the priority of salient meanings: Studies of literal and figurative language. Journal of Pragmatics 31 (7): 919–929. Giora, Rachel 2002 Literal vs. figurative language: Different or equal? Journal of Pragmatics 34: 487–506. Giora, Rachel 2003 On our Mind. Salience, Context, and Figurative Language. Oxford/New York: Oxford University Press. Goffman, Erving 1981 Forms of Talk. Oxford: Basil Blackwell. Grady, Joseph 1997 Foundations of meaning: Primary metaphors and primary dcenes. PhD thesis. Berkeley: University of California. Grady, Joseph, Todd Oakley and Seana Coulson 1999 Blending and metaphor. In: Raymond W. Gibbs and Gerard J. Steen (eds.), Metaphor in Cognitive Linguistics, 101–124. Amsterdam/Philadelphia: John Benjamins. Grice, H. Paul 1975 Logic and conversation. In: Peter Cole and Jerry Morgan (eds.), Speech Acts, 41–58. (Syntax and Semantics 3). New York: Academic Press.

Conceptual principles and relations

147

Hampe, Beate (ed.) 2005 From Perception to Meaning: Image Schemas in Cognitive Linguistics. Berlin/ New York: Mouton de Gruyter. Hanks, William F. 2011 Deixis and indexicality In: Wolfram Bublitz and Neal R. Norrick (eds.), Foundations of Pragmatics, 315–346. (Handbooks of Pragmatics Series, Vol. I). Berlin/New York: Mouton de Gruyter. Herman, Vimala 1999 Deictic Projection and Conceptual Blending in Epistolarity. Poetics Today 20 (3): 523–542. Johnson, Mark 1987 The Body in the Mind. Chicago: University of Chicago Press. Kecskes, Istvan (ed.) 2007 Explorations in Pragmatics. Linguistic, Cognitive and Intercultural Aspects. Berlin/New York: Mouton de Gruyter. Kosecki, Krzysztof (ed.) 2007 Perspectives on Metonymy. Frankfurt/Main: Peter Lang. Kotarbin´ski, Tadeusz 1961 [1929]Elementy teorii poznania, logiki formalnej i metodologii nauk [Elements of Cognitive Theory, Formal Logic and Methodology of Science]. Wrocław: Ossolineum. Kövecses, Zoltán 2006 Language, Mind and Culture. A Practical Introduction. Oxford: Oxford University Press. Kövecses, Zoltán and Gunter Radden 1998 Metonymy: Developing a cognitive linguistic view. Cognitive Linguistics 9 (1): 37–71. Krzeszowski, Tomasz Paweł 1997 Angels and Devils in Hell. Elements of Axiology in Semantics. Warszawa: Energeia. Lakoff, George 1987 Women, Fire, and Dangerous Things: What Categories Reveal about the Mind. Chicago: University of Chicago Press. Lakoff, George 1993 The contemporary theory of metaphor. In: Andrew Ortony (ed.), Metaphor and Thought, 202–251. Cambridge: Cambridge University Press. Lakoff, George 1996 Moral Politics: What Conservatives Know that Liberals Don’t. Chicago: University of Chicago Press. Lakoff, George and Mark Johnson 1980 Metaphors we Live by. Chicago: University of Chicago Press. Lakoff, George and Mark Johnson 1999 Philosophy in the Flesh. The Embodied Mind and its Challenge to Western Thought. New York: Basic Books. Langacker, Ronald 1987 Foundations of Cognitive Grammar, Vol. 1. Stanford, CA: Stanford University Press.

148

Małgorzata Fabiszak

Langacker, Ronald 1991 Foundations of Cognitive Grammar, vol. 2. Stanford: Stanford University Press. Langacker, Ronald 1993 Universals of construal. Proceedings of the Annual Meeting of the Berkeley Linguistics Society 19: 447–463. Leech, Geoffrey 1983 Principles of Pragmatics. London: Longman. Liddell, Scott K. 1995 Real, surrogate and token space: grammatical consequences in ASL. In: Karen Emmorey and Judy Reilly (eds.), Language, Gesture and Space, 19–41. Hillsdale, NJ: Erlbaum. Liddell, Scott K. 2000 Blended spaces and deixis in sign language discourse. In: David McNeill (ed.), Language and Gesture, 331–357. Cambridge: Cambridge University Press. Lloyd, Barbara B. and Eleanor Rosch (eds.) 1978 Cognititon and Categorization. Hillsdale, NJ: Erlbaum. Markman, Ellen M. 1989 Categorization and Naming in Children. Cambridge, MA: MIT Press. Marmaridou, Sophia S.A. 2000 Pragmatic Meaning and Cognition. Amsterdam/Philadelphia: Benjamins. Mervis, Caroline B. and Eleanor Rosch 1981 Categorization of natural objects. Annual Review of Psychology 32: 89–115. Musolff, Andreas 2004 Metaphor and Political Discourse: Analogical Reasoning in Debates about Europe. London/New York: Palgrave Macmillan. Nerlich, Brigitte, Craig A. Hamilton and Victoria Rowe 2002 Conceptualising Foot and Mouth Disease: The socio-cultural role of metaphors, frames and narratives. Metaphoric.de 2: 90–113. Ortony, Andrew (ed.) 1993 Metaphor and Thought. Cambridge: Cambridge University Press. Panther, Klaus-Uwe 2005 The role of conceptual metonymy in meaning construction. In: Francisco-José Ruiz de Mendoza-Ibáñez and Sandra Peña (eds.), Cognitive Linguistics. Internal Dynamics and Interdisciplinary Interaction, 353–386. Berlin/New York: Mouton de Gruyter. Panther, Klaus-Uwe and Linda Thornburg 1998 A cognitive approach to inferencing in conversation. Journal of Pragmatics 30: 755–769. Papafragou, Anna 2000 Modality: Issues in the Semantics-Pragmatics Interface. Oxford: Elsevier. Pascual, Esther 2002 Imaginary Trialogues: Conceptual Blending and Fictive Interaction in Criminal Courts. Utrecht: LOT. Pascual, Esther 2006 Questions in legal monologues: Fictive interaction as argumentative strategy in a murder trial. Text & Talk 26(3): 383–402.

Conceptual principles and relations

149

Pascual, Esther 2008 Fictive interaction blends in everyday life and courtroom settings. In: Todd Oakley and Anders Hougaard (eds.), Mental Spaces in Discourse and Interaction, 79–108. Amsterdam/Philadelphia: John Benjamins. Regier, Terry 1996 A model of the human capacity for categorizing spatial relations. Cognitive Linguistics 6: 63–88. Richardson, Daniel C., Michael J. Spivey, Lawrence W. Barsalou and Ken McRae 2003 Spatial representations activated during real-time comprehension of verbs. Cognitive Science 27: 767–780. Rohrer, Tim 2005 Image schemata in the brain. In: Beate Hampe (ed.), From Perception to Meaning: Image Schemas in Cognitive Linguistics. 165–196. Berlin/New York: Mouton de Gruyter. Rosch, Eleonor 1978 Principles of categorization. In: Barbara B. Lloyd and Eleanor Rosch (eds.), Cognition and Categorization, 27–48. Hillsdale: Erlbaum. Ruiz de Mendoza-Ibáñez, Francisco-José 1997 Cognitive and pragmatic aspects of metonymy. Cuadernos de Filología Inglesa 6(2): 161–178. Ruiz de Mendoza Ibáñez, Francisco-José 2007 High level cognitive models: in search of a unified framework for inferential and grammatical behaviour. In: Krzysztof Kosecki (ed.), Perspectives on Metonymy, 11–30. Frankfurt/Main: Peter Lang. Ruiz de Mendoza-Ibáñez, Francisco-José and Annalisa Baicchi 2007 Illocutionary constructions: Cognitive motivation and linguistic realization. In: Istvan Kecskes (ed.), Explorations in Pragmatics. Linguistic, Cognitive and Intercultural Aspects, 95–127. Berlin/New York: Mouton de Gruyter. Ruiz de Mendoza-Ibáñez, Francisco-José and Lorena Pérez-Hernández 2001 Metonymy and the grammar: Motivation, constrains and interaction. Language and Communication 21: 321–357. Saeed, John 2003 Semantics. Oxford: Blackwell. Schank, Roger and Robert Abelson 1977 Scripts, Plans, Goals, and Understanding. Hillsdale: Erlbaum. Schön, Donald A. 1993 Generative metaphor: A perspective on problem setting in social policy. In: Ortony Andrew (ed.), Metaphor and Thought, 137–163. Cambridge: Cambridge University Press. Searle, John 1975 Indirect speech acts. In: Peter Cole and Jerry Morgan (eds.), Syntax and Semantics 3: Speech Acts, 59–82. New York: Academic Press. Sinha, Chris 2005 Blending out of the background: Play, props and staging in the material world. Journal of Pragmatics 37: 1537–1554. Sperber, Dan and Deirdre Wilson 1986/95 Relevance Theory: Communication and Cognition. Oxford: Blackwell. [Second edition with Postface 1995]

150

Małgorzata Fabiszak

Sweetser, Eve 1990 From Etymology to Pragmatics. Cambridge: Cambridge University Press. Szwedek, Aleksander 2007 An alternative theory of metaphorisation. In: Małgorzata Fabiszak (ed.), Language and Meaning. Cognitive and Functional Perspectives, 312–327. Frankfurt am Main: Peter Lang. Szwedek, Aleksander 2008 Ontological sources of structural and orientational metaphors. In: Zdzisław Wa˛sik and Tomasz Komendzin´ski (eds.), Metaphor and Cognition, 9–15. (Philologica Wratislaviensia: From Grammar to Discourse 1.) Frankfurt am Main: Peter Lang. Talmy, Leonard 1988 Force dynamics in language and cognition. Cognitive Science 12: 49–100. Talmy, Leonard 2000 Toward a Cognitive Semantics, Vols. 1 and 2. Cambridge, MA: MIT Press. Taylor, John R. 2002 Cognitive Grammar. Oxford: Oxford University Press Taylor, John R. 2004 Linguistic Categorization. Prototypes in Linguistic Theory, 3rd ed. Oxford: Clarendon. Thornburg, Linda and Klaus-Uwe Panther 1997 Speech act metonymies. In: Wolf-Andreas Liebert, Gisela Redeker and Linda Waugh (eds.), Discourse and Perspective in Cognitive Linguistics, 205–219. Amsterdam/Philadelphia: John Benjamins. Turner, Mark 2002 The cognitive study of art, language, literature. Poetics Today 23(1): 9–20. Van Dijk, Teun 1998 Ideology: An Interdisciplinary Approach. London: Sage. Vlassova, Tatiana Aleksandrovna and Tatiana Stiepanovna Vlassova 1971 Sovremennuj razgovornuj anglijskij jazyk [Contemporary Conversational English]. Moscow: Izdatelstvo Vysshaya Shkola. Wa˛sik, Zdzisław and Tomasz Komendzin´ski (eds.) 2008 Metaphor and Cognition (Philologica Wratislaviensia: From Grammar to Discourse 1). Frankfurt am Main: Peter Lang. Wodak, Ruth and Michał Krzy˙zanowski 2008 Qualitative Discourse Analysis for the Social Sciences. Basingstoke: Palgrave. Zanotto, Mara Sophia, Lynne Cameron and Marilda C. Cavalcanti (eds.) 2008 Confronting Metaphor in Use: An Applied Linguistic Approach. Amsterdam/ Philadelphia: John Benjamins. Zinken, Jörg 2004 Metaphor practices in the German Wende discourse. Journal of Multilingual and Multicultural Development 25(5–6): 424–436.

6.

Contextual salience, domains, and active zones John R. Taylor

1.

Introduction

As an introduction to the subject matter of this chapter, consider the following sentence and how it might be interpreted.1 (1) I left the university a short time ago. A person who utters (1) could be referring to one of two different kinds of event. On the one interpretation the sentence refers to a departure event; the person in question moved from a place in or at the university to a place away from the university. On the other interpretation the sentence refers to the severing of a relation; the person terminated their relation with the university, for example, by graduating or by resigning their position. The two interpretations are associated with different understandings of the component parts of the sentence. Leave can refer to locomotion from inside an enclosed space or to the severing of a relation. The university can refer to the physical premises – the buildings and the surrounding grounds – or it can refer to an institution. The subject pronoun also refers to subtly different entities. In the one case the speaker presents herself as a physical object, capable of moving from one place to another; in the other the speaker is construed as a cognitive entity, able to contract and to terminate relations with an institution. Even the adjunct phrase a short time ago gets interpreted differently. On the motion interpretation, a short time ago would probably be taken to be a matter of minutes or hours; on the severing of a relation interpretation, the time scale would probably be larger, likely to be a matter of weeks or months. The two interpretations are also associated with different entailments. On the severing of a relation interpretation the subject referent would have to have been associated with the university in some capacity, for example, as a student or as an employee. On the departure interpretation no such entailment holds; the person could have been a member of the public who was visiting the premises or even someone who had taken a short cut through the university grounds. The departure interpretation also entails that the person was located on the university premises prior to the designated event. Moreover, a witness could attest to the truth of the sentence on this interpretation by declaring that he saw the event with his own eyes. The severing of a relation interpretation carries no such entailment. You can “leave the university” without actually moving at all.

152

John R. Taylor

The different ways in which sentence (1) can be interpreted raise a number of long-standing issues in linguistic semantics and the relation of semantics to pragmatics. The semantics/pragmatics distinction is based on the observation that the contextualized interpretation of a sentence may go beyond the linguistically encoded meaning. For example, the fact that the subject referent in (1) is taken, on one interpretation, to be a student or employee of the university is not encoded in the subject nominal, nor indeed in any of the other constituents of the sentence. Accounting for this kind of additional meaning is the province of pragmatics, while stating the linguistically encoded meaning is the subject matter of semantics. Where to draw the line between the two enterprises has been subject to much debate (Huang 2007). Irrespective of how the issue is resolved, however, it is generally assumed that a pragmatic interpretation is dependent on a prior determination of a semantic representation, the latter deriving from knowledge that is permanently stored in a person’s mental grammar. In this chapter I focus on just one aspect of the semantics-pragmatics relation, namely, word meanings and how they might be represented in the mental lexicon, with an eye on the interpretations that words receive in specific contexts of use. The role of pragmatic interpretation intersects with a recurring theme in lexical semantics, namely, the question of polysemy and the possibility that items in the mental lexicon are associated with more than one stored meaning, perhaps even with a large number of distinct meanings. Lakoff (1987: 416) observes that window can refer to an opening in a wall or to a glass-filled frame which fits into the opening. There are additional subtleties. Break a window refers (probably) only to the glass, paint a window refers (probably) only to the wooden parts, while open the window refers to the moveable portion. Pursuing the matter further, we might note, with Lakoff, that opening a window involves different activities from opening a present, opening the curtains, or opening your arms (Taylor 2006: 52). Rather than assuming that these different understandings are a matter of pragmatic interpretation, Lakoff proposes a polysemy account. The different readings of window and of open constitute “categor[ies] of senses” (Lakoff 1987: 418), permanently stored in a person’s mental lexicon. It will be appreciated that this kind of approach, postulating polysemy on the basis of sometimes very small differences in contextual interpretation, if pursued systematically, would tend to minimize the discrepancy between a word’s stored semantic representation and its contextual interpretation. Contextual interpretation would be simply a matter of selecting one of the stored meanings. On more conservative polysemy proposals, such as that argued for by Tyler and Evans (2001), the contextualized meaning need not coincide with any of the stored meanings. Different polysemes are proposed only to the extent that a contextualized reading cannot be derived from contextual information or on the basis of real-world knowledge. As a matter of fact, even the extreme polysemy approach cannot entirely eliminate the possibility of a discrepancy between a stored meaning and a contex-

Contextual salience, domains, and active zones

153

tualized interpretation, as is the case when a word is used creatively in a way not already sanctioned by the mental lexicon. Indeed, it is not at all unusual, when studying the use of a word in a corpus, to come across citations which cannot be reconciled with any of the definitions listed in even voluminous paper dictionaries. Fillmore and Atkins (2000) discuss the matter using the example of the word crawl, while Kilgarriff (1997) makes a similar point with respect to the word handbag. It would seem that even the most fine-grained polysemy analysis will never be able to account for all the potential uses of a word. There are no a priori reasons for limiting the extent of polysemy in the mental lexicon; one cannot therefore dismiss Lakoff’s extreme polysemy approach by appealing to the need for economy of storage. Nevertheless, the postulation of polysemy, especially rampant polysemy of the kind entertained by Lakoff, raises a number of problems, over and above those presented by speakers’ creative usage. If a word is polysemous, then any sentence containing that word will be, in principle, ambiguous. If two or more words in a sentence are polysemous, the number of possible sentence readings will increase exponentially. Most of these readings will, no doubt, be pragmatically inappropriate or conceptually incoherent and will need to be discarded. Even so, the assumption that polysemy is both widespread and extensive would mean that the disambiguation of a sentence of even modest length would require a vast amount of cognitive processing. Against this is the fact that in natural contexts sentence disambiguation is rarely an issue in language comprehension. One of the reasons for this is that the different polysemes of a word are typically associated with different lexico-syntactic contexts (Kishner and Gibbs 1996; Hoey 2005; Gries 2006; Miller and Leacock 2000). Indeed, contrary to the expectation that increasing the length of a sentence will increase the number of possible readings, it is a matter of common experience that the greater the amount of sentence context, the easier it is to access the intended meaning of a word. A window of up to about two words on either side is usually sufficient for disambiguation to take place (Choueka and Lusignan 1985; Miller and Leacock 2000). How would a polysemy account fare with respect to the sentence cited above? It might have some plausibility in the case of the verb leave, perhaps also for the noun university. We might suppose that leave is associated in the mental lexicon with two distinct, albeit related meanings, one having to do with an entity’s motion from an enclosed space, the other with the termination of an institutional relation. The difference, it will be noted, concerns the nature of the landmark entity – a spatially enclosed area or an institution. The difference, in other words, concerns the spatial vs. the abstract-metaphorical understanding of the landmark. On the other hand, we could also imagine an account of the verb which is unspecified for the spatial/abstract contrast. How the verb’s meaning is instantiated in a given context would be dependent on the conceptual domain – physical motion or institutional relations – against which the utterance is interpreted. Neither is the verb leave unique in this respect. Approach, enter, and remain, amongst others, exhibit simi-

154

John R. Taylor

lar contrasts. Thus, you can approach a building (move towards it in physical space) or approach an organization (with a view to membership, perhaps). The different senses of leave, if such they are, are therefore symptomatic of a more general polysemization process. A polysemy account might also be entertained for university. Let us suppose that the concept is understood primarily as the name for an institution, with its associated aims, practices, personnel, and so on. By metonymy, the term comes to denote the buildings which house the institution, and, by a further metonymy, the grounds in which the buildings are situated. Again, several other institution-type words exhibit the same meaning variation: museum, institute, college, and school, to name just a few. How these terms are interpreted – as names for institutions or names for buildings – follows once again from the domain of discourse. In the case of these nouns, however, it is difficult to imagine a schematic representation which covers both the institution and the building interpretations. A building and an institution are ontologically different kinds of entities, which cannot reasonably be brought under a single, more general concept. It is less plausible to suppose that the other items in (1) are associated in the mental lexicon with distinct meanings. It would surely be bizarre to claim that the personal pronoun I (along with all other person-denoting words) is associated with both a physical entity and a cognitive agent meaning. The case for polysemy is equally weak with respect to the two readings of a short time ago. The term is understood to refer to a time span which is less than the expected norm. The time units in which the norm is measured are not part of the semantic meaning of the expression. Rather, what counts as a short time depends on the activity in question. We walk in and out of buildings many times each day. To say that someone walked out of a building “a short time ago” imposes a fine-scale measurement. On the other hand, we quit jobs, or graduate, only a handful of times, perhaps even only once during a lifetime; “a short time” therefore suggests a more extended time span. Taking up a point made earlier, it is worth noting that even though there might be grounds for regarding both leave and university as, at a minimum, two-ways polysemous, it is not the case that leave the university turns out to have four possible readings. (Neither, on the assumption that I is two-ways polysemous, does I left the university turn out to be 8-ways ambiguous.) What we observe, instead, is that the readings of the terms co-select each other, with the result that we are not able to identify any specific item in the sentence as triggering the ambiguity. It is as if the domain of interpretation – motion or institutional relations – constitutes a kind of “semantic prosody”2 which determines the interpretation of each of the constituents. I have discussed the contextualized interpretations of some of the constituent words of sentence (1). The discussion raises the following questions:

Contextual salience, domains, and active zones

155

a. What constitutes the stored meaning (or meanings, in case the words are taken to be polysemous) associated with the words in the mental lexicon? b. Assuming we have a satisfactory answer to the above question, we need to enquire into the process, or processes, by which a stored meaning is instantiated in a given context. c. A further issue concerns the compositionality of the complex expression. According to the compositionality principle, the meaning of an expression is a function of the meanings of its parts and the manner of their combination. The question arises whether compositionality operates over the stored meanings of the constituent words or over the contextualized readings. The above discussion suggests that the latter option may be the correct one. We might also note in this connection that the logical properties of an expression – its entailments and presuppositions and the inferences that it generates – appear to derive from the contextualized interpretation, not from the stored meanings of its constituent words. In the following, I review a number of proposals for addressing these questions, returning at the end of the chapter to the legitimacy of supposing that words are indeed associated with “stored meanings”, or at least with meanings which are fixed and stable.

2.

Active zones and metonymy

One approach to meaning variation – or at least to some aspects of it – was discussed by Langacker under the rubric of the active zone phenomenon (Langacker 1984; see also Taylor 2003a: ch. 7). The term refers to the fact that when an entity participates in a semantic relation typically only some facet of the entity is actively involved in the relation. Suppose I am bitten by a dog. From one point of view, it is the dog, as an autonomous creature, which is responsible for the action and I, as an individual, who suffer from it. At the same time, it is evident that only some part of the dog was actively involved in the biting, namely, its teeth, while only some part of me, such as my left calf, was the location of the biting. The dog’s active zone is implicit in the meaning of the verb bite. Biting inherently involves using the teeth; thus, to say that the dog bit me with its teeth would be, at best, redundant. My active zone, however, is not implicit in the sentence. If need be it could be spelled out in an adjunct phrase: I was bitten on my left calf. The active zone phenomenon is ubiquitous, so much so that Langacker notes that it is quite difficult to find a predication where all parts of an entity are equally involved in a relation (a possible example might be a statement that the Earth moves round the Sun). For example, if you wash your car, you wash only the exterior surface; if you vacuum-clean your car, you clean only the upholstered in-

156

John R. Taylor

terior; if you have your car serviced, only the mechanical parts are attended to; if you fill up your car with books, you are filling up the space in the boot or on the seats, while if you fill up your car with petrol, the petrol goes into the tank. An active zone account might also be considered with respect to the different interpretations of the personal pronoun I in (1). One facet of a person is her physical body, which can move from place to place; another is the person as a social and cognitive agent, which can contract relations with institutions. Langacker (1993) presents the active zone phenomenon as an aspect of a more general process. The entity named in a noun phrase serves as a reference point which guides the interpreter towards the identification of the entity involved in the predication. The reference point is a cognitively salient entity, which provides mental access to entities to which it is closely related. In this respect, the active zone phenomenon gradually shades into what is traditionally referred to as metonymy. Compare the following: (2) a. I graduated from the university last year. b. I am in the phone book. c. I am parked round the corner. Whereas (2a) refers to the person as a psycho-social entity, (2b) refers to a unique property of the person, namely, their name and telephone number. (As a matter of fact, the sentence exhibits a further metonymy, in that it is not the person’s name as such which appears in the telephone book, but a written representation of it.) (2c) refers not to the person at all, but to one of their possessions, namely their car. Since this last example involves a transfer of reference, involving the pragmatically salient relation of a person and their possessions (Barcelona 2002), it would probably be regarded as a clear case of metonymy.3 Some metonymies are conventionalized and the extended senses are plausibly represented in the mental lexicon. Such, for example, is the case with nouns such as Burgundy, which can designate both a place and a wine produced in that place. (One can easily imagine a person learning the name of the wine without being aware of the other sense of the word; to this extent the wine-denoting sense has an independent status.) Other metonymies are one-off, computed in the act of utterance interpretation. The interpretations of the personal pronoun are cases in point. I have discussed the active zone with reference to nominal concepts. The question arises whether the notion might also be applicable to verbal, adjectival, and other kinds of concepts. In pursuing this possibility we need to address a topic which was implicit in the above discussion. With respect to the example of my being bitten by a dog, the primary reference of the nominals I and dog is evident. It is the two beings understood as integrated, autonomous biological entities, only some facets of which are actively involved in the event. Similarly, with respect to the car example, we understand the noun car in terms of a certain kind of vehicle, with its various parts, their functions, and their arrangement. It is this concept

Contextual salience, domains, and active zones

157

which, we might suppose, is linked up with the word in the mental lexicon. But when we turn to a verbal concept – that designated by leave, for example – it is by no means obvious what the “basic” sense of the verb is, the sense that might be stored in the mental lexicon. Restricting ourselves to spatial uses only, the verb could refer to movement out of an enclosed space (as in leave the room) or to movement away from an entity, typically another person, as in leave me now. Admittedly, the two senses are closely related and both may sometimes be implicated. When you leave the room you both exit from an enclosed space and move away from the space; when you leave a person you both move away from the person and exit their personal space. In other cases it might be difficult to decide which of the two senses is involved. If you leave a country the country could be construed both as a containing space and as a geographical location. These few examples by no means exhaust the spatial senses of leave. For example, if you leave something in a room you not only move away from the object, you fail to take it with you. If asked to leave something alone you are supposed not to approach it or to touch it (Taylor 1995: 14–15). These examples suggest that the ‘leave’ concept may be better regarded as a cluster of closely related senses, some of which mutually entail each other, some of which highlight certain facets which are downplayed or are irrelevant in other cases, none of which, however, can be regarded as constituting the “basic”, or “true” meaning of the verb. These considerations lend support to Lakoff’s (1987: 416) notion that the mental representation of a word constitutes a “family of related senses”. A similar situation arises with prepositions. One of the issues which has driven the still ongoing “story of over”4 has been uncertainty as to what the basic, or central sense of the preposition is, the sense from which other uses can be most plausibly derived, whether by the active zone phenomenon, by metonymy, by metaphor, or whatever. Is over basically a place preposition, designating the higher location of the trajector, as argued by Tyler and Evans (2001), or does it designate a complex configuration, combining the senses of above and across, as proposed by Brugman (1981) and Lakoff (1987)? If the latter, is the preposition indifferent as to the shape of the trajectory, or it is associated with a distinctive arc-like up-down configuration, as proposed by Dewell (1994)? An alternative possibility is that the basic sense of the word is to be understood in terms of the location of an entity intervening between a usually unnamed observer and a landmark (Deane 2005). As a matter of fact, similar questions arise in connection with nominal concepts. What, for example, is the “basic” meaning of photograph? Is it an image or the representation of the image in a certain medium? Consider the following: (3) That photograph of me is torn. Photograph of me requires the photograph to be understood as an image. To state that the photograph is torn, however, requires that the photograph be construed as a

158

John R. Taylor

material substance of a certain kind, such as a sheet of paper. An image and a piece of paper imprinted with the image are different ontological kinds. Nevertheless, sentence (3) displays not the slightest incongruity, as might be expected when types are mixed. It is as if the concept ‘photograph’, while by no means schematic for both the image and the medium interpretations, is nevertheless able to encompass both. How this might be achieved is addressed in the next section.

3.

Domain activation

It is axiomatic in cognitive-linguistic approaches to semantics – and indeed in encyclopaedic approaches more generally – that concepts do not exist as encapsulated, self-contained chunks of semantic content. Any conceptualization presupposes a conceptual network of interrelated knowledge. A crucial notion is that of domain (Langacker 1987; Croft 1993). A domain is any coherent network of knowledge which provides the necessary background for understanding a concept. Typically, a concept will need to be characterized against more than one domain, while the domains themselves often presuppose other, more abstract, general domains. For example, the concept ‘uncle’ presupposes a kinship network; kinship, in turn, rests on notions of gender, marriage, procreation, family, and parenting. The concept is also understood against notions of a human being, and indeed of a physical object, with properties of size, mass, shape, and so on. In addition, there are the cultural norms pertaining to the role of an uncle in the life of the nephew/ niece. Specific uses of the word uncle may bring one or more of these domains into perspective. Thus, a statement that someone was an uncle to me perspectivizes the cultural domain while downplaying, or even overriding, the kinship domain. Consider the concept ‘television’. The concept is a complex one and needs to be understood against a large number of equally complex and multifaceted domains. To speak of “the invention of television” focuses on the science and technology of the transmission of visual images; to say that someone “works in television” activates our understanding of the activities surrounding the production of the transmitted content; if a play is “adapted for television” the focus is on the artistic medium; while if you “buy a television” you are referring to the receiving device. As was the case with the photograph example, the word television can refer to ontologically distinct kinds of entities. The concept cannot be regarded as an abstraction over these different kinds, neither is it evident which of the different readings might be regarded as basic, from which the others could be derived, by processes of metonymy, for example. The mental construct associated with the word is better regarded as a region in a conceptual network, only some of whose links are activated when the word is encountered in context.5 This kind of approach may also be indicated with respect to the concept ‘university’ in (1). On the one hand, the concept is understood against the domain of in-

Contextual salience, domains, and active zones

159

stitutions, specifically, institutions for teaching and research; it is also understood against the knowledge of buildings and their function. (There are other aspects too, of course. The institution domain includes the knowledge that individuals work for the institution and promote its agenda. This knowledge is activated when we learn, for example, that “the university has just telephoned”.) As we saw in the last section, an active zone (or metonymy) account may well be able to accommodate the different readings of the word. On the other hand, a focus on domains, and domain activation, has the advantage that it removes the need to propose a basic, unified account of a word’s meaning. Instead, we can say that a word provides access to different links in a conceptual network. The domain-based approach brings other benefits. One of these concerns the possibility that some of the network links may be strongly entrenched, having been activated many times on previous occasions.6 These entrenched associations are likely to constitute the default interpretation of a word, the interpretation that is invoked when a word is contemplated in isolation or with minimal supporting context. It is in such terms that we might want to explain what Evans (2009: 207) has called “the illusion of semantic unity”. This is the belief, strengthened no doubt by the existence of print dictionaries, that words do have a fixed and clearly delineated meaning (or several such meanings, in the case of polysemous items), permanently stored in the mental lexicon.7 There have been a number of theoretical perspectives addressing the role of domain-based knowledge. For Evans (2009), a lexical item merely provides access to parts of a conceptual network; the item itself is virtually empty of semantic content, beyond that supplied by its morphosyntactic status. (Thus, a noun names a thing, an adjective names a property, and so on.) In Evans’s theory, the building up of a semantic representation for a sentence is handled by the stepwise selection and integration of conceptual knowledge, whose progress is determined, in large part, by the syntactic constituency of the sentence. Others, in contrast, have proposed semantic representations rich in conceptual content. For Wierzbicka (1996), the content is dictated by the uses of a word in the language. The fact that we can refer to a person as a grey mouse requires that the ‘mouse’ concept include the information that mice are small inconspicuous creatures (Wierzbicka 1996: 340– 341). On the other hand, encyclopaedic knowledge about the prolific breeding habits of mice does not feature in the concept, since this aspect, according to Wierzbicka, has no linguistic repercussions. It is different, of course, with rabbits. Even so, it cannot be ruled out that speakers might come up with the expression breed like mice, thus requiring that the dictionary definition should be expanded so as to include this information.8 The example demonstrates the impossibility, in principle, of drawing a line between linguistic-semantic and encyclopaedic knowledge. It also confirms the open-ended nature of domain-based knowledge, a matter which has been emphasized by Langacker (1987: 158). In principle, any piece of knowledge that a speaker might associate with an entity is available for activation in a suitable linguistic context.

160

John R. Taylor

Some scholars have attempted to formalize the activation of domain-based knowledge. Thus, Bierwisch (1983) proposed the following as the semantic representation of ‘university’. (4) X [PURPOSE[X,Y] & ADVANCED STUDY AND TEACHING[W]] The value of X is determined by conceptual interpretation. To this end, he proposed the following functions: (5) a. X [INSTITUTION[X] &PURPOSE[X, W]] b. X [BUILDING[X] & PURPOSE[X,Y]] In this way, the ‘university’ concept is able to be instantiated as either an institution or a building. (Presumably, an additional function would be required in order to accommodate the personnel interpretation of the word, as in the earlier cited example The University has just telephoned.) Pustejovsky (1991) attempted to systematize the knowledge associated with nominal concepts in terms of an entity’s qualia, or essences. Four kinds of qualia are proposed: constitutive (what the thing is made of), formal (what the thing looks like and how it is identified), telic (its purpose or function), and agentive (how it was created or came about). Information associated with the qualia is selectively activated by context. Consider what it means to “begin a book”. Strictly speaking, the verb begin subcategorizes for a process. Book, however, does not designate a process, but a thing. In order to make sense of the expression, we need to make reference to a process in which books participate. Since books are meant to be read (their telic quale), begin a book can be interpreted as ‘begin to read a book’. Since books have to be written by someone (their agentive quale), an alternative interpretation is ‘begin to write a book’. There are other possibilities. Consider what begin a book might mean to a translator, a proof-reader, or a book-binder. Here, the interpretation, by a process of co-composition, has as much to do with the telic qualia of translators, proof-readers, and book-binders, as it has with the qualia of books. The compositional process is, however, subject to various constraints. For example, a feature of begin, as noted by Pustejovsky and Bouillon (1995: 136–137), is that the invoked process has to involve a change of state and/or have a definite (intended or actual) end-point. Begin books, with an indefinite plural, is hardly acceptable at all, since the process is unbounded. Similarly, begin the highway is likely to mean ‘begin to build the highway’ rather than ‘begin to drive along the highway’ (an unbounded process), even though highways are meant to be driven along (their telic quale). There is one type of expression whose interpretation draws heavily on background, domain-based knowledge, namely noun-noun compounds.9 Structurally, the compounds consist simply of two juxtaposed nouns. In keeping with general principles of English morphology, the final noun constitutes the head while the initial noun functions as a kind of restrictive modifier. The semantic relation between

Contextual salience, domains, and active zones

161

the two nouns is, however, left open. Nevertheless, and in spite of the fact that noun-noun compounds can be freely created (Benczes 2006), their interpretation is usually fairly clear. A water pistol shoots out water, water skis are for skiing on water, water fowl live on water, water crystals are made of water, while a water tank is for storing water. These interpretations depend on our knowledge of the function and purpose of pistols, skis, and storage tanks, the constituency of crystals, and the possible habitats of birds. Jackendoff (2010) points to the sometimes complex interaction between the knowledge structures associated with the component nouns. A cheese shop is a shop which sells cheese; the interpretation derives from the fact that the function of a shop is to sell things, and cheese is a saleable commodity. On the other hand, a barber shop does not sell barbers, since barbers are not a saleable commodity. The relevant point for the interpretation of the compound is the function of a barber, which is to shave people. What a barber shop sells is the service which barbers provide. The selective activation of domain-based knowledge has been studied experimentally using the technique of cued recall. Anderson and Ortony (1975) presented subjects with sentences like the following: (6) a. Televisions need expert repairmen. b. Televisions look nice in family rooms. It turned out that appliance was a better cue for the recall of (a), while (b) was better cued by the word furniture. It would seem that subjects, on interpreting these sentences and then committing them to memory, conceptualized the televisions in slightly different ways, focusing on the different domains against which the receiving apparatus can be understood. Greenspan (1986) found that there were asymmetries in the effectiveness of recall cues. The first sentence in (7) focuses on the piano as a musical instrument, while the second construes it merely as a heavy, bulky object. (7) a. The young man tuned the piano. b. The young man lifted the piano. Heavy proved to be a better cue for (b) than for (a). However, music was equally effective for both sentences. Greenspan discussed the findings in terms of central vs. peripheral properties of an entity. A piano is inherently a musical instrument; it is only incidentally a heavy object. The account is in line with Langacker’s (1987: 159) contention that some domains are more central to the understanding of a concept while others are more peripheral or circumstantial. Schoen (1988) addressed this issue by having subjects judge the importance of properties of a nominal concept. In one condition, the nouns were presented in isolation. For the noun walnut, the property ‘is edible’ was rated higher than ‘is hard’ and ‘is round’. The property ‘is edible’ was still highly rated when walnut was embedded in contexts which highlighted other facets of the concept. On the other

162

John R. Taylor

hand, the sentence The walnut wouldn’t break dramatically increased the salience of the property ‘is hard’, while The walnut rolled down the hill increased the rating of the property ‘is round’.

4.

Instantiation

The lexicon of a language is structured in a number of ways. One relation is the syntagmatic relation between items which typically occur in close proximity; this is the relation between, for example, bark and dog, or between unmitigated and its typical collocate, disaster. Another is the paradigmatic relation between a more general term, the hypernym, and more specific terms, its hyponyms. Langacker (1987: 371) discusses the matter in terms of the relation between a schema and its instances, while in Hudson’s (2007) Word Grammar the relation is represented by means of the “isa” relation. Thus, musical instrument is a hypernym of (or, in Langacker’s terminology, is schematic for) stringed instrument, woodwind instrument, and so on, while violin and viola are instances of stringed instrument. Although taxonomies are often discussed as relations amongst nominals, the relation applies equally to words of other categories. Thus, do is schematic for go, walk is an instance of go (or, in Hudson’s terminology, walking isa going and going isa doing). As the examples show, a more specific term may itself function as the hypernym for a set of even more specific hyponyms, a fact which can result in taxonomies of considerable depth. A point stressed by Hudson is that the isa hierarchy extends right down to individual instances. The unique referent of The violin he is now playing isa violin; the unique temporally and spatially located walking event designated by Mary walked to the store this morning isa walking. One of the properties of a taxonomy (or isa hierarchy) is that each term, barring cases of override,10 inherits all the features of each higher term. A further property, one that is of particular relevance to the present topic, is that any term, other than uniquely referring terms at the very lowest level, is necessarily vague with respect to the range of its instantiations. It follows that any linguistically encoded information will always be underspecified with respect to the situation it refers to. If you learn that a person plays “a musical instrument”, or that they own “a vehicle”, you cannot say with any certainty what kind of musical instrument, or what kind of vehicle, is involved. To be sure, default interpretations might come to mind. The topic has been investigated in connection with prototype theories of categorization (Rosch 1978; Taylor 2003a). Thus, mention of a vehicle suggests, in the first instance, a car, rather than a tractor or a motorized wheel-chair. Not only this, but subjects are able to construct a mental image of the prototypical instance, they can make a drawing of it, they can state its parts and their arrangement, and can describe its appearance and how one typically interacts with it (Rosch and Mervis 1975). As mentioned, the prototype is a default interpretation, and, as Barsalou

Contextual salience, domains, and active zones

163

(1987) documented, the default can easily be overridden by more specific knowledge and expectations. The vehicle purchased by a farmer might well be assumed to be a tractor, or, in the case of a handicapped person, a motorized wheel-chair. Instantiation has been studied experimentally, again using the technique of cued recall. Anderson and Ortony (1975) presented subjects with sentences like the following: (8) a. The container held the apples. b. The container held the cola. Basket turned out to be a more effective retrieval cue for the first sentence, bottle for the second. Evidently, subjects had formed a mental image of the situation described in (8a), conceptualizing the container as one which is capable of holding apples, such as a basket, while (8b) requires that the container be one which can hold a liquid, prototypically a bottle. Interestingly, Anderson et al. (1976) found that a hyponym could be a better recall cue than the general term that had been used in the stimulus sentence. The researchers constructed pairs of sentences containing a general word, that is, a word with a highly schematic meaning, such as woman, fruit, or utensil. One of the pairs suggested a specific instantiation of the general term, the other did not. Examples of their materials are the following: (9) a. The woman was outstanding in the theatre. b. The woman worked near the theatre. (10) a. The fruit was made into wine. b. The fruit was served with the wine. It was found that actress and grapes were better recall cues for the (a) sentences than the words actually used in the sentences, namely woman and fruit. These findings, along with those of Anderson and Ortony (1975), suggest that when we process a sentence we are inclined to go beyond the inherently schematic information encoded by its constituents and attempt to construct a “situation model” (van Dijk and Kintsch 1983; Zwaan and Radvansky 1998), that is, a plausible scenario with respect to which the sentence would be judged to be true. In constructing the model, the interpreter draws on multiple sources of information, including memories of the visual and sensori-motor experiences associated with the objects and events designated by the sentence (Barsalou 2008), as well as memories of previous uses of the words (a matter I take up below). Situation models built up by previous discourse will also play a part in constraining the interpretation of an incoming sentence. Inevitably, the situation model is much richer in detail than the content supplied by the words and their “dictionary” definitions. Zwaan (2004) invites us to compare the sentences in (11):

164

John R. Taylor

(11) a. The ranger saw the eagle in the sky. b. The ranger saw the eagle in the nest. Although the sentences are syntactically parallel the designated situations are likely to be conceptualized rather differently. Mention of an eagle in the sky activates a rather different image than the notion of an eagle in its nest. Not only this, but the interpreter vicariously simulates, or mentally re-enacts, the activity of the subject referent. Thus, the first sentence has the ranger gazing upwards at a faraway eagle with outspread wings, while the second has the ranger peering at close quarters at an eagle with drawn-in wings. The schematic notion of ‘seeing’ receives two very different “embodied” interpretations. In light of these remarks let us re-consider some of the examples discussed earlier. The occurrence of the verb lift in (7b), in association with a human subject nominal, is likely to activate the sensori-motor experience of tensing one’s abdominal muscles in an attempt to raise a heavy object from the ground, thereby perspectivizing the weight and bulk of the piano. Sentence (6b) is likely to invoke a scene of a television set prominently displayed in a living room, whereas (6a) conjures up an image of a technician with his implements reaching into the mass of cables inside the device. It is these simulated scenes which invoke, respectively, the notions of the television as a piece of furniture or as an appliance, concepts which are not explicitly encoded in the stimulus sentences. Finally, with respect to our initial example sentences in (1), the interpretation of a short time ago is dependent on imaginary narratives of a person leaving a building and leaving an institution, the former being a common everyday event while the latter is comparatively infrequent. As regards the choice between the physical displacement vs. the institutional reading of the sentence, this is likely to be determined by the situation model built up in preceding discourse.

5.

A phonological excursus

Saussure (1964) famously characterized the linguistic sign – prototypically a word – as the association of a concept with an acoustic image. The Saussurian characterization was endorsed by Langacker (1987) on the opening pages of his Foundations of Cognitive Grammar. Indeed, it is axiomatic in Cognitive Grammar that linguistic structures can be exhaustively analyzed in terms of symbolic units, each of which links a semantic representation with a phonological representation. Langacker’s (and Saussure’s) characterization of the linguistic sign (or symbolic unit) might suggest a rigid one form-one meaning relation. Indeed, the biunique relation has often been invoked by theorists as an ideal state to which language systems aspire (Taylor 2003b). The ideal is, of course, rarely met. It is frustrated, most obviously, by the existence of polysemy (one form, several meanings)

Contextual salience, domains, and active zones

165

and synonymy (one meaning, several forms). The discussion in this chapter has highlighted further aspects of variation at the semantic pole of the linguistic sign. Not to be overlooked, however, is variation at the phonological pole. In fact, phonological variation is no less pervasive and no less remarkable than semantic variation. There are several reasons why we should pay attention to phonological variation. Pronunciations are observable and subject to inter-observer validation in ways in which meanings are not. Meanings can only be studied indirectly, through introspection, acceptability judgements, and observations of context of use, supplemented, whenever appropriate, by psycholinguistic experimentation. The study of phonological variation, and theoretical approaches to it, can thus anchor the study of the inherently more difficult topic of semantic variation. It turns out that the issues for phonological theory are very similar to those encountered with respect to word meanings. In view of this, we need to be open to the possibility that theories of phonological variation might provide a model for the treatment of semantic variation. As stated, variation at the phonological pole of the linguistic sign is no less extensive than variation at the semantic pole. Consider a few examples. The word upper may be pronounced, in careful speech, as [pə], though it is more likely to come out as [φə], with a bilabial fricative. The word because can be pronounced

kɒs], amongst many others. The sequence because if [bikɔz], [bikɒs], or [b may come out as [kzif], where the word because is represented solely by the [kz] cluster. The province of Canterbury, in New Zealand, is often heard as two syllables [ka˜bɹi]. Many more examples are cited in Johnson (2003). There are, in addition, the well-studied differences in regional and social accent, which affect not only prosodic aspects of an utterance, but the acoustic quality of the constituent segments, above all, the vowels. There are many factors which impinge on a word’s pronunciation. These include (a) the phonological properties of adjacent items, triggering assimilation and sandhi (or linking) phenomena; (b) prosodic features of the utterance, which may affect elisions and reductions; (c) characteristics of the speaker, including gender, ethnicity, and social and regional background, responsible for the acoustic-phonetic properties of the vowel and consonant segments; and (d) discourse factors, include rate of speaking and the degree of formality. There are also textual factors, having to do with the predictability of a word in context (Bell et al. 2009). Thus, the more predictable a word or word sequence (the lower its information content), the more likely it is to receive a compressed or reduced pronunciation. Phonological variation raises much the same kinds of questions as semantic variation. One such question concerns the way in which a word’s pronunciation is stored in the mental lexicon. One possibility is that each word-form is associated with a unique phonological representation which is subject to modification according to the contexts, both speaker- and discourse-dependent, of its use. The “ideal”

166

John R. Taylor

pronunciation might be thought of in terms of a broad phonemic transcription of a citation form. This, certainly, appears to have been the view of Jakobson and Halle (1968: 413). What they refer to as “elliptical”, “slurred”, and even “slovenly” speech forms are, they claim, “derivative of the explicit clear-speech form that carry the highest amount of information”. Accordingly, the study of the sound pattern of a language needs to be based on the “fullest, optimal codes at the command of the given speakers” (Jakobson and Halle 1968: 414). On what I take would be the standard approach to the matter, the word upper would thus be associated in the mental lexicon with the representation /pə/ (or in some other equivalent format, such as a matrix of the distinctive features which make up the component segments), while the word’s “slurred” pronunciation as [φə] would be the output of a fairly general process of stop lenition. Other “slurring” processes, however, are likely to be specific to certain lexical items, or groups of items. Bybee (2007: 41) contrasts memory and mammary, noting that the former, due no doubt to its higher usage frequency, is more likely to be reduced to two syllables. For some words, to be sure, there might be reasons for assuming that they are associated with two or more unique representations. Thus, given the entirely ad hoc and item-specific nature of any rule which would derive the one form from the other, either would be stored as both /aiðə/ and /iðə/, data as both /deitə/ and /datə/. The situation, it will be noted, is akin to proposing that a word is polysemous, that is, it is associated with two or more distinct meanings.11 An account such as the above would need to be augmented in two ways. First, there would need to be a set of processes whereby a unique underlying form is realized in accordance with various context- and speaker-related factors. Second, the hearer would have to “undo”, or reverse, the word’s contextualization, in order to retrieve the underlying form and to match it with the form stored in the mental lexicon. In the cases cited above – [kz] as the pronunciation of because, [ka˜bɹi] as the pronunciation of Canterbury – the discrepancy between the encountered pronunciation and the presumed maximally explicit form is considerable, and could be bridged only by a substantial battery of phonological rules, some of which might even be specific to a single lexical item. Complicating the matter further is the issue of dialectal variation, and the question whether speakers competent in two or more varieties of their language are able to relate these to unique, pan-dialectal representations. In light of the problems associated with notion of unique phonological representations, it is perhaps not surprising that phonologists have been exploring an alternative and rather radical approach to the issue (Lachs, McMichael, and Pisoni 2000; Port 2007, 2010). This is the exemplar model of phonological storage, according to which the mental representation of a word’s pronunciation consists of a memory store of all previously encountered instances, indexed for their contextual properties, such as the properties of the speaker, the formality of the discourse, and the lexical and prosodic context. The stored exemplars cluster according to their

Contextual salience, domains, and active zones

167

similarity along various dimensions, with some clusters acquiring the status of prototypes, others being outliers. Speaking would be a matter of selecting from an appropriate region in the phonological space; listening would involve matching the incoming signal with one of the stored exemplars (Pierrehumbert 2001). There is circumstantial evidence for the correctness of this approach. Consider, for example, the fact that speakers of a language are exquisitely sensitive to dialectal and regional variation, being able to “place” a speaker on the basis of sometimes minute acoustic-phonetic cues (Hay and Drager 2007). This ability would be difficult to explain, except on the assumption that language users have mentally registered the pronunciation variants and indexed them with the characteristics of their speakers. There is, moreover, experimental evidence that subjects are able to remember speaker-specific aspects of how a word was pronounced, an effect which can persist for several weeks (Goldinger 1996).

6.

Conclusion

Developments in exemplar phonology cast doubt on the notion that words are associated in the mental lexicon with rigid phonological representations from which the full range of pronunciation variants can be derived by rule. Rather, speakers accumulate a memory store of pronunciations, indexed for their contextualized properties. The question arises whether a comparable state of affairs might not obtain with regard to the mental representation of word meanings.12 It is worth noting in this connection that exemplar models of category representation have a distinguished pedigree in cognitive psychology (Smith and Medin 1981; Ross and Makin 1999); they have also been successfully implemented in computer simulations of learning and recall (Hintzman 1986; Kruschke 1992). We would be led to question the notion that words are associated in the mental lexicon with a fixed chunk of semantic information, which is then modified in accordance with various contextual factors. Although I started out this chapter by problematizing meaning variation in precisely such terms, the ensuing discussion proposed alternative approaches to word meaning, more in keeping with the variation found in actual contexts of use. Word meanings, it was suggested, constitute a region in a conceptual network, whereby a contextualized use of the word activates only some of the links in the network. In addition, understanding a word in context involves the simulation of the designated scene, with interpreters drawing on their memories of sensori-motor experiences associated with the objects and events. It is likely that speakers and interpreters also draw on their memories of previous uses of the lexical items in question. Indeed, an exemplar approach to word representation underlies the well-documented phenomenon of collocation, that is, the tendency of a word to co-occur with other words in its near vicinity. Speakers are well aware of these tendencies. Consider the two sentences below:

168

John R. Taylor

(12) a. The trip was a total failure. b. The trip was a total success. While both sentences are equally ‘grammatical’, speakers are likely to judge (a) to be more ‘idiomatic’ than (b). The reason lies not so much in the meaning we might want to attribute to total, but in the fact that total failure is a more frequent collocation than total success13 and that language users appear to have learned this property of the word on the basis of their previous exposure to the language. It may therefore be inappropriate to enquire into the meaning(s) of words like leave, open, or over, even less, to assume a basic or central meaning, permanently stored in the mental lexicon, from which contextual variants are derived. Many of the contextual variants will already have been stored, not, however, in the manner suggested by Lakoff (1987), in terms of rampant polysemy, but as collocations, phraseologies, and pre-formed chunks, whose semantic representations are not built up compositionally, but which correspond to recurring patterns of experience. Taylor (2006) compares Lakoff’s (1987) and Searle’s (1983) accounts of the semantics of open. Lakoff, as we have noted, favours extensive polysemy while Searle pleads for a unitary representation. What both authors ignore, however, is the possibility that speakers have simply learned the phraseologies associated with the verb, e.g. open the window, open your mouth, open your arms, open the newspaper, etc., each of which is understood against a cultural practice along with its associated sensori-motor patterns. The question of the “real”, or “basic” meaning of open does not arise. Neither do we need to appeal to a supposed basic meaning in order to understand novel combinations. Suppose you are told to “open the television”. This is not an accepted collocation (a fact confirmed by corpus searches) and it is by no means obvious what you are supposed to do. You are likely to interpret the request, not by a compositional process operating over the stored senses of the constituent words, but by drawing analogies with other uses of the verb, while respecting the constraints of the current situation model. In the first section of this chapter I briefly mentioned the division of labour between semantics and pragmatics. My discussion of words and their meanings suggests that the distinction may well be an artificial one (see also Gibbs 1994). Word meanings are the creations of context, where “context” ranges over conceptual structure, the linguistic environment, and embodied cultural practices.

Notes 1. The example is adapted from Bierwisch and Schreuder (1992: 31). 2. The term was popularized by Louw (1993) and refers to the phenomenon whereby a semantic feature, rather than inhering in a particular item, spreads over a larger unit. 3. The example is not without its problems, however. Consider the translation equivalent of (2c) in a language such as Italian, which marks grammatical gender on adjectivals. The

Contextual salience, domains, and active zones

4. 5.

6. 7.

8.

9. 10. 11.

12.

13.

169

choice between Io sono parcheggiato/a all’angolo depends on the gender of the speaker, not on the gender of the intended referent (presumably la macchina, ‘the car’, a feminine noun). Although the personal pronoun appears to refer to the car, its meaning, for the purposes of the morphosyntactic form of the sentence, is the speaking subject (Nunberg 1995). The term alludes to Brugman’s (1981) eponymous thesis, summarized in Brugman and Lakoff (1988). Brugman’s analysis is elaborated in Lakoff (1987: Case study 2). Note that the different understandings of television correlate with different lexico-semantic contexts (work in television, adapt for television, etc.) and even with different syntactic statuses, concerning, for example, the status of television as a mass or a count noun. As pointed out earlier in this chapter, the linguistic context goes some way towards disambiguating the possible readings of the word. For the notion of entrenchment, see Langacker (1987: 59). Matching the illusion of semantic unity is the illusion of phonological unity, that is, the belief that words are associated with a fixed and stable phonological representation. I take up this matter later in this chapter. As a matter of fact, internet searches turned up many examples of breed like mice, used to denote prolific reproduction, though no examples are attested in the BNC (British National Corpus) or in the COCA (Corpus of Contemporary American English; Davies 2008). For the role of domain-based knowledge in conceptual combination more generally, see Murphy (2004: Ch. 12). Thus, a feature of birds is that they can fly. This feature is overridden by specific knowledge pertaining to penguins and kiwis. As is well-known, Chomsky and Halle (1968) went so far as to propose unique phonological representations, not only for the words of a language, but also for the morphemes, even for those which appear in several phonological guises. Thus, the morpheme which is common to both divine and divinity is supposed to have a unique underlying representation which, on Chomsky and Halle’s account, is distinct from each of its surface manifestations. There are, of course, precedents for basing semantic theory on concepts adapted from phonology. For example, the semantic features of componential analysis (Lyons 1977: 317) were modelled on the distinctive features of phonological theory. According to both the BNC and the COCA, total failure is about four times more frequent than total success.

References Anderson, Richard C., James W. Pichert, Ernest T. Goetz, Diane L. Schallert, Kathleen V. Stevens and Stanley R. Trollip 1976 Instantiation of general terms. Journal of Verbal Learning and Verbal Behavior 15: 667–679. Anderson, Richard C. and Andrew Ortony 1975 On putting apples into bottles: A problem in polysemy. Cognitive Psychology 7: 167–180.

170

John R. Taylor

Barcelona, Antonio 2002 Clarifying and applying the notions of metaphor and metonymy within Cognitive Linguistics: An update. In: René Dirven and Ralf Pörings (eds.), Metaphor and Metonymy in Comparison and Contrast, 207–77. Berlin: Mouton de Gruyter. Barclay, J. R., John D. Bransford, Jeffery J. Franks, Nancy S. McCarrell and Kathy Nitsch 1974 Comprehension and semantic flexibility. Journal of Verbal Learning and Verbal Behavior 13: 471–481. Barsalou, Laurence 1987 The instability of graded structure: Implications for the nature of concepts. In: Ulrich Neisser (ed.), Concepts and Conceptual Development: Ecological and Intellectual Factors in Categorization, 101–140. Cambridge: Cambridge University Press. Barsalou, Laurence 2008 Grounded cognition. Annual Review of Psychology 59: 617–645. Bell, Alan, Jason Brenier, Michelle Gregory, Cynthia Girand and Dan Jurafsky 2009 Predictability effects on content and function words in conversational English. Journal of Memory and Language 60: 92–111. Benczes, Réka 2006 Creative Compounding in English. Amsterdam: John Benjamins. Bierwisch, Manfred 1983 Semantische und konzeptuelle Repräsentation lexikalischer Einheiten [Semantic and conceptual representation of lexical units]. In: R. Ru˚zˇicˇka and W. Motsch (eds.), Untersuchungen zur Semantik (= Studia grammatica XXII), 61–99. Berlin: Akademie-Verlag. Bierwisch, Manfred and Robert Schreuder 1992 From concepts to lexical items. Cognition 42: 23–60. Brugman, Claudia 1981 Story of over. MA thesis, University of California, Berkeley. Reproduced by the Indiana University Linguistics Club, Bloomington. Brugman, Claudia and George Lakoff 1988 Cognitive topology and lexical networks. In: Steven L. Small, Garrison W. Cottrell and Michael K. Tanenhaus (eds.), Lexical Ambiguity Resolution: Perspectives from Psycholinguistics, Neuropsychology, and Artificial Intelligence, 477–508. San Mateo, CA: Morgan Kaufmann. [Reprinted in: Dirk Geeraerts (ed.), Cognitive Linguistics: Basic Readings (2006), 109–139. Berlin: Mouton de Gruyter.] Bybee, Joan 2007 Frequency of Use and the Organization of Language. Oxford: Oxford University Press. Chomsky, Noam and Morris Halle 1968 The Sound Pattern of English. New York: Harper & Row. Choueka, Yaacov and Serge Lusignan 1985 Disambiguation by short contexts. Computers and the Humanities 19: 147–157. Croft, William 1993 The role of domains in the interpretation of metaphors and metonymies. Cognitive Linguistics 4: 335–370.

Contextual salience, domains, and active zones

171

Davies, Mark 2008 The Corpus of Contemporary American English (COCA): 385 million words, 1990–present. Available online at http://www.americancorpus.org. Deane, Paul 2005 Multimodal spatial representation: On the semantic unity of over. In: Beate Hampe (ed.), From Perception to Meaning: Image Schemas in Cognitive Linguistics, 235–282. Berlin: Mouton de Gruyter. Dewell, Robert 1994 Over again: Image-schema transformations in semantic analysis. Cognitive Linguistics 5: 351–380. Evans, Vyvyan 2009 How Words Mean: Lexical Concepts, Cognitive Models, and Meaning Construction. Oxford: Oxford University Press. Fillmore, Charles and B. T. S. Atkins 2000 Describing polysemy: The case of ‘crawl’. In: Yael Ravin and Claudia Leacock (eds.), Polysemy: Theoretical and Computational Approaches 91–110. Oxford: Oxford University Press. Gibbs, Raymond 1994 The Poetics of Mind: Figurative Thought, Language, and Understanding. Cambridge: Cambridge University Press. Goldinger, Stephen 1996 Words and voices: Episodic traces in spoken word identification and recognition memory. Journal of Experimental Psychology: Learning, Memory, and Cognition 22: 1166–1183. Greenspan, Steven L. 1986 Semantic flexibility and referential specificity of concrete nouns. Journal of Memory and Language 25: 539–557. Gries, Stefan Th. 2006 Corpus-based methods and cognitive semantics: The many senses of to run. In: Stefan Th. Gries and Anatol Stefanowitsch (eds.), Corpora in Cognitive Linguistics: Corpus-based Approaches to Syntax and Lexis 57–99. Berlin: Mouton de Gruyter. Hay, Jennifer and Katie Drager 2007 Sociophonetics. Annual Review of Anthropology 36: 89–103. Hintzman, Douglas L. 1986 “Schema abstraction” in a multiple-trace memory model. Psychological Review 93: 411–28. Hoey, Michael 2005 Lexical Priming: A New Theory of Words and Language. London: Routledge. Huang, Yan 2007 Pragmatics. Oxford: Oxford University Press. Hudson, Richard 2007 Language Networks: The New Word Grammar. Oxford: Oxford University Press. Jackendoff, Ray 2010 The ecology of English noun-noun compounds. In: Meaning and the Lexicon: The Parallel Architecture 1975–2010, 413–451. Oxford: Oxford University Press.

172

John R. Taylor

Jakobson, Roman and Morris Halle 1968 Phonology in relation to phonetics. In: Bertil Malmberg (ed.), Manual of Phonetics, 411–449. Amsterdam: North-Holland. Johnson, Keith 2003 Massive reduction in conversational American English. Proceedings of the Workshop on Spontaneous Speech: Data and Analysis, Tokyo. Accessed November 2011 at http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.142.5012& rep=rep1&type=pdf. Kilgarriff, Adam 1997 I don’t believe in word senses. Computers and the Humanities 31: 91–113. Kishner, Jeffrey and Raymond Gibbs 1996 How “just” gets its meanings: Polysemy and context in psychological semantics. Language and Speech 39: 19–36. Kruschke, John K. 1992 ALCOVE: An exemplar-based connectionist model of category learning. Psychological Review 99: 22–44. Lachs, Lorin, Kipp McMichael and David Pisoni 2000 Speech perception and implicitmemory: Evidence for detailed episodic encoding of phonetic events. Research on Spoken Language Processing, Progress Report No. 24, 149–67. Indiana University. Accessed November 2011 at http://www.iu.edu/~srlweb/pr/24/Lachs-149.pdf. Lakoff, George 1987 Women, Fire and Dangerous Things. Chicago: Chicago University Press. Langacker, Ronald W. 1984 Active zones. Proceedings of the Berkeley Linguistics Society 10: 172–188. [Reprinted in: Concept, Image, and Symbol (1991), 189–201. Berlin: Mouton de Gruyter.] Langacker, Ronald W. 1987 Foundations of Cognitive Grammar, vol. i: Theoretical Prerequisites. Stanford: Stanford University Press. Langacker, Ronald W. 1993 Reference point constructions. Cognitive Linguistics 4, 1–38. [Reprinted in: Grammar and Conceptualization (1999), 171–202, Berlin: Mouton de Gruyter.] Louw, Bill 1993 Irony in the text or insincerity in the writer? The diagnostic potential of semantic prosodies. In Mona Baker, Gill Francis and Elena Tognini-Bonelli (eds.), Text and Technology: In Honour of John Sinclair, 157–76. Amsterdam: Benjamins. Lyons, John 1977 Semantics. Cambridge: Cambridge University Press. Miller, George and Claudia Leacock 2000 Lexical representations for sentence processing. In: Yael Ravin and Claudia Leacock (eds.), Polysemy: Theoretical and Computational Approaches, 152–160. Oxford: Oxford University Press. Murphy, Gregory 2004 The Big Book of Concepts. Cambridge, MA: MIT Press.

Contextual salience, domains, and active zones

173

Nunberg, Geoffrey 1995 Transfers of meaning. Journal of Semantics 12: 109–132. Pierrehumbert, Janet 2001 Exemplar dynamics: Word frequency, lenition and contrast. In: Joan Bybee and Paul Hopper (eds.), Frequency and the Emergence of Linguistic Structure 137–157. Amsterdam: Benjamins. Port, Robert 2007 How are words stored in memory? Beyond phones and phonemes. New Ideas in Psychology 25: 143–170. Port, Robert 2010 Rich memory and distributed phonology. Language Sciences 32: 43–55. Pustejovsky, James 1991 The generative lexicon. Computational Linguistics 17: 409–441. Pustejovsky James and Pierrette Bouillon 1995 Aspectual coercion and logical polysemy. Journal of Semantics 12, 133–162. Rosch, Eleanor 1978 Principles of categorization. In: Eleanor Rosch and Barbara Lloyd (eds.), Cognition and Categorization, 27–48. Hillsdale, NJ: Lawrence Erlbaum. Rosch, Eleanor and Carolyn Mervis 1975 Family resemblances: Studies in the internal structure of categories. Cognitive Psychology 7: 575–605. Ross, Brian. H. and Valarie S. Makin 1999 Prototype versus exemplar models. In: R. J. Steinberg (ed.), The Nature of Cognition, 205–241. Cambridge, MA: MIT Press. Saussure, Ferdinand de 1964 Cours de linguistique générale. Paris: Payot. First published 1916. Translated by Roy Harris as Course in General Linguistics. London: Duckworth (1983). Schoen, Lawrence M. 1988 Semantic flexibility and core meaning. Journal of Psycholinguistic Research 17: 113–23. Searle, John 1983 Intentionality. Cambridge: Cambridge University Press. Smith, Edward E. and Douglas L. Medin 1981 Categories and Concepts. Cambridge, MA: Harvard University Press. Taylor, John, R. 1995 Approaches to word meaning: The network model (Langacker) and the twolevel model (Bierwisch) in comparison. In: René Dirven and Johan Vanparys (eds.), Current Approaches to the Lexicon, 3–26. Frankfurt: Lang. Taylor, John R. 2003a Linguistic Categorization. Oxford: Oxford University Press. [1st edition: 1989]. Taylor, John R. 2003b Polysemy’s paradoxes. Language Sciences 25: 637–655. Taylor, John R. 2006 Polysemy and the lexicon. In: Gitte Kristiansen, Michel Achard, René Dirven and Francisco Ruiz de Mendoza Ibáñez (eds.), Cognitive Linguistics: Current Applications and Future Perspectives, 51–80. Berlin: Mouton de Gruyter.

174

John R. Taylor

Tyler, Andrea and Vyvyan Evans 2001 Reconsidering prepositional polysemy networks: The case of over. Language 77: 724–65. [Reprinted as “The case of over” in: Brigitte Nerlich, Zazie Todd, Vimala Herman and David D. Clarke (eds.), Polysemy: Flexible Patterns of Meaning in Mind and Language, 2003, 99–159. Berlin: Mouton de Gruyter.] van Dijk, Teun and Walter Kintsch 1983 Strategies of Discourse Comprehension. New York: Academic Press. Wierzbicka, Anna 1996 Semantics: Primes and Universals. Oxford: Oxford University Press. Zwaan, Rolf 2004 The immersed experiencer: Toward an embodied theory of language comprehension. In: B. H. Ross (ed.), The Psychology of Learning and Motivation, 35–62. New York: Academic Press. Zwaan, Rolf and Gabriel Radvansky 1998 Situation models in language and memory. Psychological Bulletin 123: 162–185.

7.

Encyclopaedic knowledge and cultural models Istvan Kecskes

1.

Introduction

The term encyclopaedic knowledge is used to refer to the knowledge of the world as distinguished from the knowledge of the language system. The encyclopaedic view represents a model of the system of conceptual knowledge that underlies linguistic meaning. This system plays a profound role in how human beings make sense in communication. Traditionally, the division between the ontology and the lexicon illustrates the distinction between encyclopaedic and dictionary knowledge. Dictionary knowledge is supposed to cover the idiosyncrasies of particular words, whereas encyclopaedic knowledge covers everything regarding the underlying concepts. In cognitive linguistics, however, meaning, emerging from language use, is a function of the activation of conceptual knowledge structures as guided by context. Consequently, there is no principled distinction between semantics and pragmatics (e.g. Fauconnier 1997; Evans 2006). In cognitive approaches, practically no sentence encodes a complete thought. Certain processes of contextual filling-in are required before anything of a propositional nature emerges at all (Carston 1998). Encyclopaedic knowledge is considered to be a structured system of knowledge, organized as a network. Rosch’s (1977) view of category structure and its implications for reasoning (e.g. Lakoff and Johnson 1980) have had an important impact on the emergence of schema theory. Schema theory is a theoretical framework which has been developed in the cognitive sciences over the last two decades. The central concept, “schema”, has gone by a number of other names, including “cultural model” (Holland and Quinn 1987; D’Andrade 1992), “frame” (Minsky 1975; Fillmore, 1982), “mental model” (Johnson-Laird 1983), “idealized cognitive model” (Lakoff 1987), “folk model” (D’Andrade 1987) and “script” (Schank and Abelson 1977). This chapter will discuss encyclopaedic knowledge and its structure and role in the communicative process in the framework of a socio-cognitive approach to communication and pragmatics (Kecskes 2002, 2008; Kecskes and Zhang 2009). In this paradigm, culture is seen as a socially constituted set of various kinds of knowledge structures that individuals turn to as relevant situations permit, enable, and usually encourage. Encyclopaedic knowledge is mostly represented in cultural models that provide scenarios or action plans for individuals of how to interpret and behave in a particular situation or how to interpret the behavior of others in one or another situation.

176 2.

Istvan Kecskes

The socio-cognitive approach (SCA)

The socio-cognitive approach unites the societal and individual features of interaction and considers communication a dynamic process in which individuals are not only constrained by societal conditions but also shape them at the same time. Speaker and hearer are equal participants in the communicative process. They both produce and comprehend speech relying on their most accessible and salient knowledge expressed in their private contexts in production and comprehension. Consequently, only a holistic interpretation of utterances from both the perspective of the speaker and the perspective of the hearer can give us an adequate account of language communication. The socio-cognitive approach to communication and knowledge transfer (Kecskes 2008, 2010; Kecskes and Zhang 2009) emphasizes the complex role of cultural and private mental models, how these are applied categorically and/or reflectively by individuals in response to socio-cultural environmental feedback mechanisms and how this leads to and explains different meaning outcomes and knowledge transfer. In meaning construction and comprehension individuals rely both on pre-existing encyclopaedic knowledge and knowledge created in the process of interaction. 2.1.

A synthesis of positivist and social constructivist perspectives

The socio-cognitive approach tries to make a dialectical synthesis of positivism and social constructivism. According to the positivist epistemology, knowledge consists of objective facts that can be measured independently of the inquiring, interpreting, and creative mind. Bernstein (1983: 8) argued that “there is some permanent, ahistorical matrix or framework to which we can ultimately appeal in determining the nature of rationality, knowledge, truth, reality, goodness, or rightness”. In this paradigm, research focuses on procedural measures rather than interpretive perspectives. It is usually assumed that stored knowledge provides templates for thinking as well as acting (cf. Alvesson and Kärreman 2001). Meaning is embedded in words and symbols rather than in the mind that perceives them. In contrast to the positivist approach, the social constructivist perspective holds that knowledge and meaning are socially constructed. They are constituted and transferred through practices and activities (e.g. Wittgenstein 1953; Gherardi 2000, 2001; Brown and Duguid 2001). According to Vygotsky (1978), social reality and meaning only exist as we create them. Social constructivists see language use as socio-cultural construction. They put emphasis on usage and value the ways people currently use the language. Instead of looking for one self-professed authority to pronounce correct usage, constructivists would take a consensus of expert users. In sum, positivists consider words and texts as carriers of objectified meaning while for social constructivists practice (i.e. action, doing) plays that role. Rad-

Encyclopaedic knowledge and cultural models

177

ical constructivists (e.g. Ernst von Glasersfeld 1995) go one step further to say that knowledge, no matter how it is defined, is in the heads of people, and that the thinking person has no alternative but to construct what s/he knows on the basis of his or her own experience. What we make of experience constitutes the only world we consciously live in. The socio-cognitive approach argues that to equate practice with knowledge is to ignore the huge amount of pre-existing knowledge that both speakers and hearers must have in common for the hearer to infer and categorize the intended meaning of a practice. Practice can hardly work without the presence of relevant cultural mental models with which people process the observed practice, or which they use to actually create practice. Even when we pass on simple routines by sharing them in practice (e.g. how to prepare a certain dish), we rely on the presence of a large amount of pre-existing knowledge. Besides, practice does not provide semantic codes for its own decoding (i.e. sense making). These codes must already exist in the mind of the interpreter (Ringberg and Reihlen 2008). Without taking into account that meaning is mediated by people’s mental predisposed socio-cultural models, practice-based research is unable to explain creativity, innovation, and the transfer of meaning among interlocutors. The social character of communication and knowledge transfer should not put community-of-practice theory at odds with individualistic approaches to knowledge. After all, social practices pass “through the heads of people, and it is such heads that do the feeling, perceiving, thinking, and the like” (Bunge 1996: 303). While communities of practice exist, members of these communities may still interpret shared practices differently. Collective cultural models are distributed to individuals in a privatized way. In order for members to share the meaning of a particular practice, a huge amount of shared knowledge must already be present to assure common ground. Levinthal and Rerup (2006) argued that practice is similar to sentences in a text. Its grammar or structure is not meaningful apart from the meaning that is assigned by the receiver. The synthesis of the positivist and social constructivist views is a socio-cognitive approach that acknowledges the importance of both societal and individual factors in meaning creation and comprehension as well as knowledge transfer. Shared cultural models privatized through individuals’ private experience and prior practice interact with the actual situational context in social interaction and practices (Kecskes 2008). 2.2.

Communication in the socio-cognitive paradigm

In the socio-cognitive paradigm communication is driven by the interplay of “cooperation” required by societal conditions and “egocentrism” rooted in the prior experience of an individual. Consequently, egocentrism and cooperation are not mutually exclusive phenomena. They are both present in all stages of communication to a different extent because they represent the individual and societal traits

178

Istvan Kecskes

of the dynamic process of communication (Kecskes and Zhang 2009). On the one hand, speakers and hearers are constrained by societal conditions, but as individuals, on the other hand, they all have their own goals, intentions and desires, etc. which are freely expressed and recognized in the flow of interaction. In the socio-cognitive approach framed by the dynamic model of meaning (Kecskes 2008; Kecskes and Zhang 2009), communication is characterized by the interplay of two traits that are inseparable, mutually supportive and interactive: Individual trait: attention private experience egocentrism salience

Social trait: intention actual situational experience cooperation relevance

Communication is the result of the interplay of intention and attention motivated by socio-cultural background that is privatized individually by interlocutors. The socio-cultural background is composed of the encyclopaedic knowledge of interlocutors deriving from their “prior experience”, tied to the linguistic expressions they use, and their “current experience”, in which those expressions create and convey meaning. The process of privatalization through which the individual blends his prior experience with the actual situational (current) experience results in a dynamic process of meaning construction in which nothing is static. The two sides (prior and current) constantly change and affect each other. Meaning construction relies both on relatively definable cultural models and norms as well as situationally evolving features. Prior experience is represented in relatively definable cultural models and norms that are related and/or blended with actual situational experience. The socio-cognitive approach integrates the pragmatic view of cooperation and the cognitive view of egocentrism, and emphasizes that both cooperation and egocentrism are manifested in all phases of communication to a varying extent. While cooperation is an intention-directed practice and governed by relevance, egocentrism is an attention-oriented trait and governed by salience. Consequently, in communication we show our two sides. We cooperate by generating and formulating intention that is relevant to the given actual situational context. At the same time our egocentrism means that we activate the most salient information to our attention in the construction (speaker) and comprehension (hearer) of utterances. Language processing is anchored in the assumption that what is salient or accessible to oneself will also be accessible to one’s interlocutors (Giora 2003; Barr and Keysar 2005; Colston and Katz 2005; Kecskes 2007).

Encyclopaedic knowledge and cultural models

3.

179

The study of meaning in the encyclopaedic view

The study of meaning in cognitive linguistics relies on an encyclopaedic view of semantics which presupposes the following core assumptions (Evans 2006): 3.1.

“There exists no principled distinction between semantics and pragmatics”

Cognitive semanticists usually reject the idea that there is a distinction between “core” (dictionary) meaning on the one hand, and pragmatic, social or cultural meaning on the other. According to this approach, there is no autonomous mental lexicon which contains semantic knowledge separately from other kinds of (linguistic or non-linguistic) knowledge. Consequently, according to cognitive linguistics, there is no distinction between dictionary knowledge and encyclopaedic knowledge. There is only encyclopaedic knowledge, which incorporates both linguistic and extralinguistic knowledge. 3.2.

“Encyclopaedic knowledge is structured”

In cognitive linguistics, encyclopaedic knowledge is viewed as a structured system of knowledge, organized as a network. Moreover, not all aspects of the knowledge that is, in principle, accessible by a single word has equal standing (e.g. Evans 2002). Several terms have been used to denote the structured system of knowledge. These terms differ only slightly from each other. Frames are preconceived understandings of a new situation (e.g. we have a faculty meeting). Scripts are sequences of activities that we associate with a particular situation (i.e. we have procedures to follow when having a faculty meeting). Scenarios are sets of organized units in cognitive processes. They are components we anticipate for any new situation that has been given a label that we understand (i.e. we have an understanding of who and what should be present during a faculty meeting). Schemata are manifestations of higher-level knowledge that helps us understand a situation (i.e. our knowledge of practice in a faculty meeting). Mental or cultural models are logical sequences of thought that explain and give sense to a situation. There is some overlap between these terms but they give us some perspective from which to analyze our data (cf. D’Andrade 1992; Kronenfeld 2008). 3.3.

“Encyclopaedic meaning emerges in context”

Encyclopaedic meaning arises in context(s) of use. The “selection” of actual situational meaning is informed and maybe even determined by contextual factors. In the dictionary view of meaning, there is a separation of core meaning (semantics) from non-core actual meaning (pragmatics). The encyclopaedic view, however,

180

Istvan Kecskes

claims that encyclopaedic knowledge is included in semantics, and meaning is determined by context. According to this approach, there is no definable, pre-existing word meaning because the meaning of a word in context is selected and shaped by encyclopaedic knowledge. 3.4.

“Lexical items are points of access to encyclopaedic knowledge”

In the encyclopaedic view, lexical items are “points of access” to encyclopaedic knowledge (Langacker 1987). Consequently, words are not “containers” (cf. Reddy 1979) that present neat, pre-packaged bundles of information (Evans 2009). Instead, they selectively provide access to particular parts of the huge network of encyclopaedic knowledge. There are several theories in cognitive linguistics which adopt the encyclopaedic view, such as Frame Semantics (Fillmore 1982; Fillmore and Atkins 1992), the approach to domains in Cognitive Grammar (Langacker 1987), the approach to Dynamic Construal (Croft and Cruse 2004) and the Theory of Lexical Concepts and Cognitive Models – LCCM Theory (Evans 2006). One of the basic theories of the encyclopaedic view is Frame Semantics, which has been applied to several latter cognitive theories such as construction grammar and LCCM. Originally, Frame Semantics is a specific approach to natural language semantics developed by Charles Fillmore. The main idea is that we cannot understand the meaning of a word (or a linguistic expression in general) without access to all the encyclopaedic knowledge related to that word. According to Frame Semantics, meaning in natural language is not separated from other forms of knowledge. Consequently, it does not make sense to maintain a strict separation between world knowledge and knowledge of linguistic meaning. Frame semantics relies on the specific structures of encyclopedic knowledge, which are called frames. They relate to things happening and occurring together in reality. The internal structure of word meaning is not autonomous but exists against a background of our general assumptions about the world (including socio-cultural beliefs, assumptions etc.). In this approach, word meaning cannot be fully analyzed into features, since the meaning and its frame are inseparable from each other. The frame may not be part of the lexical meaning itself, but our understanding of meaning crucially involves analysis of both the frames and the lexical senses which depend on them. Frames can either be created by or reflected in the language. An example of a frame created by the language itself would be the case of grading detergent packages. The framing device allows the consumer to properly interpret the label “large” on the package; knowing that the other sizes are “economy”, “family size” and “jumbo”, the consumer is led to the correct conclusion that “large” signifies the smallest package. These core assumptions are not always maintainable in the socio-cognitive approach as we will see in the following sections.

Encyclopaedic knowledge and cultural models

4.

Cultural models and the intersection of the socio-cultural and the individual

4.1.

The nature of cultural models

181

Cultural models are cognitive frames or templates of assumed or implicit knowledge that assist individuals in interpreting and understanding information and events. Encyclopaedic knowledge includes cultural models that are usually defined as “a cognitive schema that is intersubjectively shared by a social group” (D’Andrade 1992: 99). There exist certain mental schemas which are activated when an individual experiences similar new situations or linguistic tasks. The notion of schema was first introduced by Immanuel Kant to account for the mediation between logical concepts and sensory information, which gives significance to our mental representations. Research exploring the intersection of culture and the individual claims that cognition consists of subsets of shared cultural models that organize much of how people make sense of the world (cf. D’Andrade 1992; Shore 1996; DiMaggio 1997). D’Andrade argued that a cultural model can be understood as “an interpretation which is frequent, well organized, memorable, which can be made from minimal cues, contains one or more prototypic instantiations, and is resistant to change” (1992: 29). In cognitive linguistics, the cultural models underlying reasoning and argumentation are considered to some extent idealized entities (see, for instance, the notion of “Idealized Cognitive Models” as introduced in Lakoff 1987). Geeraerts argued that actually occurring phenomena and actual situations usually differ to a smaller or greater extent from the models that act as cognitive reference points. The models themselves appear to be somewhat abstract, general, or even simplistic, because we use them to make sense of phenomena that are oftentimes more complicated (Geeraerts 2006: 274). Cultural models become internalized by individuals through everyday shared experiential processes (cf. DiMaggio 1997). These experiential processes are cognitive patterns that develop from different types of inputs, such as instruction, activities, communication, observation, practices, etc. Every human being is exposed to various aspects of the socio-cultural life, which leads to membership of a subset of socio-cultural speech communities (Shore 1996). Each speech community is identified by a variety of dominant cultural models that provide certain assumptions and a certain outlook on the world. Because cultural models are a part of a person’s cognitive resources, they influence his/her world view and behavior, as well as the way s/he interprets and reacts to other people’s behavior, information and situations. But we must be careful because, although cultural models usually create a harmonizing effect, people are not cognitive clones of culture. Collective cultural models are internalized and privatized by individuals through their own experience and developed into private mental models. However, any sharp distinction between private and collective cultural models is purely analytical. In real life, such a dis-

182

Istvan Kecskes

tinction is gradual and depends as much on an individual’s cognitive dispositions as it does on life experience. Consider the following example: (1) Car rental Clerk: Customer: Clerk: Customer:

What can I do for you, sir? I have a reservation. May I see your driver’s license? Sure. Here you are.

Most people are familiar with the cultural frame of renting a car. Certain situationbound utterances (see Kecskes 2000; 2002) such as What can I do for you?, I have a reservation, May I see your driver’s license? and the like are expected to be used in this frame. However, how exactly this frame is played off depends on the prior experience of the individuals who participate in its activation. When language is used, its unique property is activated in two ways. When people speak or write, they craft what they need to express to fit the situation or context in which they are communicating. But, at the same time, the way people speak or write these words, the expressions and utterances they use create that very situation, context and socio-cultural frame in which the given communication occurs. Consequently, two things seem to happen simultaneously: People attempt to fit their language to a situation or context that their language, in turn, helped to create in the first place (Gee 1999). This dynamic behavior of human speech and reciprocal process between language and context basically eliminates the need to ask the ever-returning question: Which comes first? The situation the speakers are in (e.g. a faculty meeting, renting a car, ordering dinner etc.) or the particular language that is used in the given situation (expressions and utterances representing ways of talking and interacting)? Is this a “car rental” because participants are acting and speaking that way, or are they acting and speaking that way because this is a “car rental”? Acting and speaking in a particular way constitutes social situations and socio-cultural frames, and these frames require the use of a particular language. “Which comes first” does not seem to be a relevant question synchronically. Social and cultural routines result in recurring activities and institutions. However, these institutions and routinized activities have to be rebuilt continuously in the here and now. The question is whether these cultural models, institutions and frames exist outside the language or not. The social constructivists insist that models and frames have to be rebuilt again and again, so it is just our impression that they exist outside language. However, the socio-cognitive approach argues that these cultural mental models are manifested as psychological reality in the individual mind, and when a concrete situation occurs, the appropriate model is recalled, which supports the appropriate verbalization of triggered thoughts and activities. Of course, building and rebuilding our world occurs not merely through language but through the interaction of language with other real-life phenomena such as non-linguistic symbol systems, objects, tools, technologies, etc.

Encyclopaedic knowledge and cultural models

183

The individual is not only constrained to some extent by collective cultural models but also participates in creating them. Private models may originate from a person’s creative (and even unintended) combination of existing cultural models as well as unique cognitive dispositions (self reflection, critical thinking, etc.). Some private models always remain idiosyncratic (i.e. private), while others may enter into the socio-cultural framework and establish new cultural trends (cf., e.g., Berger and Luckmann 1967). Both private and cultural models help people organize events, make actions easier, and, as such, free up cognitive resources that can be applied to less familiar issues and experiences. 4.2.

The “reality” of cultural models

Language and culture are usually considered “collective representations”, i.e., socially constituted systems (cf. Saussure [2002]; Durkheim 1947; Kronenfeld 2008). There are two main approaches to the debate about the actual existence of these systems. According to one of them, these systems have been considered to be merely epiphenomenal, which means that they have no actual direct existence (e.g. Kronenfeld 2008). However, they have the appearance of direct existence insofar as they are the byproducts of a group of individuals with similar minds confronting similar situations in similar contexts. The problem with this approach is that human beings usually talk about and rely upon language and culture as if they actually exist, as if they exist externally to them as individuals. Our individual understandings of language and culture are quite consistent across individuals. Generally it is more so than our sense of our own individual patterns. We have highly shared senses of the collective patterns, and each of us is capable of describing where we ourselves deviate, or are somewhat idiosyncratic. The opposed view to non-existence has been that these systems have some sort of objective existence outside the individual (cf. Simmel 1972). Culture is “real” and deals with the problem of the relationship between the individual and the given community. This approach sees a child’s socialization or enculturation as a process by which basic cultural structures and schemata are “internalized” deeply into the individual psyche. However, these cultural models and schemata keep changing both diachronically and synchronically. Definitely there is a great difference between cultural models that existed a hundred years ago and the ones that we have in our time. Besides, the internalization process is not mechanical, i.e. enculturation occurs as a bidirectional interaction between the individual and the social environment. When we talk about culture we usually mean “subjective culture” (cf. Triandis 2002), which is a community’s characteristic way of perceiving its social environment. However, there is a general distinction between two basic aspects of culture. When this distinction is not clarified, confusion may occur about whether culture exists “out there” or not. One aspect of culture is subjective culture – the psychological feature of culture including assumptions, values, beliefs and patterns of

184

Istvan Kecskes

thinking. The other is objective culture, which includes the institutions and artifacts of culture, such as its economic system, social customs, political structures and processes, arts, crafts and literature. Objective culture can be treated as an externalization of subjective culture which usually becomes reified, meaning that those institutions which are properly seen as extensions of human activity attain an independent status as external entities. They seem to exist “out there”, and their ongoing human origins are usually forgotten. The study of objective culture is well established because institutions and external artifacts of behavior are more accessible to observation and examination. Subjective culture is usually treated as an unconscious process influencing perception, thinking and memory, or as personal knowledge which is inaccessible to trainers or educators. Simmel (1972) also makes a difference between subjective and objective culture with the latter referring to the cultural level of social reality. In his view, people produce culture, but because of their ability to reify social reality, the cultural world and the social world come to have lives of their own and increasingly dominate the actors who created them. We can also think about language in this way. It has been created and is being created by people, but appears to have a life of its own as an institution “out there”. Simmel identified a number of components of objective culture, including tools, transportation, technology, the arts, language, the intellectual sphere, conventional wisdom, religious dogma, philosophical systems, legal systems, moral codes, and ideals. The size of objective culture increases with modernization. The number of different components of the cultural realm also grows. Simmel was concerned about the effect of objective culture on the individual’s subjective existence. Postmodernists have taken that concern to another level. In the past, most of the culture was produced by people situated in real social groups that interacted over real issues. This grounded culture created real meanings and morally infused norms, values and beliefs. In the postmodern era, much of the culture is produced or colonized by business using advertising and the mass media. This important historic shift implies that culture has changed from a representation of social reality to representations of commodified images. In our time culture is produced rather than created, and people have changed from culture creators to culture consumers.

5.

Cultural models at work

5.1.

Development of cultural models

Every one of us has rich individual experiences, and the cognitive structuring that pertains to them may differ, whether coded linguistically or not. When we communicate with other people through language or otherwise, we need to interrelate

Encyclopaedic knowledge and cultural models

185

our separate experience and cognitive structures. When we do things with other people, routinely and repeatedly, we usually develop some standardized way of doing so. These shared action plans may emerge as cultural models. Kronenfeld (2008) argues that language gets involved when we need to communicate verbally, and then only with regard to those aspects of the action plan that need to be discussed and talked about or coded in memory. He emphasizes that language is a socially constructed tool that can be exceedingly helpful to thought, but in no sense does it form the basis for individual thought, and it need not provide the basis for (much of) the shared or coordinated thought that makes up culture. I think language plays a more important role in culture than the one Kronenfeld assigns to it. In fact, language supports both the development and reinforcement of cultural models, mainly through formulaic language, which is the heart and soul of nativelike language use. Formulaic language generally serves as a core for language use in a speech community because prefabricated linguistic expressions usually mean the same for each member of the community. Languages and their speakers have preferred ways of saying things (cf. Wray 2002; Kecskes 2007). English native speakers shoot a film, dust the furniture, make love or ask you to help yourself at the table. The use of these expressions creates scenarios and gives a certain kind of idiomaticity to language use. For instance: (2) Jim: Let me tell you something. Bob: Is something wrong? The expression Let me tell you something usually has negative connotation, it creates a scenario that anticipates trouble. Our everyday communication is full of prefabricated expressions and utterances because we like to stick to preferred ways of saying things. Why is this so? Kecskes (2007) argues that there are three important reasons: 1. Formulas decrease the processing load: There is psycholinguistic evidence that fixed expressions and formulas have an important economizing role in speech production (cf. Miller and Weinert 1998; Wray 2002). Sinclair’s “idiom principle” says that the use of prefabricated chunks “may […] illustrate a natural tendency to economy of effort” (Sinclair 1991: 110). This means that in communication we want to achieve more cognitive effects with less processing effort. Formulaic expressions ease the processing overload not only because they are “readymade” and do not require any “putting together” from the speaker/hearer, but also because their salient meanings are easily accessible in online production and processing. 2. Phrasal utterances have a strong framing power: Frames and cultural models are basic cognitive structures which guide the perception and representation of reality (Goffman 1974). Frames help determine which parts of reality become noticed. They are not consciously manufactured

186

Istvan Kecskes

but are unconsciously adopted in the course of communicative processes. Formulaic expressions usually come with framing. Most fixed expressions are defined relative to a conceptual framework. If a policeman stops my car and says Step out of the car, please, this expression will create a particular frame in which the roles and expressions to be used are quite predictable. 3. Formulaic units create shared bases for common ground in coordinating joint communicative actions: The use of formulaic language requires shared experience and conceptual fluency. Tannen and Öztek (1981: 54) argued that “cultures that have set formulas afford their members the tranquility of knowing that what they say will be interpreted by the addressee in the same way that it is intended, and that, after all, is the ultimate purpose of communication”. Cultural models provide a kind of reference library for possible plans of action for oneself or possible interpretation of actions of others. These models are not learned directly as models, but are inferred anew by each of us from what we see and experience with those other people around us. But what we see and experience are never the models themselves. What we infer from experience are pieces of information, images and features that keep a scenario together. What we infer depends directly on what parts of the given scenario are saliently and repetitively present in the messages we experience for us to pull out the regularities on which we will base our construction of the scenario behind them. Thus, systematic and repeated changes in speech or cultural behavior in one generation will be learned by the next generation as part of the givens of language or culture. The cultural models that we actually experience (that is, cultural models in the form in which we actually experience them) acquire specificity through the process of their instantiation in the concrete situations in which the models were realized. Much of our application of cultural models (instantiation and then realization) is in situations that represent some kind of extension from the prototypical, unmarked default situation. The core of cultural models shared by people in the same speech community changes diachronically through systematic and repeated shifts that can come from socio-political, technological and environmental changes and the like. The application of the core, however, changes synchronically. No situation occurs exactly the same way as we have experienced it at any previous time. In the socio-cognitive paradigm, action is always performed by individuals, and individuals are always adapting cultural forms to fit their needs. People use cultural models as devices to facilitate effective interaction with others in the various communities to which they belong. In this way, individuals not only shape cultural models but are also constrained by them. Most of these cultural models come from people’s past experiences, but they are constantly recreated in use. It is important to note that people are not

Encyclopaedic knowledge and cultural models

187

required to follow cultural conventions (whether in the use of cultural models or in other ways). At any given time, they can ignore or modify cultural models that kick in in their mind when they get into a typical situation. Certain cultural models can (and often do) show slight variations across groups to which we all belong – groups that can be formal or informal, long-lived or evanescent, imposed or voluntary, and so forth. 5.2.

Instantiating cultural models

Cultural models are abstract plans at varying degrees of specificity. They relate knowledge, goals, values, perceptions, emotional states, etc. to actions in different contexts. Kronenfeld (2008) argues that these conceptual models do not directly or automatically apply to any specific situation. First they have to be “instantiated” by having their general generic details replaced with the specific details of the situation at issue. The instantiated cultural model is still only a conceptual structure, and several different (even mutually contradictory) ones may be taken into consideration for any given situation. Finally, one particular instantiated model is “realized” in the actual concrete situation. This can be an action plan for one’s own behavior, or a device for interpreting the behavior of some other people. It is important to note that at any given moment only one instantiated model can be realized. But we can quickly jump back and forth between different realizations. The relationship between the abstract collective cultural model and the private realization of the model by interlocutors in a concrete situation is the same as in linguistics between phonemes and phones or between morphemes and morphs. We consider phones as the actual phonetic realization of phonemes, and morphs as the actual forms used to realize morphemes. In his pragmatic acts theory, Mey (2001) also speaks about “pragmemes” which are instantiated in pragmatic acts in speech situations. A particular pragmeme can be substantiated and realized through individual pragmatic acts. In other words, a pragmatic act is an instance of adapting oneself to a context, as well as adapting the context to oneself. For instance: (3) She is after my money. Like I care. Like I care is a pragmatic act that expresses the pragmeme “I do not care”, which can be also substantiated by several other concrete pragmatic acts such as I do not care, I do not mind, It’s none of my business etc. According to Mey, pragmatic acts are situation-derived and situation-constrained. There is no one-to-one relationship between speech acts and pragmatic acts because the latter does not necessarily include specific acts of speech. Consider for instance:

188

Istvan Kecskes

(4) Mother: Joshua, what are you doing? Joshua: Nothing. Mother: Will you stop it immediately. (Mey 2001: 216) The pragmeme represented by the pragmatic act Nothing can be described as “trying to get out (opt out) of a conversation” that may lead too far. But pragmemes in the sense that Mey uses the term are not cultural models. They are more like scenarios within cultural models. However, the process of instantiation happens similarly both in the case of pragmemes and cultural models. 5.3.

Practices

Culture includes many practices or routines. Feldman and Pentland (2003) argue that routines (i.e. practices) consist of two elements: the ostensive and the performative. The ostensive element comprises individuals’ cognitive understanding of the processes, while the performative element consists of actual behavior in the actual situational context. From a socio-cognitive perspective, both of these processes should be of interest to us. Cognitive understanding relies both on cultural and private models, and on how these models are applied by cognitive processing, spanning from excessive automatic (as in categorical) thinking to self-reflective (as in reflective) thinking (see Ringberg and Reihlen 2008). When categorical thinking is applied, people establish meaning by automatically integrating incoming stimuli based on existing cultural and private models. Kecskes (2008) argues that in the process of communication, a speaker’s private context generated by intention gets encoded in lexical units and formulated in an utterance (actual linguistic context) that is uttered (or written) “out there” in the world by a speaker in a situation (actual situational context), and is matched (“internalized”) to the private cognitive contexts “inside” the head of the hearer (prior knowledge). Meaning is the result of interplay between the speaker’s private context and the hearer’s private context in the actual situational context as understood by the interlocutors. Research in social cognition indicates that several epistemic factors can affect the applicability of categorical thinking (cf. Ringberg and Reihlen 2008). For instance, people usually apply categorical thinking in everyday routines when they have high cognitive load, are under pressure to make quick decisions, have limited cognitive capacity, and/or are distracted. Categorical thinking generally leads to efficient processing of regular socio-cultural interactions and stimuli. As a consequence of relying on categorical thinking, environmental stimuli are ‘pushed’ into existing mental models. This may prevent the person from adjusting to divergent inputs and unusual circumstances. In communication, this phenomenon is observable in the use of prefabricated linguistic units and situation-bound utterances. For instance:

Encyclopaedic knowledge and cultural models

189

(5) Shop Assistant: Can I help you, Madame? Customer: Thank you. I’m just looking. In this conversation Can I help you? and I’m just looking function as plain situation-bound utterances (Kecskes 2000, 2002). The customer is distracted because she is busy looking at clothes and wants to get rid of the assistant. Reflective processing works in a different way. It requires the ability of people to sustain a high level of cognitive responsiveness and combine or broaden internalized cultural and private models in thoughtful and creative ways to improve their sense making. The degree of application of categorical versus reflective thinking varies across situational contexts, and life experience and general acumen are also important variables. In reflective thinking, cultural and private models are applied in non-automatic fashions. Ringberg and Reihlen (2008: 923) argued that reflective thinking is a proactive process that occurs when a person has the cognitive capacity and need for deliberate cognition to engage with stimuli that are not easily or usefully made sense of by a categorical application of private and/or cultural models. Category-inconsistent information may activate reflective thought processes among some people through which they recombine cultural and private models in deliberate ways to improve the relevance of their sense-making of a particular situation (cf. Bodenhausen, Macrae and Garst 1998; Wilson and Sperber 2004). This can be demonstrated in the following conversation: (6) Jill: Jane: Jill: Jane: Jill: Jane:

I met someone today. Good for you. He is a police officer. Are you in trouble? Oh, no, I liked the man. We met in a cafe. He was nice and polite. Not all of them are …

In this conversation there is a clear difference between the two women’s private context tied to the word police officer. The collective cultural model attached to police officer has been changed in Jill’s privatized model as a result of the positive experience while this is not the case with Jane. She appears to have a private model that is close to the collective cultural model. The socio-cognitive approach incorporates cultural and private models into categorical and reflective processing. This means that most of the time a cognitive system is neither fully closed nor fully open, it is neither fully determined nor independent of external sensory inputs, and people are neither autonomous processors nor cultural dopes. Consequently, meaning creation and knowledge transfer are located somewhere on a continuum between fully automatic and fully idiosyncratic. This depends on several variables that include the nature of people’s private and cultural models, the level of categorical and reflective thinking, and environmental feedback mechanisms. The socio-cognitive approach broadens traditional

190

Istvan Kecskes

positivist and social-constructionist positions by situating sense-making within the mind (and body) that may be influenced but rarely determined by environmental feedback mechanisms (Bandura 1986; Bunge 1996).

6.

Cultural models and common ground

Cultural models make up a “container” of scenarios that individuals can choose from in any given situation. They provide options for possible plans of action for the individual or possible interpretation of actions of other individuals. The choice is driven by a number of factors including prior experience, the possible alternative outcomes different choices lead to, the degree of similarity of the application situation to the prototypic situation for the given cultural model, etc. Kronenfeld (2008) argues that there is no guarantee that different individuals involved in a given situation will all be working off the same cultural models. However, they all will have to mutually adjust (or negotiate) their separate models in order to ensure the desired degree of mutual interpretability and interactive reciprocity. This is where the interplay of cooperation and egocentrism becomes important from the perspective of common ground. The main problem is that our individual experience and, as a consequence, our individual knowledge is not necessarily shared with the other members of our speech community in the way and to the extent that pragmatic theories have assumed. This fact should be taken into account when describing common ground, which is the main facilitator of cooperation-based communication in general (cf. Horton, this volume). However, current pragmatic theories seem to have failed to describe common ground in its complexity because they usually retain a communication-as-transfer-between-minds view of language (e.g. Arnseth and Solheim 2002), and disregard the fact that disagreement and egocentrism of speaker-hearers are as fundamental parts of communication as agreement and cooperation (e.g. Barr and Keysar 2005; Colston and Katz 2005). On the other hand, current cognitive research has overestimated the egocentric behavior of the dyads and argued for the dynamic emergent property of common ground while devaluing the overall significance of cooperation in the process of verbal communication. The socio-cognitive approach proposed by Kecskes (2008) and Kecskes and Zhang (2009) eliminates this conflict and intends to combine the two views into an integrated concept of common ground, in which both core common ground (assumed shared knowledge, “a priori” mental representation) and emergent common ground (emergent participant resource, “a post facto” emergence through use) converge to construct a dialectical socio-cultural background for communication. Based on this view, common ground is perceived as an effort to converge the mental representation of shared knowledge present as memory that we can activate, with shared knowledge that we seek, and knowledge that we can create in the communicative process. The

Encyclopaedic knowledge and cultural models

191

socio-cognitive approach emphasizes that common ground is a dynamic construct that is mutually constructed by interlocutors throughout the communicative process. The core and emergent components join in the construction of common ground in all stages, although they may contribute to the construction process in different ways, to different extents and in different phases of the communicative process. Common ground is a cooperatively constructed mental abstraction. It is assumed by interlocutors in a sense that no one will know for sure that it exists. On the time dimension, common ground derives from the interlocutors’ information gained from prior communicative experience and current communicative experience (actual situational context). On the range dimension, common ground derives from the interlocutors’ shared information that belongs to a community (a macroconcern), and that pertains to their individual experiences (a micro-concern). The distinction between macro-concern and micro-concern resembles Clark’s (1996: 100) distinction between communal and personal common ground. However, there are some differences. According to Kecskes (2008: 390), prior experiences of speakers create private contexts which are encoded in lexical items in the minds of speakers of a particular speech community. These private contexts incorporate core knowledge (tied to prior experience), which is the public part of the private context, and individual-specific knowledge that may not be shared by the other members of the speech community because it is the individualized reflection of the socio-cultural context. The public context, that is to say, the public part of the private context is, however, available to each speaker of that speech community because it refers to relatively similar conceptual content that is conventionalized. Clark (1996: 116) says that “common ground isn’t just there, ready to be exploited. We have to establish it with each person we interact with.” Similar thoughts are formulated by Stalnaker (2002). The difference between Clark’s contribution theory (Clark 1996) and the socio-cognitive approach is that while the former considers communication as a constant search for common ground following a contribution-by-contribution sequence in an idealized way, the latter underscores the “untidy”, chaotic nature of communication, which is not just recipient design and intention recognition as most theories that have grown out of Grice’s approach claim. The socio-cognitive approach emphasizes that speakers do not always seek common ground, and they are both egocentric and cooperative in the same communicative process. Kecskes and Zhang (2009) argue that there are two sides to common ground: “Core common ground” and “emergent common ground”. Core common ground refers to the relatively static (synchronically), generalized and common knowledge that belongs to a certain speech community as a result of prior interaction and experience, whereas emergent common ground refers to the relatively dynamic, particularized and “private” knowledge activated and/or created in the course of communication that belongs to the individual(s). The former is a repertoire of

192

Istvan Kecskes

knowledge and cultural models that can be assumed to be shared among individuals of a speech community independent of the actual situational circumstances, such as when and where the conversation occurs, between whom it occurs, etc. In contrast, the actual contextual part is knowledge that is aroused or involved as shared enterprises in the particular situational context that pertains to the interlocutors exclusively. This contingent circumstance draws the interlocutors’ attention to the same entities or states and, with the formation of particular intentions therein, activates some of their prior individual experiences that join in this intention-directed action. There is a dialectical relationship between core common ground and emergent common ground. First, the core part derives from the macro socio-cultural information of a community (or any groups of people divided by nations, regions, etc.) that has been referred to as cultural model. This is accessible to all individuals in that community, whereas the actual part derives from micro socio-cultural information that pertains to individuals solely. Second, the core part changes “diachronically”, whereas the actual part changes “synchronically”. Third, the core part may affect the formation of the actual part in that it partly restricts the way the actual part occurs. In most cases, the actual part is an instance of information that is predictable in the core part. On the other hand, the actual part may contribute to the core part in that the contingent actual part in a frequent ritual occurrence potentially becomes public disposition that belongs to the core part. In other words, they are different components of assumed common ground, which have internal connections.

7.

Conclusion

In this chapter, the focus has been on encyclopaedic knowledge, its structure, instantiation and role in meaning creation and comprehension. A socio-cognitive approach was used to interpret and discuss the issues raised in connection with the subject matter. This theory helps us understand how encyclopaedic knowledge can, through its cultural models, function both as a repository of knowledge that changes diachronically and as synchronically changing emergent knowledge created in the process of communication. Encyclopaedic knowledge is usually instantiated in cultural models which provide scenarios, scripts or action plans for individuals to interpret and behave in a particular situation, or process and interpret the behavior of others in various life situations. The socio-cognitive approach also changes the view of how meaning creation is understood in the cognitive paradigm. Instead of emphasizing the decisive role of context in meaning, it proposes that both prior experience encapsulated partly in lexical items and actual situation experience occurring in the process of communication participate in meaning creation. Both the linguistic system and the system of cultural models are collective representations which may be differently interpreted

Encyclopaedic knowledge and cultural models

193

by each individual in actual situational contexts. The dynamism of communication is ensured by the dialectical relationship between collective cultural models instantiated and realized in an individual manner in language production and comprehension in actual situational contexts. Common ground shows very clearly how this dynamism works in life. In the socio-cognitive view both core common ground (assumed shared knowledge, “a priori” mental representation) and emergent common ground (emergent participant resource, “a post facto” emergence through use) converge to construct a dialectical socio-cultural background for communication. While attention (through salience, which is the cause for interlocutors’ egocentrism) explains how emergent property unfolds, intention (through relevance, which is expressed in cooperation) explains how presumed shared knowledge is used. The core and emergent components join in the construction of common ground in all stages, although they may contribute to the interaction in different ways, to different extents, and in different phases of the communicative process.

References Alvesson, Mats and Dan Kärreman 2001 Odd couple: making sense of the curious concept of knowledge management. Journal of Management Studies 38: 995–1018. Arnseth, Hans Christian and Ivar Solheim 2002 Making sense of shared knowledge. In: GerryStahl (ed.), Proceedings of CSCL 2002, Computer Support for Collaborative Learning: Foundations for a CSCL Community, 102–110. Boulder, Colorado: University of Colorado. Bandura, Albert 1986 Social Foundations of Thought and Action: A Social Cognitive Theory. Englewood Cliffs, NJ: Prentice-Hall. Barr, Dale and Boaz Keysar 2005 Making sense of how we make sense: The paradox of egocentrism in language use. In: Herbert L. Colston and Albert N. Katz. (eds.), Figurative Language Comprehension, 21–43. Mahwah, NJ: Lawrence Erlbaum. Berger, Peter. L. and Thomas Luckmann 1967 The Social Construction of Reality: A Treatise in the Sociology of Knowledge. Garden City, NY: Doubleday. Bernstein, Richard. J. 1983 Beyond Objectivism and Relativism: Science, Hermeneutics, and Praxis. Philadelphia, PA: University of Pennsylvania Press. Bodenhausen, Galen, C. Neil Macrae and Jennifer Garst 1998 Stereotypes in thought and deed: Social-cognitive origins of intergroup discrimination. In: Constantine Sedikides, John Schopler and Chester A. Insko (eds.), Intergroup Cognition and Intergroup Behavior, 311–335. Mahwah, NJ: Lawrence Erlbaum.

194

Istvan Kecskes

Brown, John S. and Paul Duguid 2001 Knowledge and organization: A social-practice perspective. Organization Science 12: 198–213. Bruner, Jerome. S. 1996 The Culture of Education. Cambridge, MA: Harvard University Press. Bunge, Mario 1996 Finding Philosophy in Social Science. New Haven/London: Yale University Press. Bunge, Mario 1998 Social Science under Debate. Toronto/Buffalo/London: University of Toronto Press. Carston, Robyn 1998 Negation, ‘presupposition’ and the semantics/pragmatics distinction. Journal of Linguistics 34(2): 309–350. Clark, Herbert 1996 Using Languages. Cambridge: Cambridge University Press. Colston, Herbert L. and Albert Katz (eds.) 2005 Figurative Language Comprehension: Social and Cultural Influences. Hillsdale, NJ: Lawrence Erlbaum. Croft, William and Alan Cruse 2004 Cognitive Linguistics. Cambridge: Cambridge University Press. D’Andrade, Roy 1987 A folk model of the mind. In: Dorothy Holland and Noami Quinn (eds.), Cultural Models in Language and Thought, 113–147. Cambridge: Cambridge University Press. D’Andrade, Roy 1992 Schemas and motivation. In: Roy D’Andrade and Claudia Strauss (eds), Human Motives and Cultural Models, 23–44. New York: Cambridge University Press. D’Andrade, Roy and Claudia Strauss (eds.) 1992 Human Motives and Cultural Models. Cambridge: Cambridge University Press. DiMaggio, Paul 1997 Culture and cognition. Annual Review of Sociology 23: 263–287. Durkheim, Emile 1947 The Division of Labor in Society [Translated by George Simpson]. New York: The Free Press. Evans, Vyvyan 2006 Lexical concepts, cognitive models and meaning-construction. Cognitive Linguistics 17(2): 491–534. Evans, Vyvyan 2009 How Words Mean: Lexical Concepts, Cognitive Models and Meaning Construction. Oxford: Oxford University Press. Evans, Vyvyan and Melanie Green 2006 Cognitive Linguistics: An Introduction. Edinburgh: Edinburgh University Press.

Encyclopaedic knowledge and cultural models

195

Fauconnier, Gilles 1997 Mappings in Thought and Language. Cambridge: Cambridge University Press. Feldman, Michael. S. and Brian T. Pentland 2003 Reconceptualizing organizational routines as a source of flexibility and change. Administrative Science Quarterly 48: 94–118. Fillmore, Charles 1982 Frame semantics. In: The Linguistic Society of Korea (ed.), Linguistics in the Morning Calm, 111–137. Seoul: Hanshin Publishing Co. Fillmore, Charles and B. T. Sue Atkins 1992 Toward a frame-based lexicon: The semantics of RISK and its neighbors. In: Adrienne Lehrer and Eva F. Kittay (eds.), Frames, Fields and Contrasts, 75–102. Hillsdale, NJ: Lawrence Erlbaum. Gee, James. P. 1999 An Introduction to Discourse Analysis: Theory and method. New York: Routledge. Geeraerts, Dirk (ed.) 2006 Cognitive Linguistics: Basic Readings. Berlin/New York: Mouton de Gruyter. Gherardi, Silvia 2000 Practice-based theorizing on learning and knowing in organizations. Organization 7: 211–23. Gherardi, Silvia 2001 From organizational learning to practice-based knowing. Human Relations 54: 131–139. Giora, Rachel 2003 On Our Mind: Salience, Context and Figurative Language. New York: Oxford University Press. Glasersfeld, Ernst von 1995 Radical Constructivism: A Way of Knowing and Learning. London/Washington, D.C.: The Falmer Press. Goffman, Erving 1974 Frame Analysis: An Essay on the Organization of Experience. New York: Harper and Row. Holland, Dorothy and Naomi Quinn (eds.) 1987 Cultural Models in Language and Thought. Cambridge: Cambridge University Press. Johnson-Laird, Phillip. N. 1983 Mental Models. Cambridge: Harvard University Press. Kecskes, Istvan 2000 A cognitive-pragmatic approach to situation-bound utterances. Journal of Pragmatics 32(6): 605–625. Kecskes, Istvan 2002 Situation-Bound Utterances in L1 and L2. Berlin/New York: Mouton de Gruyter. Kecskes, Istvan 2007 Formulaic language in English Lingua Franca. In: Istvan Kecskes and Laurence Horn (eds.), Explorations in Pragmatics: Linguistic, Cognitive and Intercultural Aspects, 191–219. Berlin/New York: Mouton de Gruyter.

196

Istvan Kecskes

Kecskes, Istvan 2008 Dueling contexts: A dynamic model of meaning. Journal of Pragmatics 40(3): 385–406. Kecskes, Istvan 2010 Situation-Bound Utterances as Pragmatic Acts. Journal of Pragmatics 42(11): 2889–2897. Kecskes, Istvan and Fenghui Zhang 2009 Activating, seeking and creating common ground: A socio-cognitive approach. Pragmatics & Cognition 17(2): 331–355. Kronenfeld, David 2008 Cultural models. Intercultural Pragmatics 5(1): 67–74. Lakoff, George 1987 Women, Fire and Dangerous Things: What Categories Reveal About the Mind. Chicago: University of Chicago Press. Lakoff, George and Mark Johnson 1980 Metaphors we live by. Chicago: University of Chicago Press. Langacker, Ronald 1987 Foundations of Cognitive Grammar, Volume I. Stanford: Stanford University Press. Levinthal, Daniel and Claus Rerup 2006 Crossing an apparent chasm: bridging mindful and less-mindful perspectives on organizational learning. Organization Science 17: 502–513. Mey, Jacob 2001 Pragmatics. London: Blackwell. Miller, Jim and Regina Weinert 1998 Spontaneous Spoken Language: Syntax and Discourse. Oxford: Clarendon Press. Minsky, Marvin 1975 A Framework for Representing Knowledge. In: Patrick H. Winston (ed.), The Psychology of Computer Vision, 211–277. New York: McGraw-Hill. Reddy, Michael J. 1979 The conduit metaphor: A case of frame conflict in our language about language. In: A. Ortony (ed.), Metaphor and Thought, 284–310. Cambridge: Cambridge University Press. Ringberg, Torsten and Markus Reihlen 2008 Towards a socio-cognitive approach to knowledge transfer. Journal of Management Studies 45 (5): 912–935. Rosch, Eleanor 1977 Human categorization. In Neil Warren (ed.), Advances in Cross-cultural Psychology (Vol. 1), 1–72. London: Academic Press. Saussure, Ferdinand de [2002] Écrits de linguistique générale (edition prepared by Simon Bouquet and Rudolf Engler). Oxford: Oxford University Press. Schank, Roger. C and Robert P. Abelson 1977 Scripts, Plans, Goals, and Understanding: An Enquiry into Human Knowledge Structures. Hillsdale, NJ: Lawrence Erlbaum.

Encyclopaedic knowledge and cultural models

197

Shore, Bradd 1996 Culture in Mind: Cognition, Culture, and the Problem of Meaning. New York: Oxford University Press. Simmel, Georg 1972 On Individuality and Social Forms. [Edited by and with an introduction by Donald Levine] Chicago: University of Chicago Press. Sinclair, John 1991 Corpus, Concordance, Collocation. Oxford: Oxford University Press. Tannen, Deborah and C. Oztek Piyale 1981 Health to our mouths. Formulaic expressions in Turkish and Greek. In: Florian Coulmas (ed.), Conversational Routine, 37–54. The Hague: Mouton. Triandis, Harry C. 2002 Subjective culture. In: Walter J. Lonner, Dale L. Dinnel, Susanna A. Hayes and David N. Sattler (eds.), Online Readings in Psychology and Culture (Unit 15, Chapter 1), Center for Cross-Cultural Research, Western Washington University, Bellingham, Washington. Vygotsky, Lev. S. 1978 Mind in Society. Cambridge, MA: Harvard University Press. Vygotsky, Lev. S. 1987 The Collected Works of L. S. Vygotsky: Vol. 1, Problems of General Psychology. New York: Plenum. Wilson, Deirdre and Dan Sperber 2004 Relevance Theory. In: Laurence Horn and Gregory Ward (eds.), The Handbook of Pragmatics, 607–632. Oxford: Blackwell. Wittgenstein, Ludwig 1953 Philosophical Investigations. Oxford: Blackwell. Wray, Alison 2002 Formulaic Language and the Lexicon. Cambridge: Cambridge University Press.

Part III: The psychology of pragmatics Processing and acquisition

8.

The processing of pragmatic information in discourse Ted J.M. Sanders and Anneloes R. Canestrelli

1.

Introduction

A cognitive view on pragmatics should not only account for a cognitively plausible analysis of pragmatic phenomena from a linguistic point of view, it should also deal with language users’ online processing, as well as with the cognitive mechanisms and representations involved. This chapter focuses on the psycholinguistic research on mental operations and representations: how do humans process pragmatic information? What is the role of crucial pragma-linguistic characteristics in the cognitive representation people make of the information they have processed? We will focus on issues at the discourse level. After all, people use language to communicate, and even though we sometimes use short messages of one word (Dirk! Okay, Stop!) or one sentence (I declare the games opened.) to get the message across, language users usually communicate through discourse. The importance of the discourse level for the study of language and linguistics can hardly be overestimated: “Discourse is what makes us human” (Graesser, Millis, and Zwaan 1997). It is not surprising, therefore, that the study of text and discourse has become an increasingly important area over the past decades, in both linguistics and psychology.

2.

The connectedness of discourse: cohesion and coherence

Within the field of linguistics, pragmatics and semantics have a significant role to play in the study of discourse. A central objective of linguists working on the discourse level is to characterize discourse connectedness. Linguists have traditionally approached this problem by looking at overt linguistic elements and structures, thereby characterizing it in terms of cohesion (Halliday and Hasan 1976). Connectedness is localized in the discourse itself because of explicit linguistic clues like pronouns referring to previously mentioned subjects, or conjunctions, such as because, expressing a causal relation. It is clear by now that the cohesion approach to connectedness is inadequate (Sanders and Pander Maat 2006, and many others). Instead, the dominant view in the field has come to be that the connectedness of discourse is a characteristic of the cognitive representation of the discourse rather than of the discourse itself (Sanders, Spooren, and Noordman 1992; Graesser, McNamara, and Louwerse

202

Ted J.M. Sanders and Anneloes R. Canestrelli

2003). The connectedness thus conceived is often called coherence. Language users establish coherence by actively relating the different information units in the text. Generally speaking, there are two respects in which discourse representations cohere: 1. Referential coherence: smaller linguistic units (often nominal groups) may relate to the same mental referent; 2. Relational coherence: discourse segments (say clauses) are connected by coherence relations like CAUSE - CONSEQUENCE between them. Although coherence phenomena are of a cognitive nature, their reconstruction is often based on linguistic signals in the text itself. Both types of coherence phenomena under consideration – referential and relational coherence – have clear linguistic indicators that can be taken as processing instructions that the speaker/writer provides to the listener/reader. In this respect, the cohesive ties identified in cohesion accounts have a role to play: they serve as the explicit linguistic signals that help readers and listeners to construct coherence in their mental representations. For referential coherence these are devices such as pronouns and demonstratives, and for relational coherence these are connectives and (other) lexical markers of relations, such as cue phrases and signalling phrases. A major research issue is the relation between the linguistic surface code (what Givón calls “grammar as a processing instructor”, 1992: 5) and cognitive representation of the discourse.

3.

Referential coherence

Repeated reference to the same entity often constitutes coherence between discourse segments. This observation has a central position in several theoretical frameworks. 3.1.

Varying amounts of activation: accessibility

For instance, a discourse topic may be referred to quite elaborately in the first sentence (e.g. The well-known linguistics professor from UC Berkeley), but once the referent has been identified, names (Sweetser) and later pronominal forms (she) suffice. This is not a coincidence. Many linguists have noted this regularity (e.g. Ariel 1990; Givón 1992; Gundel, Hedberg, and Zacharski 1993). Ariel (1990), for instance, has argued that this type of pattern in grammatical coding should be understood to guide processing. In her accessibility theory, “high accessibility markers” use little linguistic material and signal the default choice of continued activation. By contrast, “low accessibility markers” contain more linguistic material and signal the introduction of a new referent.

The processing of pragmatic information in discourse

203

The accessibility of referential information is one of the most important challenges at the intersection of linguistic and psycholinguistic studies of text and discourse processing (Sanders and Gernsbacher 2004). Linguistic indicators of referential coherence behave according to a systematic pattern: longer linguistic forms (such as full lexical NPs) tend to be used when accessibility to the referents is relatively low, shorter forms (pronouns and zero anaphors) are used when referents are highly accessible. This linguistic theory fits in nicely with a dynamic view of text and discourse processing: when a reader proceeds through a text, the activation of concepts as part of the reader’s representation fluctuates constantly. Making use of reading time paradigm and eye-movement recording, experimental work has indeed shown how activation patterns play a role during online processing. Speakers and writers can be considered experts in predicting activation states of recipients. They deploy referring expressions in tune with computed levels of accessibility. In fact, quite some work has been done in experimental research on text processing, which can be taken to demonstrate the ‘psychological reality’ of linguistic indicators of referential coherence. Online studies of pronominal reference have resulted in the formulation of cognitive parsing principles for anaphoric reference (cf. Garrod and Sanford 1994; Sanford and Garrod 1994). A classical finding is that it is easier to resolve a pronoun with only one possible referent, and it is easier to resolve pronouns with proximal referents than those with distant ones. As for the time course, eye-tracking studies have repeatedly shown that anaphoric expressions are resolved immediately (e.g., Carpenter and Just 1977; Duffy and Rayner 1990; Arnold et al. 2000). (1) a. The guard mocked one of the prisoners in the machine shop. b. He had been at the prison for only one week. In a classical eye-tracking experiment, Carpenter and Just (1977) found that when readers process sentence pairs like (1) and come upon ambiguous pronouns such as he, they frequently look back in the text. More than 50 % of these regressive fixations land on one of the two nouns in the text preceding the pronoun, suggesting that readers indeed attempt to resolve the pronoun immediately. As for the meaning representation, it has been shown that readers have difficulty understanding the text correctly when the antecedent and referent are too far apart and reference takes the form of a pronoun. 3.2.

The interplay of several linguistic factors: Centering Theory

One of the leading theories in the field of referential coherence is Centering Theory (see Walker, Joshi, and Prince 1998 for an overview). Centering Theory makes precise predictions about the referent that is “in focus” at a certain moment in a discourse. It even predicts that the degree of discourse coherence is determined by the

204

Ted J.M. Sanders and Anneloes R. Canestrelli

extent to which it conforms to Centering constraints. Given an utterance in which referential antecedents are presented, Centering Theory predicts the likelihood that an antecedent will be a central referent – which is “in focus” – in the next utterance. The salience of a discourse entity is determined by a combination of syntactic, lexical and contextual factors, such as grammatical role (subject or not), expression type (zero, pronoun or NP), but also its occurrences in the preceding discourse segments. An illuminating case in point is the resolution of ambiguous pronouns like he in (2a). (2) a. George hit Al. He … Several studies have shown how grammatical role seems to be a crucial factor in this resolution process: The referent that is intended by the author is usually the grammatical subject of the preceding sentence (Pander Maat and Sanders 2009). Such a subject preference would lead us to expect a continuation of (2a) as in (2b). (2) b. George hit Al. He had been wanting to do this for a long time. Brennan, Friedman, and Pollard (1987) and Kameyama (1996) have suggested a ranking in which the subject precedes the object(s) and in which the object(s) in turn rank higher than other grammatical functions; that is, when a referent is placed in subject position, it stands a better chance of being the central referent in the subsequent discourse than when it is an object in the preceding utterance. In a corpus study of oral reports on a basketball game, Brennan (1995) has shown that speakers tend to introduce participants as subject referents more often when they expect this participant referent to remain at the centre of attention than when subsequent events are uncertain. She also points out a difference between the ways in which referents of subject and object NPs are referred to in subsequent utterances: subject participants were more often pronominalized than object participants. Her explanation is that object referents are not yet fully established as the centre of attention, and can more easily be pronominalized after they have filled the subject position once. The prominence of subject referents has been suggested by other theorists as well. Givón (1992: 10–11; 1995: 65–67) has suggested a similar grammatical hierarchy for what he calls topicality, and the corpus-statistical evidence presented in Givón (1995) supports the claim that subject referents are more persistent in subsequent discourse. Arnold (1998) has arrived at similar conclusions. Tomlin (1997) even states that – at least in many European languages – the subject role grammatically encodes topic status, and that the notion of topic by itself is superfluous. Moreover, the subject preference in the interpretation of pronouns has received empirical support in experimental work testing Centering Theory (Gordon, Grosz, and Gilliom 1993; Gordon and Scearce 1995; Hudson-D’Zmura and Tanenhaus

The processing of pragmatic information in discourse

205

1998). For instance, when a subject referent is resumed by means of a name instead of a pronoun, the reading process is slowed down; this phenomenon has been called the “repeated name penalty”. In the reading time experiments carried out by Gordon, Grosz, and Gilliom (1993) and the eye-tracking experiments by Kennison and Gordon (1997), this penalty did not occur when the intended referent was in non-subject position. 3.3.

Topic continuity

Pander Maat and Sanders (2009) conducted corpus studies to assess the plausibility of the theoretical ideas in Centering Theory. They systematically examined the occurrence of potentially ambiguous pronouns in language use in cases like (2). The corpus consisted of newspaper fragments in which he is preceded by a sentence presenting two male singular participants. It turned out that next to, among others, grammatical role (being a subject), the status of a referent as discourse topic also played a significant independent role in increasing the chances that a referent is the intended one for a potentially ambiguous pronoun, while expression type (NP or not) makes no difference. Vonk, Hustinx, and Simons (1992) showed the relevance of discourse context for the interpretation of referential expressions in an unexpected way. Sometimes anaphors are more specific than would be necessary for their identificational function (for instance, full NPs are used rather than pronominal expressions: The recently elected US President and former senator rather than Barack Obama or he). The authors convincingly argue that this phenomenon can be explained in terms of the thematic development of discourse: if a character is referred to by a proper name after a run of pronominal references, then the name itself serves to indicate that a shift in topic is occurring. Indeed, readers process these referential expressions differently, as becomes apparent from reading times: it takes them longer to process these unexpected full mentions. Whereas anaphoric reference modulates the availability of previously mentioned discourse entities, cataphoric devices change the availability of entities for the text that follows. Gernsbacher and Shroyer (1989) have demonstrated readers’ sensitivity to this type of linguistic indicator of reference. The two researchers contrasted cataphoric reference by way of the indefinite a(n) versus the definite this to refer to a newly-introduced referent in a story. For example, the new referent egg was introduced either as an egg or as this egg. It was hypothesized that the cataphor this would signal that a referent is likely to be mentioned again in the following story and that therefore the this-cataphor results in a higher activation. Subjects listened to texts and were then asked to continue the text after the critical discourse entity. They appeared to refer sooner and more often to an entity introduced by this than by an. These and other results show that entities that are marked as a potential discourse topic by this are more strongly activated, more resistant to being sup-

206

Ted J.M. Sanders and Anneloes R. Canestrelli

pressed in activation, as well as more effective in suppressing the activation of other referents (Gernsbacher and Jescheniak 1995). It is this type of finding that provides the psycholinguistic underpinning for the idea of ‘grammar as a processing instructor’. 3.4.

Parsing principles for referential coherence

By now, the results of online studies of pronominal reference enable the formulation of cognitive parsing principles for anaphoric reference. Person, number, and gender obviously guide pronominal resolution. More interestingly, data from reading time (Clark and Sengul 1979), eye-tracking (e.g. Duffy and Rayner 1990; Arnold et al. 2000), and priming (McDonald and MacWhinney 1995) studies show that it takes less processing cost to – resolve pronouns with only one possible referent than pronouns with several referents; – resolve pronouns with proximal referents than pronouns with distant ones; – resolve reference to topical concepts than to less topical ones. One obvious explanation for these findings is accessibility: anaphors are instructions to connect incoming information with already mentioned referents, and the referent nodes can be more or less accessible. As a result, it takes more or less processing time to understand anaphors. The overview we have presented so far shows a logical division of labour in the study of referential coherence: (text and discourse) linguists identify the relevant signals that guide the interpretation, develop theories on how the linguistic realization of information systematically varies so as to ‘instruct’ the interpreters, and – ideally – check the validity of their theoretical work in natural language corpora. Psycholinguists develop cognitive theories on how the actual processing occurs and test these theories in psycholinguistic experiments. Such a division of labour is ideal, and we are indeed coming close to realizing this in recent research, which aims at gathering converging evidence from theory, corpus studies and processing experiments. For instance, Ariel (2004) has further elaborated on her Accessibility theory. She describes the linguistic means of reference to discourse entities, and in so doing distinguishes between discourse profiles and discourse functions. She then formulates specific hypotheses she encourages psycholinguists to test. In processing studies, Gernsbacher et al. (2004) used Gernsbacher’s Structure Building Framework to explain new experimental data about readers’ mental representations during narrative comprehension. It was specifically investigated whether readers’ access to their mental representations of the main character in a narrative becomes enhanced (producing a ‘benefit’) when the character is re-mentioned, and whether readers’ access to the main character in a narrative becomes weakened or interfered with (producing a ‘cost’) when a new character is introduced. Reading

The processing of pragmatic information in discourse

207

experiments indeed showed increased accessibility to the main character when it was re-mentioned in the narrative, and reduced accessibility to the main character when it was newly introduced. This work suggests that successful narrative comprehension involves managing mental representations of salient and quite often interfering characters. This study is a convincing example of a dynamic processing theory of accessibility in discourse. 3.5.

Markers of topic discontinuity

We have now seen how a global discourse factor like discourse topichood has an important role to play in discourse processing. However, next to topic continuity, there is also such a thing as topic shifts: moving to another topic. Languages have different means of marking topic discontinuity. A fair number of researchers have argued that, similar to referential expressions, temporal segmentation markers such as next and at two o’clock should be regarded as signals for a topic shift. It has been assumed (Bestgen and Vonk 1995, 2000) that language processing relies on strategies of continuity, referred to as the “nextness principle” (Ochs 1979), or “principle of continuity” (Segal, Duchan, and Scott 1991: 32). According to these principles, readers assume by default (following Grice’s Maxim of Relevance) that continuity is maintained, hence interpreting subsequent sentences in a continuous fashion. It follows that if continuity is violated, which occurs in the case of a topic shift, it should be marked in some way (Bestgen and Costerman 1994). Indeed, many linguists have observed that clause initial adverbials of time and place, such as one evening, at twelve o’clock, and at the bar, function as markers for thematic discontinuity (e.g. Virtanen 1992). In language use, speakers use temporal adverbs like then or next to introduce important shifts in the narrative (Bestgen and Costermans 1994). Evidence for the cognitive relevance of temporal segmentation markers comes from psycholinguistic studies that show how overt marking of discontinuity affects processing. Bestgen and Vonk (1995), for example, have demonstrated that temporal segmentation markers reduce the availability of previous information, in both two-paragraph sentences and narrative texts. It turns out that the words preceding continuity markers (and) are more accessible than those preceding temporal segmentation markers (then), resulting in shorter probe recognition latencies to the former compared to the latter. These results support the idea that temporal segmentation markers function as a signal that prompts the reader to construct a new substructure. As a result, the information from the previous segment becomes less available (Gernsbacher 1990). Additional evidence comes from a reading experiment by Bestgen and Vonk (2000) in which the effect of initial temporal adverbials on discourse processing was tested by manipulating continuous and discontinuous texts. The authors found

208

Ted J.M. Sanders and Anneloes R. Canestrelli

that unmarked discontinuous discourses (3a) result in longer processing times of the target sentence (in bold) compared to its continuous (3b) version. This difference was cancelled out by the use of temporal adverbials such as around eleven o’clock. Continuous (3) a. I went into the kitchen to prepare the dinner. I peeled the potatoes. I put the roast in a saucepan. (Around eleven o’clock) I cut up a slice of cooked ham. I gave it to the cat […] Discontinuous (3) b. This Monday, I got up very late. I had a full breakfast. I decided to go for a trip in the country. I dressed myself warmly. (Around eleven o’clock) I cut up a slice of cooked ham. I gave it to the cat […] If, however, the temporal expression was inserted at the end of the target sentence, it did not lead to such facilitation effects. Moreover, the facilitation was not due to sentence initial adverbials in general. The presence of adverbials like as usual, which are assumed not to have a segmentation function, did not affect the processing of discontinuous texts. In addition to topic continuity, chronological distance between events also has an impact on the online processing of a discourse. It has been assumed that, by default, people tend to interpret subsequent events as temporally successive (Chafe 1979). In a strong version of this iconicity principle, this preference is taken even further, assuming that the second event is by default interpreted as immediately following the first one. Evidence for such a preference comes from Anderson, Garrod and Sanford (1983), who tested the online effects of narrative shifts. The results of their study indicated that processing is faster after a temporal adverbial that expresses a small chronological distance between events (ten minutes) compared to those that express a longer distance (seven hours). In addition, words from the text preceding the adverbial were more accessible when the chronological distance expressed by the adverbial was small. The authors explain their data in terms of a scenario account. According to this theory, readers interpret a discourse in scenarios. Information within the boundaries of this scenario becomes less accessible when the story moves on beyond those boundaries. Temporal markers that express a longer distance prompt the reader to construct a new event and deactivate the previous one, resulting in longer probe recognition times to information from the previous event. A disadvantage of this account, however, is that it predicts that readers can only make use of the temporal information in well-known scenarios because it relies on world knowledge of the prototypical duration of certain events. In addition, it is not clear why information within the boundaries of a scenario should be easier to process compared to the same words when the story has moved beyond

The processing of pragmatic information in discourse

209

that boundary. After all, in the former case, the scenario account implies that more information has to be retained in working memory because the scenario is kept activated. This may just as well result in an increased processing load, inducing longer reading times. Zwaan (1996) conducted a set of self-paced reading experiments which provide further evidence for the strong iconicity principle. The author manipulated short stories that described events which last longer than an hour but are completed within a day, see example (4). The target sentence, presented here in bold, was an unrelated event that was preceded by temporal markers. These markers varied the chronological distance from the first event from close (a moment later), to intermediate (an hour later), to distant (a day later). (4) The grand opening1 Today was the grand opening of Maurice’s new art gallery. He had invited everybody in town, who was important in the arts. Everyone who had been invited, had said that they would come. It seemed like the opening would be a big success. At seven o’clock, the first guests arrived. Maurice was in an excellent mood. He was shaking hands and beaming. A moment / an hour / a day later, he turned very pale. He had completely forgotten to invite the local art critic. And sure enough, the opening was very negatively reviewed in the weekend edition of the local newspaper. Maurice decided to take some Advil and stay in bed the whole day. Recognition probe: beaming Reading times for the target sentence were shortest in the close condition but the intermediate and the distant condition did not differ from each other. In other words, discontinuity within the time interval of the described events caused a disruption, but the exact duration of the chronological distance did not seem to matter. Also, response latencies to probe words from the first event significantly increased with chronological distance. Note that although the target sentence he turned very pale is not a plausible continuation for a day later because the event of turning pale is something that happens in a rather delimited period of time, this does not explain the results. In the case of the marker an hour later there is no such infelicity, yet this condition still led to processing difficulties. Further control experiments excluded the possibility that these results were due to the wording of the temporal markers, differences in base-rate reading times of the temporal phrases or the plausibility of the time shifts. These results thus demonstrate that temporal discontinuity within the time interval of an event also causes

210

Ted J.M. Sanders and Anneloes R. Canestrelli

processing disruption, which is not predicted by the scenario account. According to Zwaan (1996: 1205) any chronological discontinuity triggers the set-up of a new time interval, which is costly in terms of processing load. Results like the ones described above form an illustration of how pragmatic information that marks referential coherence and topic continuity affects online discourse processing. If a topic discontinuity is left unmarked, readers are faced with substantial processing difficulties. Moreover, chronological distance conveyed by temporal adverbials affects the way in which language users process and represent information. When this distance increases, processing becomes more demanding and memory of preceding information decays.

4.

Relational coherence

As stated earlier, a second type of coherence is constituted by coherence relations such as additive, causal and contrastive relations (cf. Hobbs 1979; Sanders, Spooren, and Noordman 1992; Kehler 2002), which are also known as rhetorical relations (Mann and Thompson 1988; Taboada and Mann 2006). 4.1.

Coherence relations and their linguistic markers

Coherence relations are meaning relations connecting two or more discourse segments. A defining characteristic for these relations is that the interpretation of the related segments needs to provide more information than is provided by the sum of the segments taken in isolation. Examples are relations like CAUSE - CONSEQUENCE , LIST and PROBLEM - SOLUTION . These relations are conceptual and they can, but need not, be made explicit by linguistic markers, so-called connectives (because, so, however, although) and lexical cue phrases (for that reason, as a result, on the other hand). In pragmatics, the term discourse marker is often used to refer to the same items. Next to the linguistic markers of referential coherence, the lexical markers of relational coherence can be regarded as a second type of ‘processing instructions’ to language users. Although the lexical markers vary from connectives to cue phrases and signalling phrases, and thus vary with respect to the grammatical categories to which they belong, they do have a very similar function when it comes to establishing discourse coherence. This does not imply that explicit linguistic marking is a prerequisite for coherence, however. As stated earlier, coherence is a cognitive phenomenon, whereas cohesion directly refers to the textual items. Consider example (5), which is perfectly coherent even in the absence of a connective: (5) It’s raining. My hair is going to be totally messed up!

The processing of pragmatic information in discourse

4.2.

211

Linguistic markers of relational coherence: connectives and cue phrases

In a cognitive approach to discourse coherence, linguistic markers of coherence relations can be seen as explicit processing instructions on how one part of a text is related to another (Britton 1994; Sanders and Spooren 2007). As such, connectives are assumed to have an integration function (Noordman and Vonk 1997). They facilitate the integration of two text segments because they signal the relation between them. Over the past decades, a fair amount of work has studied the contribution of these markers to the online processing and the representation of information. Several studies have reported faster processing times for information preceded by a connective and faster recognition times to information followed by a connective. Haberlandt (1982) observed that readers benefit from the presence of a connective. Using short narratives like (6), the author presented phrases that were preceded by a connective, such as therefore, so and however, and compared reading times to reading times of an implicit version of the same clause. (6) The jet had just taken off. The left engine caught fire. The passengers were terrified. They thought the plane would crash. However, the pilot made a safe landing. The results demonstrated that the first words of the target sentence, the pilot, were read faster when it was preceded by a connective. Haberlandt concluded that the connective makes the relation between two clauses explicit, hence explaining the facilitation effect. However, because the facilitation was found only at the first words following the connective, this effect might just as well reflect another function of connectives. Connectives not only inform about the relation between two text segments, they also tell us something about the structure of the incoming material. This segmentation function may have an impact on sentence processing, since we know that the first words of a new sentence are processed more slowly compared to the same words if they do not occur sentence-initially (Haberlandt 1984). A second issue concerns the connection between the clauses without the connective. The example makes it clear that without the connective however it is very difficult to relate the two sentences. It is therefore possible that the facilitation effect observed here is not due to the fact that the connective makes the contrast relation explicit, but simply because the connective is a prerequisite for constructing the relation at all. Besides the segmentation and the integration function, another property has been ascribed to connectives. According to Noordman and Vonk (1997), connectives trigger the inference of extra-textual information; this is also referred to as the inference function. There is some experimental evidence that suggests that these different functions associated with connectives are cognitively real.

212

Ted J.M. Sanders and Anneloes R. Canestrelli

Investigating the processing of sentences such as (7), Millis and Just (1994) found that the presence of the connective because initially leads to faster processing, but delays the processing of the final word of the second clause. Moreover, the authors report faster recognition times to a probe word from the previous clause presented at the end of the second segment. According to Millis and Just, these faster responses are a result of the reactivation of the first clause due to integration processes. (7) The elderly parents toasted their only daughter at the party [because] Jill had finally passed the exams at the prestigious university. Probe: toasted Cozijn (2000) studied the different functions of the Dutch connective omdat (‘because’) in a series of eye-tracking experiments. The materials consisted of short expository texts with four introductory sentences followed by a target sentence (8) that was presented either with or without a connective. (8) On his way to work he experienced a long delay, (because) there was a large traffic jam on the highway. Reading times for the target sentence were compared across conditions. Like Millis and Just, Cozijn found a speeding-up effect on the words following the connective compared to an implicit version of the same relation. Because the effect was not restricted to the first words, the author concluded that this is not merely a segmentation effect but a reflection of the integration function. In addition, an inference effect was observed. The final words of the second clause were read more slowly in the presence of a connective. Such integration effects are not limited to connectives per se; they have also been observed when coherence relations are marked by means of signalling devices such as The problem is and Another aspect might be. Sanders and Noordman (2000) found that such signals speed up the processing of both problem-solution and list relations. 4.3.

The role of various coherence relations in discourse processing and representation

But how about the effect of different coherence relations, irrespective of the role of linguistic markers? The idea that coherence relations affect text processing does indeed get support from results in processing studies (Sanders and Spooren 2007). Several studies suggest a processing difference between CAUSAL and NON -CAUSAL relations, such as additive ones. For one thing, causally related events in short narratives are recalled more easily than non-causally related events (Black and Bern 1981; Trabasso and Van den Broek 1985). Keenan, Baillet and Brown (1984) and

The processing of pragmatic information in discourse

213

Myers, Shinjo and Duffy (1987) demonstrated that the effect of causal connectedness on memory for sentences is greatest for moderate levels of causality. Also, causally related sentences are read faster (Haberlandt and Bingham 1978), and reading times decrease when the causality increases (Keenan, Baillet, and Brown 1984; Myers, Shinjo, and Duffy 1987). Fewer studies exist for expository text. Sanders and Noordman (2000) embedded a similar text segment in two different contexts. In one case it was a solution to a problem, in the second case the same segment was part of an addition. It was found that PROBLEM - SOLUTION relations lead to faster processing, better verification and superior recall. The authors conclude that the processing of a text segment depends on the relation it has with preceding segments. Such findings illustrate the relevance of causal versus additive coherence relations: causally related information is encoded faster and retrieved better than additively related information, suggesting a strong processing and representational advantage for causal relations. So far, we have looked at studies of discourse processing and representation which provide a window on the human mind at work. Studies of language acquisition provide another window on the mind, especially with respect to cognitive complexity. In areas like morphology and syntax, it has repeatedly been argued that less complex structures are acquired before more complex ones. We focus on discourse here. Most children build their first multi-clause discourse before the age of three; instead of uttering one clause at a time, they start producing combined clauses (cf. Clark 2002). At first, the coherence relations between these clauses remain implicit. Later, children learn how to use connectives to make coherence relations explicit. With regard to relational coherence, we may expect the relational categories that are distinguished to be relevant for the way in which children acquire connectives and relations. More precisely, different conceptual categories can be expected to show a different pattern in acquisition (Evers-Vermeul and Sanders 2009). For instance, learning the conceptual relation of causality as well as acquiring the linguistic tools to express it only seems to be possible on the basis of an understanding of simple additive relations: children have to understand that events can be related, that they can be ordered in time and, subsequently, that the first event, which precedes the second, can be the cause of the second. Indeed, the hypothesis that cognitive complexity sets the pace for the acquisition (Clark and Clark 1977: 338) of causals seems to fit available data. Both naturalistic and experimental studies of first language acquisition show that additive relations are acquired before causals, and that additive connectives are acquired before causal connectives, which appear at around age two (Clark 2002). More precisely, Bloom et al. (1980) found that English children follow the same route in acquiring coherence relations: (9) additive < temporal < causal < adversative

214

Ted J.M. Sanders and Anneloes R. Canestrelli

Studying the transcripts of spontaneous speech data of twelve Dutch-speaking children in the CHILDES database (MacWhinney 2000), Evers-Vermeul and Sanders (2009) show that the first causal connective (as in their example 10b) does not appear before an additive connective (as in 10a) has occurred. (10) a. Kim is een meisje. En jij bent een jongetje he? (Thomas, 2;8.23)2 ‘Kim is a girl. And you are a boy, aren’t you?’ b. Ik heb (een) beetje griep. Want ik ben laat gaan slapen. (Abel, 3;3.8) ‘I’ve got a touch of flu. Because I went to bed late.’ All Dutch children use the additive en before they come up with the causal connective want. Hence, in terms of first emergence, empirical studies of English and Dutch corroborate the idea that causal relations are more complex than additive relations. 4.4.

Connectives and relations that vary in subjectivity

In pragmatics and (text) linguistics, further categorizations of connectives as well as coherence relations have been made between objective (content or semantic) relations and subjective (or epistemic and speech act, pragmatic) relations. These two types of relations differ with respect to the degree of speaker involvement or, following Langacker (1990), subjectivity (Pander Maat and Sanders 2001; Pander Maat and Degand 2001; Verhagen 2005), and are associated with specific markers. Examples of objective causal markers are as a result and consequently, examples of subjective markers are therefore, so and that’s why. Recently, much (cross-)linguistic work has been done on these categories and on the way in which they are realized in connectives across languages (Stukker, Sanders, and Verhagen 2008). Here, however, our main interest is in the cognitive processes and the representation of these categories of relations and connectives. There is reason to believe that subjective causal relations are cognitively more complex than objective ones. Subjective relations carry a higher overall degree of uncertainty than objective causals because they involve inferences about causes and effects (Degand and Pander Maat 2003). In addition, in order to understand subjective relations, you need to be able to follow the reasoning of the speaker. Objective causal relations, on the other hand, do not require such inferences, but simply relate two events in the physical world, and as such they may be easier to process. In language acquisition it has been found that children acquire at least some subjective causal relations systematically later than others; for example, corpora of Dutch and English child language data show that subjective causal relations appear later than objective relations in both English and Dutch children (Evers-Vermeul and Sanders 2010), suggesting that the former category can indeed be considered as being more complex. This difference in complexity is paralleled by online processing data with adults. Studying on-line text processing in English, Traxler, Bybee and Pickering (1997) found that subjective causal relations (like

The processing of pragmatic information in discourse

215

11b) produced longer reading times on the prefinal region, bit him, compared to objective causal relations (11a). (11) a. Rick almost died from shock because a poisonous snake bit him on the leg b. Rick was walking in a remote area because a poisonous snake bit him on the leg Noordman and de Blijzer (2000) obtained very similar results for Dutch. With a self-paced reading experiment they found that subjective causal relations took more time to process compared to objective relations. Traxler, Bybee, and Pickering propose that the difficulty of these subjective causals is explained by the inferences that readers have to make in order to establish the relation. Because readers have to infer that the first clause is a speaker’s belief rather than a fact in the world, subjective relations would require an extra inference. One explanation for this processing difference could be that the subjective relations manipulated in such experiments are relatively far-fetched, see for example (11a, b). Another, arguably more interesting one might be that the subjective relation can only be interpreted adequately when readers understand that they are interpreting causal reasoning of the speaker or author (see Sanders, Sanders, and Sweetser [2009] for an account in the framework of Mental Spaces Theory). Once they are aware of this, subjective relations might not be so complex after all (Canestrelli, Mak, and Sanders, to appear). Moving away from the dominant text type studied – that of informative texts – Kamalski et al. (2008) performed a self-paced reading experiment to test the effects of different markers on processing short persuasive texts. Different types of connectives and cue phrases were manipulated and each experimental item was presented with either a subjective marker or an objective marker, or remained without a marker. Example (12) illustrates an objective manipulation (a) and a subjective manipulation (b) in this study. (12) a. Such a test will assess students’ writing ability, reasoning and computer skills. Consequently, the board can ensure that students don’t graduate without possessing basic skills. b. Although some state officials worried that such a test would cause professors to gear their classes toward the test, this will ultimately have a positive effect on the quality of the state college system. Therefore, an exit exam for college students needs to be implemented as soon as possible. All students will benefit from such a system. In the objective condition, Kamalski et al. found a tendency for speeding-up for the entire clause following the connective. These data were interpreted as reflecting an integration effect without an inference effect since no slow-down could be observed. In the subjective condition an interaction between segment and marker was observed, meaning that the difference between the conditions shows variation with

216

Ted J.M. Sanders and Anneloes R. Canestrelli

respect to the position in which it is measured. The authors conclude that a subjective marker invokes an integration effect on subsequent material, but results in a slow-down in the final region. So what could cause readers to slow down when they have encountered subjective coherence markers? The off-line data of this experiment revealed that only in the case of subjective markers did subjects give faster responses to questions about the author’s intent, for example to persuade or to inform. This result suggests that readers are more aware of the author’s intentions when coherence is marked by subjective markers, which may explain the observed slow-down. Limitations of this study involve the statistical analysis of the results (see Mulder 2008 for details) and the high degree of variability between the segments for which reading times are measured, due to the experimental set-up. Nevertheless, studies like these suggest that the ‘subtle’ linguistic categories (subjective versus objective causal relations), as distinguished by text linguists are indeed relevant in text processing as well. Perhaps even more importantly, they show this to be the case in a genre other than that of the well-researched informative text. 4.5.

Signalling global text structure

Relational coherence markers are not the only signals that affect the reading and interpretation process. Global coherence marking, such as signals and cue phrases which inform about the organization of a text, may be just as relevant in discourse processing, since they establish coherence as well. It has been consistently reported in the educationally-oriented literature that headings and topic overviews that cue the structure of a text contribute to the retention and retrieval of information and to the way in which the information is processed. Hyönä and Lorch (2004), for example, investigated the effect of headings on online comprehension. Using the eye-movement technique they presented readers with texts in which the presence of headings was manipulated. These headings consisted of noun phrases such as “pandas” and “wind power” which mention the new topic. The results showed that text summaries contained more topics when headings were present. As reported in many other studies (Murray and McGlone 1997; Sanchez, Lorch, and Lorch 2001), the online data revealed that reading times of topic sentences decreased when they are preceded by a heading. These data suggest that topic headings can indeed be seen as effective signals that facilitate the online comprehension processes. An important limitation to these studies is that little or no attention is paid to the actual wording of these headings. Moreover, many mixed signal studies involve the manipulation of structural and relational signals simultaneously, which makes it rather difficult to say anything independently about the independent effects of the manipulations (cf. Degand and Sanders 2002).

The processing of pragmatic information in discourse

4.6.

217

Faster processing, sloppy representation?

As we have seen so far, there is substantial evidence that (at least certain types of) connectives and cue phrases speed up the processing of the subsequent information. Still, one could question how faster processing affects the representation of the textual information afterwards. After all, if a given relation is already made clear by overt marking, the language processor has to do less work, which may result in a sloppier representation. Several researchers have investigated this matter; the empirical data so far do not yet provide a decisive picture. It has been shown in a number of studies that overt marking of coherence relations improves the mental representation of a text. This results in better recall performance (e.g. Lorch and Lorch 1986), more complete summaries (Hyönä and Lorch 2004), faster and more accurate responses to comprehension questions (Millis and Just 1994; Degand and Sanders 2002), faster responses on verification tasks (Sanders and Noordman 2000) and a higher overall quality of recalled information (Meyer, Brandt, and Bluth 1980). Other results, on the other hand, suggest that such markers do not have this facilitating function. Spyridakis and Standal (1987) found only a limited effect of marking on text comprehension. In addition, some studies report that linguistic markers of coherence relations do not contribute to the amount of information that can be recalled. Sanders and Noordman (2000), as discussed previously, found that overt marking decreased the reading time for both problem solution and list relations, however, this manipulation did not lead to a better recall of the target information as measured with a free recall task. So how can we account for the different results? A number of factors could explain this diverging evidence. The first is that different studies investigate different types of signals. For instance, in many experimental studies, additive, contrastive and causal relations were made explicit and subsequently compared to an implicit version of the same text. It is not clear whether different types of signals have a different impact on text representation. A second issue is related to the experimental texts used in different studies. In many cases, connectives are plugged in between two sentences without considering the context and the plausibility of the relation. A third concern is that the operationalization of text comprehension varies. It may be the case that not all experimental methods are sensitive enough to tap into the representation of a text. Experiments that have applied the free recall method have failed to report effects of relational marking on comprehension (Meyer 1975; Sanders and Noordman 2000). It may very well be the case that other measures, such as answering and sorting tasks (Kintsch 1998), can provide very different results. In fact, question-answering may be a more suitable test to assess text comprehension. Degand and Sanders (2002) tested the effects of causal connectives (want,

218

Ted J.M. Sanders and Anneloes R. Canestrelli

omdat, doordat, which are all translated by because) and cue phrases (the reason was, a consequence of this was) on the comprehension of expository texts in both L1 and L2 of Dutch- versus French-speaking readers. It was found that both groups benefit from causal markers in their first as well as their second language. Overt marking led to higher accuracy on comprehension questions that concerned not only the causal relation but also other aspects of the text. Contrary to what was found in Sanders and Noordman (2000), who used free recall, these data suggest that linguistic signals do improve the representation of information. Clearly, different methods provide different results. A fourth explanation for the differences between studies concerns the interaction of the text under investigation with reader characteristics. We know by now that readers with a high degree of knowledge about a given topic do not benefit from signals in the way that low-knowledge readers do. McNamara and Kintsch (1996), McNamara et al. (1996) and Kamalski, Sanders and Lentz (2008) found that high-knowledge readers benefit more from an implicit version of a text. For these readers, explicit marking of local and global coherence results in an inferior representation of the text. In addition, Kamalski et al. (2008) demonstrated that this interaction depends on the specific text genre that is manipulated. The pattern in persuasive texts differed from that in informative texts. It is still an open question, though, whether and how high-knowledge readers differ from low-knowledge readers when it comes to the online processing of information. With respect to topic headings it has been found that facilitation effects are modulated by reader characteristics as well. More advanced readers seem to benefit less from headings, presumably because they already apply systematic topic-processing strategies and do not rely on the instructions provided by these signals (Meyer, Brandt, and Bluth 1980; Sanchez, Lorch, and Lorch 2001). Also, other factors such as readers’ goals (Noordman, Vonk, and Kempff 1992) and verbal ability (Meyer, Young, and Bartlett 1989) seem to influence the effects of relational markers. 4.7.

Some pitfalls of experimental studies

Our overview demonstrates that a substantial amount of experimental work has been done in order to test the effects of pragmatic markers, such as connectives, on discourse processing. Unfortunately, it is not always possible to draw clear conclusions from such results. One problem that arises in many experiments concerns the materials used. In a typical set-up, the processing of unmarked relations is compared to a marked version of the same relation. The idea is that, by doing so, researchers are able to measure reading times and eye-movements, or, in recent years, even neural activation, in response to identical target information in both conditions. An important point is that too often those conditions are not really comparable, because of a mismatch between the marker and the relation.

The processing of pragmatic information in discourse

219

Maury and Teisserenc (2005), for instance, investigated the effects of connectives on the processing of scientific texts. The materials in this study consisted of items in which an implicit causal relation was marked by either a causal connective, an additive connective, or no connective (13). It is irrelevant to discuss all their results, but one outcome was that the additive connective and had a negative impact on the processing of the second clause compared to the implicit condition, while so only had a marginally facilitating effect. However, it would be wrong to jump to the conclusion that and disrupts processing and so does not really help either. What the authors have not taken into account is that there is an intricate relation between connectives on the one hand and the properties of the relation on the other. Maury and Teisserenc compared marked relations to an implicit version, and that implicit version is clearly of a causal nature. Although the additive connective and allows for a causal interpretation, insertion of this connective changes the interpretation processes compared to its implicit counterpart. In order to draw sensible conclusions about the effect of the connective and, the implicit condition has to be of an additive nature as well.3 (13) The chemist heats the plastic to a high temperature, the plastic spreads over the mould. The chemist heats the plastic to a high temperature, so the plastic spreads over the mould. The chemist heats the plastic to a high temperature, and the plastic spreads over the mould.4 A more practical limitation to the online study of connectives concerns the operationalization of how reading times are measured. The different functions of causal connectives operate on different moments in time. Recall that the integration function leads to an initial speed-up, but that processing slows down at the end of the clause, due to the inference function. When reading times are not measured on the separate words (or phrases), but pooled in total reading times, these processing effects may be evened out. To sum up, the on- and off-line effects of linguistic markers of relational coherence have been well attested to date. Yet a general problem in some experimental studies is that the text materials are not always adequate. It seems that closer collaboration between experimental psycholinguists on the one hand and (text) linguists on the other might improve this situation. What has become clear so far is that under certain conditions – e.g. readers should not be experts on the text content – readers seem to benefit from coherence signals, even though this effect is not found with all comprehension and recall methods. Online results show that linguistic markers of coherence usually speed up processing of the immediately following information. An important question is whether it is possible to generalize over connective and relation type. For example, it is clear by now that causal connectives induce an integration effect, because they overtly signal the causal re-

220

Ted J.M. Sanders and Anneloes R. Canestrelli

lation that holds between the clauses. Adversative connectives, on the other hand, seem to be a prerequisite for constructing the relation, because without them, the adversative relation cannot be established in the first place. As such, their function in processing may be very different from that of causal connectives. In addition, the exact type of coherence relations also affects processing speed. Implicit causal relations are processed faster than other types of coherence relations. Therefore, the magnitude of the facilitation due to adversative markers is likely to be much larger compared to causal connectives because implicit causals are already processed relatively quickly. Further theoretical work on the nature of the semanticpragmatic differences between various types of relations and connectives is needed in order to clarify this discussion. Cognitive Linguistics may have a role to play here, see the final section of this chapter.

5.

Exploring some neurocognitive foundations of discourse processing

Recent data from neuroimaging studies involving methods such as fMRI and event related potentials (ERP) form a new and exciting contribution to the investigation of discourse processing. Not only do these techniques inform us about the neural correlates that are involved in the processing of pragmatic information, they also give insights into the exact time course of effects (in the case of ERP) and help us understand why they arise in the first place. Experimental work on referential coherence includes some nice examples of such studies. Recently it has been demonstrated that the repeated name penalty, as discussed at the beginning of this chapter, is reflected in a clear pattern of neural activation. Almor et al. (2007) used the same items as Gordon, Grosz and Gilliom (1993) in an fMRI experiment to investigate how repetitive reference affects the neural substrates, see example (14). (14) Susan is really into animals. The other day Susan / she gave Betsy a pet hamster. Susan / she reminded Betsy that such hamsters are quite shy and need gentle handling. Question: Does Susan like animals? It was found that, compared to pronouns, the repetition of names elicited more activation in the left and right intraparietal sulcus, a brain region that has been associated with spatial processing in tasks that also involve the representation of several entities. This finding thus suggests that the repeated name penalty arises because of the increased representational load due to the activation of a new representation. These results fit in nicely with Ariel’s Accessibility Theory, as discussed at the beginning of this chapter. In this theory, long linguistic forms are associated with ref-

The processing of pragmatic information in discourse

221

erents that are low-accessible, and should therefore trigger the construction of a new referent. With respect to global coherence, a few studies using fMRI have investigated the effects of text titles on the neural substrates involved in discourse processing. Both St George et al. (1999) and Martín-Loeches et al. (2008) studied the neural activation involved in the processing of globally coherent paragraphs with a very similar paradigm, but obtained diverging results. The authors manipulated texts that were difficult to understand without a title – a paradigm known since Bransford and Johnson (1972) – and compared the neural activation between paragraphs with and without a title. In keeping with previous research, Martín-Loeches et al. found that the presence of a title results in higher comprehension rates and better recall of the content of a paragraph. In addition, attempts to achieve global coherence, in the case of the untitled paragraphs, resulted in an increased activation of both hemispheres. However, while St George et al. observed the largest increase in activation in the right hemisphere, Martín-Loeches et al. found that the left hemisphere was primarily involved in the process. The authors suggest that incoherent paragraphs may be read as single sentences, or even as isolated words – an idea in line with the findings of St George, Mannes, and Hoffman (1994), who found an increased N400 effect for all words in a globally incoherent text. This effect has consistently been associated with semantic integration and may reflect the difficulty in relating words and sentences when the reader cannot grasp what the story is about. It is beyond the purpose of the present chapter to decide upon the exact location of the regions involved in establishing global coherence. Martín-Loeches et al. correctly suggest that the variance may be due to the fact that different narratives can result in different activations. Even though this matter remains unsettled, these studies do illustrate that globally incoherent texts result in a larger amount of brain activation compared to coherent texts, which reflects the difficulties readers face when presented with incoherence. These studies are just a few examples from a rapidly expanding field, which promises to be very significant. So far, we have only a very limited understanding of the neural basis of cognitive pragmatics, partly due to the methodological shortcomings of techniques such as fMRI. For instance, the measurement of changes in blood flow, on which this technique relies, is based on the assumption that this change reflects a change in neural activity. Yet this signal is also affected by factors that have nothing to do with neural activity (see Bartsch and Haller 2009 for discussion). Nevertheless, the contribution of neuroimaging research to the study of cognitive pragmatics is exciting and promising. The rapid developments in the field with respect to the equipment, the software, and the statistics involved will hopefully allow for more reliable and insightful experiments in the near future.

222 6.

Ted J.M. Sanders and Anneloes R. Canestrelli

Conclusion and future perspectives

When we try to oversee the field of pragmatics and discourse processing in the way in which we have just sketched it, a first key question is: can we generalize about all these different linguistic markers of coherence? Can referential and relational markers really be considered to be related? And can however and because really be expected to have similar functions? On the one hand, one could argue that all coherence markers have at least one thing in common: by providing explicit information on relational and referential coherence they enable readers to construe text representations. And this effect should be visible during online processing as well as in discourse representations. On the other hand, the various types of coherence markers are likely to have varied effects on the processing and representation of discourse. For instance, we have seen earlier in this chapter that objective causal connectives, expressing realworld causality between events, have a different influence on online processing and text representation than subjective markers that express the arguments of the author. Therefore, further research on the differences between the various markers and coherence relations is imperative. In fact, we have seen how a cognitive approach to coherence relations (Sanders and Spooren 2009) accounts for differences in acquisition, processing and representation between positive and negative and between causal and additive relations in terms of cognitive complexity. However, such an approach does not account for the intuition of the ‘necessity’ of contrastive connectives – as in Haberlandt’s “However, the pilot made a safe landing” (see section 4.2, and example 6). In addition, it does not yet sufficiently account for the differences between objective and subjective relations. Yet Cognitive Linguistics may have a role to play here, because of its determination to provide cognitively plausible analyses of linguistic phenomena. Fauconnier’s Mental Space framework (Fauconnier 1985, 1994) may be suitable for modelling this type of phenomena. He treats connectives as one type of socalled space-builders, that is, linguistic expressions that typically establish new mental spaces. Mental Spaces are mental constructs set up to interpret utterances, “structured, incremental sets […] and relations holding between them […], such that new elements can be added to them and new relations established between their elements” (Fauconnier 1994: 16). An example of a connective acting as a space-builder is the if-then conditional, as in If I were a millionaire, my VW would be a Rolls. An expression like if p then q sets up a new mental space H in which q holds. In other words, if I were a millionaire is the space-builder and in this new space my VW from the initial space is identified with the Rolls in the new space. The theory of mental spaces seems to be particularly compatible to the conceptualization of connectives, and can help to explain the differences in processing. For instance, in this approach connectives are often treated either as elements that

The processing of pragmatic information in discourse

223

block certain inferences (such as but and however) or as space-builders, i.e. linguistic expressions that typically establish new mental spaces, such as if-then conditionals. In recent years, cognitive linguists have shown how the theory of mental spaces can be used fruitfully to clarify the meaning and use of conditionals (Dancygier and Sweetser 2000, 2005) and causals (Sanders, Sanders, and Sweetser 2009; Verhagen 2005). From a methodological point of view, it can be concluded that the integration of cognitively plausible theories with empirical testing is the ultimate aim rather than a situation that has already been realized (Sanders and Spooren 2007), even though this tendency is clearly visible in book volumes such as Methods in Cognitive Linguistics (Gonzalez-Marquez et al. 2007). One way to realize this goal is to proceed with the thorough investigation of corpora of actual language use. Digital corpora enable researchers to do this on a larger scale than ever, and recent studies show how fruitful statistic and (partly) automatic analyses of corpora can be for the area of causal verbs and connectives, too. Speelman and Geeraerts (2009) is an outstanding example on causative verbs, and on connectives, Bestgen, Degand and Spooren (2006) have shown the way. Furthermore, it is especially important to extend corpus research in the direction of spoken discourse. On the whole, the field of corpus-linguistic studies is still largely based on the study of written discourse. There are at least two important questions to consider: to what extent can results be generalized to apply to spoken discourse? And what do the specific insights from the linguistic analysis of spoken discourse add to the picture we have so far? At present, we have only limited results on non-written connective use (Ford 1993; Couper-Kuhlen 1996; Gohl 2000), even though this challenge has recently been taken up by, among others: Sanders and Spooren (2007); Speelman and Geeraerts (2009), Spooren et al. (2010). Integration of text-linguistic and psycholinguistic insights is a second way to realize the goal of interaction between theory and empirical testing. The sometimes subtle semantic-pragmatic distinctions proposed by linguists, on the one hand, and the processing effects revealed by psycholinguistic research, on the other hand, still need to be linked. For instance, the general processing effects of because and and have been investigated (e.g. Millis, Golding, and Barker 1995), but there is very little experimental research into the processing instructions encoded by connectives that differ in specificity (e.g. but versus although, or dus versus daardoor). Similarly, a few processing studies have been conducted based on linguistically sophisticated analysis of the categories of relations and connectives discussed earlier in this chapter – such as the differences in subjectivity. We believe a closer integration of theoretical and corpus-linguistic work with this type of processing studies would lead to significant further progress in the research field as a whole.

224

Ted J.M. Sanders and Anneloes R. Canestrelli

Notes 1. 2. 3. 4.

This is the original experimental text, including commas. The notation “2;8;23” stands for age in years;months;days. Millis, Graesser and Haberlandt (1993) ran into similar problems. We only discuss a part of the materials, relevant to our discussion. Each item also included a backward version of the relation marked by because.

References Almor, Amit, David V. Smith, Leonardo Bonilha, Julius Fridriksson and Chris Rorden 2007 What is in a name? Spatial brain circuits are used to track discourse references. Neuroreport 18(12): 1215–1219. Anderson, Anne, Simon C. Garrod and Anthony J. Sanford 1983 The accessibility of pronominal antecedents as a function of episode shifts in narrative text. The Quarterly Journal of Experimental Psychology A: Human Experimental Psychology 35: 427–440. Ariel, Mira 1990 Accessing Noun-phrase Antecedents. London: Routledge. Ariel, Mira 2004 Accessibility marking: Discourse functions, discourse profiles, and processing cues. Discourse Processes 37: 91–116. Arnold, Jennifer E. 1998 Reference form and discourse pattern. Stanford, CA:Stanford University Dissertation. Arnold, Jennifer E., Janet G. Eisenband, Sarah Brown-Schmidt, S., and John Trueswell 2000 The rapid use of gender information: Evidence of the time course of pronoun resolution from eyetracking. Cognition 76: 13–26. Bartsch, Andreas J. and Sven Haller 2009 Pitfalls in fMRI. European Radiology 19: 2689–2706. Bestgen, Yves and Jean Costerman 1994 Time, space, and action: Exploring the narrative structure and its marking. Discourse Processes 17: 421–446. Bestgen, Yves and Wietske Vonk 1995 The role of temporal segmentation markers in processing. Discourse Processes 19: 385–406. Bestgen, Yves and Wietske Vonk 2000 Temporal adverbials as segmentation markers in discourse comprehension. Journal of Memory and Language 42: 74–87. Bestgen, Yves, Liesbeth Degand and Wilbert Spooren 2006 Toward Automatic Determination of the Semantics of Connectives in Large Newspaper Corpora. Discourse Processes 41 (2): 175–193. Black, John B., and Hyman Bern 1981 Causal coherence and memory for events in narratives. Journal of Verbal Learning and Verbal Behavior 20: 267–275.

The processing of pragmatic information in discourse

225

Bloom, Lois, Margaret Lahey, Lois Hood, Karin Lifter and Kathleen Fiess 1980 Complex sentences: Acquisition of syntactic connectives and the semantic relations they encode. Journal of Child Language 7: 235–261. Bransford, John D. and Marcia K. Johnson 1972 Contextual prerequisites for understanding: some investigations of comprehension and recall. Journal of Verbal Learning and Verbal Behavior 11: 717–26. Brennan, Susan E. 1995 Centering attention in discourse. Language and Cognitive Processes 10 (2): 137–167. Brennan, Susan E., Marilyn W. Friedman and Charles J. Pollard 1987 A centering approach to pronouns. In: Proceedings of the 25th Annual Meeting of the Association for Computational Linguistics (ACL-37), 155–162. Stanford: CA. Britton, Bruce K. 1994 Understanding expository text: Building mental structures to induce insights. In: Morton A. Gernsbacher (ed.), Handbook of Psycholinguistics, 640–674. San Diego: Academic Press. Canestrelli, Anneloes R., Willem M. Mak and Ted J. M. Sanders to appear Causal connectives in discourse processing: How differences in subjectivity are reflected in eye-movements. Language and Cognitive Processes. Carpenter, Patricia A. and Marcel A. Just 1977 Integrative process in comprehension. In: David LaBerge and S. Jay Samuels (eds.), Basic Processes in Reading: Perception and Comprehension, 217–241. Hillsdale NJ: Erlbaum. Chafe, Wallace (ed.) 1979 The Flow of Thought and the Flow of Language. Syntax and Semantics. New York: Academic Press. Clark, Eve V. 2002 First language acquisition. Cambridge: Cambridge University Press. Clark, Herbert H. and Eve V. Clark 1977 Psychology and Language; An Introduction to Psycholinguistics. New York: Harcourt Brace Jovanovich. Clark, Herbert H. and C. J. Sengul 1979 In search of referents for nouns and pronouns. Memory and Cognition 7: 35–41. Couper-Kuhlen, Elizabeth 1996 Intonation and clause combining in discourse: The case of because. Pragmatics 6: 389–426. Cozijn, Reinier 2000 Integration and inference in understanding causal sentences. Ph.D. dissertation, Tilburg University. Dancygier, Barbara and Eve Sweetser 2000 Constructions with if, since, and because: Causality, epistemic stance, and clause order. In: Elisabeth Couper-Kühlen and Bernd Kortmann (eds.), Cause, Condition, Concession and Contrast: Cognitive and Discourse Perspectives, 111–142. Berlin: Mouton de Gruyter.

226

Ted J.M. Sanders and Anneloes R. Canestrelli

Dancygier, Barbara and Eve Sweetser 2005 Mental Spaces in Grammar: Conditional Constructions. Cambridge: Cambridge University Press. Degand, Liesbeth and Henk Pander Maat 2003 A contrastive study of Dutch and French causal connectives on the Speaker Involvement Scale. In: Arie Verhagen and Jeroen van de Weijer (eds.), Usage Based Approaches to Dutch, 175–199. Utrecht: LOT. Degand, Liesbeth and Ted Sanders 2002 The impact of relational markers on expository text comprehension in L1 and L2. Reading and Writing 15: 739–757. Degand, Liesbeth, Nathalie Lefèvre and Yves Bestgen 1999 The impact of connectives and anaphoric expressions on expository discourse comprehension. Document Design 1: 39–51. Duffy, Susan A. and Keith Rayner 1990 Eye Movements and Anaphor Resolution: Effects of Antecedent Typicality and Distance. Language and Speech 33: 103–119. Evers-Vermeul, Jaqueline and Ted Sanders 2009 The emergence of Dutch connectives; how cumulative cognitive complexity explains the order of acquisition. Journal of Child Language 36: 829–854. Evers-Vermeul, Jaqueline and Ted Sanders 2010 Discovering Domains – on the acquisition of causal connectives. Journal of Pragmatics 43(6): 1645–1662. Fauconnier, Gilles 1985 Mental Spaces. Cambridge, MA: MIT Press. Fauconnier, Gilles 1994 Mental Spaces: Aspects of Meaning Construction in Natural Language. 2nd edition. Cambridge, MA: Bradford. Ford, Celia E. 1993 Grammar in Interaction: Adverbial Clauses in American English Conversations. Cambridge: Cambridge University Press. Garrod, Simon C. and Anthony J. Sanford 1994 Resolving sentences in a discourse context: How discourse representation affects language understanding. In: Morton A. Gernsbacher (ed.), Handbook of Psycholinguistics, 675–698. San Diego: Academic Press. Gernsbacher, Morton A. 1990 Language Comprehension as Structure-building. Hillsdale, NJ: Erlbaum. Gernsbacher, Morton A. and Joerg D. Jescheniak 1995 Cataphoric devices in spoken discourse. Cognitive Psychology 29: 24–58. Gernsbacher, Morton A. and Suzanne Shroyer 1989 The cataphoric use of the indefinite this in spoken narratives. Memory & Cognition 17: 536–540. Gernsbacher, Morton A., Rachel R. W. Robertson, Paola Palladino and Necia K. Werner 2004 Managing mental representations during narrative comprehension. Discourse Processes 37: 145–164. Givón, Talmy 1992 The grammar of referential coherence as mental processing instructions. Linguistics 30: 5–56.

The processing of pragmatic information in discourse

227

Givón, Talmy 1995 Coherence in text vs. coherence in mind. In: Morton Ann Gernsbacher and Talmy Givón (eds.), Coherence in Spontaneous Text, 59–115. Amsterdam: John Benjamins. Gohl, Christine 2000 Causal relations in spoken discourse: Asyndetic constructions as a means for giving reasons. In: Elizabeth Couper-Kuhlen and Bernd Kortmann (eds.), Cause, Condition, Concession and Contrast: Cognitive and Discourse Perspectives, 35–56. New York: Mouton de Gruyter. González-Márquez, Mónica, Irene Mittelberg, Seana Coulson and Michael J. Spivey (eds.) 2007 Methods in Cognitive Linguistics. Amsterdam: John Benjamins. Gordon, Peter and Kimberly A. Scearce 1995 Pronominalization and discourse coherence, discourse structure and pronoun interpretation. Memory and Cognition 23: 313–323. Gordon, Peter, Barbara Grosz and Laura Gilliom 1993 Pronouns, names and the centering of attention in discourse. Cognitive Science 17: 311–347. Graesser, Art C., Danielle S. McNamara and Max M. Louwerse 2003 What do readers need to learn in order to process coherence relations in narrative and expository text. In: Anne P. Sweet and Catherine E. Snow (eds.), Rethinking Reading Comprehension, 82–98. New York: Guilford Publications. Graesser, Arthur C., Keith K. Millis and Rolf A. Zwaan 1997 Discourse comprehension. In: Janet Spence, John Darley and Donald Foss (eds.), Annual Review of Psychology 48, 163–189. Palo Alto, CA: Annual Reviews. Gundel, Jeannette, Nancy Hedberg and Ron Zacharski 1993 Cognitive status and the form of referring expressions in discourse. Language 69: 274–307 Haberlandt, Karl 1982 Reader expectations in text comprehension. In: Jean F. Le Ny and Walter Kintsch (eds.), Language and Comprehension, 239–249. Amsterdam: North Holland Haberlandt, Karl 1984 Components of sentence and word reading times In: David E. Kieras and Marcel A. Just (eds.), New Methods in Reading Comprehension research, 219–251. Hillsdale, NJ: Erlbaum. Haberlandt, Karl, and George Bingham 1978 Verbs contribute to the coherence of brief narratives: Reading related and unrelated sentence triples. Journal of Verbal Learning and Verbal Behavior 17: 419–425. Halliday, Michael A. K. and Ruqaiya Hasan 1976 Cohesion in English. London: Longman. Hobbs, Jerry R 1979 Coherence and coreference. Cognitive Science 3: 67–90. Hudson-D’Zmura, Susan and Michael K. Tanenhaus 1998 Assigning antecedents to ambiguous pronouns: The role of the center of attention as the default assignment. In: Marilyn Walker, Aravind Joshi and Ellen

228

Ted J.M. Sanders and Anneloes R. Canestrelli

Prince (eds.), Centering Theory in Discourse, 199–226. Oxford: University Press. Hyönä, Jukka and Robert F. Lorch 2004 Effects of topic headings on text processing: Effects from adult readers’ eye fixation patterns. Learning and Instruction 14: 131–152. Kamalski, Judith, Ted Sanders and Leo Lentz 2008 Coherence marking, prior knowledge and comprehension of informative and persuasive texts: Sorting things out. Discourse Processes 45: 323–345. Kamalski, Judith, Leo Lentz, Ted Sanders and Rolf Zwaan 2008 The forewarning effect of coherence markers in persuasion: off-line and online evidence. Discourse Processes 45: 545–579. Kameyama, Megumi 1996 Indefeasible Semantics and Defeasible Pragmatics. In: Kanazawa, Makoto, Christopher Piñon, and Henriëtte de Swart (eds.), Quantifiers, Deduction, and Context, 111–138. Stanford, CA: CSLI. Keenan, Janice M., Susan D. Baillet and Polly Brown 1984 The effects of causal cohesion on comprehension and memory. Journal of Verbal Learning and Verbal Behavior 23: 115–126. Kehler, Andrew 2002 Coherence, Reference, and the Theory of Grammar. Stanford, CA: CSLI Publications. Kennison, Sheila.M. and Peter C. Gordon 1997 Comprehending referential expressions during reading: Evidence from eye tracking. Discourse Processes 24: 229–252. Kintsch, Walter 1998 Comprehension: A Paradigm for Cognition. Cambridge: Cambridge University Press. Langacker, Ronald W. 1990 Subjectification. Cognitive Linguistics 1: 5–38. Lorch, Elisabeth P. and Robert F. Lorch 1986 The on-line processing of summary and importance signals in reading. Discourse Processes 9: 489–496. MacWhinney, Brian 2000 The CHILDES Project: Tools for Analyzing Talk. Mahwah, NJ: Lawrence Erlbaum Mann, William C. and Sandra A. Thompson 1988 Rhetorical structure theory: towards a functional theory of text organization. Text 8: 243–281. Martín-Loeches, Manuel, Pilar Casado, Juan. A. Hernández-Tamames and Juan ÁlvarezLinera 2008 Brain activation in discourse comprehension: A 3t fMRI study. NeuroImage 41: 614–622. Maury, Pascale and Amélie Teisserenc 2005 The role of connectives in science text comprehension and memory. Language and Cognitive Processes 20: 489–512. McDonald, Janet L. and Brian James MacWhinney 1995 The time course of anaphor resolution: Effects of implicit verb causality and gender. Journal of Memory and Language 34: 543–566.

The processing of pragmatic information in discourse

229

McNamara, Danielle S. and Walter Kintsch 1996 Learning from texts: Effects of prior knowledge and text coherence. Discourse Processes 22: 247–288. McNamara, Danielle S., Eileen Kintsch, Nancy Butler Songer and Walter Kintsch 1996 Are Good Texts Always Better? Interactions of Text Coherence, Background knowledge, and levels of understanding in learning from text. Cognition and Instruction 14: 1–43. Meyer, Bonnie J. F. 1975 The Organization of Prose and its Effects on Memory. Amsterdam: North-Holland. Meyer, Bonnie J. F., David M. Brandt and George J. Bluth 1980 Use of top-level structure in text: Key for reading comprehension in ninthgrade students. Reading Research Quarterly 16: 72–103. Meyer, Bonnie J. F., Carole J. Young and Brendan J. Bartlett 1989 Memory Improved: Enhanced Reading Comprehension and Memory Across the Life Span Through Strategic Text Structure. Hillsdale, NJ: Lawrence Erlbaum. Millis, Keith K. and Marcel A. Just 1994 The influence of connectives on sentence comprehension. Journal of Memory and Language 33: 128–147. Millis, Keith K., Jonathan M. Golding and Gregory Barker 1995 Causal connectives increase inference generation. Discourse Processes 20: 29–49. Millis, Keith K., Arthur C. Graesser and Karl Haberlandt 1993 The impact of connectives on memory for expository texts. Applied Cognitive Psychology 7: 317–340. Mulder, Gerben 2008 Understanding Causal Coherence Relations. Universiteit Utrecht: LOT Murray, John D. and Chadd. McGlone 1997 Topic overviews and the processing of topic structure. Journal of Educational Psychology 89: 251–261. Myers, Jerome L., Makiko Shinjo and Susan A. Duffy 1987 Degree of Causal Relatedness and Memory. Journal of Memory and Language 26: 453–465. Noordman, Leo G. M. and Femke De Blijzer 2000 On the processing of causal relations. In: Elisabeth Couper-Kühlen and Bernd Kortmann (eds.), Cause, Condition, Concession and Contrast: Cognitive and Discourse Perspectives, 35–56. New York: Mouton de Gruyter. Noordman, Leo G. M. and Wietske Vonk 1997 The different functions of a conjunction in constructing a representation of the discourse. In: Jean Costermans and Michel Fayol (eds.), Processing Interclausal Relationships. Studies in the Production and Comprehension of Text, 75–93. New Jersey: Lawrence Erlbaum Associates. Noordman, Leo G. M., Wietske Vonk and Henk J. Kempff 1992 Causal inferences during the reading of expository texts. Journal of Memory and Language 31: 573–590

230

Ted J.M. Sanders and Anneloes R. Canestrelli

Ochs, Elinor (ed.) 1979 Planned and Unplanned Discourse. (Syntax and Semantics XII: Discourse and syntax.) New York: Academic Press. Pander Maat, Henk and Liesbeth Degand 2001 Scaling causal relations and connectives in terms of speaker involvement. Cognitive Linguistics 12: 211–245. Pander Maat, Henk and Ted Sanders 2001 Subjectivity in causal connectives: An empirical study of language in use. Cognitive Linguistics 12: 247–273. Pander Maat, Henk and Ted Sanders 2009 Local and global determinants of forward prominence of discourse referents: two corpus studies. Linguistics 47: 1273–1319. Sanchez, Rebecca P., Elizabeth P. Lorch and Robert F. Lorch 2001 Effects of headings on text processing strategies. Contemporary Educational Psychology 26: 418–428. Sanders, Ted and Morton A. Gernsbacher 2004 Accessibility in discourse and text processing. Discourse Processes 37: 79–89. Sanders, Ted J. M. and Leo G. M. Noordman 2000 The role of coherence relations and their linguistic markers in text processing. Discourse Processes 29: 37–60. Sanders, Ted and Henk Pander Maat 2006 Cohesion and Coherence: Linguistic approaches. In: Keith Brown (ed.), Encyclopedia of Language and Linguistics, Volume 2, 591–595. London: Elsevier. Sanders, Ted J. M. and Wilbert Spooren 2007 Discourse and text structure. In: Dirk Geeraerts and Hubert Cuykens (eds.), Handbook of Cognitive Linguistics, 37–60. Oxford: Oxford University Press. Sanders, Ted and Wilbert Spooren 2009 Causal categories in discourse – Converging evidence from language use. In: Ted Sanders and Eve Sweetser (eds.), Causal Categories in Discourse and Cognition, 205–246. Berlin: Mouton de Gruyter. Sanders, Ted and Eve Sweetser (eds.) 2009 Causal Categories in Discourse and Cognition. Berlin: Mouton De Gruyter Sanders, Ted, José Sanders and Eve Sweetser 2009 Causality, cognition and communication: a mental space analysis of subjectivity in causal connectives. In: Ted Sanders and Eve Sweetser (eds.), Causal Categories in Discourse and Cognition, 19–61. Berlin: Mouton de Gruyter. Sanders, Ted, Wilbert Spooren and Leo Noordman 1992 Toward a taxonomy of coherence relations. Discourse Processes 15: 1–35. Sanford, Anthony J. and Simon C. Garrod 1994 Selective processes in text understanding. In: Morton A. Gernsbacher (ed.), Handbook of Psycholinguistics, 699–720. New York: Academic Press. Segal, Erwin M., Judith F. Duchan and Paula J. Scott 1991 The role of interclausal connectives in narrative structuring: Evidence from adults’ interpretations of simple stories. Discourse processes 14: 27–54. Speelman, Dirk and Dirk Geeraerts 2009 The role of concept characteristics in lexical dialectometry. International Journal of Humanities and Arts Computing 2: 221–242.

The processing of pragmatic information in discourse

231

Spooren, Wilbert, Ted Sanders, Mike Huiskes and Liesbeth Degand 2010 Subjectivity and causality: A corpus study of spoken language. In: Sally Rice and John Newman (eds.), Empirical and Experimental Methods in Cognitive/ Functional Research, 241–255. Chicago: CSLI/University of Chicago Press. Spyridakis, Jan H. and Timothy C. Standal 1987 Signals in expository prose. Reading Research Quarterly 22: 285–298. Stukker, Ninke, Ted Sanders and Arie Verhagen 2008 Causality in verbs and in discourse connectives. Converging evidence of crosslevel parallels in Dutch linguistic categorization. Journal of Pragmatics 40: 1296–1322. St George, Marie, Suzanne Mannes and James E. Hoffman 1994 Individual differences in inference generation: An ERP analysis. Journal of Cognitive Neuroscience 9: 776–787. St George, Marie, Marta Kutas, Antigona Martinez and Marty I. Sereno 1999 Semantic integration in reading: engagement of the right hemisphere during discourse processing. Brain 122: 1317–1325. Taboada, Maite and William C. Mann 2006 Applications of Rhetorical Structure Theory. Discourse Studies 8: 567–588. Tomlin, Russell S. 1997 Mapping conceptual representations into linguistic representations: The role of attention in grammar. In: Jan Nuyts and Eric Pederson (eds.), Language and Conceptualization, 162–189. Cambridge: Cambridge University Press. Trabasso, Tom, and Paul Van den Broek 1985 Causal thinking and the representation of narrative events. Journal of Memory and Language 24: 612–630. Traxler, Matthew J., Micheal D. Bybee and Martin J. Pickering 1997 Influence of connectives on language comprehension: Eye-tracking evidence for incremental interpretation. The Quarterly Journal of Experimental Psychology 50: 481–497. Verhagen, Arie 2005 Constructions of Intersubjectivity. Discourse, Syntax, and Cognition. Oxford: Oxford University Press. Virtanen, Tuija 1992 Discourse functions of adverbial placement in English. Åbo: Åbo Akademi University Press. Vonk, Wietske, Lettica G. M. M. Hustinx and Wim H. G. Simons 1992 The use of referential expressions in structuring discourse. Language and Cognitive Processes 7: 301–333. Walker, Marilyn, Aravind Joshi and Ellen Prince (eds.) 1998 Centering Theory in Discourse. Oxford: Clarendon Press. Zwaan, Rolf 1996 Processing narrative time shifts. Journal of Experimental Psychology: Learning, Memory, and Cognition 22: 1196–1207.

9.

Happy New War1 – The role of salient meanings and salience-based interpretations in processing utterances Rachel Giora

Whether our cognitive machinery is adept at swiftly and accurately homing in on a single, contextually appropriate interpretation, as assumed by The Direct Access View (Gibbs 1986, 1994) and Relevance Theory (Sperber and Wilson 1986/1995), or whether it is less efficient at sieving out interpretations based on salient (coded and prominent) word and phrase meanings which might be activated irrespective of contextual information and speakers’ intent, as proposed by The Graded Salience Hypothesis (Giora 1997, 1999, 2003) is an unresolved issue within pragmatics and psycholinguistics. Within the field of literal and non-literal language, this translates into whether accessible but incompatible meanings and messagelevel interpretations are involved even when contextual information is highly supportive of an alternative interpretation. More specifically, the debate concerning the processing, for instance, of ironic, metaphoric, optimally innovative as well as negated utterances revolves around the role of salient meanings and salience-based yet incompatible interpretations in shaping contextually compatible but non-salient interpretations in contexts strongly benefitting such interpretations. This chapter reviews state-of-the-art research involved in this debate, looking into the role of salient but incompatible meanings and salience-based but incompatible interpretations in verbal and non-verbal communication.

1.

Salient but incompatible meanings

What makes a response or a meaning salient? According to The Graded Salience Hypothesis (Giora 1997, 1999, 2003), a response to a stimulus, be it verbal (a word or a phrase) or non-verbal (perceptual), is salient if it is coded in the mental lexicon/repository2 and foremost on our mind due to factors such as extensive and repetitive exposure (e.g. frequency of occurrence, experiential familiarity, conventionality), cognitive priority (e.g. prototypicality, animacy), or (undisclosed) preoccupation (e.g. sex); it is less salient if it is coded but low on these variables; it is non-salient if it is not coded (e.g. novel or derived. On tests for degree of salience, see Giora 2003: chapter 2). A coded response will always be accessed when the relevant stimulus is encountered, irrespective of context; salient responses cannot, then, be preempted. Rather, access is exhaustive (as suggested by Fodor 1983) yet ordered: Salient responses are activated faster than less-salient ones (see also

234

Rachel Giora

Duffy, Morris, and Rayner 1988; Rayner, Pacht, and Duffy 1994). If outputs of response-access and contextual processes accidentally match, speedy integration processes are anticipated; if, however, they accidentally mismatch, integration processes are expected to be slower. The involvement of irrelevant responses in interpretation processes is, thus, a consequence of a modular mechanism – an automatic and encapsulated process whose course of action is blind to information outside the module (as suggested by Fodor 1983). Information outside the module (e.g. outside the mental lexicon) – such as contextual information – may also affect interpretation at a very early stage. If it is strong and supportive of a specific meaning or interpretation, it can activate this meaning via predictive mechanisms, especially if this information appears in segment (e.g. sentence) final position, which allows information to accumulate. Such contextual mechanisms can, nevertheless, not block access of salient meanings (Peleg, Giora, and Fein 2001). To illustrate the unconditional activation of salient responses, consider the title of this chapter – Happy New War – ironically praising the Israeli war on Gaza, which took place in January 2009. Although its meaning is literal and, literally, spelled out, one cannot escape accessing another collocation (Happy New Year), which instantly springs to mind, on account of its familiarity, conventionality, and frequency. Here, the salient meaning of the familiar but implicit collocation plays an important role in shaping the full interpretation of the novel collocation. Or consider the infamous caricature, by Fathi, of Benjamin Netanyahu, the current Israeli Prime Minister, originally published in Ahram Weekly, which (for those familiar with the relevant information) must bring to mind both Pinocchio, the ultimate liar, who, as a result of his lies, grew a long nose and donkey ears,3 and the anti-Semitic stereotypical image of the long-nosed deceitful Jew: (1)

Figure 1: Caricature of Israeli Prime Minister Benjamin Nethanyahu (http://www.haaretz.com/hasen/spages/1116468.html)

Although these are random and somewhat complex examples, evidence supporting the claim that salient meanings cannot be blocked comes from a great number of empirical studies (for extensive reviews see Giora 1997, 2003: chapter 3;

Happy New War

235

Peleg, Giora, and Fein 2001, 2004, 2008). The following sections, however, focus on more recent evidence from studies of ambiguity (1.1.) and polysemy (1.2.) resolution. 1.1

Ambiguity resolution

A stimulus (word, collocation) is considered ambiguous if the various responses (meanings) it evokes are conceptually unrelated to each other (bank, tear). Under such circumstances, it is vital to resolve the ambiguity as quickly as possible and home in on the contextually appropriate meaning (Frazier and Rayner 1990). There is, however, ample evidence which demonstrates that salient meanings of ambiguous words and fixed expressions cannot be blocked but, instead, are always activated irrespective of context bias (Giora 2003: chapter 3). Furthermore, it has recently become clear that they are also always activated even irrespective of type of homograph or visual field as shown by Peleg and Eviatar (2008, 2009, 2012). Peleg and Eviatar’s recent studies examine not only homophonic homographs (e.g. bank) – the prevailing ambiguity in most of the (English based) literature – but also heterophonic homographs (e.g. tear), which have been largely ignored. Additionally, they take into account type of context (neutral, strongly biased toward the salient meaning, strongly biased towards the less-salient meaning) and early (150 ms, 250 ms) and late (1000 ms) SOAs (stimulus onset asynchrony). This section, however, focuses on the early stages of comprehension only (150 ms – 250 ms) in order to show that activation of salient meanings cannot be preempted even by contexts strongly supportive of the alternative, less-salient meaning. Although support for The Graded Salience Hypothesis should come from contexts biased toward the less-salient meaning, we start by considering, as a reference point, the activation pattern of salient meanings in contexts biased in favour of the salient meaning. Findings in Peleg and Eviatar’s studies show that, in such contexts, salient meanings are always activated and always exclusively regardless of visual field, type of homograph or SOA. Compared to contexts biased toward the salient meaning, findings in Peleg and Eviatar further show that effects of contexts biased toward the less-salient meaning (‘river’, ‘rip’) of the homograph (bank, tear) are sensitive to type of homograph. For instance, in the case of homophonic homographs (bank), context bias has no effect on lexical access; where orthography and phonology are unambiguously related (collaborating to intensify degree of salience), salient meanings (‘money’) are immediately (150 ms SOA) activated in both the left (LH) and the right (RH) hemisphere. Similarly, in the case of heterophonic homographs (tear), where orthography and phonology are ambiguously related, salient meanings (‘eye’) are also activated very early on. Their activation, however, takes place a bit sooner (150 ms SOA) in the RH than in the LH (250 ms SOA).

236

Rachel Giora

Recent findings thus replicate previous results showing that salient meanings of homographs, both contextually appropriate and inappropriate, are always activated, regardless of contextual bias. They further show that salient meanings, both contextually appropriate and inappropriate, cannot be blocked, regardless also of type of homograph. 1.2.

Polysemy resolution

A stimulus (word, collocation) is considered polysemous if the various responses (meanings) it evokes are conceptually related to each other. Under such circumstance, it is not always necessary to resolve the ambiguity by blocking the inappropriate sense, because, most often, this sense does not interfere with the process of homing in on the contextually appropriate meaning and may even be instrumental in shaping it (Giora 2003). In this respect, metaphors, ironies, and optimal innovations are paradigmatic (Giora et al. 2004). Williams (1992), for instance, showed that salient, contextually compatible and incompatible meanings (‘strict’; ‘solid’) of familiar metaphors (The schoolteacher was criticized for not being firm) and their literal counterparts (The couple wanted a bed that was firm) were facilitated initially in both the metaphoric and the literal contexts. Selecting the appropriate meaning (‘solid’) occurred later on and only in the context (firm bed) where the alternative, metaphorical (‘strict’) meaning was not conducive to the appropriate meaning. Findings that salient but contextually incompatible meanings of metaphors become activated and cannot be blocked even when context strongly favors the metaphoric meaning have gained more recent support. For instance, in Peleg, Giora, and Fein (2001), responses to a lexical decision task show that a strong context supportive of the metaphorical interpretation (Sarit’s sons and mine went on fighting continuously. Sarit said to me: These delinquents won’t let us have a moment of peace) did not inhibit salient but contextually incompatible meanings (‘criminals’) of targets (delinquents) which were as accessible as contextually compatible meanings (‘kids’). This was true even when targets were placed in sentence final position (which benefits contextual mechanisms) and probed immediately afterwards (Sarit’s sons and mine went on fighting continuously. Sarit said to me: A moment of peace won’t let us have these delinquents). For similar results, see also Ferretti, Patterson, and Katz 2006; Rubio Fernández 2007; for counterevidence, see Ferretti, Schwint, and Katz 2007. Similarly, in Hasson and Glucksberg (2006), contextually inappropriate metaphoric meanings (‘fast’) of negated conventional metaphors (The train to Boston was no rocket) were accessed initially in spite of a contextual cue (negation) to the contrary. Consequently, when probed later on, they facilitated targets embedded in a coherent discourse (Giora, Fein, Aschkenazi and Alkabets-Zlozover 2007).

Happy New War

237

Even literal meanings of highly conventional idioms (kick the bucket) were shown to play a role in shaping their idiomatic meaning and in how they can be used (Gibbs 1994; Gibbs, Nayak, and Cutting 1989; Hamblin and Gibbs 1999). Recent findings, thus, replicate previous results showing that salient meanings of multiple-meaning stimuli, both contextually appropriate and inappropriate, are always activated, regardless of type of context and type of relations (ambiguity/ polysemy) obtaining between a stimulus’ multiple meanings. Are contextually incompatible meanings discarded so that they do not interfere with the appropriate interpretation (Gernsbacher 1990) or are they retained in case they can play a role in affecting the final message-level interpretation, in spite of their inappropriateness? Clearly, our cognitive machinery is not adept at swiftly and accurately pinpointing a single, contextually appropriate meaning. Can this “disadvantage” still advantage some utterance-level interpretations?

2.

Salience-based interpretations

Do salient but incompatible meanings play a role in utterance interpretation? In general, salient meanings are the building blocks of salience-based interpretations: Salience-based interpretations are message-level interpretations, derived on the basis of the salient meanings of the stimulus components. When salient but incompatible meanings are involved (as shown in section 1), they may result in saliencebased but incompatible interpretations (e.g. the literal interpretations of metaphors embedded in a metaphor-inviting context). But will salience-based but incompatible interpretations be also represented in utterance end-product interpretation? For instance, will the salience-based literal interpretation of ironies, whose nonironic interpretation is literal (This one’s really sharp, said of a dull pair of scissors), or the salience-based metaphoric interpretation of ironies whose non-ironic interpretation is metaphoric (This one’s really sharp, said of a dull student; see Colston and Gibbs 2002) be instrumental in deriving the appropriate (ironic) interpretation and form part of the end-product interpretation? Admittedly, there is no reason to assume that a stimulus’ contextually inappropriate meanings should not be suppressed so as to effect a smooth integration of end-product interpretations with contextual information (Fodor 1983; Gernsbacher 1990). But is such a suppression mechanism automatic, deactivating any contextually incompatible meanings or interpretations, or is it selective – sensitive to both contextual cues as well as the potential role of such incompatible meanings and interpretations in utterance’s processing and representation, despite their contextual misfit? Whether salience-based contextually inappropriate interpretations are involved in utterance interpretation or whether they are blocked or discarded can be resolved empirically. The various models of language comprehension and interpretation

238

Rachel Giora

have different predictions about the involvement (or lack of it) of salience-based but incompatible interpretations in a stimulus representation. 2.1.

The Direct Access View

According to The Direct Access View (Gibbs 1986, 1994, 2005; Sperber and Wilson 1986/1995), our cognitive machinery can swiftly and accurately home in on a single, contextually appropriate interpretation: if context is strongly supportive of the appropriate reading, end product interpretations will not involve contextually inappropriate interpretations. This should be true even if lexical processes do not obstruct access of salient but contextually incompatible meanings (Gibbs 1983, 1984, 1994, 2002, 2005). According to this view, rich context should activate the contextually appropriate interpretation exclusively so that only that interpretation becomes available for further processes. Thus, in a strong context, the question as to whether salience-based contextually inappropriate interpretations are involved in utterance interpretation is irrelevant. Under such circumstances, inappropriate, salience-based outputs should not be fleshed out at all; instead, the appropriate interpretation must be derived exclusively, affecting a seamless integration with contextual information. 2.2.

The Standard Pragmatic Model

The Direct Access View argues against the traditional serial model – The Standard Pragmatic Model (Grice 1975; Searle 1979) – according to which understanding language involves activating the literal interpretation of the utterance first. In case that interpretation fails to meet contextual fit, an additional stage of monitoring and adjustment will follow, affecting suppression and replacement of the literal interpretation by appropriate outputs. Such a model thus predicts that arriving at the utterance-appropriate interpretation will not be seamless; however, it will result in an exclusive appropriate interpretation. On both, The Direct Access View and The Standard Pragmatic Model, then, utterance final output contains the contextually appropriate interpretation only. Whereas on The Direct Access View inappropriate interpretations are not expected to be derived, on The Standard Pragmatic Model they are, but are also expected to undergo suppression. In spite of their differences concerning early context effects, both views predict the same end-product interpretations, free of inappropriate message-level interpretations. 2.3.

The Graded Salience Hypothesis

Unlike The Direct Access View and The Standard Pragmatic Model, The Graded Salience Hypothesis has different predictions. According to The Graded Salience

Happy New War

239

Hypothesis (Giora 2003, 2006; Giora and Fein 1999), activated meanings are not discarded unconditionally (The Suppression/Retention Hypothesis). They are discarded if they interfere with the interpretation processes, as found for contextually inappropriate (but not for appropriate) meanings of ambiguities, which lost initial levels of activation between 250 – 1000 ms SOA (see Peleg and Eviatar 2007, 2008, 2012); they are retained if they are deemed undisruptive or conducive to the final interpretation, as found for inappropriate senses of polysemies such as the literal meanings of metaphors, which did not lose initial levels of activations even at 1000 ms ISI (see Rubio Fernández 2007; Williams 1992; see also Brisard, Frisson and Sandra 2001; Giora and Fein 1999). Given that salient meanings are accessed initially and need not be suppressed (Peleg and Eviatar 2008, 2009, 2012), deriving salience-based interpretations must occur quite early on. Their activation must be speedy because it relies on the salient meanings of the utterance’s components. Salience-based interpretations may therefore either accidentally match contextual information or result in contextual misfit. Regardless of this, according to The Graded Salience and The Suppression/Retention Hypotheses, even when salience-based interpretations are contextually inappropriate, they may still play a role in shaping the utterance’s final interpretation (as shown for polysemous utterances, such as metaphors and ironies, see Giora 2003). Indeed, a contextual mismatch need not be a disadvantage; even saliencebased contextually inappropriate interpretations may be instrumental in shaping final interpretations and effect humor or pleasure (Giora et al. 2004; Giora, Fein, Laadan et al. 2007). Based on The Graded Salience Hypothesis (Giora 2003, 2006; Giora and Fein 1999), salience-based interpretations should not be sensitive to degree of literality. Given that they are derived on the basis of the salient meanings of the utterance components, they may be either literal or non-literal, as found for the ironic This one’s really sharp, whose salience-based interpretation is metaphorical (see Colston and Gibbs 2002). To sum up, the Graded Salience Hypothesis predicts that salience-based interpretations will be derived swiftly, regardless of context or degree of literality. They will be retained, even when inappropriate, if deemed conducive to the final interpretation of the utterance. Diverging from The Direct Access View and The Standard Pragmatic Model, then, The Graded Salience Hypothesis predicts that final message-level outputs may involve contextually incompatible interpretations even in strong contexts favoring the appropriate interpretation.

240 3.

Rachel Giora

Message-level outputs involve contextually incompatible interpretations – Empirical findings

Reviewing the predictions of the various models (section 2) reveals that whereas The Graded Salience Hypothesis predicts possible involvement of contextually incompatible interpretations in an utterance end-product interpretation, The Direct Access View and The Standard Pragmatic Model anticipate no such involvement. In this section, I will review empirical findings in order to test the various models’ predictions (section 3.2). First, however, I review the measures that can be used to tease apart the various views. 3.1.

Measures suitable to test suppression and retention of salience-based, though incompatible, interpretations

3.1.1.

Reading times

Reading times of whole utterances may disclose something about their interpretation processes. If contextually inappropriate interpretations comprise a part of the final output, then reading times of these utterances should take longer than when these inappropriate interpretations are not involved in the process. For instance, in Brisard, Frisson and Sandra (2001) and Giora and Fein (1999), novel metaphors took longer to read when the context was biased toward the metaphorical than toward the literal (salience-based) interpretation, suggesting that deriving non-salient interpretations involved activating the salience-based interpretation. In Meytes and Tamir (2005; see Giora 2006), negative idioms (I don’t know my right from left) took less time to read than their affirmative counterparts (I know my right from left), even though the latter are shorter. The extra processing time incurred by the less familiar version might be explained by the need to revisit the idiomatic interpretation. It is not, however, the case that salient non-literal meanings (I don’t know my right from left) should always interfere with utterance salience-based (e.g. literal) interpretation; at times they may fit in the representation of that interpretation. For instance, in Ferretti, Patterson and Katz (2006), although literal (salience-based) interpretations of familiar proverbs (Lightning never strikes the same place twice) and their salient proverbial meanings took equally long to read, literal (saliencebased) interpretations, invited by a literally biasing contexts (a storm), were easier to incorporate into the text representation than figurative meanings embedded in figuratively biasing contexts (financial market), as testified by ERP measures. Indeed, in this literal condition, both the literal and the non-literal interpretations could be both true and contextually compatible, which was not the case when the figurative context was considered. Even if reading times of whole utterances may not be sensitive enough to disclose differences or difficulties, these difficulties may surface later on and slow

Happy New War

241

down reading times at the space that follows the utterance or for the first words of the next utterance. For instance, in Pexman, Ferretti and Katz (2000), although supportive contexts speeded reading times at the last non-literal word of novel metaphorical and ironic utterances (relative to a neutral context), reading times at the space following these utterances and for the first word of the next utterance were longer, with longer reading times in the irony than in the metaphor condition. Such “spill-over effects” demonstrate difficulties at these locations, originating in the previous utterance. 3.1.2.

Lexical decisions

Although reading times may occasionally indicate processing difficulties, it is also reasonable to suspect that the observed difficulties do not necessarily reflect involvement of inappropriate interpretations in utterances’ representation. After all, these longer reading times could also be explained by novelty of the items. Additionally, reading times cannot tease apart interference of lexical meanings from that of salience-based interpretations. In contrast, lexical decisions to probes related to the appropriate and inappropriate message-level interpretations of an utterance can be highly revealing in so far as (both early and) late processes are concerned. For instance, I see that you picked the ideal meal today embedded in an irony-inviting context induced faster response times to a probe-word related to its salience-based (often literal) interpretation (‘healthy’) compared to the (equally familiar/salient) contextually appropriate probe-word (‘harmful’) related to the non-salient ironic interpretation of the utterance. Such patterns of results were obtained even when probes were presented 1000 ms following offset of the ironic utterance, thus allowing lengthy processing time (Giora, Fein, Laadan et al. 2007; see section 3.2.1.). Lexical decisions to such probes occurring at a late processing stage, at interstimulus intervals (ISIs) of 750 to 1000 ms, can thus disclose underlying processes involved in utterance end-product interpretation. 3.1.3.

Resonance

There are, in addition, other ways to index the accessibility of inappropriate message-level interpretations by looking at the manner in which natural discourse unfolds. For instance, measuring the amount of resonance obtaining between neighboring utterances and the way their compatible and incompatible interpretations shape the ongoing discourse may be one such way. “Resonance” pertains to the “activation of affinities across utterances”, whether they are the speaker’s or her interlocutors’, resulting in an environment that aligns with utterances’ interpretations (Du Bois 2004). Looking at utterances’ contexts can be thus revealing about which interpretations are entertained by the speaker, whether resonating with

242

Rachel Giora

herself or with her interlocutors. For instance, in (2), the ironic “splendid job of our fine pilots” resonates with its salience-based albeit inappropriate literal interpretation (Hooray to the Israeli Airforce pilots doing a splendid job) appearing earlier in the context. In (3), the Figure 2 – KNOW HOPE – resonates with the Ground featuring its salience-based incompatible opposite – ‘no hope’. In (4), the ironic Welcoming the liberators at Basrah resonates with the familiar – Manneken Pis – the Pissing little boy in Brussels (5); whereas the latter conveys a salience-based interpretation of joy, innocence, and pride, the former gets across an additional message of contempt. (2) Hooray to the Israeli Airforce pilots doing a splendid job effused Brigadier General Avi Benayahu, the IDF spokesperson, talking to Yonit Levy – white turtleneck against a background of tanks, vis à vis hundreds of funerals in Gaza – a token of the splendid job of our fine pilots (Levy 2008). Such instances of resonance obtaining between messages are consistent with the view that salience-based interpretations are activated and retained even when contextually incompatible. (3)

Figure 2.

KNOW HOPE (http://www.flickr.com: 80/photos/idanska/247228762/)

Happy New War

(4)

Figure 3. Welcoming the liberators at Basrah (http: www.oocities.org/telavivdiary/diaryapr7.html)

(5)

Figure 4. Manneken Pis (http://www.gadling.com/2007/03/28/ manneken-pis-brussles-pissing-boy/)

243

244

Rachel Giora

3.2.

Empirical findings

3.2.1.

Irony

Interpreting ironic utterances in highly supportive contexts can be an ideal testcase for the involvement (or lack of it) of salient but incompatible meanings and salience-based but incompatible interpretations in utterance processing. According to The Graded Salience Hypothesis and The Suppression/Retention Hypothesis (Giora 2003; Giora and Fein 1999), salient but incompatible meanings and salience-based but incompatible interpretations which are conducive to the appropriate interpretation of the utterance will make up part of the final representation of the utterance. Whereas the final ironic output is heavily dependent on contextual information, the incompatible salient meanings and salience-based interpretations rely on the salient meanings of the components that make up the ironic statement. Recent research reveals that both incompatible meanings as well as incompatible interpretations are represented in the products of irony processing. Salient but incompatible meanings Previous online research has demonstrated that the products of irony interpretation included salient but incompatible meanings. Such meanings were shown to be activated initially and to remain accessible even at the late stages of the interpretation process, suggesting that such meanings did not undergo suppression. For instance, in Giora and Fein (1999), allowing comprehenders another 1,000 ms processing time following offset of target utterances revealed that both familiar and unfamiliar ironies retained rather than suppressed incompatible meanings activated initially on account of their salience. In Giora, Fein and Schwartz (1998), even when probed 2,000 ms following offset of the ironic statements, such salient meanings were still as active. Such results argue against the replacement hypothesis of The Standard Pragmatic Model (Grice 1975) and The Modular View (Swinney 1979). Salience-based but incompatible interpretations To argue against both The Direct Access View and Relevance Theory, as well as against The Standard Pragmatic Model, it is necessary to show that salience-based but incompatible message-level interpretations make up part of the products of irony when targets are embedded in a strong context, inducing an expectation for an ironic utterance (Gibbs 1986, 2002). In our more recent studies we have attempted to show this. In these experiments we included contexts raising an expectation for an ironic target and measured both reading times of literal and ironic utterances and response times to probes related to the utterances’ interpretation rather than to the lexical meanings of their key words.

Happy New War

245

Findings in Giora, Fein, Laadan et al. (2007), replicated also in Giora, Fein, Kaufman et al. (2009), showed that utterances took longer to read when embedded in dialogues shown to induce an expectation for an ironic utterance than when embedded in very similar dialogues inviting, however, a salience-based interpretation of the same target utterances.4 For instance, reading times of target statements such as Sounds like you are going to have a really interesting evening were longer following contexts biased toward the ironic interpretation (6) than toward the salience-based (often literal) interpretation (7): (6) Barak: I finish work early today. Sagit: So, do you want to go to the movies? Barak: I don’t really feel like seeing a movie. Sagit: So maybe we could go dancing? Barak: No, at the end of the night my feet will hurt and I’ll be tired. Sagit: You’re a really active guy … Barak: Sorry, but I had a rough week. Sagit: So what are you going to do tonight? Barak: I think I’ll stay home, read a magazine, and go to bed early. Sagit: Sounds like you are going to have a really interesting evening. Barak: So we’ll talk sometime this week. (7) Barak: I was invited to a film and a lecture by Amos Gitai. Sagit: That’s fun. He is my favorite director. Barak: I know, I thought we’ll go together. Sagit: Great. When is it on? Barak: Tomorrow. We will have to be in Metulla1 in the afternoon. Sagit: I see they found a place that is really close to the center. Barak: I want to leave early in the morning. Do you want to come? Sagit: I can’t, I’m studying in the morning. Barak: Well, I’m going anyway. Sagit: Sounds like you are going to have a really interesting evening. Barak: So we’ll talk sometime this week. Additionally, in Giora, Fein, Laadan et al. (2007), contextual expectation for an ironic utterance was induced via presenting participants only contexts which all ended in an ironic utterance (8). These were compared to a condition in which only half of the contexts ended in an ironic utterance; the other half ended in a nonironic utterance (9). Using lexical decisions to probes related to utterance-level interpretation showed facilitation only of the salience-based but incompatible interpretation; non-salient albeit compatible interpretations, namely ironically related probes, were not facilitated even when 1000 ms extra processing time was allowed. This pattern of results was not different from that obtained in the control design, which did not induce expectation for an ironic interpretation. Such results suggest

246

Rachel Giora

that a strong context, inducing expectation for an ironic interpretation, does not affect a seamless interpretation process, as predicted by The Direct Access View. Overall, these results argue against the assumption that supportive contextual information can effect contextually appropriate interpretation exclusively, even when non-salient (see Gibbs 2002; Grice 1975). (8) Yuval and Omry went out for their lunch break after a morning of work. They went to the cafeteria in their office building and each filled a platter with food. They stood in line for a long while and were eager to start the meal. When they had sat down, Yuval saw that his colleague chose fried sausage, chips, a glass of coke for a drink, and a sugar-glazed doughnut for dessert. Then Yuval said: I see that you picked the ideal meal today!” (9) Yuval and Omry went out for their lunch break after a morning of work. They went to the cafeteria in their office building and each filled a platter with food. They stood in line for a long while and were eager to start the meal. When they had sat down to eat, Yuval saw that his colleague filled his platter with salad, tofu, and sprouts and chose natural carrot juice for a drink. Then Yuval said: I see that you picked the ideal meal today!” Probes: Salience-based related – healthy; ironically related – harmful Importantly, strengthening the ironically biasing design used in Giora, Fein, Laadan et al. (2007) by introducing an additional constraint did not affect the patterns of results obtained earlier. In Giora (2011), in addition to the biasing design which induced an expectation for an ironic statement via repetitive use of such statements, the expectation for ironic utterances was strengthened by informing participants that the aim of the experiment was to test irony interpretation. The control group, whose experimental design was mixed, raising no expectation, were not informed about this specific aim of the experiment; their contextual information was thus weaker compared to that in which expectation for ironic utterances was made obvious both implicitly and explicitly. Still, in spite of strongly biasing contextual information in favor of the ironic interpretation, the pattern of results obtained earlier did not change. Even this multiple constraints condition did not facilitate irony interpretation, not even when 1000 ms processing time was allowed. In contrast, only incompatible salience-based interpretations were made available in both conditions. This pattern of results obtained even when longer processing time was allowed (1500, 2000 and 3000 ms). Under all these conditions, the salience-based literal interpretation was never less accessible than the ironic interpretation, despite its incompatibility. Such results demonstrate that understanding utterances in a context strongly biasing their interpretation toward a non-salient but appropriate (ironic) interpretation does not unconditionally involve dispensing with the salience-based but in-

Happy New War

247

compatible interpretation. The involvement of salience-based but incompatible interpretations in irony interpretation supports The Graded Salience Hypothesis but argues against both The Standard Pragmatic Model and The Direct Access View. Corpus-based studies Can lab results showing that salience-based but incompatible interpretations feature in irony final outputs be supported by corpus-based studies looking into the environment of natural (spoken and written) uses? Indeed, if the natural environment of irony discloses salience-based but incompatible interpretations, this will support the view that the online processes involved in irony interpretation can be mirrored in irony production. In Giora et al. (forthcoming), we examined (Hebrew) newspaper editorials and op-ed articles (taken from Haaretz, Ynet, NRG, Walla, The Seventh Eye, and Mouse) written by well-known ironists during 2008 and 2009. This corpus comprises 70 347 words and includes 1597 ironic utterances (15 466 words). To find out which interpretations help shape the environment of irony, we looked at how the context of irony resonates with its various compatible and incompatible interpretations. Specifically, we looked at (i) whether the environment did not resonate with any of the interpretations of the ironic utterance; (ii) whether it resonated with both its salience-based and ironic interpretations; (iii) whether it included extended ironies – namely, ironies extending their salience-based interpretation and consequently their ironic interpretation as well, thus creating an environment which resonates with both the compatible and incompatible interpretations of the utterance; (iv) whether it resonated with the ironic interpretation only (10a–b), (v) or whether it resonated with its salience-based interpretation only (11a–b). The latter two are exemplified below (the ironic key word/phrase – in bold and the way it is resonated with – in italics). They count as the data which we analyzed: (10) Environment manifesting ironic resonance (a) A modest studio of less than 140 square meters with two enormous rooms outfitted like the most luxurious hotel (Kashua 2008). (b) The man [Olmert] who made a number of courageous statements about peace late in his tenure has orchestrated no fewer than two wars. Talking peace and making war, the “moderate” and “enlightened” prime minister [Olmert] has been revealed as one of our greatest fomenters of war (Levy January 2009). (11) Environment manifesting salience-based resonance (a) All this really could have been peachy if not for the fact that blindness is dangerous and the not-so-good ending is known in advance (Levy July 2009).5

248

Rachel Giora

(b) Most of the tycoons that have borrowed, in recent years, tens of billions of NIS, in bonds, from the public, find it hard to adjust to the harsh conditions in the First Class of commercial airlines (Rolnik 2008).6 Results show that most of the ironies (43 %) did not affect their environment in that they were not reflected in their contexts; very few (3 %) were addressed by relating to both their ironic and salience-based interpretations and some (10 %) were extended ironies. However, a great number of ironies (36 %) resonated with their salience-based interpretations only; significantly less so (8 %) were echoed via their ironic interpretations only. Such results are more consistent with The Graded Salience Hypothesis than with The Direct Access View and The Standard Pragmatic Model. They show that, as predicted by The Graded Salience Hypothesis, saliencebased albeit incompatible interpretations are retained in irony production, allowing speakers to resonate with (their own) contextually inappropriate interpretations on account of their relative accessibility. In a similar manner, corpus-based studies of spoken interaction also afford evidence supporting the involvement of salience-based but incompatible interpretations in processing ironic utterances. For instance, in Giora and Gur (2003), 75 % of the conversational ironies exchanged among friends were responded to by reference to their salience-based but incompatible interpretations. In Kotthoff (2003), dinner-table conversations among friends abounded in responses to “what is said” – the salience-based but incompatible interpretations of the ironies; responses to “what is meant” – the non-salient compatible interpretation – were significantly less frequent (the reverse was true of such exchanges between adversaries participating in TV talk-shows). In Partington (2007), too, evidence supports the view that speakers retain salience-based but incompatible interpretations of irony (termed The surface narrative of irony), which could be further exploited to induce effects such as laughter and even strategic advantage. To sum up, online studies as well as natural language use demonstrate that, as predicted by The Graded Salience Hypothesis, comprehenders activate and retain salient but incompatible meanings and salience-based but incompatible interpretations of ironic utterances. Such results cannot be accommodated by either The Direct Access View and Relevance Theory or by The Standard Pragmatic Model. 3.2.2.

Metaphor

Interpreting metaphors, particularly novel metaphorical utterances embedded in contexts supportive of the metaphorical interpretation, can also allow an insight into the involvement or lack of salient but incompatible meanings and saliencebased but incompatible interpretations in utterance processing. According to The Graded Salience Hypothesis and The Suppression/Retention Hypothesis (Giora 2003; Giora and Fein 1999), salient but incompatible meanings and salience-based

Happy New War

249

but incompatible interpretations, conducive to the appropriate interpretation of the utterance, will be accessed and make up part of the final representation of the utterance. Whereas the final metaphoric output of such metaphors is heavily dependent on contextual information, incompatible salient meanings and salience-based interpretations rely on the salient meanings of the components that make up the metaphoric utterance. Recent research reveals that both incompatible meanings as well as incompatible interpretations are represented in the products of a metaphor. Salient but incompatible meanings Recent research demonstrates that the products of a metaphor include salient but incompatible meanings. Such meanings were shown to be activated initially and to remain accessible even at the late stages of the interpretation process, suggesting that they were neither sieved out nor suppressed by contextual information supportive of the appropriate interpretation. For instance, in Rubio Fernández (2007), following contexts supportive of (literally meaningless) novel metaphoric interpretations (John doesn’t like physical contact. Even his girl friend finds it difficult to come close to him. John is a cactus.), salient, incompatible literal meanings (‘spike’), conducive, though, to the interpretation of the (novel) metaphor, were accessed immediately and remained accessible even at a 1000 ms delay. In Kacinik and Chiarello (2007), conventional metaphors (bright), having salient (literal) and somewhat less salient (metaphorical) senses, were embedded in literally and metaphorically biasing contexts (It’s the building with the bright colors vs. The teacher praised the bright student). Using the divided visual field paradigm, Kacinik and Chiarello demonstrated priming effects for both senses in both hemispheres, regardless of degree of salience. Importantly, both hemispheres retained both senses, showing that salient and less-salient but incompatible meanings which are conducive to the appropriate interpretation made up part of the final representation. In Giora, Fein, Aschkenazi and Alkabets-Zlozover (2007), negated (literally sensible) metaphors (no rocket) primed their salient but incompatible metaphoric meaning (‘fast’) even as long as 1000 ms following their offset (The train to Boston was no rocket. The trip to the city was *fast*, though). Such results attest to the presence of salient but incompatible meanings in the products of metaphor. Salience-based but incompatible interpretations To argue against The Standard Pragmatic Model, The Direct Access View and The Relevance Theory, it is necessary to show that salience-based but incompatible interpretations make up part of the products of metaphor processing. Indeed, findings from a number of methodologies support the involvement of such incompatible in-

250

Rachel Giora

terpretations in interpretation processes and products. For instance, results of reading times show that novel metaphors took longer to read than their literal interpretations, suggesting that salience-based but incompatible interpretations might have been involved in the process (Brisard, Frisson, and Sandra 2001; Giora and Fein 1999; but see Ortony et al.1978). In Kacinik and Chiarello (2007), the salience-based but incompatible (literal) interpretations of metaphorical utterances were accessed and retained in spite of a sentence context supportive of the metaphorical interpretation. Specifically, when allowed enough processing time (950 ms), literal, message-level interpretations (‘wilted’) of novel metaphoric utterances such as Henry thought her eyes were petals were primed and retained in the right hemisphere alongside appropriate message-level interpretations (‘lovely’). Moreover, this effect was equivalent to that obtained when the literal targets followed literal utterances such as That plant keeps losing its petals. Such findings demonstrate that even when the metaphorical interpretation was accomplished, incompatible interpretations conducive to its interpretation still featured in the final metaphorical output. Such results support the view that salience-based interpretations may be retained even when incompatible and make up part of the output of metaphors even in contexts highly supportive of the metaphorical interpretation. In addition to lab results, evidence from corpus-based research shows that metaphors (in bold), both novel and familiar, can prime and be primed by their literal interpretation (in italics) (see Giora 2011). For instance, in the following example, novel metaphors (Denying food is fire, Denying water is fire) are primed by a reference to their salience-based but incompatible literal interpretation (ceasefire) which precedes them: (12) Let’s be clear about this: Israel’s fire at Gaza has not ceased. There is no Israeli ceasefire in Gaza. There is no Israeli ceasefire even when Israel’s soldiers aren’t shooting a single bullet in Gaza … There are food shortages in Gaza. Israel is denying Gaza food … Food shortages kill. Denying food is fire. There’s a shortage of potable water in Gaza. … Water shortages kill. Denying water is fire (Mazali 2006). Conventional metaphors too can be primed by their salience-based incompatible (literal) interpretation. For instance, in “The ceasefire will go up in flames” (Levy 2006), the literal, salience-based but incompatible interpretation (ceasefire) primes a familiar metaphor (go up in flames) which follows it; in “To expel them [the Israeli children of foreign workers] is disgraceful – it is this disgrace that we must expel” (Aloni 2009), the metaphorical expel is primed by its salience-based incompatible (literal) interpretation (‘send away’). Conventional metaphors can also prime their salience-based but incompatible interpretation. For instance, Glitterati in “Glitterati at Leviev’s New York Gala Stunned by Palestinian Rights Protest” resonates with diamonds, related to its sa-

Happy New War

251

lience-based but incompatible interpretation, appearing in the utterance that follows it: (13) Glitterati at Leviev’s New York Gala Stunned by Palestinian Rights Protest. Leviev’s diamonds fund repression in Angola, and violations of international law in Palestine (Adalah-NY 2007). Similarly the familiar metaphor “thirsting for justice” resonates with its saliencebased but incompatible literal interpretation in “control of water” which follows it: (14) Palestine thirsting for justice – Israel’s control of water as tool of apartheid and means of ethnic cleansing (Barghouti 2009). A corpus-based study of metaphors in a conversation between intimates, including 2856 words (taken from Du Bois et al. 2000), discusses the 23 metaphors appearing in it. Whereas 5 did not prime any of their interpretations, the other 18 either primed their literal interpretation or elaborated on it by extending the metaphor. Of these 18, 17 were familiar metaphors (Giora 2003; see also Giora and Balaban 2001). Needless to say, novel metaphors can also prime their salience-based but incompatible interpretation. For instance, in (15), the metaphoric igniting under fire and shooting, resonate with their salience-based but incompatible (literal) interpretation related to kindle, fire and shooting, which follow them: (15) Both Gaza and the West Bank will go on igniting under fire, till they kindle Sderot [an Israeli town] again too. The bullet-less fire that Israel is shooting at the dispossessed of Gaza is fire that it is also shooting, by proxy, at the dispossessed of Sderot (Mazali 2006). Metaphors (darkness) can also prime other metaphors (light) to which they are related via their incompatible literal interpretation: (16) A war in Iraq will soon break out, and with it a great darkness will descend on events in the territories … This is the time to caution us all that, under the cover of that darkness, grave things may come to pass. Not that there is much light there now, either (Levy 2003). Findings from various methodologies, then, demonstrate that both salience-based but incompatible meanings as well as salience-based but incompatible interpretations make up part of the products of a metaphor. Such findings cannot be explained by The Standard Pragmatic Model (Grice 1975), The Direct Access View (Gibbs 1994) or The Relevance Theory (Sperber and Wilson 1986/1995). Instead, they support The Graded Salience Hypothesis and The Suppression/Retention Hypothesis (Giora 2003; Giora and Fein 1999) which predict that salient meanings and salience-based interpretations will not be inhibited, nor will they be suppressed by a strong context supportive of the appropriate interpretation in case they are conducive to that interpretation.

252 3.2.3.

Rachel Giora

Visual negation

Does processing negation involve salient but incompatible meanings and salience-based but incompatible interpretations? Note that what would count as incompatible salient meanings and incompatible salience-based interpretations in processing negation is related to the affirmative meaning of the concept within the scope of negation which, if retained, will affect the final representation of the negated concept. The fact that negated information involves incompatible salient meanings is widely acknowledged: Negation does not inhibit access of salient meanings of concepts within its scope (Hasson and Glucksberg 2006; Giora, Balaban et al. 2005; Kaup et al. 2007; MacDonald and Just 1989: Experiments 1 – 2 reading phase). However, according to the received view, these meanings should undergo suppression and be replaced by an available alternative (MacDonald and Just 1989). Thus, when presented in isolation, the salient metaphorical meaning of rocket (‘fast’) in The train to Boston was no rocket was discarded from the mental representation (Hasson and Glucksberg 2006); in the door is not open, presented in isolation, the concept within the scope of negation (open) was deactivated and replaced by its alternative opposite (‘closed’; Kaup, Lüdtke and Zwaan 2006). In contrast, Giora and colleagues (Giora 2006, 2007; Giora, Balaban et al. 2005; Giora, Fein, Aschkenazi et al. 2007) argued that suppression following negation is not obligatory; instead, it is sensitive to discourse goals and requirements. It will not be induced unless invited by context, nor will it operate when contextual information motivates retention of the negated concept (see also Kaup and Zwaan 2003). According to this view, then, negation might have a number of consequences rather than a single contrastive outcome. In addition to testing this pragmatically oriented view by examining linguistic negation (Giora 2006; Giora, Fein, Aschkenazi et al. 2007), Giora, Heruti et al. (2009) examined visual negation. The working hypothesis was that visual negation should operate along the lines proposed for linguistic negation (Giora 2006, 2007). Visually negated information was therefore not expected to be unconditionally discarded from the mental representation (as proposed for linguistic negation by Hasson and Glucksberg 2007; Kaup, Lüdtke and Zwaan 2006; MacDonald and Just 1989). Rather, it was expected to be sensitive to discourse goals and requirements and retain visual information within the scope of negation when this was required (Giora 2006). For instance, in road-signs such as (17), (17a), which indicates end of highway by means of a visual negation marker, invites discarding information within the scope of negation (17b) and replacing it with an alternative (17c), which indicates an urban area:

Happy New War

253

(17)

However, in (18) by Gal (1996), wiping The Israel Museum of Art off an aerial photo has a different effect. Using a censoring technique used by the Israeli military Censor to wipe military bases off aerial photos, this visual erasure draws the viewer’s attention to the erased object, which, as a result, is represented in a way that likens it to a military base. This indeed was the artist’s way of expressing his criticism of the complicity of institutional art in Israel in the occupation and the oppression of the Palestinians (Gal 2006). (18)

Figure 5. Erasing the Israel Museum (http://archiv.ok-centrum.at/english/ausstellungen/promise/gal.html)

These examples suggest that information within the scope of visual negation is not discarded automatically. It is suppressed if it needs to be replaced by alternative information; it is retained in case it plays a role in the end-product interpretation.

254

Rachel Giora

To test this, Giora, Heruti et al. (2009) ran an offline task, in which participants were presented visual instructions and assertions such as (19), containing a visual negation marker (X) imposed on a percept: (19)

The participants were asked to rate the extent to which each of three different interpretations was an appropriate interpretation of the image. One such interpretation retained the negated information (Don’t leave the door open!); another presented an alternative opposite (Close the door!); and another conveyed irrelevant but still somehow related information (There is a strong lock on the door). Results showed that, as found for linguistic negation (see Giora 2006; Giora, Fein, Laadan et al. 2007; Kaup and Zwaan, 2003), comprehenders did not discard negated information unconditionally. Rather, they preferred the interpretation which preserved the information within the scope of negation (Don’t leave the door open!), suggesting that the so-called inappropriate interpretation (‘leave the door open!’) was retained and featured in the products of visual negation.7

4.

Conclusions

Is our cognitive machinery adept at swiftly and accurately homing in on a single, contextually appropriate interpretation, as assumed by The Direct Access View (Gibbs 1986, 1994) and Relevance Theory (Sperber and Wilson 1986/1995), or is it less efficient at sieving out interpretations based on salient (coded and prominent) word and phrase meanings, as proposed by The Graded Salience Hypothesis (Giora 1997, 1999, 2003)? A review of many of the recent findings in irony, metaphor, and negation challenges The Direct Access View, The Relevance Theory and The Standard Pragmatic Model. Instead, it supports The Graded Salience Hypothesis, shedding light on the role salient meanings play in inducing salience-based interpretations which are involved in utterance interpretation even when incompatible and even when contextual information is highly supportive of the compatible interpretation. Indeed, converging evidence from multiple experimental methodologies, both online and offline, testifies to the significant role salient but incompatible meanings and salience-based but incompatible interpretations play in shaping

Happy New War

255

contextually compatible interpretations in contexts strongly benefitting such interpretations. Our cognitive machinery, then, does not swiftly and accurately home in on a single, contextually appropriate interpretation, because it is sensitive to the pragmatic role inappropriate meanings and interpretations play in affecting the products of processing utterances.

Acknowledgments This paper is supported by a grant to the author by THE ISRAEL SCIENCE FOUNDATION (grant No. 652/07).

Notes 1. http://forecasthighs.com/2009/01/02/happy-new-war/ 2. A coded response to a stimulus involves not only its semantics but also its default contexts of use and co-occurrences. For instance, a stimulus such as doctor involves its semantics (‘medical expertise’) as well as concepts with which it co-occurs (such as ‘nurse’ or ‘hospital’). 3. http://en.wikipedia.org/wiki/The_Adventures_of_Pinocchio#Plot 4. Both types of dialogues included an ironic speaker uttering an ironic utterance in context mid-position. However only in the ironically biasing context that speaker uttered another ironic utterance – the target utterance. On the immediate effect of such pragmatic information on irony comprehension, see Regel, Coulson and Gunter (2010). 5. On negative understatements being ironic, see Giora, Fein, Ganzi et al. (2005). 6. In the original Hebrew version hard and harsh were described by using the same (polysemous) root (translated by Elad Livnat). 7. An additional experiment ruled out the possibility that this preference was motivated by a politeness principle.

References Adalah-NY 2007 Glitterati at Leviev’s New York Gala Stunned by Palestinian Rights Protest. http://nyc.indymedia.org/en/2007/11/92735.shtml Aloni, Shulamit 2009 Shas ‘racial purity’ policy toward migrant kids is a disgrace. http://www.haaretz.com/hasen/spages/1123067.html Barghouti, Omar 2009 Palestine thirsting for justice – Israel’s control of water as tool of apartheid and means of ethnic cleansing. Loyola Law School, Los Angeles. Brisard, Frank, Steven Frisson and Dominiek Sandra 2001 Processing unfamiliar metaphors in a self-paced reading task. Metaphor and Symbol 16: 87–108.

256

Rachel Giora

Colston, L. Herb and Raymond W. Jr. Gibbs 2002 Are irony and metaphor understood differently? Metaphor and Symbol 17: 57–80. Du Bois, John W. 2004 The dialogic locus: Where language, mind, and culture engage. Paper presented at Conference on Language, Culture, and Mind, July 17–20 2004. Portsmouth, England. Du Bois, John W., Wallace L. Chafe, Charles Meyer and Sandra A. Thompson 2000 Santa Barbara Corpus of Spoken American English Part I. Linguistic Data Consortium, Philadelphia. http://www.ldc.upenn.edu/Catalog/CatalogEntry.jsp?catalogId=LDC2000S85 Duffy, Susan A., Robin K. Morris and Keith Rayner 1988 Lexical ambiguity and fixations times in reading. Journal of Memory and Language 27: 429–446. Ferretti, Todd R., Courtney Patterson and Albert N. Katz 2006 Context-Independent Influence of Explicit Markers on Proverb Interpretation. Proceedings of the Twenty-Eighth Annual Meeting of the Cognitive Science Society: 2481. Mahwah, NJ: Erlbaum. Ferretti, Todd R., Christopher A. Schwint and Albert N. Katz 2007 Electrophysiological and behavioral measures of the influence of literal and figurative contextual constraints on proverb comprehension. Brain and Language 101: 38–49. Fodor, Jerry 1983 The Modularity of Mind. Cambridge, MA: MIT Press. Frazier, Lyn and Keith Rayner 1990 Taking on semantic commitments: Processing multiple meanings vs. multiple senses. Journal of Memory and Language 29: 181–200. Gal, Meir 1996 Erasing the Israel Museum (from the series: Erasing the Major Museums), processed photograph. Gal, Meir, 15 June 2006 Untitled (Erasing the Major Museums) 1996. Maarv 4 (In Hebrew). http://www.maarav.org.il/archive/classes/PUItem17f2.html. Gernsbacher, Morton Ann 1990 Language Comprehension as Structure Building. Hillsdale, NJ: Lawrence Erlbaum Associates. Gibbs, Raymond W. Jr. 1983 Do people always process the literal meanings of indirect requests? Journal of Experimental Psychology: Learning, Memory, and Cognition 9: 524–533. Gibbs, Raymond W. Jr. 1984 Literal meaning and psychological theory. Cognitive Science 8: 575–304. Gibbs, Raymond W. Jr. 1986 On the psycholinguistics of sarcasm. Journal of Experimental Psychology: General 115: 3–15. Gibbs, Raymond W. Jr. 1994 The Poetics of Mind. Cambridge: Cambridge University Press.

Happy New War

257

Gibbs, Raymond W. Jr. 2002 A new look at literal meaning in understanding what is said and implicated. Journal of Pragmatics 34: 457–486. Gibbs, Raymond W. Jr. 2005 The psychological status of image schemas. In: Beate Hampe and Joseph E. Grady (eds.), From Perception to Meaning: Image Schemas in Cognitive Linguistics, 113–136. Berlin: Mouton de Gruyter. Gibbs, Raymond W. Jr., Nandini. P. Nayak and Cooper Cutting 1989 How to kick the bucket and not decompose: Analyzability and idiom processing. Journal of Memory and Language 28: 576–593. Giora, Rachel 1997 Understanding figurative and literal language: The graded salience hypothesis. Cognitive Linguistics 7: 183–206. Giora, Rachel 1999 On the priority of salient meanings: Studies of literal and figurative language. Journal of Pragmatics 31: 919–929. Giora, Rachel 2003 On our Mind: Salience, Context, and Figurative Language. New York: Oxford University Press. Giora, Rachel 2006 Anything negatives can do affirmatives can do just as well, except for some metaphors. Journal of Pragmatics 38: 981–1014. Giora, Rachel 2011 Will anticipating irony facilitate it immediately? In: Marta Dynel (ed.), The Pragmatics of Humour across Discourse Domains, 19–31. Amsterdam: John Benjamins. Giora, Rachel 2011 Your baby is no longer an infant: On metaphor as context. In: Michael Gluzman and Orly Lubin (eds.), Intertextuality in Literature and Culture: A Festschrift in Honor of Ziva Ben-Porat, 235–247. Tel Aviv: Hakibbuz Hameuchad. [In Hebrew] Giora, Rachel and Noga Balaban 2001 Lexical access in text production: On the role of salience in metaphor resonance. In: Ted Sanders, Joost Schilperoord and Wilbert Spooren (eds.), Text Representation, 111–124. Amsterdam: John Benjamins. Giora, Rachel, Noga Balaban, Ofer Fein and Inbar Alkabets 2005 Negation as positivity in disguise. In: Herb L. Colston and Albert N. Katz (eds.), Figurative Language Comprehension: Social and Cultural Influences, 233–258. Hillsdale, NJ: Erlbaum. Giora, Rachel and Ofer Fein 1999 Irony: Context and salience. Metaphor and Symbol 14: 241–257. Giora, Rachel, Ofer Fein, Keren Aschkenazi and Inbar Alkabets-Zlozover 2007 Negation in context: A functional approach to suppression. Discourse Processes 43: 153–172. Giora, Rachel, Ofer Fein, Jonathan Ganzi, Natalie Alkeslassy Levi and Hadas Sabah 2005 On negation as mitigation: The case of irony. Discourse Processes 39: 81–100.

258

Rachel Giora

Giora, Rachel, Ofer Fein, Ronie Kaufman, Dana Eisenberg and Shani Erez 2009 Does an Ironic situation favor an ironic interpretation? In: Geert Brône and Jeroen Vandaele (eds.), Cognitive Poetics. Goals, Gains and Gaps, 383–399. (Applications of Cognitive Linguistics series.) Berlin/New York: Mouton de Gruyter. Giora, Rachel, Ofer Fein, Ann Kronrod, Idit Elnatan, Noa Shuval and Adi Zur 2004 Weapons of mass distraction: Optimal innovation and Pleasure Ratings. Metaphor and Symbol 19: 115–141. Giora, Rachel, Ofer Fein, Daphna Laadan, Joe Wolfson, Michal Zeituny, Ran Kidron, Ronie Kaufman and Ronit Shaham 2007 Expecting irony: Context vs. salience-based effects. Metaphor and Symbol 22: 119–146. Giora, Rachel, Ofer Fein and Tamir Schwartz 1998 Irony: Graded salience and indirect negation. Metaphor and Symbol 13: 83–101. Giora, Rachel and Inbal Gur 2003 Irony in conversation: Salience and context effects. In: Brigitte Nerlich, Zazi Todd, Vimala Herman and David Clarke (eds.), Polysemy: Flexible Patterns of Meanings in Language and Mind, 297–316. Berlin: Walter de Gruyter. Giora, Rachel, Vered Heruti, Nili Metuki and Ofer Fein 2009 Visual negation “When we say no we mean no”: Interpreting negation in vision and language. Journal of Pragmatics 41: 2222–2239. Giora, Rachel, Moshe Raphaely, Ofer Fein and Elad Livnat forthcoming On the involvement of salience-based but inappropriate interpretation in irony – a corpus-based study. Cognitive Linguistics. Grice, Herbert Paul 1975 Logic and Conversation. In: Peter Cole and Jerry Morgan (eds.), Speech Acts. Syntax and Semantics 3, 41–58. New York: Academic Press. Hamblin, Jennifer L. and Raymond W. Jr. Gibbs 1999 Why you can’t kick the bucket as you slowly die: Verbs in idiom comprehension. Journal of Psycholinguistic Research 28: 25–39. Hasson, Uri and Sam Glucksberg 2006 Does understanding negation entail affirmation? An examination of negated metaphors. Journal of Pragmatics 38: 1015–1032. Kacinik, Natalie A. and Christine Chiarello 2007 Understanding metaphors: Is the right hemisphere uniquely involved? Brain and Language 100: 188–207. Kashua, Sayed, September 21 2008 Manhattan transfer. http://www.haaretz.com/hasen/spages/986181.html Kaup, Barbara and Rolf A. Zwaan 2003 Effects of negation and situational presence on the accessibility of text information. Journal of Experimental Psychology: Learning, Memory, and Cognition 29: 439–446. Kaup, Barbara, Jana Lüdtke and Rolf A. Zwaan 2006 Processing negated sentences with contradictory predicates: Is a door that is not open mentally closed? Journal of Pragmatics 38: 1033–1050.

Happy New War

259

Kaup, Barbara, Richard H. Yaxley, Carol J. Madden, Rolf A. Zwaan, Jana Lüdtke 2007 Experiential simulations of negated text information. Quarterly Journal of Experimental Psychology 60: 976–990. Kotthoff, Helga 2003 Responding to irony in different contexts: cognition and conversation. Journal of Pragmatics 35: 1387–1411. Levy, Gideon 2003 The IDF’s ‘Permissiveness’ in the Territories. http://lists.econ.utah.edu/pipermail/rad-green/2003-February/006661.html Levy, Gideon 2006 The cease-fire will go up in flames. http://www.haaretz.com/print-edition/ opinion/the-cease-fire-will-go-up-in-flames-1.206065 Levy, Gideon 2008 Cast lead operation. (In Hebrew. Translated by Elad Livnat). http://www.mouse.co.il:80/CM.television_articles_item,790,209,31362,.aspx Levy, Gideon 2009 Things One Sees from The Hague. http://www.haaretz.com/hasen/spages/1054578.html Levy, Gideon 2009 Israelis don’t pay price for injustice of occupation. http://www.haaretz.com/hasen/spages/1101176.html MacDonald, C. Maryellen and Marcel A. Just 1989 Changes in activation levels with negation. Journal of Experimental Psychology: Learning, Memory, and Cognition 15: 633–642. Mazali, Rela 2006 Talk for December 2 Demo-Coalition to Stop the Siege of Gaza. http://www.jewishvoiceforpeace.org/publish/article_778.shtml Meytes, Dalia and Ariella Tamir 2005 Negation is sometimes faster than affirmation. Unpublished manuscript. Tel Aviv University. Ortony, Andrew, Diane L. Schallert, Ralph E. Reynolds and Stephen J. Antos 1978 Interpreting metaphors and idioms: Some effects of context on comprehension. Journal of Verbal Learning and Verbal Behavior 17: 465–477. Partington, Alan 2007 Irony and reversal of evaluation. Journal of Pragmatics, 39: 1547–1569. Peleg, Orna and Zohar Eviatar 2008 Hemispheric sensitivities to lexical and contextual constraints: Evidence from ambiguity resolution. Brain and Language 105(2): 71–82. http://dx.doi.org/10.1016/j.bandl.2007.09.004 Peleg, Orna and Zohar Eviatar 2009 Semantic asymmetries are modulated by phonological asymmetries: Evidence from the disambiguation of heterophonic versus homophonic homographs. Brain and Cognition 70: 154–162. Peleg, Orna and Zohar Eviatar 2012 Understanding written words: Phonological, lexical and contextual effects in the two cerebral hemispheres. In: Miriam Faust (ed.), The Handbook of the Neuropsychology of Language, 59–76. Malden, MA: Wiley-Blackwell.

260

Rachel Giora

Peleg, Orna, Rachel Giora and Ofer Fein 2001 Salience and context effects: Two are better than one. Metaphor and Symbol 16: 173–192. Peleg, Orna, Rachel Giora and Ofer Fein 2004 Contextual strength: The whens and hows of context effects. In: Ira Noveck and Dan Sperber (eds.), Experimental Pragmatics, 172–186. Basingstoke: Palgrave. Peleg, Orna, Rachel Giora and Ofer Fein 2008 Resisting contextual information: You can’t put a salient meaning down. Lodz Papers in Pragmatics 4(1): 13–44. Pexman, Penny M., Todd Ferretti and Albert Katz 2000 Discourse factors that influence irony detection during on-line reading. Discourse Processes 29: 201–222. Rayner, Keith, Jeremy M. Pacht and Susan A. Duffy 1994 Effects of prior encounter and global discourse bias on the processing of lexically ambiguous words: Evidence from eye fixations. Journal of Memory and Language 33: 527–544. Regel, Stefanie, Seana Coulson and Thomas C. Gunter 2010 The communicative style of a speaker can affect language comprehension? ERP evidence from the comprehension of irony. Brain Research 1311: 121–135. Rolnik, Guy 2008 We have 3 safety nets for you. (In Hebrew) http://www.themarker.com/tmc/article.jhtml?ElementId=ea20081202_78683 &origin=ibo&strToSearch=%EE%F9 %E1 %F8 Rubio Fernandez, Paula 2007 Suppression in Metaphor Interpretation: Differences between Meaning Selection and Meaning Construction. Journal of Semantics 24(4): 345–371. Searle, John 1979 Expression and Meaning. Cambridge: Cambridge University Press. Sperber, Dan and Deirdre Wilson 1986/95 Relevance: Communication and Cognition. Oxford: Blackwell. [Second edition with Postface 1995] Swinney, David A. 1979 Lexical access during sentence comprehension: (Re)consideration of context effects. Journal of Verbal Learning and Verbal Behavior 18: 645–659. Williams, John. N. 1992 Processing polysemous words in context. Evidence from interrelated meanings. Journal of Psycholinguistic Research 21: 193–218.

10.

Components of pragmatic ability and children’s pragmatic language development Daniela K. O’Neill

1.

Introduction

Differing definitions and the lack of a coherent theory of pragmatics have led to differing views of what abilities and topics of study reside within its domain, and also what kinds of knowledge are implicated. At a purely academic level it may not be urgent to resolve these differences, but from the viewpoint of a child who may be encountering difficulties with the acquisition of pragmatics, the lack of agreement as to which abilities to consider as pragmatic has made assessment more difficult and time-consuming (Gallagher 1991a) and resulted in a deplorably small number of standardized measures to assess pragmatics (but see O’Neill 2007, 2009). Furthermore, a fuller and more precise conception of what kinds of knowledge underlie different pragmatic abilities has important implications for directing intervention when faced with a child experiencing delay or impairment in pragmatic development. This chapter will present an overview of a number of different types of pragmatic abilities acquired in early and later childhood within the context of three domains of knowledge that are discussed in both the child language research and child clinical literature as underlying pragmatic ability, namely, social, social-cognitive, and cognitive knowledge. The literature on pragmatics is vast, even if limited only to children’s acquisition. Just as there are numerous definitions of pragmatics and lists of pragmatic phenomena, the lists are even longer when considering the types of knowledge suggested as underlying these phenomena. These underlying knowledge bases are also discussed in many different ways. Some accounts identify several main types of knowledge for communicative competence. Gallagher (1991a), for example, identifies language structural knowledge, presuppositional knowledge, and conversational knowledge as three types of knowledge required for communicative competence. Other accounts are more process-oriented. For instance, Verscheuren (1999) argues that language use involves the continuous making of choices and defines adaptability as a key notion to explain, for example, how speakers choose among the ingredients of the communicative context (e.g., physical surroundings, social relationships, states of mind) influencing linguistic choices. At present, I would argue, however, that the literature, especially with respect to children, offers little in the way of a comprehensive framework for considering the acquisition of pragmatics or the large number of different abilities subsumed under the general term pragmatics, and how these skills may or may not be related to other types of knowledge a child is acquiring.

262 2.

Daniela K. O’Neill

Approaches to pragmatics

The complex domain of pragmatics has naturally led to different models and approaches in the literature. Among early influential models, Bloom and Lahey’s (1978) depicts language as the intersection between the three domains of content rules, form rules, and use rules. Prutting’s model (1982) introduced and emphasized the dyad as the minimal unit of analysis and depicted, in a venn style diagram, three areas of knowledge – social and cognitive knowledge, linguistic rules, and pragmatic rules – “that operate together in the flow of discourse” (Prutting 1982: 130). From the beginning, early definitions and pragmatic language models acknowledged the social nature of language (Austin 1962; Bates 1976; Bloom and Lahey 1978; Bruner 1975; Searle 1969) as captured in this statement by Gallagher (1991b: 11–12): “when [the] language code is used to communicate, it is an inherently social phenomenon. Pragmatics is the study of language as it is used and when language is used in conversation it is a social behavior”. With a specific focus on children’s development, several approaches have aimed to abandon a view of pragmatics as simply an inventory of abilities and sought a broader interpretation in which pragmatics is viewed as a more dynamic influence on language acquisition (Rees 1982; Tomasello 1992; Wetherby 1991). Functionalist models have placed at the forefront an emphasis on what children are trying to communicate in naturalistic settings rather than maintaining a strict distinction between language form, meaning, and use (cf. Bates 1976; Bates and MacWhinney 1982; Nelson 1981; Prutting 1982; Ninio and Snow 1996; Perkins 2007; Verscheuren 1999). Prominent among authors who consider pragmatics from a clinical or neuropsychological standpoint is the viewpoint that pragmatics quickly extends beyond the realm of language structure and linguistics, and involves knowledge that is more aptly viewed as social, cultural, cognitive, or even sensorimoter (Adams 2005; Fujiki and Brinton 2009; Ninio and Snow 1996; Perkins 2007). Fujiki and Brinton (2009), for example, build on the functionalist position and advocate for an expanded view of pragmatics in a wider context of social communication that allows a wider range of interactive behaviours to be considered as part of pragmatics and intervention (e.g., behaviours leading to successful peer group entry). Newer emergentist and neuropsychological/neurolinguistic approaches, such as that of Perkins (2007, 2009), build on earlier interactionist approaches and seek to abandon a generic view of pragmatic impairment in favour of providing a taxonomy of pragmatic impairments (e.g., cognitive, linguistic, non-verbal, sensorimoter) based on differences in component abilities underlying these deficits (e.g., impaired lexical access versus theory of mind). The latter clearly aim not just to describe pragmatic impairment, but to explain it in a way that provides clinicians with clear targets for intervention (cf. Perkins 2007). With this landscape in mind, I would like to propose a framework in which the domain of pragmatics, as shown in Figure 1, is comprised of three core compo-

Components of pragmatic ability and children's pragmatic language development

263

nents, namely, social pragmatics, mindful pragmatics, and cognitive pragmatics, which draw, respectively, on knowledge that is social, social-cognitive or cognitive in nature. Two caveats should be noted with respect to Figure 1. First, this depiction of the domain of pragmatics should not be interpreted as implying static knowledge. Rather, the highlighted domain of pragmatics is where these three types of knowledge come to be used in conversation or dialogue. In other words, these three components constitute pragmatic competence and the active, ongoing, dynamic event of communicating with others. One may be best able to think of this in the context of multiple iterations of Figure 1, each representing one person present in a communicative interaction. Another way to describe this is that social, social-cognitive and cognitive knowledge are used in other settings and for other purposes that have nothing to do with pragmatics and dialogue or conversation. Thus, the highlighted oval represents when these domains of knowledge are used specifically in conversation or dialogue in an active, ongoing manner. Indeed, a picture of three children in conversation serves as a reminder of this in Figure 1. The second caveat is that Figure 1 is also meant to convey that not all social, social-cognitive or cognitive knowledge will be relevant to pragmatics. Rather, I will define pragmatics at present as the utilization of social, social-cognitive or cogni-

Figure 1. Three components of the domain of pragmatics drawing on social, social-cognitive and cognitive knowledge respectively

264

Daniela K. O’Neill

tive knowledge in interaction with others that is specifically related to how individuals or groups may possess differing experiential or conceptual mental perspectives and expectations that must be taken into account in order to communicate or understand another communication effectively. This definition is meant to capture, as will be discussed in more detail in this chapter, that an interlocuter must take into account that the social, social-cognitive or cognitive perspectives or points of view of conversational interactants may differ. This approach can be viewed as similar to other approaches, albeit these are usually discussed with respect to just one of these types of knowledge separately. For example, with respect to cognitive knowledge, Sperber and Wilson’s relevance theory emphasizes participants’ shared “cognitive environments” (Sperber and Wilson 1986/1995). Here I have attempted to be more inclusive with respect to different types of knowledge that children may be utilizing to assess the potentially differing experiential, conceptual mental perspectives and expectations of communicative interactants. Space constraints, and the goal of this chapter to provide an overview of pragmatic development in childhood, do not permit a detailed discussion of this approach in relation to other preceding approaches and models. The discussion to follow will proceed directly to considering the three different types of knowledge represented in Figure 1 and their relation to pragmatic competence. In discussing each of these types of knowledge, findings from the literature on children’s acquisition of pragmatics will be incorporated both to illustrate examples of the different domains of pragmatic competence and to show how these domains can capture age differences in types of pragmatic competencies acquired. The particular developmental acquisitions highlighted are meant to underscore the compatibility of this approach with functionalist approaches that stress that development may be propelled by what children “suppose language to be useful for” (Nelson 1981: 186). Or as Ervin-Tripp and Mitchell-Kernan have said, “work on children’s discourse pushes back to the earliest stages of interaction” and can enable us “to find through comparative studies those facets of human interaction which are fundamental and universal” (1977a: 23). It should be noted that several authors have produced excellent, detailed overviews of the acquisition of pragmatic abilities in childhood. The reader is referred, in particular, to the works of Foster (1990), McTear and Conti-Ramsden (1992), Hoff-Ginsburg (1996), Nino and Snow (1996) and Clark (2009). The focus of this chapter is not to reproduce these (cf. Reboul, Manificat and Foudon, this volume), but rather to try to seek a coherent means whereby to classify a variety of these acquisitions (rather than an exhaustive set) as one type of pragmatic ability versus another by considering the knowledge potentially being recruited by a language user as suggested by empirical studies in the literature. One outcome of this may be greater clarity with respect to grouping certain pragmatic abilities as similar or different. A second outcome may be a reconsideration of some abilities that are commonly treated as pragmatic, but may be more appropriately regarded as residing outside the domain of pragmatics.

Components of pragmatic ability and children's pragmatic language development

265

Throughout this discussion, non-verbal means of communication such as eye gaze, facial expressions, and gesture are meant to be included and viewed as simply alternative means by which persons can convey information in lieu of speech (either alone or in conjunction with speech to express similar or additional information). And just as with spoken speech, the pragmatic use of gestures can recruit different types of knowledge. For example, a gesture to indicate that you would like to pay in a restaurant might recruit social knowledge pertaining to cultural conventions, while a gesture to indicate to a listener that you are trying to remember something on the tip of your tongue might recruit social-cognitive knowledge relevant to an understanding of mental states.

3.

Social cognitive knowledge and mindful pragmatics

That a child’s language use is influenced by growth in social-cognitive knowledge is well recognized (e.g., Bates et al. 1979; Faerch and Kasper 1984; Schiefelbusch and Pickar 1984; Smith and Leinonen 1992; Ninio and Snow 1996; O’Neill 1996, 2007; Thompson 1996; Verschueren 1999; Abbeduto and Short-Meyerson 2002; Tomasello, 2003) and reflected in communicative competence model approaches (Gumperz and Hymes 1964; Hymes 1972) and approaches that view language use and pragmatics as social competence (Prutting 1982). The interplay between social-cognitive knowledge and language use has been the topic of much research activity, especially with respect to one particular area of social-cognitive knowledge, namely, children’s developing understanding of mind. Increasingly, this research has focused on very young children, including infants. I will begin this overview of children’s acquisition of pragmatic competence with respect to social cognitiveknowledge and mindful pragmatics. Children’s developing understanding of the mind can be defined as their understanding of their own and other people’s behaviours, mental states, and differing perspectives (i.e., theory of mind; cf. Astington, Harris, and Olson 1988). Viewing pragmatic competence as recruiting an understanding of mind is a relatively new approach, but similar to approaches that have viewed pragmatics as encompassing “the study of understanding intentional human action” (Green 1989: 2) and reflecting “the achievement of understanding the interlocutor’s state of mind” (Ninio and Snow 1996: 191). Note, however, that an understanding of mind should be equated neither with social cognition nor pragmatics. Social cognition is a broader construct that has been defined as “the way individuals perceive, interact with, and organize knowledge about people” (Sherrod and Lamb 1981) and includes topics such as stereotype knowledge or relationship knowledge that would generally be considered to lie outside of the domain of theory of mind. The term mindful pragmatics is used to capture pragmatic competence reliant on an understanding of differing conceptual mental perspectives and expectations as they apply at a

266

Daniela K. O’Neill

more individual or personal level with respect to interactants in a dialogue or conversation (to be contrasted with a more general, conventional-level understanding when considering cognitive pragmatics). This understanding will encompass a wide range of abilities, progressing from the earliest understandings in infancy of joint attention and intentions to communicate via gestures and single words to the developing understanding during the toddler and preschool years of other people as beings with desires, wants, expectations, emotions with whom you can communicate to regulate their actions, request or share information and with whom you may have to adapt your communication to accommodate different perspectives, beliefs and knowledge (for comprehensive overviews of these early developments in understanding of mind see, e.g., Dunn 1988; Perner 1991; Astington 1993; Shatz 1994; Bartsch and Wellman 1995; Carpendale and Lewis 2006). Beyond the preschool years, this understanding will incorporate more sophisticated concepts such as mixed emotions, second- and third-order false beliefs, and an understanding of deception and irony (Perner and Wimmer 1985; Chandler, Fritz, and Hala 1989; Pons, Harris, and de Rosney 2004; Fillipova and Astington 2008). Already at 9 to 10 months of age, children use their gestures and vocalizations to request, label, answer, greet and protest (Bates, Camaioni, and Volterra 1975; Dale 1980). With respect to children’s understanding of mind, it has been realized, largely via work with children with autism spectrum disorders, that the distinction between gestures used to request something (i.e., protoimperatives) and gestures used to share an attentional focus of interest such as showing or pointing at an object or event of interest (i.e., protodeclaratives) is critical. A lack of progression from imperative to declarative pointing is now well-recognized as an important early ‘red flag’, potentially marking an impairment in children’s developing understanding of mind and constituting an early indication that a child may be on the autism spectrum (Wetherby, Prizant, and Hutchinson 1998). The use of non-verbal and verbal indications of a speaker’s intentions has also been demonstrated in very young children, and the evidence is especially abundant for word learning (Tomasello 1988, 1999, 2003; Baldwin 1991). For example, 16-month-old infants will consult the eye gaze of an adult speaker in order to determine an intended referent in a situation in which a child is playing with an object but an adult was looking at a different object (cf. also Baldwin 1993). Indeed, Tomasello (2003: 3) has posited “intention-reading (theory of mind, broadly conceived)” to be one of two key skills (the other being pattern-finding) underlying language acquisition. Similarly, Clark (2009: 129) has argued that “the study of word acquisition should be the study of the pragmatic inferences children make about language use.” By age two, children will also take account of repairs (Uh-oh, it’s not an X, it’s a Y) and utterances distinguishing intentional from unintentional actions (e.g., Oops!) when inferring the intended referent of a new word (Tomasello and Kruger 1992; Tomasello and Barton 1994; Clark and Grossman 1998).

Components of pragmatic ability and children's pragmatic language development

267

Examination of children’s two-word utterances (Wieman 1976) has revealed that children will place stress on what is new (versus given) in the conversational context and that their hierarchy of stress assignments is almost identical to Chafe’s (1970) hierarchy of the position of new information in English utterances. Twoyear-olds have also been shown to be intent on achieving understanding in addition to their expressed goal by the fact that they will repair misunderstandings regardless of whether their goal is achieved or not (Shwe and Markman 1997). Two-yearolds will also adapt their request for a toy to take into account whether their mother was present or absent when the identity and location of the toy was revealed to the child (O’Neill 1996). By three years of age, children’s intrusions in conversations also become increasingly relevant and on-topic (Dunn and Shatz 1989). Threeyear-olds have also demonstrated the ability to assess the expertise of speakers, as knowledgeable or ignorant for example, and to use this information in deciding whether to trust the information from a speaker. For example, three-year-olds attend more to speakers who express certainty (This is a spoon) than speakers who express uncertainty (I think this is a spoon) (Jaswal and Malone 2007, see also Koenig and Harris 2005). By four years of age, children will assess expertise even under conditions where two speakers are differentially accurate, but neither is 100 % accurate (Pascquini et al. 2007). At four years of age, children also speak of desires and beliefs in a contrastive manner. That is, children will, for example, contrast one person’s desires or beliefs with another person’s (e.g., The people thought Dracula was mean. But he was nice; Bartsch and Wellman 1995: 53). Four- and five-year-olds have also been shown to explicitly question and track the knowledge possessed by a conversational partner (Short-Meyerson and Abbeduto 1997; Saylor, Baird, and Gallerani 2006; Short-Meyerson 2010). By this age, in peer-to-peer conversation in non-play settings (e.g., snacktime), children have demonstrated an ability to initiate conversations about topics related to the mental states of their conversational partners (e.g., Did you know Pokeman has this many arms?), even when no adults are present to help prompt or scaffold the conversation (O’Neill, Main, and Ziemski 2009). Beyond the preschool years, humour, deceit, irony and sarcasm are among some of the pragmatic abilities viewed as relying on concurrent developments in children’s understanding of the mind. It is now well recognized that children’s playfulness with conventional wordings and other forms of verbal humour (McGhee 1979; Dunn 1988; Reddy 1991; Bergen 2001; Cameron, Kennedy, and Cameron 2008) begins early, as young as 18 months of age in some cases and certainly by age two. These findings suggest that as children’s understanding of the beliefs and expectancies of other people develops, so too does their ability to exploit these beliefs and expectancies to produce humour (cf. Reddy 1991). Attempts to produce lengthier jokes and riddles have been commonly observed among preschool-aged children (Garvey 1977; Dunn and Kendrick 1982; Dunn 1988; O’Neill, Main, and Ziemski 2009). The link between many forms of verbal humour

268

Daniela K. O’Neill

and children’s growing understanding of the mind is also supported by impairments in such forms of humour among individuals with autism displaying deficits in their understanding of mind (St. James and Tager-Flusberg 1994). Not all forms of verbal humour may rely on such an understanding, however. Among school-age children, appreciation of verbal jokes and riddles has been studied more often with respect to comprehension rather than production, and discussions have focused on the contribution of linguistic and cognitive developments (e.g., lexical ambiguity, phonological ambiguity), abstract reasoning ability, and factual knowledge (Nippold 1998). The literature on the acquisition and understanding of other forms of non-literal language in children, such as indirect requests, idioms, sarcasm and irony, is relatively large and a detailed review is beyond the scope of this chapter (for review, see Nippold 1998). These non-literal language forms are often studied together, in the context of testing predictions regarding gradations of difficulty in their comprehension (Bucciarelli, Colle, and Bara 2003; Bernicot, Laval, and Chaminaud 2007). Although these predictions are based on considering the nature and difficulty of the inferences involved (e.g., “ … the difficulty of the cognitivesocial inference required …”, Bernicot, Laval, and Chaminaud 2007: 2128), the descriptions of these inferences are actually quite different at times and it could be questioned whether treating these non-literal language forms as more of a continuum of acquisitions is appropriate. For example, Bernicot et al. (2007: 2118) describe idioms as resting on “ … a linguistic convention” and implicatures requiring a sarcastic inference as “ … based on linguistic cues (namely, the contradiction between the two terms) and shared knowledge” (Bernicot, Laval, and Chaminaud 2007: 2119). The acquisition of idioms was predicted to occur earlier than that of sarcasm (although this could be debated, or vary, depending on the type of linguistic convention or shared knowledge required). In the framework of pragmatics I am presenting here, idioms would therefore appear to rely much less on social-cognitive knowledge than sarcasm. That is, although idioms require a distinction between what is said and what is meant, a listener could be viewed as largely relying on learned, memorized information (e.g., ‘encyclopedic entries’; Vega-Moreno 2005) for opaque idioms (e.g., beating around the bush), or on more cognitive knowledge in the case of more transparent idioms (e.g., hold your tongue) (Gibbs 1987; Nippold 1998; cf. Bucciarelli et al. 2003 regarding conventional indirect forms). The same situation would not apply for sarcasm, however, where it is difficult to see how the appropriate meaning could be deciphered without an understanding of mind that allows a child to recognize the speaker’s intended purpose with respect to him/herself and the listener in using sarcasm (i.e., to be funny by teasing the listener). When children’s acquisition of non-literal language is studied, it is also most often with respect to comprehension and not production. Children are also usually required to provide two types of responses: (1) a response (that is sometimes ac-

Components of pragmatic ability and children's pragmatic language development

269

complished verbally or via a point gesture) that indicates they understand the discrepancy between what a speaker said and what the speaker meant (e.g., she said the room was clean but she meant it was messy); and (2) a verbal explanation that indicates this understanding more explicitly (e.g., the child must answer why they chose a particular meaning or, for example, why a speaker used sarcasm, for instance to tease). Some authors refer to these two dependent measures as “pragmatic skill” and “metapragmatic knowledge”, respectively (e.g., Bernicot, Laval, and Chaminaud 2007). Often, quite large age differences with respect to performance on these two different dependent measures are found as well as differences in the order of acquisition of competence across different forms of non-literal language. One development that shows consistency, however, across several studies is the acquisition of sarcasm. Sarcasm is viewed as an instance of a violation of Grice’s (1975, 1989) maxim of relation (i.e., a sarcastic violation, Bernicot, Laval, and Chaminaud 2007). Evidence suggests that it is only by about 11 years of age that children achieve a full understanding of sarcasm (Bernicot, Laval, and Chaminaud 2007; Demorest et al. 1984). The fact that sarcasm requires a child to understand a speaker’s purpose for using sarcasm certainly requires a more sophisticated understanding of mind. However, it is also clear that there are other factors contributing to children’s understanding of sarcasm, such as the use of story context cues and intonational cues (Capelli, Nakagawa, and Madden 1990) that could be argued to rely less on social-cognitive knowledge. Irony also appears to be a similarly late-emerging communicative form (Winner 1988; Lucariello and Mindolovich 1995; Bucciarelli et al. 2003; Filippova and Astington 2008). As a last example of non-literal language forms, I will consider the most heavily studied one – indirect speech acts. With respect to indirect speech acts and, in particular, indirect requests, somewhat uniquely, comprehension has not been the sole focus, and much production data exists as well (e.g., Bates 1976; ErvinTripp and Mitchell-Kernan 1977b; Ervin-Tripp et al. 1986; Ninio and Snow 1996; Ryckebusch and Marcos 2004; Bucciarelli et al. 2003; Bernicot, Laval, and Chaminaud 2007). Production and comprehension data reveal a protracted and nuanced acquisition process that begins as early as at two years of age but that also demonstrates that a full understanding is not achieved until around eight years of age, when children can provide verbal explanations and spontaneously produce polite forms quite appropriately in interaction with others (for reviews, see Foster 1990; Clark 2009). Indirect speech acts also highlight a case where social-cognitive knowledge and social-cultural knowledge are both playing important roles. Indeed, it may be more appropriate to place the acquisition of indirect requests in the domain of social pragmatics. Young children must be explicitly taught (Foster 1990) more polite forms. For example, an English-speaking child would be taught that I want juice! is better expressed as I would like some juice or Can I have some juice?, or that the more polite Could you get me my shoes? will be more effective than the direct command Get me my shoes. Thus, a first step for children in the ac-

270

Daniela K. O’Neill

quisition of polite speech is to learn the appropriate linguistic forms, which are a matter of convention and will vary from culture to culture. Children must also learn what forms are appropriate for given settings and given interlocutors who may vary in age, status or sex. And, third, they must also come to understand that polite forms are essential to “maintain[ing] the co-operative nature of language interchange” (Allan 1986: 10) and keep things equal or balanced within an exchange (Brown and Levinson 1987) (i.e., the notion of gaining, maintaining or losing face in relation to others). Cross-cultural research has confirmed that children receive explicit instruction from adults as to the use of polite forms and the reasons for doing so (Hollos and Beeman 1978; Erbaugh 1992; Nakamura 2001). For example, in a study of Japanese mother-child interactions (Clancy 1985), mothers talked about the feelings and thoughts of others, and cited such reasons as: because that is the way things are done, because someone else needs the compliance, because otherwise other people will ridicule the child. This research on parental input thus also supports the notion that the acquisition of politeness and the indirect forms involved are heavily dependent on a child taking into account the mental states of others and considering the perspectives of others. Overall, the study of indirect requests in children highlights, I believe, the importance of studying these forms in more naturalistic, non-laboratory settings in order to get a full picture of children’s acquisition and use in communicative settings that may prove to be simpler or more complex and demanding than laboratory settings and tasks. As a final note, in my own work, my interest in what children realize they can do with language as a result of their growing understanding of people and their own and other people’s minds has led to the development of the Language Use Inventory (LUI, O’Neill 2007, 2009), a standardized parent-report questionnaire to assess early pragmatic language use, designed for 18- to 47-month-old children. The items on the LUI are based on the view that a large portion of children’s early language use is driven by their growing social cognitive competence, their interest in people, minds, and perspectives, and a desire to communicate about these things with people. And indeed the developmental growth trends and data from the standardization study of the LUI (O’Neill 2009) support the existence of consistency in the acquisition of such forms of talk across children during the period from 18 to 47 months of age.

4.

Cognitive knowledge and cognitive pragmatics

What developing pragmatic abilities might constitute examples of cognitive pragmatics? I would like to propose that situations in which a communicator is taking into account the minds of other interactants in a very individual, subjective manner (e.g., what he/she wants, desires, intends, knows about x in the current situation) be distinguished from situations in which a communicator is taking into account cog-

Components of pragmatic ability and children's pragmatic language development

271

nitive stances that apply more generally across situations and across different interactants (e.g., inferences regarding what is conventional or relevant). The former would represent social-cognitive mindful pragmatics, whereas the latter would represent instances of cognitive pragmatics. Thus pertinent to developmental cognitive pragmatics is the fact that children are not learning to use language with just the one mind of a communicative partner in mind, but with other minds in mind. In interaction and conversation, children are learning what other minds also find salient, relevant or interesting to talk about. To illustrate, consider the findings of a study in which typically developing 22-month-old children, at only the one- and two-word stage, were given a series of toys that featured notable changes in their identities or properties, such as a change in size or weight (O’Neill and Happé 2000). Children directed their attention in a remarkably similar way to these different changes in the toys, vocalized at very similar times about these changes, and noted these differences via showing gestures and in their talk (e.g., big duck; heavy) in a very similar way. How should one explain the fact that by 22 months of age children appear to already possess and demonstrate a common understanding of what is noteworthy to share and talk about? Perhaps, as some authors have argued, infants “learn where to attend” (Donald 2001: 228) and this is achieved by parents and family “train[ing] infants to share attention with them” (Donald 2001: 205). In O’Neill (2005), I speculated that perhaps children learn ‘what to talk about’ by noting such cues as another person’s line of regard, referential gestures, or actions with objects. It is worth keeping in mind that this ability that appears to come so naturally to 22-month-old typically developing children does not come so easily to other children. For example, in our study (O’Neill and Happé 2000) the same consistency in behaviours and vocalizations was not observed among a group of children with autism. Indeed, a profound difficulty in acquiring and making use of conventional knowledge has been argued to underlie the social deficits in autism (Bruner and Feldman 1993; Capps, Kehres, and Sigman 1998). Some readers may find the pragmatic ability just described to nevertheless be better housed within the domain of mindful pragmatics than cognitive pragmatics, so I will provide another possible illustration of cognitive pragmatics from early word learning. Clark (1987, 1990, 1993) has argued that children’s inferences about the meanings of new terms can be explained via their reliance on two pragmatic principles, namely, conventionality and contrast. Conventionality is defined as “for certain meanings, speakers assume that there is a conventional form that should be used in the language community” (Clark 2009: 143). Thus speakers will give priority to conventional forms. The principle of contrast is defined as “[s]peakers assume that any difference in form signals a difference in meaning” (Clark 2009: 144). Thus, if a speaker does not use a conventional term (e.g., uses hare instead of bunny), their addressee would assume they are trying to express some other meaning than that captured by the conventional term. That is, a child

272

Daniela K. O’Neill

will interpret a speaker’s use of a different word as signalling a difference in meaning (for review, see Clark 1993, 1995). The use of these principles has been described by Clark (2009: 128) as “ … making use of pragmatically licenced inferences” and is an alternative to the proposed use of a priori constraints on children’s hypotheses about word meanings (e.g., Mervis 1987; Markman 1989). I would also propose that children’s developing ability to use deictic terms correctly falls within the domain of cognitive pragmatics given that the perspectivedependent, point-of-view information that must be taken into account is not specific to individual persons, but rather to their conversational roles, locations and times when the utterance is spoken. For example, given that the pronouns I and you shift referent with every change in speaker, determining who is in the role of speaker and listener at the time of the utterance governs the use of I and you, independent of who the speaker or listener is. It takes until about three to four years of age for children to acquire the deictic meanings of I and you and here and there, and even longer to acquire the deictic pairs this and that and come and go (Girouard, Ricard, and Decarie 1977; Clark 1978, 1990). Before this age, children’s errors reflect a tendency to use themselves, rather than the speaker, as the reference point and so, for example, here is used with the meaning ‘near me’ rather than near the speaker (Clark and Garnica 1974; Clark and Sengul 1978). Interestingly, recent studies have utilized children’s comprehension and production of deictic terms within a narrative to infer the use of situation models in narrative comprehension (e.g., Rall and Harris 2000). How situation models may play a role in the acquisition of deixis has not yet been examined. The potential importance of considering that different types of knowledge may underlie different pragmatic abilities is perhaps most evident when considering narrative ability and its development. Narrative ability is always included in lists and overviews of children’s acquisition of pragmatics. But narrative ability is not always recognized for the many different kinds of methodologies and dependent measures it encompasses, as well as very different kinds of understandings and abilities (O’Neill, Pearce, and Pick 2004). For example, a child’s ability to adopt and use the “voices” of characters in a narrative (e.g., O’Neill and Holmes 2002) and the ability of preschoolers to “step into the shoes and minds” of story characters (Fecica and O’Neill 2010) appears reliant on social-cognitive knowledge and to fall within the realm of mindful pragmatics. However, social knowledge could be viewed as underlying the ability of children to produce narratives that are considered “good” narratives by teachers once children enter school, given that children must adopt a format that is valued in school but not necessarily compatible with the narrative style valued by their family or culture (Heath 1983; Gee 2004). What might be a representative narrative ability housed within cognitive pragmatics? The ability to produce clear referents within a story may be one such ability. The main consequence of an inability to establish clear referents is judged to be

Components of pragmatic ability and children's pragmatic language development

273

difficulty on the part of a listener or listeners to comprehend a child’s story. In one such study (O’Neill & Holmes 2002), four-year-olds were found to produce clear referents when narrating the picture storybook “Frog goes to Dinner” about 50 % of the time. Interestingly, three-year-olds were also able to produce clear referents an equivalent proportion of the time, but only if their pointing gestures and use of voicing was considered in addition to their ability to produce a clear referent verbally. If only verbal instances of clear referencing were considered, then three-yearolds’ proportion of clear referents stood only at 13 %. Interestingly, though, studies of children’s referencing ability cannot discern whether referencing is being impacted or not by the presence of an audience. In many studies, the child must narrate a story back to an experimenter known to the child to already know the story, thus obscuring somewhat the motivation for clarity on the part of the child. Clear evidence does not exist as to whether narratives produced for the self display different (i.e., lower) levels of cohesion or referential adequacy than narratives produced when an audience is present. It seems plausible that if one considers narrative cognition to be a fundamental form of human cognition and, as Bruner (1990, 1991) has argued, the means by which humans make sense of and understand the world around them, then the creation of a coherent and clear narrative may be the intended aim of all human minds, for themselves primarily (see also Fecica and O’Neill 2010). Thus, it may be more appropriate to view anaphora as an outcome of how the human mind attempts to deal with longer instances of communication describing events and interactants (text, discourse, conversational stories) and for which the audience can be other people or the self. Anaphora would be independent of whether the individual is physically interacting communicatively with others (but see Levy 2003). One might conclude then that, as such, anaphora should fall outside of the domain of pragmatics and be housed within the domain of cognition. However, I think a more interesting approach is to consider whether, when defining pragmatics, to include the self as a possible interlocuter in a dialogue or conversation. With respect to children’s development, the expansion of a definition of conversational interactant in this way could also be helpful when instances of pretend play or private speech share features typically deemed pragmatic (e.g., perspective-dependent voicing; talk about one’s mental states) when they occur in conversations with people.

5.

Social knowledge and social pragmatics

Social knowledge, often also construed as cultural knowledge, is commonly included in discussions of pragmatic competence. Indeed, pragmatics is argued, not infrequently, to be a process of social development and part of children’s enculturation (Ninio and Snow 1996). Pragmatics has also been argued to be a component of social competence (Goffman 1981; Prutting 1982; Gallagher 1991b).

274

Daniela K. O’Neill

In a very insightful review of pragmatic models, Ninio and Snow (1996) describe the many tricky and confounding boundary issues when distinguishing pragmatics from other domains of knowledge. They highlight in particular the difficulty of distinguishing culturally appropriate social behaviours (e.g., table manners) from culturally appropriate linguistic behaviours (e.g., polite etiquette) and conclude that “we lack guidelines to distinguish rules for speech from societal regulation of interpersonal behavior in general” (Ninio and Snow 1996: 5). I think it is possible however, to say more than this about the type of social knowledge that could underlie social pragmatics. One could view children’s developing pragmatic competence in this area as involving an appreciation of knowledge having to do with people and groups that a speaker must take into account to better align with the perspective of the person or group. The ‘perspective’ in question is different from that of domains of social cognition and cognition and would include taking into account, when communicating with others, such factors as biological features (e.g., gender, age), aspects of physical appearance (e.g., status indicators such as professional attire), aspects of behaviour (e.g., goes to work), the type of social setting or activity (e.g., receiving a gift or compliment, playing a game), or conventions that relate to maintaining harmonious communication with people (e.g., don’t order others around). As a result of this approach, this domain of social pragmatics would contain the adaptations of language most often discussed as rulebased (e.g., polite forms, use of formal terms such as Mrs or Dr to address adults), but would not be limited to this type of rule-based knowledge. And it would contrast with mindful and cognitive pragmatics in that the knowledge invoked is less about the shared/non-shared mental state or cognitive perspective of speaker and listener(s) and more about the shared/non-shared communicative expectations of other people – be they a single interactant, a small peer group, or a much larger unit such as a country – and the adaptations that may be necessary as a result. Thus the ‘perspective’ to be taken is one that is, for the most part, not related to knowledge of any specific person per se, but rather features of persons and their behaviour and conventions surrounding different communicative interactants and settings. A number of empirical findings indeed illustrate how children are able to take into account such factors in their communication with others in the early years. As an early example of children’s talk as influenced by their observations about people’s behaviour, Dunn (1988) noted, in a hallmark longitudinal series of studies involving 52 children observed at home, four types of inquiries that children in all the families made about other people and that emerged as early as at two years of age: inquiries about their whereabouts, their actions, their inner states, and rules that people must follow. Dunn argues that children’s questions, such as Where Glynnis gone? were not born of anxiety. Rather, “the interest of the children in knowing where their friends and family members were, and what they were doing was evident, as was their pleasure in recounting with the mother the familiar routines these other people follow” (Dunn 1988: 130–131). And it is of note that Dunn

Components of pragmatic ability and children's pragmatic language development

275

did not find that one child was interested in talking about these behaviours of people but another child was not (if typically developing). All the children of a given age were found to talk about these topics. Dunn’s findings regarding common types of questions and common content or topics within these types concur well with those of other researchers examining children’s early questions (e.g., Wode 1971; Ervin-Tripp 1977; Tyack and Ingram 1977; Bloom, Merkin, and Wooten 1982; Nelson 1989; Shatz 1994). By preschool age, a number of developments indicate that children are becoming quite competent at adjusting their communication to different speech communities. Gumperz (1982) defined a speech community as a group of speakers that shares rules and norms for using language. For example, Kyratzis (2001) has found differences in preschool-aged boys’ and girls’ expressions of “emotion talk,” with boys, for example, evolving norms against the expression of scaredness. Nakamura (2001) found differences in communicative style and use of specific language forms among boy and girl same-sex peer interactions. In a study examining preschoolers’ topic initiations with peers, O’Neill, Main and Ziemski (2009) found that some categories were used only by one sex. For example, only boys used jokes to initiate communication; girls were not observed over 21 sessions to ever do so. Kyratzis, Marx and Wade (2001) found that a peer’s expertise and competence at a task was taken into account by preschoolers who then adjusted their use of assertive or deferent control acts accordingly. And studies of role play (e.g., Andersen 1990) have clearly demonstrated that four- to six-year-olds take the relative power of speakers into account by, for example, producing less polite and more direct requests for more powerful speakers (e.g., father, doctor, teacher) and more polite and more indirect requests for lower status speakers (e.g., child). Children also adjusted their speech depending on the age of the speaker. When speaking for a child, children simplified word forms and omitted function words. As a final example, Fujiki and Brinton’s (2009) study of behaviours leading to successful peer group entry would fit well within the domain of social pragmatics. That is, children must acquire the knowledge that direct entry strategies such as the verbal request Can I play? are often unsuccessful and that a better strategy is to observe the activities of the group and find a way to make their actions or utterances similar to or concordant with those of the group they wish to join. I believe these group entry skills can thus be viewed as requiring the child to assess the perspective of the group with respect to the activity or talk in question in order to adapt and align their own behaviour to be successful. This overview has not been exhaustive, of course, and the interested reader is directed, for example, to Slobin et al.’s (1996) excellent edited book in honour of Susan Ervin-Tripp. Not only does this collection highlight the seminal work of Ervin-Tripp, but it includes contributions by, and reviews the work of, many other seminal researchers in the field of developmental sociolinguists (e.g., Slobin 1985; Ochs 1988).

276

Daniela K. O’Neill

As a last point, of the three areas of pragmatics, it is social pragmatics that lends itself most easily to a representation in terms of memorized rules of sorts that pertain to both non-verbal (e.g., bowing; hand gestures) and verbal communication. However, rather than rules, a more appropriate characterization of social pragmatics, I believe, is that it involves the identification and characterization of “the norms that underlie the spontaneous use of language of a given social group” (Escandell-Vidal 2004: 2). That is, users of a language are performing a type of statistical generalization in which, by observing the practices of speakers in a given community, speakers’ preferences with respect to such communicative forms as requests, apologies, or refusals, and communicative behaviours such as turn-taking, they try to establish the norms that can then be contrasted and compared with those of different cultural groups (Blum-Kulka, House, and Kasper 1989; EscandellVidal 2004). In some approaches, this characterization of social pragmatics has been contrasted with a characterization of cognitive pragmatics as aiming to identify principles involving more general, universal causal, mechanical explanations (Escandell-Vidall 2004), such as the principles set forth by Grice (1975, 1989), Sperber and Wilson (1986/1995, 1987) and Levinson (2000).

6.

Limitations of this approach

The approach to pragmatics that I have presented in this chapter is only a beginning working model and starting point. The aim of my approach was to see whether one might be able to offer some structure to the domain of pragmatics, and the myriad of abilities acquired by children labelled as pragmatic, by considering in more depth how the types of knowledge underlying these abilities might be different. The reason for believing they might differ is increasing empirical work documenting in more detail the kinds of understandings that may underlie children’s (and adults’) pragmatic abilities. Moreover, the goal was also not to simply list a host of different kinds of knowledge underlying different pragmatic abilities, but to explore whether one might be able to discern some key types of knowledge. The end result was a framework that focuses on the contribution of social, social-cognitive, and cognitive knowledge to children’s pragmatic language development and divides the pragmatic domain into social pragmatics, mindful pragmatics, and cognitive pragmatics, respectively. One limitation of the current approach is that by carving the domain of pragmatics into three specific components, one feels compelled to house a particular ability into one component only, when in fact an ability may draw on knowledge across several components. That is, in some cases it may be more a matter of gradation, rather than a yes-no classification. For example, the ability of 22-month-old children to recognize what is relevant to talk about, discussed at the beginning of the section on cognitive pragmatics, may be such a case. Some readers may feel

Components of pragmatic ability and children's pragmatic language development

277

that this ability relies equally on social-cognitive and cognitive knowledge, or more on the former than on the latter. Nevertheless, regardless of the decision, I believe the process of considering abilities like this from the vantage point of different key types of knowledge that may underlie them is a valuable exercise. Even if the outcome is that an ability is viewed as relying on social-cognitive and cognitive abilities, this may represent a further refinement and deeper understanding of this ability, which can then lead to a better understanding of the nature of impairment and more targeted and effective intervention activities. The main definition of pragmatics, and the definitions of social, social-cognitive and cognitive pragmatics, presented here are also working definitions. Definitions of pragmatics are notoriously varied. I have attempted, via an overall focus on perspective-taking, to define the three components in terms of three specific and different types of perspective-taking relating to individual conceptual perspectives (mindful pragmatics); to more generally applicable conceptual perspectives (cognitive pragmatics); and to aligning with conversational expectations (social pragmatics). Finally, the scope of this paper has not permitted more in-depth discussion related to abilities housed in the three domains of knowledge, but not within the domain of pragmatics (the non-highlighted areas in Figure 1). These areas might be best viewed as representing knowledge that has been acquired in each of these domains up to the present by an individual. Representative abilities housed within each of these three areas could be stereotype knowledge and social etiquette rules (social knowledge); recognition of emotional expressions and an understanding of mental state concepts (social-cognitive knowledge); and script knowledge and our understanding of time (cognitive knowledge). These four areas of knowledge should not be viewed as isolated. Rather, connections can (and will) certainly exist between these different domains of knowledge that are currently not depicted in Figure 1 given its intended aim to focus on pragmatics. Relevant to the focus of this chapter however, is that although the acquisition of linguistic terms and concepts housed within these types of knowledge may be dependent on communication with others, it may not be warranted to regard the process of acquiring such knowledge as within the domain of pragmatics based on empirical evidence. For example, temporal terms such as tomorrow and yesterday or scalar implicatures – whether an utterance implies some versus all or a part of a whole (e.g., painting the roof versus painting the house; cf. Papafragou and Tantalou 2004; Huang and Snedeker 2009) – may be linguistic terms whose acquisition (albeit via exposure in conversation) rests on more general cognitive and learning processes outside the domain of pragmatics. These examples would stand in stark contrast to the earliest stages of word learning already discussed, where evidence suggests the existence of an intricate relation between infants’ successful acquisition of new terms with the child’s developing social-cognitive conceptual perspective-taking abilities (e.g., understanding focus of attention).

278

Daniela K. O’Neill

It is clear that much remains to be understood about the relation between children’s developing pragmatic abilities and concurrent developments in other domains of knowledge. But understanding these relations better is crucial to understanding key pragmatic outcomes, such as being an effective communicator. Exactly what abilities underlie being an effective communicator is not well defined in the literature. Ninio and Snow (1996: 140–141) suggest that improving effectiveness as a communicator “probably depends on the mastery of a host of cognitive principles that at present are not well understood and whose developmental history has not yet been charted. Some candidates for this list include the conversational maxims and perspective-taking”. I believe the provision of a more detailed and principled way to organize different pragmatic abilities based on the type of knowledge they may recruit, and the type of perspective-taking that may be involved, is a valuable step in defining more clearly the domain of pragmatics and more specifically the course of pragmatic language development as children become effective, successful and enthusiastic communicators.

References Abbeduto, Leonard and Katherine J. Short-Meyerson 2002 Linguistic influences on social interaction. In: Howard Goldstein, Louise A. Kaczmarek, and Kristina M. English (eds.), Promoting Social Communication: Children with Developmental Disabilities from Birth to Adolescence, 27–54. Baltimore: Paul H. Brookes. Adams, Catherine 2005 Social communication intervention for school-age children: Rationale and description. Seminars in Speech and Language 26: 181–188. Allan, Keith 1986 Linguistic Meaning. London: Routledge & Kegan Paul. Andersen, Elaine S. 1990 Speaking with Style: The Sociolinguistic Skills of Children. London: Routledge. Anglin, Jeremy M. 1993 Vocabulary development: A morphological analysis. Monographs of the Society for Research in Child Development 58 (serial no. 258). Astington, Janet W. 1993 The Child’s Discovery of the Mind. Cambridge, MA: Harvard University Press. Astington, Janet W., Paul L. Harris and David R. Olson 1988 Developing Theories of Mind. New York: Cambridge University Press. Austin, John L. 1962 How to Do Things with Words. Oxford: Oxford University Press. Baldwin, Dare A. 1991 Infants’ contributions to the achievement of joint reference. Children Development 62: 875–890.

Components of pragmatic ability and children's pragmatic language development

279

Baldwin, Dare A. 1993 Early referential understanding: Infants’ ability to recognize referential acts for what they are. Developmental Psychology 29: 832–843. Bartsch, Karen and Henry M. Wellman 1995 Children Talk about the Mind. New York: Oxford University Press. Bates, Elizabeth 1976 Language and Context: The Acquisition of Pragmatics. New York: Academic Press. Bates, Elizabeth, Laura Benigni, Inge Bretherton, Luigia Camaioni and Virginia Volterra 1979 The Emergence of Symbols: Cognition and Communication in Infancy. New York: Academic Press. Bates, Elizabeth, Luigia Camaioni and Virginia Volterra 1975 The acquisition of performatives prior to speech. Merrill-Palmer Quarterly 21: 205–226. Bates, Elizabeth and Brian MacWhinney 1982 Functionalist approaches to grammar. In: Eric Wanner and Lila Gleitman (eds.), Language Acquisition: The State of the Art, 173–218. Cambridge, MA: MIT Press. Bergen, Doris 2001 Finding the humor in children’s play. In: Jaipaul L. Roopnarine (ed.), Conceptual, Social Cognitive, and Contextual Issues in the Fields of Play, 209–222. Westport, CT: Ablex Publishing. Bernicot, Josie, Virginia Laval and Stephanie Chaminaud 2007 Nonliteral language forms in children: In what order are they acquired in pragmatics and metapragmatics? Journal of Pragmatics 39: 2115–2132. Bloom, Lois and Margaret Lahey 1978 Language Development and Language Disorders. New York: John Wiley and Sons. Bloom, Lois, Susan Merkin and Janet Wootten 1982 Wh-questions: Linguistic factors that contribute to the sequence of acquisition. Child Development 53: 1084–92. Blum-Kulka, Shoshana, Juliane House and Gabriele Kaspar (eds.) 1989 Cross-cultural Pragmatics: Requests and Apologies. Norwood: Ablex. Brown, Penelope and Stephen C. Levinson 1987 Politeness: Some Universals in Language Usage. Cambridge: Cambridge University Press Bruner, Jerome 1975 From communication to language: A psychological perspective. Cognition 3: 255–287. Bruner, Jerome 1990 Acts of Meaning. Cambridge, MA: Harvard University Press. Bruner, Jerome 1991 The narrative construction of reality. Critical Inquiry 18: 1–21. Bruner, Jerome and Carol Feldman 1993 Theories of mind and the problem of autism. In: Simon Baron-Cohen, Helen Tager-Flusberg and David. J. Cohen (eds.), Understanding other Minds: Perspectives from Autism, 267–291. Oxford: Oxford University Press.

280

Daniela K. O’Neill

Bucciarelli, Monica, Livia Colle and Bruno G. Bara 2003 How children comprehend speech acts and communicative gestures. Journal of Pragmatics 35: 207–241. Cameron, E. Leslie, Katherine M. Kennedy and Catherine A. Cameron 2008 “Let Me Show You a Trick!”: A Toddler’s Use of Humor To Explore, Interpret, and Negotiate Her Familial Environment During a Day in the Life. Journal of Research in Childhood Education 23: 5–18. Capelli, Carol. A., Noreen Nakagawa and Cary M. Madden 1990 How children understand sarcasm: The role of context and intonation. Child Development 61: 1824–1841. Capps, Lisa, Jennifer Kehres and Marian Sigman 1998 Conversational abilities among children with autism and children with developmental delays. Autism 2: 325–344. Carpendale, Jeremy and Charlie Lewis 2006 How Children Develop Social Understanding. Oxford: Blackwell. Chafe, Wallace 1970 Meaning and the Structure of Language. Chicago: University of Chicago Press. Chandler, Michael J., Anna S. Fritz and Suzanne Hala 1989 Small scale deceit: Deception as a marker of 2-, 3- and 4-year-olds’ theories of mind. Child Development 60: 1263–1277. Clancy, Patricia M. 1985 The acquisition of Japanese. In: Daniel I. Slobin (ed.), The cross-linguistic study of language acquisition, Vol. I, 373–524. Hillsdale, NJ: Lawrence Erlbaum. Clark, Eve V. 1978 From gestures to word: On the natural history of deixis in language acquisition. In: Jerome S. Bruner and A. Garton (eds.), Human Growth and Development: Wolfson College Lectures 1976, 340–408. Oxford: Clarendon Press. Clark, Eve V. 1987 The principle of contrast: A constraint on language acquisition. In: Brian MacWhinney (ed.), Mechanisms of Language Acquisition, 1–33. Hillsdale, NJ: Lawrence Erlbaum. Clark, Eve V. 1990 The pragmatics of contrast. Journal of Child Language 17: 417–431. Clark, Eve V. 1993 The Lexicon in Acquisition. Cambridge, England: Cambridge University Press. Clark, Eve V. 1995 Later lexical development and word formation. In: Paul Fletcher and Brian MacWhinney (eds.), The Handbook of Child Language, 393–412. Oxford: Blackwell. Clark, Eve V. 2009 First Language Acquisition. 2nd ed. Cambridge, England: Cambridge University Press. Clark, Eve V. and Olga K. Garnica 1974 Is he coming or going? On the acquisition of deictic verbs. Journal of Verbal Learning and Verbal Behaviour 13: 559–572.

Components of pragmatic ability and children's pragmatic language development

281

Clark, Eve V. and James B. Grossman 1998 Pragmatic directions and children’s word learning. Journal of Child Language 11: 611–625. Clark, Eve V. and C. J. Sengul 1978 Strategies in the acquisition of deixis. Journal of Child Language 5: 457–475. Dale, Philip 1980 Is early pragmatic development measurable? Journal of Child Language 7: 1–12. Demorest, Amy, Christine Meyer, Erin Phelps, Howard Gardner and Ellen Winner 1984 Words speak louder than actions: Understanding deliberately false remarks. Child Development 55: 1527–1534. Donald, Merlin 2001 Origins of the Modern Mind: Three Stages in the Evolution of Culture and Cognition. Cambridge, MA: Harvard University Press. Dunn, Judy 1988 The Beginnings of Social Understanding. Cambridge, MA: Harvard University Press. Dunn, Judy and Carol Kendrick 1982 Siblings: Love, Envy and Understanding. Cambridge, MA: Harvard University Press. Dunn, Judy and Marilyn Shatz 1989 Becoming a conversationalist despite (or because of) having an older sibling. Child Development 60: 399–410. Erbaugh, Mary S. 1992 The acquisition of Mandarin. In: Daniel I. Slobin (ed.), The cross-linguistic study of language acquisition, Vol. III, 373–455. Hillsdale NJ: Lawrence Erlbaum. Ervin-Tripp, Susan M. 1977 Wait for me roller skate. In: Susan M. Ervin-Tripp and Claudia Mitchell-Kernan (eds.), Child Discourse, 165–188. New York: Academic Press. Ervin-Tripp, Susan M. and Claudia Mitchell-Kernan 1977a Introduction. In: Susan M. Ervin-Tripp and Claudia Mitchell-Kernan (eds.), Child Discourse, 1–26. New York: Academic Press. Ervin-Tripp, Susan, and Claudia Mitchell-Kernan (eds.) 1977b Child Discourse. Academic Press, New York. Ervin-Tripp, Susan M., Amy Strage, Martin Lampert and Nancy Bell 1986 Understanding requests. Linguistics 25: 107–143. Escandell-Vidal, Victoria 2004 Norms and principles: Putting social and cognitive pragmatics together. In: Rosina Márquez-Reiter and María Elena Placencia (eds.) Current Trends in the Pragmatics of Spanish, 347–371. Amsterdam – Philadelphia: John Benjamins. Faerch, Claus and Gabriele Kasper 1984 Pragmatic knowledge: rules and procedures. Applied Linguistics 5: 214–25. Fecica, Agnieszka M. and Daniela K. O’Neill 2010 A step at a time: Preliterate children’s simulation of narrative movement during story comprehension. Cognition 116: 368–381.

282

Daniela K. O’Neill

Filippova, Eva and Janet W. Astington 2008 Further development in social reasoning revealed in discourse irony understanding. Child Development 79: 126–138. Foster, Susan H. 1990 The Communicative Competence of Young Children. Harlow, UK: Longman Group. Fujiki, Martin, and Bonnie Brinton 2009 Pragmatics and social communication in child language disorders. In: Richard G. Schwartz (ed.), Handbook of Child Language Disorders, 496–423. New York: Psychology Press. Gallagher, Tanya M. 1991a A retrospective look at clinical pragmatics. In: Tanya M. Gallagher (ed.), Pragmatics of Language: Clinical Practice Issues, 1–9. New York: Chapman and Hall. Gallagher, Tanya M. 1991b Language and social skills: Implications for assessment and intervention with school-age children. In: Tanya M. Gallagher (ed.), Pragmatics of Language: Clinical Practice Issues, 11–41. New York: Chapman and Hall. Garvey, Catherine 1977 Play. Cambridge, MA: Harvard University Press. Gee, John Paul 2004 Situated Language and Learning: A Critique of Traditional Schooling. London: Routledge. Gibbs, Raymond W. 1987 Linguistic factors in children’s understanding of idioms. Journal of Child Language 14: 569–586. Girouard, Pascale, C., Marcelle Ricard and Therese Gouin Decarie 1997 The acquisition of personal pronouns in French-speaking and English-speaking children. Journal of Child Language 24: 311–326. Goffman, Erving 1981 Forms of talk. Oxford: Blackwell. Green, Georgia M. 1989 Pragmatics and Natural Language Understanding. 2nd edition. Hillsdale, New Jersey: Erlbaum. Grice, H. Paul 1975 Logic and conversation. In: Peter Cole and Jerry L. Morgan (eds.), Syntax and semantics 3: Speech acts. 225–242. New York: Academic Press. Grice, H. Paul 1989 Studies in the Way of Words. Cambridge, MA: Harvard University Press. John J. Gumperz 1982 Discourse Strategies. Cambridge, UK: Cambridge University Press. Gumperz, John J. and Dell Hymes 1964 The ethnography of communication. American Anthropologist 66b, Pt. 2 [Special Publication.] Harris, Paul L., Marc de Rosnay and Francisco Pons 2005 Language and children’s understanding of mental states. Current Directions in Psychological Science 14: 69–73.

Components of pragmatic ability and children's pragmatic language development

283

Heath, Shirley Brice 1983 Ways with Words: Language, Life and Work in Communities and Classrooms. Cambridge, UK: Cambridge University Press. Hoff-Ginsberg, Erika 1996 Language Development. Pacific Grove, CA: Thompson Brooks/Cole Hollos, Marida and William Beeman 1978 The development of directives among Norwegian and Hungarian children: An example of communicative style in culture. Language in Society 7: 345–355. Huang, Yi Ting and Jesse Snedeker 2009 Semantic meaning and pragmatic interpretation in 5-year-olds: Evidence from real-time spoken language comprehension. Developmental Psychology 45: 1723–1739. Hymes, Dell 1972 On communicative competence. In: John B. Pride and Janet Holmes (eds.), Sociolinguistics, 269–293. London: Penguin. Jaswal, Vikram J. and Lauren S. Malone 2007 Turning believers into skeptics: 3-year-olds’ sensitivity to cues to speaker credibility. Journal of Cognition & Development 8: 263–283. Koenig, Melissa A. and Paul H. Harris 2005 Preschoolers mistrust ignorant and inaccurate speakers. Child Development 76: 1261–1277. Kyratzis, Amy 2001 Emotion talk in preschool same-sex friendship groups: Fluidity over time and context. Early Education and Development 12: 359–392. Kyratzis, Amy, Traci Marx and Evelyn Wade 2001 Preschoolers’ communicative competence: register shift in the marking of power in different contexts of friendship group talk. First Language 21: 387–431. Levinson, Stephen C. 2000 Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge, MA: Harvard University Press. Levy, Elena T. 2003 The roots of coherence in discourse. Human Development 46: 169–188. Lucariello, Joan and Catherine Mindolovich 1995 The development of complex metarepresentational reasoning: The case of situational irony. Cognitive Development 10: 551–576 Markman, Ellen M. 1989 Categorization and Naming in Children: Problem of Induction. Cambridge, MA: MIT Press. McGhee, Paul E. 1979 Humor: Its Origin and Development. San Francisco: W. H. Freeman and Company. McTear, Michael F. and Gina Conti-Ramsden 1992 Pragmatic Disability in Children. London: Whurr. Mervis, Carolyn B. 1987 Child-basic object categories and early lexical development. In: Ulrich Neisser (ed.), Concepts and Conceptual Development: Ecological and Intellectual Factors in Categorization, 201–233. Cambridge, UK: Cambridge University Press.

284

Daniela K. O’Neill

Nakamura, Keiko 2001 The acquisition of polite language in Japanese children. In Keith E. Nelson and A. Aksu-Koç (eds.), Children’s Language, Vol X, 93–112. Mahwah, NJ: Lawrence Erlbaum. Nelson, Katherine 1981 Individual differences in language development: Implications for development and language. Developmental Psychology 17: 170–187. Nelson, Katherine 1989 Narratives from the Crib. Cambridge, MA: Harvard University. Ninio, Anat and Catherine E. Snow 1996 Pragmatic Development. Boulder, CO: Westview Press. Nippold, Marilyn A. 1998 Later Language Development: The School-Age and Adolescent Years. 2nd edition. Austin, TX: Pro-Ed. Ochs, Elinor 1988 Culture and Language Development: Language Acquisition and Language Socialization in a Samoan Village. Cambridge, UK: Cambridge University Press. O’Neill, Daniela K. 1996 Two-year-olds’ sensitivity to a parent’s knowledge state when making requests. Child Development 67: 659–677. O’Neill, Daniela K. 2005 Talking about “new” information: The given/new distinction and children’s developing theory of mind”. In: Janet W. Astington and Jodie Baird (eds.), Why Language Matters for Theory of Mind, 84–105. Oxford: Oxford University Press. O’Neill, Daniela K. 2007 The Language Use Inventory for Young Children: A parent-report measure of pragmatic language development for 18–47-month-old children”. Journal of Speech, Language and Hearing Research 50: 1–15. O’Neill, Daniela K. 2009 Language Use Inventory: Manual. Waterloo, ON, Canada: Knowledge in Development. O’Neill, Daniela K. and Francesca Happé 2000 Noticing and commenting on what’s new: differences and similarities among 22-month old typically developing children, children with Down syndrome, and children with autism. Developmental Science 3: 457–478. O’Neill, Daniela K. and Amanda Holmes 2002 Young preschoolers’ ability to reference story characters: the contribution of gesture and character speech. First Language 22: 73–103. O’Neill, Daniela K., Rebecca M. Main and Renata Ziemski. 2009 “I like Barney.” Preschoolers’ spontaneous conversational initiations with peers. First Language 29: 401–425. O’Neill, Daniela K., Michelle J. Pearce and Jennifer L. Pick 2004 Predictive relations between aspects of preschool children’s narratives and performance on the Peabody Individualized Achievement Test – Revised: Evidence of a relation between early narrative and later mathematical ability. First Language 24: 149–183.

Components of pragmatic ability and children's pragmatic language development

285

Papafragou, Anna and Niki Tantalou 2004 The computation of implicatures by young children. Language Acquisition 12: 71–82. Pasquini, Elisabeth S., Kathleen H. Corriveau, Melissa A. Koenig and Paul H. Harris 2007 Preschoolers monitor the relative accuracy of informants. Developmental Psychology 43: 1216–1226. Perkins, Michael 2007 Pragmatic Impairment. Cambridge: Cambridge University Press. Perkins, Michael R. 2009 Pragmatic impairment as an emergent phenomenon. In: Martin J. Ball, Michael R. Perkins, Nicole Muller and Sara Howard (eds.), The Handbook of Clinical Linguistics, 79–91. Oxford: Blackwell. Perner, Josef 1991 Understanding the Representational Mind. Cambridge, MA: MIT Press. Perner, Josef and Heinz Wimmer 1985 “John thinks that Mary thinks that …”: Attribution of second-order beliefs by 5- to 10-year-old children. Journal of Experimental Child Psychology 39: 437–471. Pons, Francisco, Paul L. Harris and Marc de Rosnay 2004 Emotion comprehension between 3 and 11 years: Developmental periods and hierarchical organization. European Journal of Developmental Psychology 1(2): 127–152 Prutting, Carol A. 1982 Pragmatics as social competence. Journal of Speech and Hearing Disorders 47: 123–134. Reddy, Vasudevi 1991 Playing with others’ expectations: teasing and mucking about in the first year. In: Andrew Whiten (ed.), Natural Theories of Mind, 143–158. Oxford: Blackwell. Rall, Jaime and Paul L. Harris 2000 In Cinderella’s slippers? Story comprehension from the protagonist’s point of view. Developmental Psychology 36: 202–208 Rees, Norma S. 1982 An overview of pragmatics or what is in the box? In: J. V. Irwin (ed.), Pragmatics: The Role in Language Development, 1–13. LaVerne, CA: Fox Point. Ryckebusch, Celine and Haydee Marcos 2004. Speech acts and context in young children: influence of parent gender and type of play from 1;5 to 2;3. Journal of Pragmatics 36: 883–897. Saylor, Megan M., Jodie A. Baird and Catherine Gallerani 2006 Telling others what’s new: Preschoolers’ adherence to the given-new contract. Journal of Cognition & Development 7: 341–379. Schiefelbusch, Richard L. and Joanne Pickar 1984 The Acquisition of Communicative Competence. Baltimore, MD: University Park Press. Searle, John R. 1969 Speech Acts: An Essay in the Philosophy of Language. Cambridge: Cambridge University Press.

286

Daniela K. O’Neill

Shatz, Marilyn 1994 A Toddler’s Life: Becoming a Person. New York: Oxford University Press. Sherrod, Lonnie and Michael Lamb 1981 Infant social cognition: An introduction. In: Michael Lamb and Lonnie Sherrod (eds.), Infant social cognition: Empirical and Theoretical Considerations, 1–10. Hillsdale, NJ: Erlbaum. Short-Meyerson, Katherine J. 2010 Preschoolers’ establishment of mutual knowledge during scripted interactions. First Language 30: 219–236. Short-Meyerson, Katherine J. and Leonard Abbeduto 1997 Preschoolers’ communication during scripted interactions. Journal of Child Language 24: 469–493. Shwe, Helen and Ellen M. Markman 1997 Young children’s appreciation of the mental impact of their communicative signals. Developmental Psychology 33: 630–636. Slobin, Dan I. 1985 Cross-linguistic evidence for the language-making capacity. In: Dan I. Slobin (ed.), The Crosslinguistic Study of Language Acquisition, Vol. 2: Theoretical Issues, 1157–1256. Hillsdale, NJ: Lawrence Erlbaum. Slobin, Dan I., Julie Gerhardt, Amy Kyratzis and Jiansheng Guo (eds.) 1996 Social Interaction, Social Context, and Language: Essays in Honor of Susan Ervin-Tripp. Mahwah, NJ: Lawrence Erlbaum. Smith, Benita R. and Eva Leinonen 1992 Clinical Pragmatics: Unravelling the Complexities of Communicative Failure. London: Chapman and Hall. Sperber, Dan and Deidre Wilson 1986/95 Relevance: Communication and Cognition. Cambridge: Harvard University Press. [Second edition with Postface 1995] Sperber, Dan and Deidre Wilson 1987 Precis of Relevance: Communication and Cognition. Behavioral and Brain Sciences 10: 697–754. St. James, Paula and Helen Tager-Flusberg 1994 An observational study of humor in autism and down syndrome. Journal of Autism and Developmental Disorders 24: 603–617. Thompson, Linda 1996 The development of pragmatic competence: Past findings and future directions for research. In: Linda Thompson (ed.), Children Talking: The Development of Pragmatic Competence, 3–21. Philadelphia, PA: Multilingual Matters. Tomasello, Michael 1988 The role of joint attentional process in early language acquisition. Language Sciences 10: 69–88. Tomasello, Michael 1992 The social bases of language acquisition. Social Development 1: 67–87. Tomasello, Michael 1999 The Cultural Origins of Human Cognition. Cambridge, MA: Harvard University Press.

Components of pragmatic ability and children's pragmatic language development

287

Tomasello, Michael 2003 Constructing a Language: A Usage-Based Theory of Language Acquisition. Cambridge, MA: Harvard University Press. Tomasello, Michael and Michelle E. Barton 1994 Learning words in nonostensive contexts. Developmental Psychology 30: 639–650. Tomasello, Michael and Ann C. Kruger 1992 Joint attention on actions: Acquiring verbs in ostensive and non-ostensive contexts. Journal of Child Language 19: 311–334. Tyack, Dorothy and David Ingram 1977 Children’s production and comprehension of questions. Journal of Child Language 4: 211–224. Vega-Moreno, Rosa Elena 2005 Idioms, transparency and pragmatic inference. UCL Working Papers in Linguistics 17: 389–425. Verschueren, Jeff 1999 Understanding Pragmatics. New York: Arnold. Wetherby, Amy M. 1991 Profiling pragmatic abilities in the emerging language in young children. In: Tanya M. Gallagher (ed.), Pragmatics of Language: Clinical Practice Issues, 249–281. New York: Chapman and Hall. Wetherby, Amy M., Barry M. Prizant and Thomas A. Hutchinson 1998 Communicative, social/affective, and symbolic profiles of young children with autism and pervasive developmental disorders. American Journal of SpeechLanguage Pathology 7: 79–91. Wieman, Leslie A. 1976 Stress patterns of early child language. Journal of Child Language 3: 283–286. Winner, Ellen 1988 The Point of Words: Children’s Understanding of Metaphor and Irony. Cambridge, MA: Harvard University Press Wode, Henning 1971 Some stages in the acquisition of questions by monolingual children. Word 27: 261–310.

Disorder

11.

Pragmatic disorders Louise Cummings

1.

Introduction

Pragmatic disorders have been an active area of clinical linguistic investigation for nearly forty years. During that time, the field has undergone considerable expansion and transformation, so much so in fact that it is doubtful if early investigators would even recognize the range of clinical populations and pragmatic phenomena that are now routinely investigated under its auspices. Notwithstanding the considerable research activity that has been directed towards the study of pragmatic disorders, there is much that has been problematic in the way investigators have examined these disorders. Clear theoretical frameworks and sound conceptual understanding have often been lacking from the investigations which have been undertaken. Studies have typically concentrated on particular clinical populations to the almost total exclusion of other populations. Certain pragmatic phenomena (e.g. speech acts and implicatures) have been extensively studied while other pragmatic behaviours (e.g. deictic and presuppositional phenomena) have been largely neglected. These various weaknesses in clinical pragmatic studies are not just academic points. For they have significant consequences in terms of how we assess and treat child and adult clients with pragmatic disorders (Cummings 2007a). There is thus an overriding clinical imperative to address these difficulties in a timely fashion. For our current purposes, we explore them in further detail with a view to delineating the content of the present chapter. For much of the relatively brief history of clinical pragmatics, the field has been characterized by a lack of theoretical direction and conceptual confusion concerning what constitutes a pragmatic disorder. Much of this conceptual disarray stems from a similar confusion within pragmatics itself about the scope of this discipline (Cummings 2007a). To the extent that workers in pragmatics have failed to circumscribe this field of study, it is unsurprising that clinical practitioners and researchers should display certain misunderstandings about the type of linguistic and non-linguistic impairments that can appropriately be called “pragmatic”. Many studies of pragmatic disorders have also exhibited a grievous lack of theoretical direction. More often than not, studies have been undertaken simply because they can be done, rather than for their ability to address theoretically interesting questions. The reason so many clinical pragmatic studies have not been theoretically motivated is related in large part to the fact that the hitherto dominant philosophical approach to theorising in pragmatics is poorly equipped to characterize the cognitively oriented considerations that are at issue in pragmatic disorders (Cummings

292

Louise Cummings

2009). In recent years, there is evidence of a growing appreciation on the part of investigators of the need to use theoretical frameworks to steer clinical pragmatic studies. In this way, Sperber and Wilson’s (1986/1995) relevance theory has been used in investigations of children with specific language impairment1 (Schelleter and Leinonen 2003), children with Asperger syndrome and high-functioning autism (Loukusa et al. 2007) and adults with right-hemisphere damage (Dipper et al. 1997), amongst other clinical populations (see Cummings 2010 for further discussion). Yet, such theoretically motivated studies are still very much in a minority. The basis of the lack of theoretical direction in clinical studies of pragmatics and possible responses to it have been examined in Cummings (2009), to which the reader is referred for further discussion. Although a large range of clinical populations experience pragmatic disorders, not all these populations have been the focus of equally intensive investigation. As is often the case in clinical linguistic studies, investigations of pragmatic disorders in child clients have tended to outnumber those that have been conducted in adults.2 By far the most studied populations of child clients are those with autistic spectrum disorders3 (ASDs; see Reboul, Manificat and Foudon, this volume) and children with developmental language disorders (see Cummings 2008 for an overview of the communication features of both populations). For autistic children who develop verbal communication skills,4 pragmatics can be severely disordered and a source of considerable disability. Much is now known about the impaired pragmatic performance of autistic clients, as will be seen from the studies that are reviewed in subsequent sections. Developmental language disorders have been the focus of intensive clinical pragmatic study since Rapin and Allen (1983) in the US and Bishop and Rosenbloom (1987) in the UK first characterized a group of children who exhibited pragmatic deficits in the presence of relatively intact structural language skills. Labelled as having semantic-pragmatic disorder,5 these children have continued to raise difficult diagnostic issues for researchers and clinicians (see Bishop 2000 and Bishop 2003 for discussion of these issues). The predominance of these clinical populations has served to overshadow a number of other child clients with significant pragmatic impairment. In this way, pragmatic deficits in children with a range of emotional and behavioural disorders are only now beginning to be identified and characterized (see Cummings 2009 for an extended discussion of pragmatic deficits in these disorders). Clearly, the traditionally narrow focus of research on certain children with pragmatic disorders must now be extended to include other clinical populations. Pragmatic disorders in adult clients have been somewhat less intensively studied than those in children. The relative neglect of adult populations has been motivated in no small part by a number of rather unhelpful assumptions about pragmatic functioning in adult clients. Chief amongst them is the idea that pragmatic deficits are secondary to problems in structural language, most notably in aphasia,6 and that these deficits will resolve (and hence require no special consideration) as impairments in syntax and semantics show improvement. However,

Pragmatic disorders

293

studies are increasingly showing that this assumption is ill-founded. In some cases, pragmatic language impairments have been shown to persist despite improvements in structural language in aphasic clients (Coelho and Flewellyn 2003). In other cases, pragmatic impairments have been demonstrated in the extralinguistic communication of subjects with left-hemisphere damage (Cutica et al. 2006). A further complicating factor is that aphasia batteries and other language assessments for use with adults are often poorly equipped to assess features such as context and mutual knowledge that are integral to the pragmatic interpretation of utterances. Adult clients may be seen to succeed in the single word and sentence testing formats used in language assessments and yet still exhibit pragmatic impairment. This is particularly true in the case of high-functioning adults with traumatic brain injury7 who may pass standardized language assessments and yet exhibit subtle communicative, primarily pragmatic, deficits (Coelho et al. 1991). This recognition has led to a growing interest in assessment and treatment methods that focus on discourse and conversation (see Beeke, this volume). With these distorting assumptions about pragmatics in adult clinical populations exposed, investigators can begin to see the real extent of pragmatic deficits in adult clients (see Cummings 2007b for further discussion of these deficits). Finally, not all pragmatic aspects of language have received the attention of clinical investigators to equal extents. Speech acts, implicatures and the comprehension of non-literal language (irony, metaphor, idiom and proverb) rank amongst the most extensively investigated aspects of pragmatics in clinical studies. However, investigators have seldom studied the use and comprehension of deictic expressions or presuppositional phenomena in language. One of the principal reasons for these discrepancies is that some pragmatic phenomena are more readily assessed than others in clinical settings. For example, a creative clinician can easily prompt a child to produce the speech act of request by placing a favourite toy out of the child’s reach but within his or her view. An adult’s understanding of irony can be assessed by asking a client to select an interpretation of an utterance that best captures a depicted scenario (e.g. the speaker who says What a delightful child! in the presence of a disruptive young boy). It is altogether more difficult to assess the use of presuppositional phenomena using similar methods. Shared knowledge in conversation can be presupposed by a speaker and his listener through the use of certain lexical choices and syntactic constructions. A speaker who fails to make use of these linguistic devices and proceeds to communicate explicitly to his listener that which can be presupposed will produce verbose messages that are peculiarly under-informative. Presuppositional failure of this type can best be gleaned from qualitative analyses of conversation. However, the additional insight that one gains into a client’s pragmatic disorder through these analyses is offset by the fact that they are particularly burdensome in terms of clinical time and effort.8 These practical issues are difficult to overcome and almost certainly guarantee that the inequalities in investigation of pragmatic phenomena to date will continue.

294

Louise Cummings

With these various caveats in place, I proceed to discuss what is known about pragmatic disorders in children and adults. In this discussion, we will examine deficits in some well recognized pragmatic phenomena (e.g. speech acts, implicatures) as well as phenomena that are less often discussed in relation to pragmatic disorders (e.g. context, shared knowledge, inferencing). These pragmatic aspects of language have been the focus of experimental procedures as well as studies of discourse and conversation. In what follows, we report findings from all three main types of investigation. For further discussion of each of these pragmatic features, the reader is referred to Cummings (2005) and Cummings (2009).

2.

Speech acts

Speech acts exemplify the complex linguistic and cognitive processes that are involved in pragmatic interpretation. The speaker who produces a request, issues a warning or commits himself in a promise to undertake some action must assess the social context in which he is communicating in order to decide if an indirect speech act is warranted over a direct speech act. He must then make certain lexical and syntactic choices to reflect his assessment of the formality of the context in which he is operating. For example, in order to make an indirect request in a formal context, the speaker must choose a syntactic construction in which the subject pronoun and auxiliary verb have been inverted (Can you open the window?). He must satisfy himself and his addressee that he fulfils all the felicity conditions that attend the performance of the speech act in question. In this way, he can only felicitously promise to undertake a certain action if he entertains a sincere intention to execute that action (a sincerity condition on the speech act of promising). Finally, as the recipient of a speech act, he must possess the cognitive skills, particularly theory of mind (ToM) skills (cf. Reboul, Manificat and Foudon, this volume), which are needed to establish the communicative intention that lies behind a speaker’s use of a certain utterance. He must be able to ascertain, for example, that the speaker who utters Can you open the window? is making a request to have the window opened and not asking a question about someone’s ability to open the window. With such an array of linguistic and cognitive skills to coordinate, it is unsurprising that the use and understanding of speech acts are impaired in clients with pragmatic disorders. The child with specific language impairment (SLI) and the adult with non-fluent aphasia, for example, may lack the syntactic and semantic skills that are necessary to produce an indirect request. The adult with schizophrenia9 may fail to adequately assess the politeness constraints at work in a particular context or the formality of a situation and may choose a more direct speech act than is warranted in the circumstances. The adult with mental retardation may not appreciate the obligations to which he or she becomes committed through the speech act of promising and may consequently misuse this speech act. The autistic

Pragmatic disorders

295

child with impaired theory of mind skills may be unable to recognize the communicative intention that lies behind an utterance and may as a result treat an indirect request as a question or a warning as a threat. The impaired capacities that underlie deficits in the use and/or understanding of speech acts are numerous and must be considered within an assessment of this particular pragmatic disorder. However, it is all too often the case that studies of speech act use and understanding in clients are limited in this regard. One study that does seek to relate speech act performance in children with ASD to underlying cognitive functioning was undertaken by Ziatas et al. (2003). These investigators examined assertive speech acts in autistic children and children with Asperger’s syndrome. Children with specific language impairment and normally developing children acted as comparison groups. It was found that autistic children used significantly lower proportions of assertions involving explanations and descriptions than the SLI or normally developing children. Autistic children also used significantly lower proportions of assertions involving internal states and explanations than the children with Asperger’s syndrome. When mental assertions were analyzed further, it was found that children with autism and Asperger’s syndrome referred predominantly to desire and made few references to thought and belief. SLI and normally developing children, however, used a higher proportion of references to thought and belief. Ziatas et al. relate these findings to ToM impairments in the autistic children. Some studies have revealed relatively intact speech act performance in child subjects. Abbeduto et al. (1988) examined the development of speech act comprehension in mentally retarded and nonretarded school age children. Despite having serious deficits in receptive linguistic competence, retarded subjects were able to use contextual and linguistic clues to achieve comprehension of target speech acts and follow an answer obviousness rule (where a question had an obvious answer, it was treated as a directive). Rollins et al. (1994) examined speech act use in five SLI children who were matched to their younger siblings on mean length of utterance. For both groups of children, the mean percentage of communicative acts for directives, statements and questions was roughly equivalent. However, SLI children used a lower percentage of commitments and a higher percentage of acts in an “other” category than their normal language siblings.10 Speech act performance has been investigated in a number of adult clinical populations. Drummond and Simmons (1995) examined the use of speech acts in two fluent and two nonfluent female aphasic subjects during group interaction. A relatively greater number of speech acts were produced by the nonfluent subjects than by the fluent subjects. The “answer” and “comment” speech act categories were used more often than other categories by both types of aphasic subject. Each subject in the study used at least four to seven types of speech acts. Drummond and Simmons (1995: 53) state that “[t]his finding corroborates previous reports indicating the use of a greater variety of speech acts in group interactions”. Ripich et al.

296

Louise Cummings

(1991) examined speech act patterns in the dyadic interactions between an examiner and eleven subjects with senile dementia of Alzheimer’s type (SDAT).11 More requestives and fewer assertives were used by SDAT subjects than by eleven normal elderly subjects. McDonald and Pearce (1998) studied the performance on a request production task of fifteen adult male subjects with acquired brain damage. This task required subjects to produce nonconventional requests that would overcome reluctance on the part of the listener. Not only were brain-injured subjects less likely than controls to produce requests that addressed the obstacle to listener compliance, but they were also less likely to produce “other” kinds of strategies and were more inclined to produce requests that involved counterproductive comments. The performance of these brain-injured subjects was shown to correlate with measures of executive function. Vanhalle et al. (2000) found that subjects with right-hemisphere damage processed speech acts as efficiently as controls in a natural or pseudo-natural situation. However, their processing of direct and indirect speech acts was significantly worse than controls when the communicative context was non-natural.

3.

Implicatures

Implicatures are another extensively investigated phenomenon in clinical pragmatic literature. Although a range of different types of implicatures exist (see Cummings 2005, and Moeschler, this volume, for discussion), particularized conversational implicatures are more often than not the focus of investigation by clinical researchers. These implicatures are generated through the operation of Grice’s principle of cooperation, in which an assumption of speaker and listener cooperation in a conversational exchange, combined with the flouting of one or more maxims, permits a conversational participant to derive the intended implicature of an utterance. In this way, the speaker who asks Are you coming to the pub tonight? may be greeted by the response Paul’s parents are visiting this evening. The speaker knows that the listener is being cooperative in the response she has produced and he uses this assumption of cooperation to determine the relevance of this response. Clearly, if someone’s parents-in-law are visiting, one cannot simultaneously join friends at the pub. The intended implicature of the listener’s response, and the implicature that the questioner will derive, is that the listener will not be going to the pub. This type of implicature is typically tested in studies through the presentation of conversational vignettes which are often accompanied by pictures. Subjects are asked to indicate the intended implication of a speaker’s utterance or are required to select from a number of interpretations that which is most likely to be intended by the speaker. Test materials can be manipulated in order that the full range of Gricean maxims – quality, quantity, relation and manner – can be examined either individually or in combination. A similar Gricean framework can be

Pragmatic disorders

297

used to capture various expressive pragmatic problems such as the tangential linguistic output of schizophrenic speakers, which can be explained as a failure of conversational relevance. The recognition in recent years of a severe and persistent ToM deficit in autistic children, and a growing awareness of the role of mindreading skills in pragmatic interpretation, have combined to make implicature the pragmatic concept of choice for ToM researchers. Surian (1996) examined the detection of utterances that violate Grice’s maxims by children with autism. While most of the autistic children in the study performed at chance on this detection task, SLI and normal children all performed above chance. Moreover, the performance of autistic children on the detection task was related to their ability to attribute false beliefs (false belief tests are typically used to assess ToM performance in autistic children). Scalar implicatures were examined by Pijnacker et al. (2009) in a study of high-functioning adults with autistic disorder and Asperger syndrome.12 The ASD subjects in this study were just as likely as controls to derive scalar implicatures in response to underinformative sentences like Some sparrows are birds. Ryder et al. (2008) examined the comprehension of implicatures in 27 children with SLI, nine of whom were pragmatically impaired. These investigators found that when the answer was provided by pictorial context, the SLI children performed similarly to their typically developing peers in using context to derive implicatures. The SLI children were not yet competent at using verbal context to answer questions that required implicatures to be drawn. The subgroup of SLI children with pragmatic language impairment performed significantly more poorly than the rest of the SLI group on these questions. The children with pragmatic language impairment were found to be accurately identified from the rest of the SLI group on the basis of performance scores on the implicature questions. Several studies have been conducted of implicature processing in adult subjects. Kasher et al. (1999) examined the processing of implicatures in 31 patients with left-hemisphere damage (LHD) following a stroke. Implicatures of quantity, quality, relation and manner were examined by means of two-sentence conversational vignettes which were literally problematic. Famous paintings, which were also literally problematic, were used to examine non-verbal implicatures. Subjects with LHD were significantly impaired relative to age-matched normal controls in implicature processing. Verbal and non-verbal implicatures intercorrelated highly in LHD subjects. Tényi et al. (2002) examined the ability of schizophrenic subjects to recognize the intended meaning behind violations of Gricean maxims. Twentysix paranoid schizophrenic subjects and 26 normal controls were presented with four question and answer vignettes in which the maxim of relevance was violated. Subjects had to identify the speaker’s intended meaning in each case. Tényi et al. found that schizophrenic subjects made significantly more errors than controls in identifying the communicative intentions that lay behind violations of this maxim. Corcoran and Frith (1996) examined politeness and appreciation of the Gricean

298

Louise Cummings

maxims of quantity, quality and relation in schizophrenic patients with different symptom profiles. Subjects had to select an appropriate final piece of speech for one of the characters in a series of stories. One piece of speech adhered to the rule under question, while the other flouted the rule. Control subjects, schizophrenic subjects with paranoid delusions and schizophrenic subjects with negative symptoms adhered to the maxim of relation. However, all other maxims were flouted by subjects with negative symptoms. Subjects with paranoid delusions often failed to respond in a polite fashion, but performed at a similar level to controls on stories involving the Gricean maxims.

4.

Deixis

Deictic expressions are an integral component of the pragmatic competence of language users. Expressions in language identified as deictic “describe entities within the wider social, linguistic or spatiotemporal context of an utterance” (Cummings 2005: 22). In this way, a speaker may signal that a particular person is the intended recipient of his remarks through uttering you will stay after class (personal deixis), that a certain physical location is indicated in John lives here (spatial deixis), that an event will take place at a specific point in time in Susie leaves next week (temporal deixis), that a particular extract of language is indicated in I rebutted this argument in the last chapter (discourse deixis) and that a certain social relationship is indicated through German Wie heißen Sie? (social deixis). As these examples demonstrate, a large range of expressions have a deictic function including personal pronouns (e.g. I, you), calendrical terms (e.g. today, tomorrow), adverbs (e.g. here, there), demonstratives (e.g. this, that), adjectives (e.g. next, last) and verbs (e.g. come, bring). A number of these expressions can serve more than one deictic function. For example, in the utterances I examine this claim in the next chapter and He departs for Rome next week, the adjective next serves a discourse deictic and temporal deictic function, respectively. In order to use these expressions appropriately in conversation, speakers must do more than simply make certain lexical selections. They must be able to establish salient features across different types of context and encode these features in language. It is to be expected, therefore, that the pragmatically impaired client will experience difficulties with the types of expressions that pragmatists discuss under deixis. Notwithstanding this expectation, deixis has been almost entirely neglected by practitioners and researchers in clinical pragmatics. This is particularly surprising given reports of pronoun problems in a clinical population with significant pragmatic disorder, namely, children with autism. Jordan (1989) examined the use and understanding of the personal pronouns you and me in 11 autistic children. There were no significant differences between these children and normally developing children and mentally handicapped children matched by receptive vocabulary age

Pragmatic disorders

299

on the comprehension of these pronouns. However, only two autistic children displayed the same pattern of responses using these pronouns as all but four of the 22 control children in the study. Autistic children used incorrect case pronouns or proper names for self or other reference. Interestingly, they displayed almost no pronoun reversal.13 Jordan (1989: 169) interprets these findings “in the light of significant difficulties in the acquisition of person deixis in autistic children”. In one of the few other clinical studies to investigate deixis, Varley (1993) examined the use of deictic forms in the conversational speech of fluent and non-fluent aphasics and subjects with right-hemisphere damage. Significantly more deictic expressions were used by fluent aphasics than by other clinical groups and by a group of nonbrain-damaged control subjects. Non-fluent aphasics used significantly fewer deictic forms than these other groups. Deictic use did not correlate significantly with performance on lexical tests. Nor was there an inverse relationship between deictic use and number of clause elements at the level of the group. Clearly, further clinical studies of deixis must be undertaken before any conclusions can be drawn about the role of this concept in pragmatic disorder.

5.

Presuppositions

Levinson (1983: 168) states that “[t]he technical sense of presupposition is restricted to certain pragmatic inferences or assumptions that seem at least to be built into linguistic expressions”. Some of these “linguistic expressions” are factive verbs (e.g. she regretted divorcing her husband presupposes that she divorced her husband), comparisons of equality (e.g. Jill is as talkative as Pat presupposes that Pat is talkative), implicatives (e.g. the students managed to solve the equation presupposes that the students tried to solve the equation) and change-of-state verbs (e.g. have you stopped fox hunting? presupposes that the addressee did at one time hunt foxes). Although presuppositions per se have seldom been the focus of clinical studies, a presuppositional ability is necessary to other pragmatic phenomena that have been examined by investigators. During discourse production, for example, a speaker must make a number of judgements about the information that should be explicitly conveyed to a listener (new information) and the information that the listener already knows (given or old information) and that it would be inefficient for the speaker to directly convey. These judgements made, presupposition becomes the speaker’s principal mechanism for dealing with given information.14 The use of the definite article the in the utterance the party will start at 8pm sharp reflects a belief on the part of the speaker that the listener knows the specific referent of the noun accompanying the definite article. The cleft construction it was Sarah who drowned in the lake presupposes the speaker’s belief that the listener knows someone drowned in the lake, while at the same time conferring prominence on the individual, Sarah, who suffered this particular fate by placing

300

Louise Cummings

her name before the tragic event in question. Speakers with intact pragmatic skills can manage given and new information through the manipulation of presuppositions. It is to be expected that these same presuppositional processes will be problematic for children and adults with pragmatic disorder. This expectation has been borne out to some degree by the results of studies, principally undertaken in children. The ability of 24 children (age 4 to 17 years) with unilateral left or right hemisphere damage to presuppose the truth of factive sentences and to infer the truth or falsity of implicative sentences was examined by Eisele et al. (1998). Presupposition and implication performance were deficient in left lesion subjects compared to age-matched controls. A more selective deficit was observed in right lesion subjects with an impairment limited to implication but not presupposition. Rowan et al. (1983) examined the presuppositional abilities of 18 language-disordered children who were functioning at a single word level of linguistic development. Like normal children, the language-disordered children in this study displayed a tendency to encode changing rather than unchanging situational elements. Presuppositional behaviours were usually in the form of word productions for both groups of children. Skarakis and Greenfield (1982) examined the marking of new and old information in the verbal expression of language-disordered children who were beyond the one-word stage. These children selectively marked new information, as did normal children. Moreover, they followed the same developmental sequence of strategies as normal children for de-emphasizing old information – at a mean length of utterance (MLU) of 3, children omitted old information, while at a MLU of 5, they pronominalized it. Over half of the language-disordered children pronominalized old information more frequently than normal subjects. McCaleb and Prizant (1985) examined the encoding of new and old information by four autistic children with MLUs of 1.96 to 2.82. Although all four subjects were able to encode new information through lexicalization in singleword utterances and contrastive stress in multiword utterances, they encoded old information almost as frequently as new information and encoded a new action or state change relatively infrequently. The use of contrastive stress by autistic children to mark the topic-comment distinction has also been examined by Baltaxe (1984).

6.

Non-literal language

Irony, metaphor, idiom and proverb are frequently topics of investigation in clinical pragmatic studies. An expression’s ironic, metaphorical, idiomatic and proverbial meaning is not the result of a compositional analysis of the literal meanings of the component parts of the expression, but instead draws upon a range of other cognitive and language skills. The speaker who utters what glorious weather we’re having in the middle of a snow storm is clearly doing so with considerable ironic

Pragmatic disorders

301

intent. To arrive at this ironic interpretation, a listener must be able to attribute a communicative intention to the speaker that has a content which is the exact opposite of the meaning expressed by his utterance (i.e. the speaker wants to communicate that he does not believe that the weather is glorious). This same mindreading ability is integral to the understanding of metaphorical language. The grandmother who says the children were angels during their stay does not intend to communicate that the children were angels, merely that they exhibited certain attributes of angels – their behaviour was blameless, they were kind, etc. To arrive at this metaphorical meaning, a listener must know something of the thoughts and feelings of this grandmother towards her grandchildren (one can envisage a scenario in which, if a different set of thoughts and feelings were attributed to the grandmother, her utterance could be interpreted as having ironic meaning rather than metaphorical meaning). Also, the speaker who utters Sally always spills the beans and as usual, Fred was last to hit the sack is intending to communicate something about Sally’s lack of discretion and Fred’s bedtime routine, respectively. Similarly, expressions such as a stitch in time saves nine and John is caught between a rock and a hard place are not communicating anything about sewing and rocks, but rather that dealing with things at the time prevents them escalating into bigger problems and that John is confronting two equally undesirable options. Perhaps unsurprisingly, these different types of non-literal language have been shown to be problematic for children and adults with pragmatic disorder. Idiom comprehension has been studied in adolescents with language-based learning disabilities (LBLD) (Qualls et al. 2004), children with communication disorders who were subdivided according to the presence of autistic features and language impairment (Norbury 2004), third- and fifth-grade children with learning disabilities (Abrahamsen and Burke-Williams 2004) and in children with spina bifida meningomyelocele (SBM), a neurodevelopmental disorder in which agenesis and hypoplasia of the corpus callosum occur (Huber-Okrainec et al. 2005). Amongst other findings, these studies revealed that children with communication disorders can use context to facilitate idiom comprehension, children with SBM can understand decomposable idioms but not non-decomposable idioms (only the former idioms are processed like literal language) and that LBLD adolescents can comprehend high and moderate familiarity idioms as well as typically developing peers in certain test contexts. Impaired proverb interpretation is a hallmark of schizophrenia (Kiang et al. 2007). Metaphor comprehension has been found to be impaired in patients with right-hemisphere damage (Rinaldi et al. 2004). Young adults with Asperger’s syndrome have been observed to have difficulty interpreting irony (Martin and McDonald 2004). Moreover, an ability to understand the belief states of others was also critical to an understanding of ironic language in these subjects. Idiom comprehension has been found to be impaired in subjects with right brain damage and in subjects with left brain damage (Papagno et al. 2006). Papagno et al. (2004) assessed idiom comprehension in ten aphasic subjects with semantic deficits.

302

Louise Cummings

Idiom comprehension was severely impaired in these subjects with a bias towards literal interpretation in evidence.

7.

Context

Context is a broad concept that has physical, linguistic, epistemic and social aspects. For example, in the exchange below between Bob and Sam, physical context includes features such as the day and time of speaking, the people who are present, the physical setting in which the exchange is conducted (office, restaurant, etc.): (1) Bob: Do you want to visit us in Rome next week? Sam: Mary’s father is undergoing hip replacement surgery. An aspect of physical context, namely time, is explicitly signalled by Bob in his use of the indexical expression next week. The immediate linguistic context that Bob uses to recover the implicature of Sam’s response (the implicature being that Sam is declining Bob’s invitation to visit him in Rome) is the question posed by Bob himself. However, earlier parts of the conversation between Bob and Sam may also contribute linguistic context that may be relevant to the implicature that Bob derives. For example, Sam may have communicated at the start of the conversation that his wife Mary is very anxious about her father’s health and that she is making daily visits to see him at home. With this piece of linguistic context in mind, Bob may even more strongly conclude that Sam is implicating that he cannot undertake a trip to Rome at this time. The epistemic context describes the shared background knowledge and beliefs between speaker and listener in an exchange. For example, Bob and Sam clearly share knowledge of who Mary is, that hip replacement surgery involves a period of recuperation, that Rome is the capital city of Italy, and that Bob and his wife own a villa in Rome. At least equally important to the success of the above exchange are the knowledge and beliefs that Bob and Sam have about each other’s mental states. In this way, Bob could not have derived an implicature from Sam’s utterance if he did not know that Sam knows that people can normally only undertake trips if they are not obliged to fulfil other duties and responsibilities. By the same token, Sam could not have implicated his rejection of Bob’s invitation if he did not know that Bob knows that family responsibilities take precedence over travel plans. Finally, it is a feature of social context, specifically mutual regard and friendship between Bob and Sam, that leads Sam to decline Bob’s invitation indirectly, rather than by means of the more direct, but altogether less polite response of simply saying no!. Studies have employed diverse methodologies to examine the processing of context by children and adults with pragmatic disorders. For the most part, context tends to be construed within these studies in terms of a single sentence or a short passage (i.e. linguistic aspects of context are examined to the exclusion of other as-

Pragmatic disorders

303

pects of context). Subjects must be able to draw upon this sentence or passage in order to achieve the disambiguation of a word or sentence or the correct pronunciation of a homograph. Jolliffe and Baron-Cohen (1999) used this method in an experiment on context processing in normally intelligent adults with either autism or Asperger’s syndrome. Subjects were asked to read aloud sentences that contained homographs (e.g. lead; read). Each homograph had a frequent and rare pronunciation. An accompanying sentence or part of a sentence favoured one pronunciation over the other and appeared either before or after the homograph. Two of the test items for the homograph lead are as follows: Rare pronunciation; before context: It was lead in the box that made it so heavy. Frequent pronunciation; after context: Mary wanted to take the dog for a walk, so she went to the cupboard and took the lead. Similar techniques have been used to investigate deficits in context processing in subjects with schizophrenia (Bazin et al. 2000; Sitnikova et al. 2002). (See Cummings 2007a and 2009 for a critical discussion of this approach to the study of context.) A substantial number of clinical investigations have addressed epistemic aspects of context such as listener knowledge. For example, Brenneise-Sarshad et al. (1991) assessed the effect of two listener conditions – a knowledgeable listener condition and a naïve listener condition – on the content of narrative discourse in aphasic subjects. In both conditions, aphasic subjects told stories based on blackand-white line drawings to a listener. Only in the knowledgeable condition did the listener view the drawings along with the aphasic subject. Referential communication tasks also require a speaker to tailor information according to a listener’s state of knowledge. In a typical referential communication task, two participants are given the same set of pictures in two different random orders A and B. The participant who received the set in the B order must attempt to replicate the A order based on a series of descriptions given by the other participant. According to one model (the collaborative model) of how this task is achieved, “when speakers try to identify a referent, they begin with a provisional clause and propose what they believe to be an adequate expression enabling the addressee to recognize it. They take into account the knowledge they assume to be shared with their partner – the common ground – in order to chose the most appropriate formulation of the message” (Feyereisen et al. 2007: 4). These tasks have been used in clinical studies of a range of clients including those with dementia related to Alzheimer’s disease (Carlomagno et al. 2005; Feyereisen et al. 2007). A range of context processing problems have been identified in child and adult clients with pragmatic disorders. Norbury (2005) examined the use of context to achieve lexical ambiguity resolution in 9- to 17-year-olds with language impair-

304

Louise Cummings

ment, ASD plus language impairment and ASD with verbal abilities in the normal range. This study examined two conditions, which involved contextual facilitation and suppression. Accuracy and response times to picture judgements were recorded in each condition. Although all groups of children responded quickly and more accurately to words that followed a biased context, children with language impairment, and ASD plus language impairment did not use context as efficiently as their peers without language impairment. Moreover, subjects with language impairment, either in isolation or with ASD, produced errors in the suppression condition that reflected poor contextual processing. Loukusa et al. (2007) examined the ability of children with Asperger’s syndrome and high-functioning autism (HFA) to use context in answering questions and in giving explanations of their answers. The younger AS/HFA group (7 to 9 years) in the study performed less well than a control group when answering contextually demanding questions. Both this younger group and an older group of AS/HFA children (10 to 12 years) had difficulty explaining their answers, which suggested that they were not always aware of how they derived answers from context. Rinaldi (2000) studied comprehension of two types of ambiguity in 64 students with specific developmental language disorder who were aged 11+ to 14+ years. The two forms of ambiguity – inconsistent messages of emotion and multiple meanings in context – are evident in a range of communicative intent (e.g. sarcasm, idiomatic expression, deceit and humour). Students with language disorder were less able to use context to understand implied meanings than either of two comparison groups. Experimental studies have repeatedly shown that schizophrenic subjects are unable to process aspects of linguistic context. Bazin et al. (2000) conducted an experiment in which 30 schizophrenic subjects and 30 control subjects were required to complete sentences using the first word(s) that came to mind. Each sentence contained an ambiguous word, the less frequent meaning of which was primed by a preceding sentence. Results showed that only control subjects were able to use the linguistic context provided by the preceding sentence to prime the less frequent meaning of the ambiguous word. Schizophrenic subjects, particularly those with thought disorder, used the most common meaning of the ambiguous word more frequently than controls. Sitnikova et al. (2002) used event related potentials to examine deficits in language comprehension in schizophrenia. Sentences that contained two clauses were read by schizophrenic and control subjects. The investigators hypothesized that the processing of target words in the second clause would be influenced by preceding linguistic context in the control subjects only. Schizophrenic subjects, by contrast, were expected to be inappropriately affected by the dominant meaning of homographs in the first clause (e.g. the “structure” meaning of bridge in the sentence The guests played bridge because the river had rocks in it). This hypothesis was confirmed. Schmitzer et al. (1997) examined the influence of context on the interpretation of denotative and connotative meanings of homographs in subjects with right-hemisphere damage. These investigators found that

Pragmatic disorders

305

right-hemisphere damaged subjects had a reduced ability to process connotative word meanings. Moreover, they were not assisted by the presence of semantically supportive linguistic information.

8.

Non-verbal communication

A range of non-verbal behaviours play a role in the pragmatic interpretation of utterances (although, as I have argued in Cummings 2007a, this does not thereby make these behaviours pragmatic in nature). These behaviours include eye contact, facial expressions, body posture, gestures, intonation and stress, amongst others. A friend’s sarcastic remark let’s try and get parked this side of Christmas to a recently qualified driver is likely to be understood as such as much for the irritated facial expression of the speaker and the intonation with which it is uttered as for the words in the utterance itself. Similarly, a fleeting lapse in eye contact may tell a speaker that his invitation to dinner is about to be rejected even before his verbal utterance is complete. These non-verbal cues are so central to the interpretation of utterances that in disorders where their processing is compromised, considerable communication difficulties can result. It is now well established that autistic individuals have reduced sensitivity to the facial expressions of others and that they are unable to assign emotional states (e.g. anger, happiness) to people on the basis of these expressions (see Clark et al. 2008 and Wright et al. 2008, for example). Although these findings are typically discussed in the context of problems with empathy in ASD subjects, they are at least as likely to have adverse implications for the ability of these subjects to engage in pragmatic interpretation of utterances. In this section, we consider several studies that have revealed deficits in non-verbal behaviours which play a role in pragmatic interpretation. The use and understanding of non-verbal behaviours are compromised in a number of clinical groups. Inappropriate non-verbal behaviours have been recorded in children, adolescents and adults who have sustained a traumatic brain injury (Aubert et al. 2004; Fyrberg et al. 2007; Sainson 2007). These investigators argue that although non-verbal behaviours are often overlooked in brain-injured clients, they are integral to the successful rehabilitation of these individuals. Bishop et al. (2000) examined non-verbal responses in children with pragmatic language impairment. These children made very little use of non-verbal responses such as nodding. Bishop et al. recorded a relatively high level of pragmatically inappropriate responses which could not be explained in terms of limited grammar or vocabulary in those children who failed to use non-verbal responses. Non-verbal expression is frequently compromised in schizophrenic patients with such patients using facial expressions less than normal subjects to invite and sustain social interaction (Brüne et al. 2009). Brüne et al. (2008) found that patients with schizophrenia spectrum disorders make less use of non-verbal behaviours during standardized interviews

306

Louise Cummings

than controls and that negative symptoms in these patients inversely correlated with several non-verbal behaviours. Luzzi et al. (2007) examined the recognition of facial expressions that revealed the emotional states of happiness and sadness in 71 patients with probable Alzheimer’s disease. Although the identification of positive and negative facial emotions was largely preserved in these subjects, some 27 per cent of patients still displayed impaired recognition of the emotions revealed through facial expressions.

9.

Conclusion

It emerges that an extensive range of linguistic and non-linguistic behaviours can be disrupted in children and adults with pragmatic disorders. These behaviours include context processing problems, difficulties with non-literal language interpretation and the contribution of irrelevant utterances during conversation in clients with schizophrenia. The individual with an autistic spectrum disorder can struggle to attribute communicative intentions to other conversational participants, a ToM deficit which makes it difficult for the ASD child and adult to recover the implicatures of utterances and to establish the illocutionary force of speech acts. These pragmatic difficulties are compounded in the ASD client on account of difficulty with the processing of facial and other non-linguistic cues. The cognitive-communication problems in the individual with traumatic brain injury make it difficult for the TBI client to produce coherent, cohesive discourse and to manage topic in conversation. These deficits are most often revealed by conversation- and discourse-analytic studies. The client with dementia may struggle initially to comprehend non-literal utterances, humour and sarcasm. As his or her condition deteriorates, other pragmatic deficits (e.g. decreased sensitivity to conversational maxims) may become increasingly evident. Finally, the aphasic adult may struggle on account of syntactic and lexical-semantic disturbances to address the informational needs of his or her conversational partner and to reformulate messages when communication breakdown occurs. In short, the full gamut of pragmatic disorders is represented by the different clinical populations examined in this chapter. Although considerable energy has been expended in the study of pragmatic disorders, their investigation has not always been undertaken in a theoretically rigorous fashion. Clinical pragmatic studies which lack a clear theoretical rationale can only tell us that a certain aspect of pragmatics is impaired. However, they are largely powerless to tell us why this impairment has occurred or to throw light on the cognitive and other factors that may contribute to this impairment. Clinical researchers are increasingly attending to this criticism. The excellent work that is examining pragmatics in clients with an autistic spectrum disorder using one or other of the prevailing cognitive theories of autism is a case in point.15 Of course,

Pragmatic disorders

307

theoretical work in pragmatics also has its part to play through the development of frameworks and models that permit of experimental testing in pragmatically impaired clients. Here, too, there is clear evidence of progress with frameworks such as relevance theory increasingly finding their way into clinical pragmatic studies. It is hoped that with further developments, the gap that still exists between theoretical endeavour in pragmatics and other disciplines (e.g. developmental psychopathology) and the investigation of pragmatic disorders in clients will become an increasingly narrow one. Theoretical issues aside, clinical pragmatics must also address certain disparities that have arisen in the study of pragmatic disorders. These disparities concern a preoccupation with some pragmatic phenomena (e.g. implicature) to the exclusion of other pragmatic phenomena (e.g. presupposition) and with certain clinical populations (e.g. autistic spectrum disorders) over other clinical populations (e.g. emotional and behavioural disorders). Here, again, there are signs that the field of clinical pragmatics is adapting to meet these challenges. There is now growing recognition that children with emotional and behavioural disorders such as conduct disorder and attention deficit hyperactivity disorder can experience significant pragmatic deficits. Moreover, studies are increasingly attempting to characterize those deficits (see chapter 2 in Cummings 2009 for further discussion). The continuing development of techniques in conversation and discourse analysis has enabled investigators to mine an increasingly rich seam of pragmatic behaviours. These behaviours include the formulation of linguistic messages to accommodate different knowledge states in listeners, the initiation, development and termination of topic across turns in conversation or during narrative discourse and the use of repair strategies during conversation to address misunderstandings on the part of a listener. To reflect the growing significance of cognitive explanations of pragmatic disorders, these conversational and discourse abilities are now also being discussed within the context of cognitive capacities such as mindreading.16 In short, while there has been much for investigators to lament in the relatively short history of the study of pragmatic disorders, it is clear that clinical pragmatics is a resourceful, dynamic field that is meeting its critics head on.

Notes 1. Specific language impairment is a diagnostic label that is applied to children who have a severe language deficit (language scores greater than 1.5 standard deviations below the mean) in the absence of hearing loss, psychiatric disturbance, craniofacial anomalies, low non-verbal intelligence or an autistic spectrum disorder. 2. A search of articles with “pragmatic” in the title of two major journals in the field supports this statement. The journals in question are the International Journal of Language and Communication Disorders (the official journal of the Royal College of Speech and Language Therapists in the UK) and the Journal of Speech, Language and Hearing Re-

308

3.

4. 5.

6.

7.

8. 9.

10.

Louise Cummings

search (published by the American Speech-Language-Hearing Association). In the former journal, 24 articles with “pragmatic” in their titles were published between 1985 and 2008, all of which were studies of children. In the latter journal, nine articles with “pragmatic” in their titles were published between 1978 and 2007, seven of which investigated child subjects. The autistic spectrum disorders (ASDs) include five diagnostic entities: (1) classic autism (autistic disorder), (2) Asperger’s syndrome, (3) Rett’s syndrome, (4) childhood disintegrative disorder and (5) pervasive developmental disorder, not otherwise specified (PDD, NOS). All these disorders involve impairments of communication, socialisation and imagination, the so-called “triad of impairments”. ASDs are diagnosed on the basis of criteria set out in the Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association 2000). The Yale Child Study Center states that speech (spoken language) is absent in about 50 per cent of autism cases. Rapin and Allen (1983) applied the term semantic-pragmatic syndrome to children with known organic aetiologies (primarily hydrocephalus), to children with no brain damage and to children with and without mental retardation. When Bishop and Rosenbloom introduced their term semantic-pragmatic disorder four years later in 1987, it was used to describe a subtype of specific language impairment. Aphasia is an acquired language disorder in adults. It is primarily caused by cerebrovascular accidents (strokes) which affect the left (language dominant) hemisphere of the brain. Although there are different forms of aphasia, it is broadly classified into fluent and non-fluent types. The main features of fluent aphasia are reduced comprehension of language in the presence of effortless, but incoherent, expressive output. In non-fluent aphasia, the comprehension of language is relatively intact, but there is a significant problem with language expression. A traumatic brain injury (TBI) can involve either an open head injury or a closed head injury. In the former, the skull is fractured by penetration from a missile, for example. In the latter, the brain is damaged while the skull remains intact. TBI is sustained most commonly in road traffic accidents, but can also be caused by trips and falls (particularly in children and the elderly), sports injuries and violent assaults. Coelho (2007: 126) states that “to transcribe and analyze a 15-minute sample of conversation may require 3 hours”. Schizophrenia is a serious mental illness that affects approximately 1 in every 100 people. A diagnosis of schizophrenia is based on the identification of positive and negative symptoms. Positive symptoms include thought disorder (disorganized and illogical thought), delusions (the holding of false and bizarre beliefs) and hallucinations (perception of things that do not exist). Negative symptoms involve the absence of normal behaviours. They include affective flattening, alogia (poverty of speech), apathy, avolition (absence of initiative or motivation) and social withdrawal. It is worth remarking that this study exemplifies the type of clinical pragmatic investigation that rather loosely interprets what constitutes a particular pragmatic feature of language, in this case speech acts. Rollins et al. include “read text aloud” and “exhibit attentiveness” within their category of “other” speech acts. In a similar fashion, in a study by Loveland et al. (1988), behaviours as diverse as nonresponses, gesture and vocalization were classified as types of speech acts. See Cummings (2007a) for further discussion of this point.

Pragmatic disorders

309

11. Dementia involves the progressive deterioration of cognitive functions including language skills. Although Alzheimer’s disease is the most common cause of dementia, other causes include vascular disease (vascular dementia), Pick’s disease (frontotemporal dementia), HIV infection, Creutzfeldt-Jakob disease and excessive alcohol consumption (Korsakoff’s syndrome or alcohol-related dementia). 12. Although autistic adults were the focus of investigation in this study, autism has its onset in the developmental period. It is for this reason that this study is discussed alongside studies involving child subjects. 13. Although pronoun reversal has been standardly included in accounts of the linguistic and communicative features of autism (Bartak et al. 1975), studies have often failed to find evidence of increased pronoun reversal by autistic children. In a study of the use and comprehension of the personal pronouns I, you and me, Lee et al. (1994) found few instances of pronoun reversal among autistic subjects. Autistic and nonautistic mentally retarded children and young adults matched for chronological age and verbal mental age were able to comprehend these pronouns in test situations. In a visual perspective-taking task, autistic subjects were significantly less likely to use the pronoun me. In certain photograph-naming tasks, lower ability subjects were more likely to use their own proper names than personal pronouns. Autistic subjects were also less likely in some circumstances to use the pronoun you to refer to the experimenter than controls. 14. Bock (1977: 723) states that “[t]he decision to treat something as given information constitutes a pragmatic presupposition on the part of the speaker”. 15. The dominant cognitive theories of autism include theory of mind (ToM) theory, weak central coherence (WCC) theory and executive function theory (EFT). ToM and WCC theories are increasingly being used in studies that examine aspects of pragmatics in ASD subjects. Examples include the role of WCC in lexical ambiguity resolution (Norbury 2005), ToM and WCC in a study of event schemas and narratives (Loth et al. 2008) and the role of ToM in referential communication (Dahlgren and Sandberg 2008). For critical discussion of all three cognitive theories of autism, the reader is referred to chapter 4 in Cummings (2009). 16. Studies by Colle et al. (2008), Hale and Tager-Flusberg (2005) and Capps et al. (2000) have established a relationship between certain discourse skills and theory of mind abilities in children and adults with autistic spectrum disorder.

References Abbeduto, Leonard, Betty Davies and Laurie Furman 1988 The development of speech act comprehension in mentally retarded individuals and nonretarded children. Child Development 59: 1460–1472. Abrahamsen, Eileen P. and Debra Burke-Williams 2004 Comprehension of idioms by children with learning disabilities: metaphoric transparency and syntactic frozenness. Journal of Psycholinguistic Research 33: 203–215. American Psychiatric Association 2000 Diagnostic and Statistical Manual of Mental Disorders. Washington: American Psychiatric Association.

310

Louise Cummings

Aubert, Sandrine, Michel Barat, Marie Campan, Patrick Dehail, Pierre Alain Joseph and Jean-Michel Mazaux 2004 Compétences de communication non verbale des traumatisés crâniens graves [Non verbal communication abilities in severe traumatic brain injury]. Annales de Réadaptation et de Médecine Physique 47: 135–141. Baltaxe, Christine A. 1984 Use of contrastive stress in normal, aphasic, and autistic children. Journal of Speech and Hearing Research 27: 97–105. Bartak, Lawrence, Michael Rutter and Antony Cox 1975 A comparative study of infantile autism and specific developmental receptive language disorder: I. The children. British Journal of Psychiatry 126: 127–145. Bazin, Nadine, Pierre Perruchet, Marie Christine Hardy-Bayle and André Feline 2000 Context-dependent information processing in patients with schizophrenia. Schizophrenia Research 45: 93–101. Bishop, Dorothy V.M. 2000 Pragmatic language impairment: A correlate of SLI, a distinct subgroup, or part of the autistic continuum?. In: Dorothy V.M. Bishop and Laurence B. Leonard (eds.), Speech and Language Impairments in Children: Causes, Characteristics, Intervention and Outcome, 99–113. Hove, East Sussex: Psychology Press Ltd. Bishop, Dorothy V.M. 2003 Autism and specific language impairment: categorical distinction or continuum? Novartis Foundation Symposium 251: 213–226. Bishop, Dorothy V.M. and Lewis Rosenbloom 1987 Classification of childhood language disorders. In: William Yule and Michael Rutter (eds.), Language Development and Disorders, 16–41. London: MacKeith Press. Bishop, Dorothy V.M., Janet Chan, Catherine Adams, Joanne Hartley and Fiona Weir 2000 Conversational responsiveness in specific language impairment: evidence of disproportionate pragmatic difficulties in a subset of children. Development and Psychopathology 12: 177–199. Bock, J. Kathryn 1977 The effect of a pragmatic presupposition on syntactic structure in question answering. Journal of Verbal Learning and Verbal Behavior 16: 723–734. Brenneise-Sarshad, Robin, Linda E. Nicholas and Robert H. Brookshire 1991 Effects of apparent listener knowledge and picture stimuli on aphasic and nonbrain-damaged speakers’ narrative discourse. Journal of Speech and Hearing Research 34: 168–176. Brüne, Martin, Claudia Sonntag, Mona Abdel-Hamid, Caroline Lehmkämper, Georg Juckel and Alfonso Troisi 2008 Nonverbal behavior during standardized interviews in patients with schizophrenia spectrum disorders. Journal of Nervous and Mental Disease 196: 282–288. Brüne, Martin, Mona Abdel-Hamid, Claudia Sonntag, Caroline Lehmkämper and Robyn Langdon 2009 Linking social cognition with social interaction: non-verbal expressivity, social competence and “mentalising” in patients with schizophrenia spectrum disorders. Behavioral and Brain Functions 5: 6.

Pragmatic disorders

311

Capps, Lisa, Molly Losh and Christopher Thurber 2000 “The frog ate the bug and made his mouth sad”: narrative competence in children with autism. Journal of Abnormal Child Psychology 28: 193–204. Carlomagno, Sergio, Anna Santoro, Antonella Menditti, Maria Pandolfi and Andrea Marini 2005 Referential communication in Alzheimer’s type dementia. Cortex 41: 520– 534. Clark, Tedra F., Piotr Winkielman and Daniel N. McIntosh 2008 Autism and the extraction of emotion from briefly presented facial expressions: stumbling at the first step of empathy. Emotion 8: 803–809. Coelho, Carl A. 2007 Management of discourse deficits following traumatic brain injury: progress, caveats, and needs. Seminars in Speech and Language 8: 122–135. Coelho, Carl A. and Laura Flewellyn 2003 Longitudinal assessment of coherence in an adult with fluent aphasia. Aphasiology 17: 173–182. Coelho, Carl A., Betty Z. Liles and Robert J. Duffy 1991 The use of discourse analyses for the evaluation of higher level traumatically brain-injured adults. Brain Injury 5: 381–392. Colle, Livia, Simon Baron-Cohen, Sally Wheelwright and Heather K. van der Lely 2008 Narrative discourse in adults with high-functioning autism or Asperger syndrome. Journal of Autism and Developmental Disorders 38: 28–40. Corcoran, Rhiannon and Christopher D. Frith 1996 Conversational conduct and the symptoms of schizophrenia. Cognitive Neuropsychiatry 1: 305–318. Cummings, Louise 2005 Pragmatics: A Multidisciplinary Perspective. Edinburgh: Edinburgh University Press. Cummings, Louise 2007a Clinical pragmatics: a field in search of phenomena? Language and Communication 27: 396–432. Cummings, Louise 2007b Pragmatics and adult language disorders: past achievements and future directions. Seminars in Speech and Language 28: 98–112. Cummings, Louise 2008 Clinical Linguistics. Edinburgh: Edinburgh University Press. Cummings, Louise 2009 Clinical Pragmatics. Cambridge: Cambridge University Press. Cummings, Louise 2010 Clinical pragmatics. In: Louise Cummings (ed.), The Pragmatics Encyclopedia, 40–43. London: Routledge. Cutica, Ilaria, Monica Bucciarelli and Bruno Bara 2006 Neuropragmatics: extralinguistic pragmatic ability is better preserved in lefthemisphere-damaged patients than in right-hemisphere-damaged patients. Brain and Language 98: 12–25. Dahlgren, Svenolof and Annika Dahlgren Sandberg 2008 Referential communication in children with autism spectrum disorder. Autism 12: 335–348.

312

Louise Cummings

Dipper, Lucy T., Karen L. Bryan and J. Tyson 1997 Bridging inference and relevance theory: an account of right hemisphere damage. Clinical Linguistics and Phonetics 11: 213–228. Drummond, Sakina S. and Tamra P. Simmons 1995 Linguistic performance of female aphasic adults during group interaction. Journal of Neurolinguistics 9: 47–54. Eisele, Julie A., Barbara Lust and Dorothy M. Aram 1998 Presupposition and implication of truth: linguistic deficits following early brain lesions. Brain and Language 61: 376–394. Feyereisen, Pierre, Joelle Berrewaerts and Michel Hupet 2007 Pragmatic skills in the early stages of Alzheimer’s disease: an analysis by means of a referential communication task. International Journal of Language and Communication Disorders 42: 1–17. ´ sa, Magda Marchioni and Ingrid Emanuelson Fyrberg, Å 2007 Severe acquired brain injury: rehabilitation of communicative skills in children and adolescents. International Journal of Rehabilitation Research 30: 153–157. Hale, Courtney M. and Helen Tager-Flusberg 2005 Social communication in children with autism: the relationship between theory of mind and discourse development. Autism 9: 157–178. Huber-Okrainec, Joelene, Susan E. Blaser and Maureen Dennis 2005 Idiom comprehension deficits in relation to corpus callosum agenesis and hypoplasia in children with spina bifida meningomyelocele. Brain and Language 93: 349–368. Jolliffe, Therese and Simon Baron-Cohen 1999 A test of central coherence theory: linguistic processing in high-functioning adults with autism or Asperger syndrome: is local coherence impaired? Cognition 71: 149–185. Jordan, Rita R. 1989 An experimental comparison of the understanding and use of speaker-addressee personal pronouns in autistic children. British Journal of Disorders of Communication 24: 169–179. Kasher, Asa, Gila Batori, Nachum Soroker, David Graves and Eran Zaidel 1999 Effects of right- and left-hemisphere damage on understanding conversational implicatures. Brain and Language 68: 566–590. Kiang, Michael, Gregory A. Light, Jocelyn Prugh, Seana Coulson, David L. Braff and Marta Kutas 2007 Cognitive, neurophysiological, and functional correlates of proverb interpretation abnormalities in schizophrenia. Journal of the International Neuropsychological Society 13: 653–663. Lee, Anthony, R. Peter Hobson and Shulamuth Chiat 1994 I, you, me, and autism: an experimental study. Journal of Autism and Developmental Disorders 24: 155–176. Levinson, Stephen C. 1983 Pragmatics. Cambridge: Cambridge University Press. Loth, Eva, Juan Carlos Gómez and Francesca Happé 2008 Event schemas in autism spectrum disorders: the role of theory of mind and weak central coherence. Journal of Autism and Developmental Disorders 38: 449–463.

Pragmatic disorders

313

Loukusa, Soile, Eeva Leinonen, Sanna Kuusikko, Katja Jussila, Marja-Leena Mattila, Nuala Ryder, Hanna Ebeling and Irma Moilanen 2007 Use of context in pragmatic language comprehension by children with Asperger syndrome or high-functioning autism. Journal of Autism and Developmental Disorders 37: 1049–1059. Loveland, Katherine A., Susan H. Landry, Sheryl O. Hughes, Sharon K. Hall and Robin E. McEvoy 1988 Speech acts and the pragmatic deficits of autism. Journal of Speech and Hearing Research 31: 593–604. Luzzi, Simona, Massimo Piccirilli and Leandro Provinciali 2007 Perception of emotions on happy/sad chimeric faces in Alzheimer disease: relationship with cognitive functions. Alzheimer Disease and Associated Disorders 21: 130–135. Martin, Ingerith and Skye McDonald 2004 An exploration of causes of non-literal language problems in individuals with Asperger syndrome. Journal of Autism and Developmental Disorders 34: 311–328. McCaleb, Peggy and Barry M. Prizant 1985 Encoding of new versus old information by autistic children. Journal of Speech and Hearing Disorders 50: 230–240. McDonald, Skye and Samantha Pearce 1998 Requests that overcome listener reluctance: impairment associated with executive dysfunction in brain injury. Brain and Language 61: 88–104. Norbury, Courtenay Frazier 2004 Factors supporting idiom comprehension in children with communication disorders. Journal of Speech, Language, and Hearing Research 47: 1179–1193. Norbury, Courtenay Frazier 2005 Barking up the wrong tree? Lexical ambiguity resolution in children with language impairments and autistic spectrum disorders. Journal of Experimental Child Psychology 90: 142–171. Papagno, Costanza, Patrizia Tabossi, Maria Rosa Colombo and Patrizia Zampetti 2004 Idiom comprehension in aphasic patients. Brain and Language 89: 226–234. Papagno, Costanza, Rita Curti, Silvia Rizzo, Franca Crippa and Maria Rosa Colombo 2006 Is the right hemisphere involved in idiom comprehension? A neuropsychological study. Neuropsychology 20: 598–606. Pijnacker, Judith, Peter Hagoort, Jan Buitelaar, Jan-Pieter Teunisse and Bart Geurts 2009 Pragmatic inferences in high-functioning adults with autism and Asperger syndrome. Journal of Autism and Developmental Disorders 39: 607–618. Qualls, Constance Dean, Jennifer M. Lantz, Rose M. Pietrzyk, Gordon W. Blood and Carol Scheffner Hammer 2004 Comprehension of idioms in adolescents with language-based learning disabilities compared to their typically developing peers. Journal of Communication Disorders 37: 295–311. Rapin, Isabelle and Doris A. Allen 1983 Developmental language disorders: nosologic considerations. In: Ursula Kirk (ed.), Neuropsychology of Language, Reading, and Spelling, 155–184. New York: Academic Press, Inc.

314

Louise Cummings

Rinaldi, Maria Christina, Paola Marangolo and Francesca Baldassarri 2004 Metaphor comprehension in right brain-damaged patients with visuo-verbal and verbal material: a dissociation (re)considered. Cortex 40: 479–490. Rinaldi, Wendy 2000 Pragmatic comprehension in secondary school-aged students with specific developmental language disorder. International Journal of Language and Communication Disorders 35: 1–29. Ripich, Danielle N., Diane Vertes, Peter Whitehouse, Sarah Fulton and Barbara Ekelman 1991 Turn-taking and speech act patterns in the discourse of senile dementia of the Alzheimer’s type patients. Brain and Language 40: 330–343. Rollins, Pamela R., Barbara A. Pan, Gina Conti-Ramsden and Catherine E. Snow 1994 Communicative skills in children with specific language impairments: a comparison with their language-matched siblings. Journal of Communication Disorders 27: 189–206. Rowan, Lynne E., Laurence B. Leonard, Kathy Chapman and Amy L.Weiss 1983 Performative and presuppositional skills in language-disordered and normal children. Journal of Speech and Hearing Research 26: 97–106. Ryder, Nuala, Eeva Leinonen and Joerg Schulz 2008 Cognitive approach to assessing pragmatic language comprehension in children with specific language impairment. International Journal of Language and Communication Disorders 43: 427–447. Sainson, Claire 2007 Communication non verbale et dysfonctionnement exécutif post-traumatique: á propos d’un cas [Non-verbal communication and executive function impairment after traumatic brain injury: a case report]. Annales de Réadaptation et de Médecine Physique 50: 231–239. Schelletter, Christina and Eeva Leinonen 2003 Normal and language-impaired children’s use of reference: syntactic versus pragmatic processing. Clinical Linguistics and Phonetics 17: 335–343. Schmitzer, A, Monica Strauss and S. DeMarco 1997 Contextual influences on comprehension of multiple-meaning words by right hemisphere brain-damaged and non-brain-damaged adults. Aphasiology 11: 447–459. Sitnikova, Tatiana, Dean F. Salisbury, Gina Kuperberg and Philip J. Holcomb 2002 Electrophysiological insights into language processing in schizophrenia. Psychophysiology 39: 851–860. Skarakis, Elizabeth and Patricia M. Greenfield 1982 The role of new and old information in the verbal expression of language-disordered children. Journal of Speech and Hearing Research 25: 462–467. Sperber, Dan and Deirdre Wilson 1986/95 Relevance: Communication and Cognition. Oxford: Blackwell. [Second edition with Postface 1995] Surian, Luca 1996 Are children with autism deaf to Gricean maxims? Cognitive Neuropsychiatry 1: 55–72. Tényi, Tamás, Róbert Herold, I.M. Szili and Mátyás Trixler 2002 Schizophrenics show a failure in the decoding of violations of conversational implicatures. Psychopathology 35: 25–27.

Pragmatic disorders

315

Vanhalle, C., S. Lemieux, S. Joubert, P. Goulet, B. Ska and Y. Joanette 2000 Processing of speech acts by right hemisphere brain-damaged patients: an ecological approach. Aphasiology 14: 1127–1141. Varley, Rosemary 1993 Deictic terms, lexical retrieval and utterance length in aphasia: an investigation of inter-relations. European Journal of Disorders of Communication 28: 23–41. Wright, Barry, Natalie Clarke, Jo Jordan, Andrew W. Young, Paula Clarke, Jeremy Miles, Kate Nation, Leesa Clarke and Christine Williams 2008 Emotion recognition in faces and the use of visual context in young people with high-functioning autism spectrum disorders. Autism 12: 607–626. Ziatas, Kathryn, Kevin Durkin and Chris Pratt 2003 Differences in assertive speech acts produced by children with autism, Asperger syndrome, specific language impairment, and normal development. Development and Psychopathology 15: 73–94.

12.

Autism from a cognitive-pragmatic perspective Anne Reboul, Sabine Manificat and Nadège Foudon

1.

Introduction

Autism is one of a group of three neuro-developmental disorders including, in addition to autism itself, Asperger Syndrome and a fairly heterogeneous group of patients who present some but not all of the symptoms of autism (see below, section 2.2). Asperger Syndrome and autism being the best described pathologies, notably in terms of language and language development, they will be the focus of our attention in what follows. Autism has been described as being to pragmatics what aphasia is to syntax, i.e., a natural testing ground for pragmatic hypotheses. This is certainly true of both Asperger Syndrome and autism, though, as will shortly be seen, autistic people are more impaired in language acquisition. The first part of the paper (section 2) will describe the pathology; the second part (section 3), the impact of the social-pragmatic deficit on language acquisition; the third part (section 4), the pragmatic deficits that remain in adulthood in Asperger and verbal autistic patients.

2.

Asperger Syndrome and autism

2.1.

A quick historical reminder

In 1947, Leo Kanner described a group of 11 children with the following symptoms: isolation, an obsession with routine behaviour, extraordinary memory performances, echolalia, sensibility to stimuli, a limited range of interests, normal intelligence. In 1944, Hans Asperger described a group of four children, presenting what he called an “autistic psychopathology” with the following symptoms: social isolation, linguistic and motor stereotypes, resistance to change, specific interests in certain objects or topics and a preserved language. Children with autism correspond to Kanner’s description (with the important proviso that they are not all normally intelligent), while children of normal intelligence and a preserved language (as well as a more or less normal acquisition) are considered to be suffering from Asperger Syndrome. It should be noted that high-functioning autistic adults (who have an IQ in the normal range of 70–130) and Asperger adults are very similar, the main difference between them being the differences in linguistic acquisition (see Frith 1991, and below, section 3).1

318 2.2.

Anne Reboul, Sabine Manificat and Nadège Foudon

Autism spectrum disorders: diagnostic criteria, prevalence and etiology

Nowadays, both Kanner’s and Asperger’s descriptions have been largely endorsed in diagnosis handbooks, such as the DSM IV (1994), with two provisos: despite their largely common symptoms, Asperger Syndrome and autism (so-called Kanner’s autism, given that it was described by Kanner), differ on the linguistic acquisition of the patients. Asperger children, despite a delay, acquire language by immersion as do typically developing children, while autistic children are not only more delayed (see below, section 3), but usually need speech therapy and about half of them remain mute. Autism is additionally fairly frequently accompanied by more or less severe learning disabilities. Here is the complete list of diagnostic criteria of autism, as given by DSM IV: symptoms include qualitative alteration of social interactions; qualitative alteration of communication; restricted, repetitive and stereotyped behaviour, interests and activities. A second criterion is the delay or the abnormality of the following functions, which has to be evident before the child is three years old: social interactions, language necessary for social communication, symbolic or make-believe play. A final criterion is to the effect that there is no alternative explanation. According to Kogan et al. (2009: 2), “Autism spectrum disorders (ASD) is a group of developmental disorders comprising autistic disorders and 2 related but less severe disorders: Asperger disorder and pervasive developmental disorder not otherwise specified (PDD-NOS)”. The main criteria of ASD are impairments in social interaction, communication (both linguistic and non-verbal), deficient symbolic play, and repetitive and restricted patterns of behaviours and interests. The inclusion of the less severe disorders (Asperger and PDD-NOS) in the autism spectrum has had the consequence of greatly increasing the prevalence of the disease from 2 to 5 in 10 000 in the 1960–1980s (when the diagnosis was limited to autism) to 110 in 10 000 in a 2007 USA survey (see Kogan et al. 2009), though the gender distribution has remained unchanged (with four boys to one girl). Similar though less dramatic (57 in 10 000) increases have been reported for the UK (see Scott et al. 2002). Data for continental Europe usually concern autism and PDD-NOS rather than the whole spectrum of disorders and, as a consequence, the figure has stayed at a lower level (27 in 10,000).2 The aetiology, while still not entirely clear, is heterogeneous, though there is no doubt that it is biological. Current hypotheses mainly target genetic anomalies which are linked to the X chromosome, but also recessive autosomal, as well as ante-, obstetric or neonatal incidents. The heavy predominance of social deficits in the symptomatology of autism spectrum disorders does not only result in difficulties in communication, it also has consequences on the acquisition of language, and, particularly (though perhaps not exclusively, see below, 3.6.3 and 3.6.4), on lexical acquisition. In what follows, the

Autism from a cognitive-pragmatic perspective

319

focus will be on the two best characterized disorders, i.e., autism and Asperger Syndrome.

3.

The impact of the social-pragmatic deficit on language acquisition in autistic people

As mentioned above, the major differences between autistic and Asperger subjects have to do with learning disabilities, which are often associated with autism but not with Asperger Syndrome, and with difficulties in linguistic acquisition, which are universal in autism and apparently absent from Asperger Syndrome. The present section will be devoted to a description of the language that both populations may achieve in adulthood as a preliminary to a description of acquisition focused on autistic children. In conclusion, I will come back to the difference in linguistic acquisition between autism and Asperger Syndrome and try to explain it. 3.1.

Language and ASD

Relative to language, the subjects suffering from ASD3 can be characterized by the following schema:

Figure 1. Language in ASD

A closer look at the chronology of acquisition between typically developing children, autistic children and Asperger children gives the results4 summed up in the following table (compiled from Bloom 2000 for typically developing children, and from Howlin 2003 for Asperger and autistic children):

320

Anne Reboul, Sabine Manificat and Nadège Foudon

Table 1. Chronology of acquisition Population

First words

First combinations

Typically developing children

11 months

17 months

Asperger children

15 months

26 months

Autistic children

38 months

52 months

As is obvious, the delay in acquisition is slight in Asperger children compared to typically developing children, while it is severe in autistic children, with important consequences for schooling, for those children who might otherwise benefit from it. Indeed, the delay for first words is more important in autistic children than in children with Specific Language Impairment (Leonard 2000), who utter their first words at around 23 months of age. Additionally, an interesting factor is the gap between first words and first combinations (≈ 2 words), which is a mere 6 months in normally developing children, 11 months in Asperger children and 14 months in both autistic and Specific Language Impairment Children. If, as is usually assumed, first combinations signal the very beginnings of syntax, this suggests not only that language onset is delayed in ASD patients, but that its development is slower than in typically developing children. Both the rather poor outcome in autistic patients (i.e., the fact that about 50 % of them never acquire language, despite speech therapy) and the enormous delay in first words production seem to single them out as particularly deficient in linguistic acquisition and we will now turn to lexical acquisition in autistic children, before coming back to the difference between Asperger and autistic children in the conclusion to this section. 3.2.

Language acquisition in autistic children

There have been very few longitudinal studies of language acquisition in autistic children, the only ones (to our knowledge) having been done by Tager-Flusberg et al. (1990) among high-functioning American autistic children and by Foudon (2008) among French autistic children. We will take this second and more recent study as an example, before discussing Tager-Flusberg’s results below (see 3.6.3). The longitudinal study of the French autistic children5 was conducted in the day care department of Hôpital Saint-Jean de Dieu (Lyon, France) accommodating autistic children from 3 to 12, between 2005 and 2008.6 We followed three groups of autistic children, assembled to correspond to three rough stages of linguistic acquisition, based on the Mean Length of Utterance (MLU)7 of each child at the beginning of the study, for three years:

Autism from a cognitive-pragmatic perspective

– – –

321

First words (MLU = 1); First combinations (MLU ≥ 2); First sentences (MLU > 2).

The idea was thus to cover the whole spectrum of language acquisition, because, given the slow development of language in autistic children, one cannot be sure that an autistic child at the first word stage, even if he8 turns out to be verbal later on, will have progressed to the final stage three years later. Children were recorded every three months (excepting the summer period), that is three times a year in three situations (work, play, meal) in the day care department they attended. The corpora were transcribed according to the prescriptions of CHILDES. Criteria for inclusion were a diagnosis of autism in accordance with the recommendations of DSM IV and an absence of auditory impairment. IQ level was not a criterion for two reasons: it is very difficult to assess in autistic children with a low linguistic development and in contrast with Tager-Flusberg’s 1990 study, we wanted to investigate language acquisition in autistic children regardless of their intellectual abilities. Thus, the children selected, apart from the obvious criteria indicated above, merely had to show an MLU corresponding to one of the three stages of language acquisition indicated above to be included in the study. The resulting group of children was composed as follows: – – –

The first words group consisted of two boys (3.3 and 3.5 years at the beginning of the study), and a girl (2.8 years); The first combinations group also consisted of two boys (4.5 and 5.5 years), and a girl (7. 9 years); the first sentences group again consisted of two boys (8.11 and 7.8 years), and a girl (5).9

One child in the first group did not utter a single word in any recording session after the first one, and we therefore did not include him in our analyses. We will quickly present the preliminary analyses which have been carried out on the corpora and which give a good idea of acquisition in autistic children. Unsurprisingly, given the heterogeneous profile of autistic children in terms of language (see also Kjelgaard and Tager-Flusberg 2001; and below 3.6.3), the progression of the children in our sample was highly diverse. We were able to separate the children into three groups depending their linguistic development: two who do not progress at all, three who do progress, but only slightly, two who have progressed fairly well, leaving aside one, whose progress curve was so chaotic as to defy description. It should be noted that the progression is very slow, by comparison with typically developing children,10 who, in six months, will progress from one word utterances to four words or more, leaving the autistic children far behind. Thus autistic children are, indeed, not only delayed in terms of language onset, they also progress extremely slowly relative to typically developing children.

322

Anne Reboul, Sabine Manificat and Nadège Foudon

We complemented the MLU by two other measures: we plotted the repartition of syntactic categories in the final sessions of our two best developing autistic children and compared it to that of TD children with the same MLU. The results were interesting in that in TD children with MLU = 3, words are distributed, with an astonishing regularity, in 30 % verbs, 10 % nouns (including proper names) and 60 % of function words, while one ASD child’s words are distributed in 35 % verbs, 15 % nouns and 50 % of function words and the other’s are distributed in 30 % verbs, 20 % nouns, and 50 % function words.11 In other words, in those two children who have progressed rather well, the syntactic distribution of words, though it is near to TD children’s, remains different from it. Very much the same conclusion can be drawn from a last measure, that of agrammaticalities: we counted the number of omitted pronouns and determiners, non conjugated verbs, missing prepositions and faulty agreements. The number of such agrammaticalities was very high in the first sessions for TD children, but diminished fairly rapidly and, indeed, completely disappeared in the last sessions. In autistic children, the results were more heterogeneous, but, again, in those two who reached MLU = 3, the number of agrammaticalities, though it diminished strongly, still remained at around 2 %. Thus, on the whole, the general picture that can be drawn from that longitudinal study is that linguistic acquisition in autistic children is fairly heterogeneous and that, even in those autistic children who do reach the sentence level (and who, hopefully, will presumably progress well over the MLU = 3 stage in the future), language still remains slightly different from that of TD children at the same level. Before we turn to an explanation of why this should be so, let us now quickly outline a currently well-accepted model of language acquisition. 3.3.

The Emerging Coalition Model of language acquisition

A child acquiring language has to solve three different problems: – Isolating words in the flux of speech; – isolating objects and events in the environment (categorization); – mapping words and objects and events. Hirsh-Pasek and Golinkoff (see, e.g., Hirsh-Pasek and Golinkoff 1996, 2006) have proposed a model of language acquisition, the emerging coalition model, that has the advantage of taking a strong developmental stance towards lexical acquisition. At a first stage (roughly concomitant with first words: 9–18 months), the child will mainly rely on spatio-temporal coincidence between a word and the corresponding and perceptually salient object or event. This first period, strongly relying on association, is followed, between roughly 18 to 24 months, by a period during which the child begins to use socio-pragmatic cues such as gaze direction, intention, etc. Finally, from 24 months on, the child adds to his repertoire morpho-syntactic cues,

Autism from a cognitive-pragmatic perspective

323

as syntax kicks in, allowing him/her, e.g., to distinguish different syntactic categories which frequently map either onto objects (e.g., names, adjectives) or onto actions (verbs). A first advantage of the model is its taking into account the development of non-linguistic abilities in the first two years. This is especially relevant in autism, which is considered as a developmental pathology because development in autism is heterochronological or disharmonious. In other words, the retarded, absent or faulty development of an ability in autistic children may block or retard the development of normal linguistic acquisition. There are reasons to think that autistic children have difficulties with all three problems. Regarding the individuation of words, recent works (Saffran, Aslin, and Newport 1996; Saffran, Newport, and Aslin 1996; Aslin, Saffran, and Newport 1998) have highlighted abilities for statistic processing of auditory stimuli in infants and have proposed that these abilities are the basis for word individuation, through the detection of acoustic regularities in linguistic stimuli. This, however, supposes that children are able to filter out what is not linguistic in the auditory environment, an ability which seems very precocious in TD infants (see, e.g., Winkler et al. 2003), who preferentially orient toward human voices (and particularly toward their mothers’ voices), but which seem absent in children with autism (see Klin 1991 and Ceponiene et al. 2003). Regarding the individuation of objects and events, autistic patients seem to have specificities in categorical perception (see Soulières et al. 2006) and to be more dependent on prototypicality in concepts than are normally developing controls (see Gastbeg 1999). Finally, the mapping between words and objects and events is usually dependent in normally developing children on so-called socio-pragmatic cues (see e.g. Bloom 2000; Tomasello 2005, 2008). Given the strong deficits in social abilities in autistic children, there is little reason to be optimistic regarding their performances in solving this specific problem. These difficulties in lexical acquisition allow an explanation of the gap between first words and first combinations in both autism and Asperger’s Syndrome. Indeed, a fairly recent hypothesis (Hirsh-Pasek and Golinkoff 1996) introduces the notion of a critical threshold in terms of a vocabulary size under which syntax does not develop. This hypothesis, which has been empirically validated, places the threshold at between 250 and 300 words in the child’s vocabulary to trigger the development of syntax. What this suggests relative to the longer gap between first words and first combinations in both Asperger and autistic children is that their main problem may be in lexical acquisition, which would proceed at a slower pace than in normally developing children, thus explaining why they are so late in producing first combinations and first sentences.12 Let us now turn to why vocabulary should be acquired later and more slowly in autistic and Asperger children.

324 3.4.

Anne Reboul, Sabine Manificat and Nadège Foudon

Socio-pragmatic abilities

On the hypothesis just described, syntax appears only when a certain threshold is reached in terms of vocabulary size. We will mainly concentrate here on sociopragmatic abilities. What socio-pragmatic cues do in lexical acquisition is help the child match the words heard to the objects or events isolated in the world. This enables her to learn words which she picks up in the adults’ utterances even when they are not directly addressed to her. The onset of socio-pragmatic abilities at around 18 months coincides with the so-called word spurt, a dramatic increase in lexical acquisition. These socio-pragmatic abilities are the ability to cooperate in a common activity with an adult (so-called shared intentionality, see Tomasello et al. 2005; Tomasello 2008), the ability to engage in shared attention (see, e.g., Baron-Cohen 1995), and, to a lesser degree, the ability to engage in joint attention (Moore and Dunham 1995; Eilan et al. 2005). The most basic and precocious one is presumably13 joint attention, which begins at around 9 months and seems fully functional by 12 months. It consists in the fact that when the adult (parent or caretaker) turns his/her head to look at an object or event, the baby will automatically turn its head in the same direction. The final stage (by 12 months) is reached when the child will follow another’s gaze to an object behind a visual obstacle. This automatic behaviour does not seem to exist in autistic children at the same age (indeed, it could be one early diagnostic sign of autism, see, e.g. Reznik et al. 2007; Bryson et al. 2007), though it may develop much later on. This is not due to an incapacity to follow glance as such, as was shown in an experimental study by Leekham et al. (1993), where autistic children performed at the same level as typically developing children in a task where children were explicitly asked to say or show what an actor was looking at. Rather, it seems that, although they are able to follow glance if explicitly asked to do so, autistic children do not do it spontaneously and, indeed, automatically. The next ability, which can be seen as a development from joint attention, is shared attention that is fully functional in typically developing children by 18 months (see Baron-Cohen 1995). Shared attention is the behaviour that the child manifests, when it not only follows an adult’s gaze but alternates its own gaze between the object and the adult’s face to check whether they are indeed sharing an attentional state toward the same object. Shared attention seems an important step toward so-called shared intentionality. As described by Tomasello and his colleagues (see Tomasello et al. 2005; Tomasello 2008), shared intentionality entails shared attention, but, in addition, according to the authors it entails the ability to represent the other’s intentions and one’s own intentions as part of a recursive shared intention (e.g., I intend that she intends that I intend … that we open this box). Shared intentionality behaviour is taken to be important in language acquisi-

Autism from a cognitive-pragmatic perspective

325

tion in that it makes it simpler to converge on a common ground, i.e., a set of pieces of information shared between interlocutors, of which each knows that the other knows it. The potential contribution of shared intentionality to lexical acquisition seems pretty obvious: it facilitates inferences as to what the adult may be talking about. However, its application seems restricted, given that it presupposes that the child and the adult are doing something together. By contrast, it is now well known that typically developing children can learn words (at least 30 % of newly acquired words) from utterances that are not addressed to them and that are not necessarily part of a joint activity (see Bloom 2000). This is where the attentional abilities, both joint attention and shared attention, come into play, allowing the child to identify what the adult is talking about on the basis of what he/she is looking at, even when his/her utterances are addressed to a third party. Autistic children are deficient, not so much in the ability to compute gaze direction, but in the motivation for using that ability. Thus they do not develop (or develop only much later) joint attention, i.e., spontaneous head turning to follow another’s gaze. Given that shared attention is a development of joint attention and presumably rests on the same motivation, it can hardly come as a surprise that they do not develop shared attention either. 3.5.

Socio-pragmatic deficits in autism: the motivation hypothesis

This seems to place motivation at the centre of the social deficits in autism, which would strengthen Tomasello’s (2008) position regarding the species-specific motivation for cooperation. His argument in defence of this hypothesis is based on TD children’s tendency to point, which appears at around 12 months and remains very strong in the first few years, before being largely superseded by language. While proto-imperative gestures (gestures produced in the intention of obtaining something from the adult) do not manifest any specific social ability and certainly no motivation for sharing or cooperation, proto-declarative gestures, which appear slightly later, do manifest it (see Bates, Camaioni, and Volterra 1975). This is because in proto-imperatives, the other can be used as a (non-intentional) tool to satisfy a wish or an intention of the child, while in proto-declaratives, the only reward is the very sharing of the attitude with the (intentional) adult. Autistic children use proto-imperative pointing, but only rarely point in a proto-declarative way (Sigman and Kasari 1995; Tomasello et al. 2005; Tomasello 2008). Thus what is crucial may be the motivation for sharing which underlies both behaviours. Being deprived of the abilities which allow typically developing children to keep track of simple epistemic states of adults (such as attention) can only hinder lexical acquisition by restricting its mechanisms to association based on coincidence between salient word and salient object, which is in part what speech ther-

326

Anne Reboul, Sabine Manificat and Nadège Foudon

apies rely on for autistic children. However, this means that, in contrast to normally developing children, the only times when autistic children will learn new words are precisely when they are specifically taught them either by their parents or during speech therapy and specialized tuition. An important argument relative to the motivation hypothesis is echolalia. Echolalia, which can be immediate or delayed, is defined as rote and literal repetition of utterances, which are reproduced without modification. It was identified by Kanner (1947) as fairly frequent in autism. Though echolalia was first seen as an obstacle to language acquisition, pioneer studies (for a review, see Schuler and Prizant 1985) have shown, first, that in fact echolalia is not a monolithic phenomenon, second, that children with echolalia have a better long range prognosis for language acquisition, and, third, that at least some kinds of echolalia, though inappropriate in as much as they may not be readily understandable by addressees, do manifest the will, but not the linguistic abilities to communicate. Thus, though some children with autism may entirely lack the motivation for social contact and social communication, this is not as general as has often been thought. If, as proposed above, the deficits in socio-pragmatic abilities hinder lexical acquisition in autism, then the often observed regression and the strong delay in language acquisition which are characteristic of autism are readily explained. Indeed, parents of autistic children often report a normal onset of single word production at the end of the first year, followed by a regression after 18 months (see Rapin and Dunn 1997). Regression in language acquisition at around 18 months corresponds to the absence in autistic children of the socio-pragmatic processes which should come in at that age and strongly increase the rate of lexical acquisition, while the general delay can be explained by the rather limited inputs which autistic children are able to take into account as compared with typically developing children. This delay and the general slowness of lexical acquisition explains, in its turn, the late syntactic development manifested by those children who finally acquire language and, as we shall shortly see, may explain some linguistic specificities and difficulties of verbal autistic patients. 3.6.

The differences and similarities between autism and Asperger Syndrome

3.6.1.

Questions

Given what is shared between autistic and Asperger Syndrome patients, i.e., all social deficits, and what is not, i.e., the strong delay and specificity of language acquisition in autism as opposed to Asperger Syndrome, how could one claim that the deficit in language acquisition which is specific to autism can be explained by the socio-pragmatic deficits which are common to both autism and Asperger Syndrome? This is of course not only a question, but a potential objection to the theory

Autism from a cognitive-pragmatic perspective

327

outlined above. An additional question is why some autistic children never acquire language (in the strong sense that they remain mute), while others do. Let us begin with the second question. It is usually considered that language acquisition and intellectual abilities are dissociated in that an intellectually deficient child can acquire language normally, which seems to indicate that linguistic acquisition can proceed despite severe intellectual disabilities. This has been shown recently by Musolino, Chunyo, and Landau (in press) who investigated syntactic and structural (syntactically-linked) semantic abilities in children with Williams’ Syndrome14 and found them to be identical with normally developing children, in contradiction to other studies (see, e.g., Karmiloff-Smith et al. 1997), that found them linguistically disabled. It may nevertheless be the case that the most severe cases of linguistic disabilities in autism also are those associated with severe intellectual disabilities, which, given the social deficits in autism, cannot be compensated through socio-pragmatics abilities, as may be the case with other intellectual disabilities such as Williams’ Syndrome or trisomy. 3.6.2.

Are Asperger Syndrome and autistic adults really so different with regard to linguistic abilities?

Turning back to the first question, i.e., the apparent discrepancy between autism and Asperger Syndrome, it is not clear that, in adulthood, the discrepancy is as strong as it seems. Indeed, in a study, Howlin (2003) investigated two groups of adults (over 18 years), the first high-functioning autistic patients, the second Asperger patients. She used a battery of standardized vocabulary tests, initially devised to test linguistic abilities in typically developing children of school age (British Picture Vocabulary Scale and Expressive One Word Picture Vocabulary). She found out that the differences between the two groups were slight. Even more surprising was the fact that more than half of the participants in each group were far below ceiling. What this suggests is that, despite the important delay in language acquisition in autistic children, the differences between the two pathologies (in those autistic patients who are not intellectually impaired) tend to level out in later age. What is more relevant here is that this suggests that, despite an apparently normal linguistic acquisition, Asperger children may in fact be limited in their linguistic development, and, more specifically, given that Howlin’s study is focused on vocabulary assessments, in lexical development. In other words, though less impaired in early linguistic acquisition, they may nervertheless be impaired to a level comparable with that of autistic children in later years. The slight discrepancy between the two groups might then be entirely explained through the delay in early acquisition in the autistic group. What remains mysterious, however, is why there is this early difference in linguistic acquisition. A first and perhaps obvious explanation may well be that, Asperger Syndrome patients being usually diagnosed much later than autistic patients,15 their social

328

Anne Reboul, Sabine Manificat and Nadège Foudon

deficits in infancy are not well described and may be much less acute than those of autistic children. Thus, they might be less impaired, for instance, in joint attention. There are however other possibilities. 3.6.3.

Is autism sometimes associated with specific language impairment?

Another possibility is that, in addition to their social deficits, children with autism have a specific language impairment of some kind (in other words, there is in autism a comorbidity between the social deficits and a specific language impairment, while Asperger children would only be impaired in social cognition and behaviour). This second possibility has some evidence in its favour. As said above, Tager-Flusberg et al. (1990) conducted a longitudinal study on American autistic children. Their conclusions at the time were that linguistic acquisition in autistic children, despite the delays and the relative slowness of the process, followed the same acquisitional steps as that of typically developing children in terms of syntactic milestones. This seemed to exclude any kind of specific language impairment. However, in a later study, Kjelgaard and Tager-Flusberg (2001) tested a group of 89 children with autism (aged 4 to 14) using standardized linguistic tasks. They got the usual result of highly heterogeneous linguistic development and further tested the less able group with a test which has been considered as an acid test for specific language impairment, i.e., non-word repetition (see TagerFlusberg and Cooper 1999). These less linguistically able autistic children, as do SLI children, fail tests of non-word repetition, which suggests that they may indeed evidence a comorbidity between the social deficits typical of autism and SLI. The results of these two studies may appear to be contradictory, though this is less obvious on reflection, given the rather high performance of the group of children in the initial longitudinal study and given that the “autism + SLI” hypothesis is specifically directed toward the less able group of autistic children. However, other hypotheses are possible. 3.6.4.

The critical period hypothesis

A recent review (see Durleman and Zufferey 2009) both documents limits in syntactic performance in autistic patients and criticizes the hypothesis that poor linguistic performance is due to a comorbidity with specific language impairments. The authors propose that the limited syntactic abilities in autistic patients is due to the delayed acquisition which has the consequence that syntax has only a restricted time in which to develop before the end of the critical period16 and, indeed, stops developing before it reaches the standard level, hence the persisting syntactic limitations, even in verbal autistic subjects, in adulthood. On that view, autistic patients will be more or less syntactically impaired, depending on their initial delay and on the rate at which their vocabulary develops,

Autism from a cognitive-pragmatic perspective

329

until it reaches the critical threshold at which syntax is triggered. There is unfortunately no study comparing Asperger and high functioning autistic adults’ syntactic performances: a discrepancy between their performances might be related to the difference between the mean dates of syntax onset in the two groups (26 months in Asperger as contrasted with 52 months in autistic children). This could bolster the critical period hypothesis, which, as just outlined, is entirely compatible with the socio-pragmatic hypothesis regarding lexical acquisition. 3.7.

Summary

To sum up, the social deficits in autism may go a long way to explain both the extreme delay in language acquisition and the remaining deficits in adulthood, by their effects on lexical acquisition, with the consequence that the critical threshold at which syntax is triggered occurs so late that syntactic acquisition proper has a limited period during which it can develop. This itself explains the remaining syntactic limitations of autistic patients in adulthood, as well as their deficits in theory of mind, with consequences on pragmatics, which are a strong characteristic of autistic communication in adulthood.

4.

Pragmatic deficits in verbally autistic patients

Although, as just noted, recent studies have listed lexical deficits in both Asperger and autistic verbal adults and syntactic deficits in autistic verbal patients, former studies mainly reported an intact language but strongly impaired communication abilities (e.g., Frith 1990). These encompass difficulties with non-literal communication (tropes, metaphor, irony, indirect speech acts, incomplete instructions relying on common knowledge, implicatures, etc.), with politeness, turn-taking in conversation as well as in subject maintenance or, equally, with persistence on subjects which are inappropriate in the context (rude or boring to conversation partners who do not share the Asperger or autistic speaker’s specific interests17). Pragmatic and discourse difficulties meet in that both may be due to some deficit in theory of mind, as we will now see. 4.1.

Theory of mind

Theory of mind is the (probably species-specific) ability which allows humans to predict and explain others’ behaviours through the mental states (beliefs, desires, intentions, emotions, etc.) which they attribute to them. A simple example is when one sees someone else going to the refrigerator, opening it, taking out a water bottle, filling a glass and drinking it. If one is a typically developing human, by age 5 at the latest, one will be able to explain that behaviour through the fact that the

330

Anne Reboul, Sabine Manificat and Nadège Foudon

agent was thirsty and believed that there was water in the refrigerator. On the prediction side, if one knows that an agent is thirsty and believes that there is water in the refrigerator, one can predict that the agent will go to the refrigerator, take the water bottle out, etc. There are a few different theories on what theory of mind is. A first family of views is to the effect that theory of mind actually is a theory (see, e.g., Fodor 1992). This first family of views can be subdivided according to whether its proponents take theory of mind to be both modular and innate (as Fodor does) or to be modular, but not innate beyond the postulation of cognitive biases (see, e.g., Gopnik and Meltzoff 1993). A second family of views defends the notion that theory of mind is, in fact, not at all a theory, but mainly operates through simulation. The simulator puts him/herself in the shoes of the person whose behaviour (s)he wants to predict, using his/her own cognitive processes “off-line”, accessing the result through introspection and transferring that result to the agent’s actions. The notion of simulation has received partial support from the discovery of mirror neurons (i.e., neurons which fire both on performing an action and on seeing someone else performing the same action: see Rizzolatti and Sinigaglia 2007) and has been developed accordingly (see Goldman 2006). Let us now shortly outline the involvement of theory of mind in pragmatic processing. 4.2.

The contribution of theory of mind to pragmatic processing

One can see the early history of pragmatics (speech act theory, see Austin 1962; Searle 1969) as a continuous endeavour to integrate language, and particularly language use, in action, i.e. to consider language use as a behaviour on a par with any other kind of public behaviour. What this means, basically, is that language use, i.e., the production of utterances, works at least in part like any kind of intentional behaviour (and notably rests on a mixture of beliefs and intentions) and is thus susceptible to the interpretive processes dedicated to the explanation and prediction of behaviour in general. In other words, the interpretation of utterances, just as the interpretation of actions in general, involves theory of mind. This view was explicitly articulated by Grice in the sixties and seventies (though not in these terms; see Grice 1989), who described conversation as a cooperative endeavour and defined meaning as depending on the speaker’s intentions and communicative success as depending on the hearer’s recovery of the speaker’s intentions. The final step was reached in the eighties with Relevance Theory (see Sperber and Wilson 1986/1995), which introduced the notion of ostensive communication (i.e., communication which advertises itself as such) and argued that utterances came with a guarantee of optimal relevance. Relevance was defined, in economic terms, as a balance between cognitive processing efforts and cognitive (roughly informational) effects. Thus a competent speaker would choose his utterance on the basis

Autism from a cognitive-pragmatic perspective

331

of what (epistemic) mental states he/she attributed to his/her hearer. In other words, from its speech acts beginnings to its latest cognitive developments, pragmatics is rife with implicit or explicit appeals to mental states attribution, both in utterance production and in utterance interpretation. A strong prediction of pragmatics is thus that someone who is impaired in terms of theory of mind will also be impaired in linguistic communication, and ASD patients seem to strongly vindicate this view, as we will shortly see, after discussing theory of mind in autism. 4.3.

Theory of mind in ASD

The notion of theory of mind was introduced in a famous paper by Premack and Woodruff (1978, in Behavioral and Brain Sciences), in which they asked whether the chimpanzee had a theory of mind. In the peer comments following the paper, a few philosophers, prominent among whom was Dennett (1978), proposed that a fair test of theory of mind would have to rely on false belief attribution. The resulting test was based on a story in which a first character places an object in a given location, leaves the room and, during its absence, a second character takes the object and places it into another location; then, the first character comes back (Wimmer and Perner 1983). The children are asked first (control question) where the object is now, and second (test question) where the first character will look for it. Though the correct answer to the first question is based on the present situation and does not tap theory of mind, the correct answer to the second question implies attributing a false belief to the character and, hence, implies theory of mind. TD children pass false belief test during their fifth year (by 5 years at the latest) (for a review see Wellmann, Cross, and Watson 2001). In autism, however, results have found that ASD patients are impaired in theory of mind (see, e.g., Baron-Cohen, Leslie, and Frith 1985). Though the hypothesis of a central theory of mind deficit in autism (first proposed, but then rejected, by Uta Frith: see Frith 1990) has by now been abandoned as failing to account for all cognitive peculiarities in autism (notably for both preserved “islands” of ability and for perceptual specificities), there is no doubt that theory of mind is usually deficient in ASD patients. Some Asperger Syndrome and high functioning autistic patients do pass false belief tasks, but it is not clear whether they do so through the same processes as typically developing subjects or through compensatory strategies, such as those often described by high functioning ASD people themselves for dealing with social situations in general (see, e.g., Grandin 1995; Williams 1992): in other words, they rely on explicit analyses of those situations rather than on the fairly quick, automatic processing which is used by TD people. Though this may lead them to pass false belief tasks, such an analytic strategy would be powerless to compensate for the fairly quick processes implicated in linguistic communication.

332 4.4.

Anne Reboul, Sabine Manificat and Nadège Foudon

Pragmatic difficulties per se

One difficulty in assessing pragmatic difficulties in ASD patients is to disentangle the respective contributions of language deficits from pragmatic deficits per se. The most common strategies are either selecting ASD patients with a normal VIQ (verbal IQ), and comparing them with TD subjects, or comparing ASD patients with both specific language impairment (SLI) patients and TD subjects. The idea behind the first strategy is to avoid the potential language deficit, while the idea behind the second strategy is to check the potential implication of a language deficit in the task by using a specifically language disabled population (the SLI patients). One area of potential interest here is disambiguation, because it entails both linguistic competence and the ability to use and process contextual information. Norbury (2005) has investigated these abilities in four populations: patients with ASD and language impairment (ALI), patients with ASD and without language impairment (ASO), patients with language impairment (LI), and typically developing children (TD). The participants were 9- to 17-year-olds. The children were asked to judge whether an image matched the meaning of an ambiguous word, where that meaning could be either the dominant or the subordinate meaning. The results were interesting in that the two groups with language impairments had difficulties in using context for disambiguation, while the two groups with normal VIQ did not. In other words, the relevant factor was not the autism, but the linguistic abilities of the subject. The poor performance of the two language-impaired groups might be explained by the poor semantic knowledge of the verbs used to disambiguate the target words. This shows the importance of disentangling the linguistic deficits from the pragmatic deficits in the poor performance of ASD patients in pragmatic tasks. As said above, high functioning autistic and Asperger patients are impaired in the interpretation of utterances, from clearly figurative ones (metaphor and irony) to clearly literal but incomplete ones. Let us begin with a few examples: (1) Can you pass the salt? (2) Put the chicken in the oven at eleven, please. (3) Has the cat got your tongue? ASD patients have problems with all of the above sentences. Numerous studies (e.g., Frith 1990; Atwood 1998) have reported that an indirect request such as (1) will usually get the answer “yes”, without the child or adult passing the salt. There are a lot of anecdotes from parents of ASD children to the effect that even teenagers or adults will indeed, in response to an order such as (2), put the chicken in the oven, but will not turn it on. Finally, an idiom such as (3) will get a bewildered, if not fearful, reaction. What happens with all these examples is that there is a gap between the literal linguistic interpretation of the utterance and the interpretation the speaker intended

Autism from a cognitive-pragmatic perspective

333

to get across. In (1), the linguistic interpretation is a question, while the utterance is intended as a request; in (2), the linguistic interpretation is that the addressee should put the chicken in the oven, but the speaker’s intention was that the chicken should be put in an oven which is switched on; finally, by the idiom in (3) the speaker intends to ask the addressee why he does not talk. Thus, what seems difficult for ASD patients is the discrepancy between the interpretation based on the linguistic code, which the patients can access, and the interpretation intended by the speaker, which the patients not only seem to have difficulty accessing, but of which they are clearly not aware. In other words, it seems that ASD patients believe linguistic communication to be a mere matter of encoding and decoding, when it is in fact, additionally, a matter of inferential interpretation. It is relevant that being aware of the gap supposes some degree of theory of mind,18 in a way that it is reminiscent of the false belief task: just as, in the false belief task, the incorrect answer to the test question rests on the situation (i.e., describes the actual location of the object) and neglects the first character’s state of mind, in utterance interpretation, ASD patients take the strictly linguistic interpretation to be the right one, ignoring the fact that the speaker’s intention may be different in one way or another. It is, indeed, one of the things reported by high-functioning ASD people (Grandin 1995; Williams 1988), that they are bewildered by the very fact that people may intend to communicate something which is different from what they said (in the non Gricean sense). From this point of view, it is hardly surprising that ASD people find metaphor and irony difficult and this has been specifically investigated by Happé (1993) following Relevance Theory’s analysis of these tropes. Very roughly, Relevance Theory links metaphor interpretation to first-order theory of mind (such as that tested in false belief tasks, e.g., Mary believes that p), while irony, considered as echoic, is supposed to necessitate second-order theory of mind (e.g., John intends Mary to believe that p). Happé contrasted three types of utterances: similes (supposed to be literal and thus interpretable without any theory of mind), metaphors (necessitating a first-order theory of mind) and ironies (necessitating a second-order theory of mind). She tested both verbal autistic children (aged 9 to 26 years) and children with mild intellectual disabilities (aged 12 to 36 years) both on theory of mind first and second-order tasks and on understanding of all three types of sentences. The results were in line with Relevance Theory’s predictions, i.e., similes were understood by all participants, metaphors were understood by participants with at least a first-order theory of mind, while irony was understood only by participants with second-order theory of mind. An additional finding of the study was that the verbal age needed for autistic patients to understand metaphor and irony was much higher than that needed for the controls who were not as delayed or impaired in terms of theory of mind. What this dissociation between language and pragmatic interpretation shows is of course not that one does not need language to interpret metaphor and irony, but rather that language is not enough. Thus for those ASD patients who

334

Anne Reboul, Sabine Manificat and Nadège Foudon

do not have first-level theory of mind (the majority of them), linguistic communication will be problematic in that the very pragmatic interpretation of even simple utterances will be impaired. Pexman and colleagues (2011) have focused on irony and tested three groups of 8-year-olds: patients with high-functioning ASD (with a verbal ability of at least 5 years of age), and two groups of typically developing children, one matched in chronological age, the other matched in verbal age. To facilitate the task, children were asked to judge the speaker’s belief, the speaker’s humour and the speaker’s intent (critical or complimentary), by respectively answering a question, choosing a face in a face scales (from very funny to very serious) and a stuffed toy (either a soft smiling duck, or sharp-teethed shark). In this less demanding task (as the responses were non-verbal and forced-choice), children with ASD performed as did TD children in judging speaker’s belief and speaker’s intent, but were less good at recognizing speaker’s humour. Further analyses of response time and eye gaze suggested, however, that the ASD group did not use the same process as did the control groups and the hypothesis is that they used an explicit, rule-based rather than a simulative strategy. The difficulty the children with ASD had with spotting the humour in Pexman et al.’s study echoes a general difficulty that autistic patients seem to have with humour in general. This was addressed in a paper by Emerich and colleagues (2003), who compared two groups of 11- to 17-year-olds, one high-functioning ASD group, and one TD group, on their ability to choose the right ending to a sequence of cartoons or to a joke. Though the two groups performed similarly on the cartoon task, the ASD group performed significantly more poorly than controls on the more abstract joke tasks. The ability to appreciate humour depends both on the ability to process context and identify the underlying ‘incongruity’, but it also depends on the ability to identify the speaker’s intention. Still another area of interest here lies in the ability to make context-based inferences. This has been investigated in the interpretation of questions by Loukusa et al. (2007) in a comparison of three groups of children: 7- to 9-year-olds and 10to 12-year-olds high-functioning ASD patients, and a control group of 7- to 9-year-old TD children. The questions used in the experiment differed in how much their interpretation relied on contextual inference: reference assignment (Who is running on the road?); enrichment (What time might it be?); basic implicature (Why is the boy holding a book over his head?); feeling (How does the boy feel?); routine (What does the mother mean?). In addition, the children were asked to justify their answers. The results indicated that the ASD group had difficulties with the contextually demanding questions, though the older age group did better than the younger. While all groups answered the reference assignment questions easily, the younger ASD group had difficulties with enrichment questions, and both ASD groups had difficulties with basic implicature, routine and feeling questions.

Autism from a cognitive-pragmatic perspective

335

Implicatures are of course the paradigmatic pragmatic process, and they were first investigated by Surian and colleagues (1996), who compared high-functioning autistic patients, SLI patients and typically developing controls (all groups 10- to 12-year-olds). The children were presented with two dolls who answered a question posed by a third doll, either in keeping with Gricean maxims or not, and asked to point out the doll who ‘said something silly’. All the controls, the SLI patients and three (out of eight) ASD patients performed significantly above chance. The five remaining patients with high-functioning autism failed the task. There was a straightforward association between success in the implicature task and success in the false belief task in autistic patients. Rather different results have been found regarding scalar implicatures. Scalar implicatures arise when the choice of a weaker term implicates the negation of a stronger term (i.e. The orchestra played some Beethoven sonata implicates The orchestra did not play all Beethoven sonata). Pijnacker and colleagues (2009) compared high-functioning autistic patients, Asperger’s syndrome and matched controls (adults of 26–27 years of age). They used two scalar triggers, some and or and asked subjects to judge whether statements were true or false (e.g. All sparrows are birds; Some sparrows are birds; All birds are sparrows; Some birds are sparrows). All participants derived the scalar implicature quite frequently and there was no significant difference between the groups. Further analyses revealed that lower verbal IQ in the autistic group was associated with less pragmatic responses. Again, the autistic group was significantly slower in answering than the other two groups. This rather unexpected result could be explained by the fact that the demands of scalar implicatures are rather less strenuous than those of other pragmatic phenomena, asking only for first-order theory of mind. A second reason might be that both groups of patients had average to high IQ and were adults and so had had time to develop strategies different from those of the control group. Finally, a further study by Chevallier and colleagues (2010) investigated the role of prosody in the derivation of scalar implicature in high-functioning autistic adolescents. The logic was that stress on the implicature trigger would facilitate the derivation of the implicature in typically developing controls, but not in the patient group, because of the well-known difficulties of autistic patients with prosody. Patients were presented with two pictures and a statement and asked to judge whether the statement is true or false (i.e. a picture of a window, a picture of a monkey and the statement There is a window or/OR a monkey). Contrary to expectations, the two groups did not differ in their abilities to make the scalar implicature. This suggests that individuals in the high-functioning range of ASD do have some pragmatic (including prosodic) competence. It should be noted, however, that these people, despite their good performance in tests, can still be handicapped in everyday conversation, which is more taxing than experimental setups.

336 4.5.

Anne Reboul, Sabine Manificat and Nadège Foudon

Discourse deficits in autism

A first investigation of discourse deficits in autism was made by Eales (1993), who compared a group of patients with a childhood diagnosis of autism to a group of patients with a childhood diagnosis of receptive language disorder (both groups adults: 23- to 25-year-olds). The study was done on a corpus recorded while the participants were interviewed using the ADOS (Autism Diagnostic Observation Schedule: a standardized questionnaire to diagnose autism). A range of factors was investigated: the rate of inappropriate utterances; the rate of empty turns; the ratio of initiations; as well as sub-categories of these. The autistic group was much worse as far as stereotyped language use and impairment of communicative intention (two sub-types of inappropriate utterances) were concerned. These were associated with greater social impairment in ADOS. In a recent paper, Hale and Tager-Flusberg (2005) investigated discourse deficits in autism by the analysis of corpora recorded in child-mother interaction and coding for noncontingent utterances (i.e., utterances not relevant to the prior speaker’s topic). The study included 57 verbal children with autism (age 4 to 13) and aimed at correlating their discourse difficulties to their score in a standardized test for autism, ADOS. This is in fact what they found, i.e., scores in noncontigent discourse and scores in ADOS were significantly related. More tellingly, “impairment in the domain of communication,19 but not in social interaction, was significantly related to noncontigent discourse, independent of more general language skills” (Hale and Tager-Flusberg 2005: 523). This is quite reminiscent of Happé’s results regarding the interpretation of metaphor and irony, given that, once again, the major factor does not seem to be language as such, but rather communication and, here, the authors suggest that discourse deficits may be related to theory of mind rather than to social abilities per se, though it should be noted that the children were not specifically tested for theory of mind. Finally, the narrative abilities of autistic patients have been investigated by Colle and colleagues (2008), who compared a group of adults with either highfunctioning autism or Asperger syndrome and matched controls (both age 27 years of age) on a range of factors: anaphoric pronouns; temporal devices; mental state expression; length (3 measures: number of words, number of episodes; number of boy/dog episodes). They asked the participants to tell a story based on a picture book. The two groups differed neither in linguistic abilities, nor in their abilities to comprehend and extract the plot. But the patient group produced less cohesive narratives than the control group: they used fewer pronominal expressions and more full noun phrases, and fewer temporal expressions than the control group. Though the two groups did not differ in the number of mental state expressions they produced, the autistic group did not produce elaboration or explanation regarding the mental states of the characters. Thus, though their linguistic abilities did not set

Autism from a cognitive-pragmatic perspective

337

them apart from the control group, the autistic patients’ pragmatic abilities did not follow suit. 4.6.

Summary

To sum up, though there is no doubt that ASD patients are impaired in pragmatic processes in everyday conversation, this may be less obvious in experimental setups that limit the demands made on them. Their pragmatic deficits are clearly related to deficits in theory of mind.

5.

Conclusion

ASD patients are impaired in social cognition and communication. Autistic patients differ from Asperger patients in that the first are more impaired in language in infancy, being either strongly delayed in language acquisition or failing to acquire language at all. However, in adulthood, high-functioning autistic and Asperger patients seem to level out in terms of linguistic abilities, at least as far as the lexicon is concerned, members from both groups, however, often remaining far below the lexical level of typically developing adults. Additionally, ASD patients are notable for pragmatic difficulties which seem partly independent of the linguistic level they may attain: in other words, even in those patients who are verbal and high functioning, difficulties in pragmatic utterance interpretation as well as in discourse management remain very prominent and are a source of major communicative difficulties, though this is more obvious in natural conversations than in experimental setups. Though it might seem that both the difficulties in acquiring language during childhood and the pragmatic deficits in adulthood stem from the same underlying cognitive deficit, it is not clear that this is the case, and, indeed, it seems that the pragmatic deficits in adulthood may be more directly related to deficits in theory of mind, while the acquisition difficulties in childhood may be more directly related to more primitive social difficulties, which may persist in adulthood without directly impacting or explaining the pragmatic difficulties themselves. This, however, should not be taken to mean that there are no relations between the social deficits and theory of mind impairments in ASD disorders. As pointed out by Povinelli and Vonk (2004), there can be no theory of mind without the lesser ability to “read behaviour”, i.e., to identify others’ actions. If this ability is deficient in autism (as seems evidenced both by the attentional deficits data in Ceponiene et al. 2003 and by the Cattaneo et al. 2007 study of mirror neurons in autistic children), then it may be the case that theory of mind fails to develop because of insufficient or faulty inputs from that less complex behaviour reading ability. Thus, though the two abilities, theory of the mind on the one hand, and social deficits in

338

Anne Reboul, Sabine Manificat and Nadège Foudon

behaviour reading on the other, are different and indeed differently impact language at different periods in time, they might nevertheless be related in that the first depends on the second for its normal development. And both have a role to play in socio-pragmatic cognition and in the development of linguistic communication.

Notes 1. It seems that in the next version of DSM (Diagnostic and Statistical Manual), the distinction between the two pathologies will disappear. Be that as it may, the differences in language acquisition among autism spectrum patients legitimate talk of two groups in the present paper. 2. There is an ongoing debate (see, e.g., Nassar et al. 2009; King and Bearman 2009) on whether the increase in prevalence merely reflects more frequent diagnoses (including milder cases) or a true increase in the occurrence of the pathology. This will not concern us here. 3. Restricting the ASD to the best characterized pathologies, that is autism and Asperger Syndrome. 4. The ages given are means. 5. The study was financed by Fondation de France, grant n° 2005010231. 6. We want to express our gratitude to the autistic children who took part in the study, their parents and the staff at the Hospital Saint-Jean de Dieu. 7. Mean Length of Utterance or MLU – calculated as mean number of words, rather than morphemes per utterance – is a rough but standard measure of linguistic acquisition in children both with and without language impairments (see Leonard 2000). 8. Given the ratio of boys to girls in autistic children (around four boys for one girl), we will refer to autistic children in general by the masculine pronoun. 9. The comparatively high prevalence of girls in the cohort (1 girl for 2 boys) relative to the general prevalence of girls among autistic children (1 girl for 4 boys) is a pure matter of chance. 10. For comparison, we used the French corpora of normally developing children available on CHILDES, matching autistic and typically developing children on the basis of their MLU in the first session recorded in the corpora. 11. The reader should remember that the corpora described here are in French. The distribution of words in syntactic categories is language dependent, and different standard distributions would prevail in other languages. 12. Indeed, the lexicon seems to remain significantly poorer in adult high-functioning autistic and Asperger patients than in typically developing subjects (see Howlin 2003). 13. One could debate whether shared intentionality is not as fundamental (see the discussion below). 14. Williams Syndrome has been said to be the reverse of autism as far as social abilities are concerned, Williams children being described as extremely social, chatty and friendly. For a less enthusiastic view, see Tager-Flusberg and Sullivan (2000), who, however, addressed the rather cognitively demanding false belief task (see below 4.3) in Williams Syndrome.

Autism from a cognitive-pragmatic perspective

339

15. This is directly linked to the linguistic acquisition delay in autism: when parents of an autistic child note that he has not begun talking (or has regressed in his acquisition) by the age of 3, they usually have the child tested for possible disabilities. Though language delay is not the only reason for medical consultation by parents of autistic children, it does account for a fair number of diagnoses. Given that Asperger children, though delayed, are well in the normal range of age for linguistic acquisition, this usual incentive for seeking medical help does not exist. As a result, Asperger children are usually diagnosed much later (often by adolescence if not later) (see, e.g., Atwood 1998). 16. The hypothesis of a critical period during which linguistic acquisition (notably syntax and phonology) can develop naturally and after which language can only be learned through more or less explicit processes such as those used to teach second languages in adolescence and adulthood in typically developing individuals is widely shared, and not restricted to the generative perspective on linguistic acquisition. 17. Specific and restricted interests are an important diagnostic criterion for all ASD (see above 2.2). They may include train time tables, calendars, highly specialized collections, etc. 18. It should be noted that young typically developing children may manifest some of the difficulties evidenced by autistic children, though these usually are much milder and less widespread, being more limited to figurative utterances and idioms. 19. As many standardized tests, including the classical WISC for IQ, the ADOS is divided in different domains, yielding both scores for those different domains and a general score indicating the severity of the pathology.

References Aslin, Richard N., Jenny R. Saffran and Elissa L. Newport 1998 Computation of conditional probability statistics by 8-month-old children. Psychological Science 9(4): 321–324. Asperger, Hans 1944 Die “autistischen Psychopathen” im Kindesalter. Archiv für Psychiatrie und Nervenkrankheiten 117: 76–136. Atwood, Tony 1998 Asperger Syndrome: A Guide for Parents and Professionals. New York/Philadelphia: Jessica Kingsley Publishers. Austin, John L. 1962 How to Do Things with Words. Oxford: Oxford University Press. Baron-Cohen, Simon 1995 Mindblindness: An Essay on Autism and Theory of Mind. Cambridge, MA.: The MIT Press. Baron-Cohen, Simon, Alan Leslie and Uta Frith 1985 Does the autistic child have a theory of mind? Cognition 21: 37–46. Bates, Elisabeth, Luigia Camaioni and Virginia Volterra 1975 The acquisition of performatives prior to speech. Merrill-Palmer Quarterly 21: 205–224. Bloom, Paul 2000 How Children Acquire the Meaning of Words. Cambridge, MA: The MIT Press.

340

Anne Reboul, Sabine Manificat and Nadège Foudon

Bryson, Susan E., Lonnie Zwaigenbaum, Jessica Brian, Wendy Roberts, Peter Szatmari, Vicki Rombough and Catherine McDermott 2007 A prospective case series of high-risk infants who developed autism. Journal of Autism and Developmental Disorders 37: 12–24. Cattaneo, Luigi, Maddalena Fabbri-Destro, Sonia Boria, Cinzia Pieraccini, Annalisa Monti, Giuseppe Cossu and Giacomo Rizzolatti 2007 Impairment of action chains in autism and its possible role in intention understanding. Proceedings of the National Academy of Sciences 104: 17825– 17830. Ceponiene, Rita, Tuulia Lepistö, Anna Shestakova, Raija Vanhala, Paavo Alku, Risto Näätänen and Kuniko Yaguchi 2003 Speech-sound-selective auditory impairment in children with autism: they can perceive but do not attend. PNAS 100(9): 5567–5572. Chevallier, Coralie, Deirdre Wilson, Francesca Happé and Ira Noveck 2010 Scalar inferences in autism spectrum disorders. Journal of Autism and Developmental Disorders 40:1104–1117. Colle, Livia, Simon Baron-Cohen, Sally Wheelwright and Heather van der Lely 2008 Narrative discourse in adults with high-functioning autism or Asperger syndrome. Journal of Autism and Developmental Disorders 38: 28–40. Dennett, Daniel 1978 Beliefs about beliefs. Behavioral and Brain Sciences 4: 568–570. Durleman, Stephanie and Sandrine Zufferey 2009 The nature of syntactic impairment in autism. Rivista di Grammatica Generativa 34: 57–86. Eales, Martin 1993 Pragmatic impairments in adults with childhood diagnoses of autism or developmental receptive language disorder. Journal of Autism and Developmental Disorders 23: 593–617. Eilan, Naomi, Christopher Hoerl, Theresa McCormack and Johannes Roessler 2005 Joint attention: Communication and other minds: Issues in Philosophy and Psychology. Oxford: Oxford University Press. Emerich, David, Nancy Creaghead, Sandra Grether, Donna Murray and Carol Grasha 2003 The comprehension of humorous materials by adolescents with high-functioning autism and Asperger’s syndrome. Journal of Autism and Developmental Disorders 33: 253–257. Fodor, Jerry 1992 A theory of the child’s theory of mind. Cognition 44(3): 283–296. Foudon, Nadège 2008 L’acquisition du langage par les enfants autistes: une étude longitudinale. Ph.D. Thesis, Doctoral school of Cognitive Sciences, Université LumièreLyon2. Frith, Uta 1990 Autism: Explaining the Enigma. Oxford: Oxford University Press. Frith, Uta (ed.) 1991 Autism and Asperger Syndrome. Cambridge: Cambridge University Press. Gastbeg, Holly Zajac 1999 Do Individuals with Autism Process Categories Differently? The Role of Typicality. B.S. in Neurosciences and Psychology, Allegheny College.

Autism from a cognitive-pragmatic perspective

341

Goldman, Alvin 2006 Simulating Minds: the Philosophy, Psychology, and Neuroscience of Mindreading. Oxford: Oxford University Press. Gopnik, Alison and Andrew N. Meltzoff 1993 Words, Thoughts and Theories. Cambridge, MA.: The MIT Press. Grandin, Temple 1995 Thinking in Pictures and other Reports from my Life with Autism. New York: Doubleday. Grice, Paul 1989 Studies in the Ways of Words. Cambridge, MA.: Harvard University Press. Hale, Courtney M. and Helen Tager-Flusberg 2005 Brief report: the relationship between discourse deficits and autism symptomatology. Journal of Autism and Developmental Disorders 35(4): 519–524. Happé, Francesca 1993 Communicative competence and theory of mind in autism: A test of relevance theory. Cognition 48(2): 101–119. Hirsh-Pasek, Kathryn and Roberta Michnick Golinkoff 1996 The Origins of Grammar: Evidence from Early Language Comprehension. Cambridge, MA.: The MIT Press. Hirsh-Pasek, Kathryn and Roberta Michnick Golinkoff 2006 Action Meets Words: How Children Learn Verbs. Oxford: Oxford University Press. Howlin, Patricia 2003 Outcome in high-functioning adults with autism with and without early language delays: implications for the differentiation between autism and Asperger syndrome. Journal of Autism and Developmental Disorders 33: 3–13. Kanner, Leo 1947 Autistic disturbances of affective contact. The Nervous Child 2: 217–250. Karmiloff-Smith, Annette, Julia Grant, Ioanna Berthoud, Mark Davies, Patricia Howlin and Orlee Udwin 1997 Language and Williams Syndrome: How intact is “intact”? Child Development 68(2): 246–262. King, Marissa and Peter Bearman 2009 Diagnostic change and the increased prevalence of autism. International Journal of Epidemiology [Advanced publication, Sept. 2009]: 1–11. Kjelgaard, Margaret M. and Helen Tager-Flusberg 2001 An investigation of language impairment in autism: Implications for genetic subgroups. Language and Cognitive Processes 16(2–3): 287–308. Klin, Ami 1991 Young autistic children’s listening preferences in regard to speech: A possible characterization of the symptom of social withdrawal. Journal of Autism and Developmental Disorders 21(1): 29–42. Kogan, Michael D., Stephen J. Blumberg, Laura A. Schieve, Coleen A. Boyle, James M. Perrin, Reem M. Ghandour, Gopal K. Singh, Bonnie B. Strickland, Edwin Trevathan and Peter C. van Dyck 2009 Prevalence of parent-reported diagnosis of autism spectrum disorder among children in the US, 2007. Pediatrics 124(4): 1–9.

342

Anne Reboul, Sabine Manificat and Nadège Foudon

Leekam, Sue, Simon Baron-Cohen, Daniel Perrett, Maarten Milders and Susan Brown 1993 Eye-direction detection. A dissociation between geometric and joint-attention skills in autism. Unpublished manuscript, Institute of Social Psychology, University of Kent. Leonard, Laurence B. 2000 Children with Specific Language Impairment. Cambridge, MA: The MIT Press. Loukusa, Soile, Eeva Leinonen, Sanna Kuusilo, Katja Jussila, Marja-Leena Mattila, Nuala Ryder, Hanna Ebeling and Irma Moilanen 2007 Use of context in pragmatic language by children with Asperger syndrome or high-functioning autism. Journal of Autism and Developmental Disorders 37: 1049–1059. Moore, Chris and Philip J. Dunham (eds.) 1995 Joint Attention: Its Origins and Role in Development. Hillsdale, NJ: Lawrence Erlbaum. Musolino, Julien, Gitana Chunyo and Barbara Landau in press Uncovering knowledge of core syntactic and semantic principles in individuals with Williams Syndrome. Language Learning and Development 6: 126–161. Nassar, Natasha, Glenys Dixon, Jeny Bourke, Carol Bower, Emma Glasson, Nick de Klerk and Helene Leonard 2009 Autism spectrum disorders in young children: effects of changes in diagnostic practices. International Journal of Epidemiology [Advance publication, Sept. 2009]: 1–10. Norbury, Courtenay Frazier 2005 Barking up the wrong tree? Lexical ambiguity resolution in children with language impairments and autistic spectrum disorders. Journal of Experimental Child Psychology 90: 142–171. Pexman, Penny, Kristin Rostad, Carly McMorris, Emma Climie, Jacqueline Stowkowy and Melanie Glenwright 2011 Processing of ironic language in children with high-functioning autism spectrum disorder. Journal of Autism and Developmental Disorders 41: 1097– 1112. Pijnacker, Judith, Peter Hagoort, Jan Buitelaar, Jan-Pieter Teunisse and Bart Geurts 2009 Pragmatic inferences in high-functioning adults with autism and Asperger syndrome. Journal of Autism and Developmental Disorders 39: 607–618. Povinelli, Daniel and Jennifer Vonk 2004 We don’t need a microscope to explore the chimpanzee’s mind. Mind and Language 19(1): 1–28. Premack, David and Guy Woodruff 1978 Does the chimpanzee have a theory of mind? Behavioral and Brain Sciences 4: 515–526. Rapin, Isabelle and Michelle Dunn 1997 Language disorders in children with autism. Seminars in Pediatric Neurology 4(2): 86–92. Reznik, J. Steven, Grace T. Baranek, Shaye Reavis, Linda R. Watson, Elizabeth R. Crais 2007 A parent-report instrument to identify one-year-olds at risk for an eventual diagnosis of autism: The first year inventory. Journal of Autism and Developmental Disorders 37: 1691–1710.

Autism from a cognitive-pragmatic perspective

343

Rizzolatti, Giacomo and Corrado Sinigaglia 2007 Mirrors in the Brain: How our Minds Share Actions and Emotions. Oxford: Oxford University Press. Saffran, Jenny R., Richard N. Aslin and Elissa L. Newport 1996 Statistical learning by 8-month-old infants. Science 274: 1926–1928. Saffran, Jenny R., Elissa L. Newport and Richard N. Aslin 1996 Word segmentation: the role of distributional cues. Journal of Memory and Language 35: 606–621. Schuler, Adriana L. and Barry M. Prizant 1985 Echolalia. In: Eric Schopler and Gary B. Mesibov (eds.), Communication Problems in Autism, 163–184. New York/London: Plenum Press. Scott, Fiona J., Simon Baron-Cohen, Patrick Bolton and Carol Brayne 2002 Brief report: prevalence of autism spectrum conditions in children aged 5–11 years in Cambridgeshire, UK. Autism 6(3): 231–237. Searle, John R. 1969 Speech Acts: an Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Sigman, Marian and Connie Kasari 1995 Joint attention across contexts in normal and autistic children. In: Chris Moore and Philip J. Dunham (eds.), Joint Attention: Its Origins and Role in Development, 189–204. Hillsdale, NJ: Lawrence Erlbaum. Soulières, Isabelle, Laurent Mottron, Daniel Saumier and Serge Larochelle 2006 Atypical categorical perception in autism: autonomy of discrimination. Journal of Autism and Developmental Disorders 37: 481–490. Sperber, Dan and Deirdre Wilson 1986/95 Relevance: Communication and Cognition. Oxford: Basil Blackwell. [Second edition with Postface 1995]. Surian, Luca, Simon Baron-Cohen and Heather Van der Lely 1996 Are children with autism deaf to Gricean maxims? Cognitive Neuropsychiatry 1: 55–71. Tager-Flusberg, Helen and Judith Cooper 1999 Present and future possibilities for defining a phenotype for specific language impairment. Journal of Speech, Language, and Hearing Research 42: 1001– 1004. Tager-Flusberg, Helen and Kate Sullivan 2000 A componential view of theory of mind: evidence from Williams syndrome. Cognition 76: 59–89. Tager-Flusberg, Helen, Susan Calkins, Tina Nolin, Therese Baumberger, Marcia Anderson and Ann Chadwick-Dias 1990 A longitudinal study of language acquisition in autistic and Down syndrome children. Journal of Autism and Developmental Disorders 20(1): 1–21. Tomasello, Michael 2005 Constructing a Language: A Usage-based Theory of Language Acquisition. Cambridge, MA: Harvard University Press. Tomasello, Michael 2008 Origins of Human Communication. Cambridge, MA: The MIT Press.

344

Anne Reboul, Sabine Manificat and Nadège Foudon

Tomasello, Michael, Melinda Carpenter, Josep Call, Tanya Behne and Henrike Moll 2005 Understanding and sharing intentions: The origins of cultural cognition. Behavioral and Brain Sciences 28: 675–735. Wellman, Henry M., David Cross and Julanne Watson 2001 Meta-analysis of theory-of-mind development: the truth about false belief. Child Development 72: 655–684. Williams, Donna 1992 Nobody Nowhere. London: Doubleday. Wimmer, Heinz and Joseph Perner 1983 Beliefs about beliefs: representation and constraining function of wrong belief in young children’s understanding of deception. Cognition 13: 103–128. Winkler, Isvan, Helena Kushnerenko, Janos Horvath, Rita Ceponiene, Vineta Fellman, Minna Huotilainen, Risto Näätanen and Elyse Sussman 2003 Newborn infants can organize the auditory world. Proceedings of the National Academy of Sciences 100: 11812–11815.

13.

Aphasia: The pragmatics of everyday conversation Suzanne Beeke

1.

Introduction

The condition of aphasia constitutes loss or impairment of language function in people with previously normal language abilities. It is acquired as a result of focal damage usually in the left cerebral hemisphere, and most commonly caused by a stroke. Each year, in England and Wales alone, over 130,000 people have a first stroke. Approximately one third are left with permanent disability in the form of aphasia.1 Most people with aphasia experience some level of impairment in all areas of language: auditory comprehension, reading comprehension, spoken language production and writing, with the severity of difficulty varying enormously between language areas and across individuals. Clinical classifications attempt to distinguish between aphasia types such as Broca’s, Wernicke’s, Anomic and Global, each with distinct symptom complexes based on key parameters such as fluency, auditory comprehension, word finding difficulty and grammar (Goodglass and Kaplan 1983). Aphasic language is commonly characterised via psycholinguistic and cognitive neuropsychological testing (Kay, Lesser, and Coltheart 1992; Goodglass, Kaplan, and Barresi 2001). The routine assessment of pragmatics in aphasia came to the fore around 30 years ago, driven by the clinical observation that people with aphasia communicate better than they speak (most famously noted by Holland 1980). Pragmatic assessments commonly focus on language output and take the form of observation checklists, interviews and role plays, with data coded for utterance function and judged for communicative adequacy. Research using such assessments has, for the most part, concluded that people with aphasia have relatively intact pragmatic skills. In recent years a qualitative, data-driven methodology called Conversation Analysis (CA) has been applied to the investigation of aphasic language use in everyday communicative contexts. It has revealed much about how aspects of language and non-verbal behaviour are used and adapted to achieve the goal of taking a turn at talk in everyday conversation, in the presence of aphasia. With its focus on the mechanics of turn taking, topic management and the collaborative achievement of mutual understanding, CA gives us an additional tool with which to explore the pragmatics of aphasia, uncover the ways in which pragmatics can be disrupted, and consider evidence for and against pragmatic impairment as a secondary consequence or a primary disorder of aphasia. In this chapter the term ‘pragmatics’ is used broadly to refer to mechanisms of language use within everyday communi-

346

Suzanne Beeke

cative settings, and particularly in conversation, “the vehicle by which we express what we know and who we are” (Ramsberger and Menn 2003: 285). This chapter begins with a summary of the findings of early pragmatic approaches to aphasic spoken language, before moving on to present an overview of the CA methodology, a review of key CA findings in relation to aphasic conversation and some insights into pragmatic impairment as a primary or secondary consequence of aphasia, based on an analysis of extracts from the conversations of two dyads where one speaker has agrammatic aphasia.

2.

Pragmatics and aphasia

It has long been recognised that the study of aphasia, and indeed of language disorders in general, needs to account for speakers’ abilities within the context of natural language use, and not just their performance on metalinguistic tasks (see for example Holland 1980; Davis and Wilcox 1985; McTear and Conti-Ramsden 1992; Lesser and Milroy 1993; Perkins 2007). Sarno (1969) was the first to highlight the need in aphasiology for a distinction between natural language use and linguistic performance in artificial testing situations. Given that the primary aim of rehabilitation is to improve communicative participation in the real world, not to gain a higher score on an assessment, the last 30 years has seen a significant interest in applying the concepts and theories of pragmatics to aphasia. Early pragmatic investigations of aphasia took a top-down, theory and analyst driven approach, based on a set of derived organisational principles for competent communicative behaviour. Many were heavily influenced by speech act theory (Austin 1962; Searle 1979), and some drew on Gricean maxims (Grice 1989), particularly to explain (in)appropriateness. For example, Holland (1980) developed a standardised test, Communicative Abilities of Daily Living (CADL; CADL-2: Holland, Frattali, and Fromm 1999), which simulates natural communicative exchange via a structured interview in the clinic. The CADL makes use of role play tasks to elicit behaviour that can then be analysed against predetermined pragmatic categories. Other approaches such as the Pragmatic Protocol (Prutting and Kirchner 1987), Profile of Communicative Appropriateness (Penn 1985) and Edinburgh Functional Communication Profile (Wirz, Skinner, and Dean 1990), avoid the artificiality of role play by instead applying a schedule of pragmatic categories to naturalistic interaction between clinician and patient. The reader is directed to Manochiopinig, Sheard, and Reed (1992) and Perkins and Lesser (1993) for an overview of pragmatic assessments. Following extensive investigation utilising such tools, there is considerable agreement that most people with aphasia have relatively intact pragmatic skills (Davis and Wilcox 1985; Holland 1991; Lesser and Milroy 1993). It appears to be people with right cerebral hemisphere brain damage or widespread bilateral

Aphasia: The pragmatics of everyday conversation

347

changes, as in dementia, who have disorders that can be characterised as primarily pragmatic in nature (Perkins, Whitworth, and Lesser 1998; Body and Parker 2005; Perkins 2007). Nevertheless, it is important to note that aphasic language impairments can have pragmatic consequences, in so far as they disrupt conversational timing, turn taking or the ability to clarify talk (Lesser and Milroy 1993; Perkins 2007). On occasion, a secondary pragmatic problem may be created by the use of a strategy deployed to compensate for an aphasic difficulty. This issue will be taken up further in the discussion of CA studies of aphasia, below. An important second strand in the application of pragmatics to aphasia uses pragmatic concepts to inform clinical interventions that aim to maximise communicative effectiveness. This approach to therapy is carried out through two main methods: teaching strategies directly to the person with aphasia (a functional communication approach), and encouraging the development of strategies in the nonaphasic communication partner (see Simmons Mackie et al. 2010 for a systematic review of conversation partner training). Both require the clinician to reinforce the use of current successful strategies, eliminate counter-productive ones, and teach ones that are not in use (see for example Green 1984; Davis and Wilcox 1985; Holland 1991; Kagan 1998; McVicker et al. 2009). Although these top-down pragmatic approaches to assessment and intervention have advanced our understanding of aphasia and its communicative consequences, they are not without their limitations. Firstly, many use procedures such as clinical interviews or role plays to elicit communicative behaviour. This is problematic in the light of research that has found that the so-called “institutional” nature of such dialogues directly affects pragmatic behaviours such as topic initiation and turn taking, as well as linguistic characteristics such as utterance length (Drew and Heritage 1992). Secondly, many early procedures did not study the behaviour of the conversation partner, merely the speaker with aphasia (Lesser and Milroy 1993). As a result, there was and still is a tendency for transactional aspects of language use – how speakers with aphasia “get the message across” – to be emphasised to the detriment of social interactional aspects, such as maintaining interpersonal relationships, yet we know that real-life communication involves both, even if one interactant has aphasia (Kagan 1995).2 Thirdly, clinical interventions such as Conversational Coaching (Holland 1991), Supported Conversation for Adults with Aphasia (Kagan 1998) and Connect partner training (McVicker et al. 2009) tend to teach generic communication techniques; they are not based on a detailed understanding of the specific effects on conversation of an individual’s often very unique pattern of linguistic strengths and weaknesses. Emerging evidence of the diversity and subtlety of conversational problems and strategies that can result from seemingly identical aphasic symptoms, e.g. word finding difficulty (Laakso and Klippi 1999; Yates 2009) or agrammatism (Beeke, Wilkinson, and Maxim 2003a, 2003b, 2007a, 2007b), suggests that the imposition of ‘one size fits all’-strategies may have therapeutic limitations. In an attempt to address this issue, a new therapeutic

348

Suzanne Beeke

approach is being developed that aims to target the conversational strategies of both a person with aphasia and their key conversation partner (Wilkinson et al. 2010; Beeke et al. 2011).

3.

Conversation Analysis

Recent recognition of the importance of everyday conversation to the long term psychosocial welfare of people with aphasia (Lyon 1992; Kagan 1995, 1998; Parr, Duchan, and Pound 2003) has led to an upsurge in the application of qualitative research methodologies, in an attempt to engage with more authentic and naturalistic data (see Damico et al. 1999 for a review). Conversation Analysis (CA) is one such methodology that is gaining attention in aphasiology (Damico, Oelschlaeger, and Simmons-Mackie 1999; Wilkinson 1999a; Beeke, Maxim, and Wilkinson 2007), as well as in other areas of communication disability (Perkins, Whitworth, and Lesser 1998; Bloch and Wilkinson 2004; Dickerson et al. 2005; Radford, Ireson, and Mahon 2006). CA is a systematic procedure for the analysis of recorded, naturally occurring talk produced in everyday human interaction. It grew out of ethnomethodology, a sociological approach originating in the 1960s, via the pioneering works of Harvey Sacks and his collaborators, Emanuel Schegloff and Gail Jefferson. The principal aim is to discover how participants understand and respond to each other via turns at talk, and how such turns are organized into sequences of interaction. The focus is on conversation as an orderly accomplishment that is negotiated between participants. Thus “CA seeks to uncover the organization of talk not from any exterior, God’s eye view, but from the perspective of how the participants display for one another their understanding of ‘what is going on’” (Hutchby and Wooffitt 1998: 15). An analysis begins with actual utterances in real contexts and is not constrained by prior theoretical assumptions; it is a bottom-up, data-driven approach. CA takes the view, from linguistics, that language is a structured system for the production of meaning, but it diverges from many branches of linguistics in that it views language primarily as a vehicle for communicative interaction, rather than as a discrete system for predication (Schegloff 1996). Thus, it believes that phonetics, lexis, grammar and prosody are primarily shaped by their use as tools to accomplish mutual understanding in turn-by-turn interactions. A CA approach makes use of recordings of naturally occurring conversations that would have taken place even if they had not been recorded. There is a preference in the study of communication disorders for video rather than audio recording, because this format captures for the analyst the multimodal resources that were available to the interactants themselves at the time of the conversation. Although proponents of CA acknowledge that video recording may make people selfconscious, it is argued that unobtrusive recording over a period of time can lessen

Aphasia: The pragmatics of everyday conversation

349

these effects (Goodwin 1981). Data are transcribed using a system developed to document the minutiae of conversation such as cut-off words and phrases, fillers (um and er, etc.), pausing and overlap (Atkinson and Heritage 1984), and analysis proceeds via repeated scrutiny of the recording alongside the transcript. 3.1.

Principles of Conversation Analysis

Wilkinson (1999a) summarizes the four main principles of CA as follows: (1) analysis is data- and participant-driven; (2) conversation is orderly; (3) sequential context is important; and (4) there is wariness of quantification. Each of these will be outlined below. The reader is directed to other texts for a detailed explication of CA methodology (Hutchby and Wooffitt 1998; Silverman 1998; Schegloff 2007). 3.1.1.

Principle 1: analysis is data- and participant-driven

An analysis begins with actual utterances in real contexts and is not constrained by prior theoretical assumptions. It aims to describe and explain how the participants display their interpretations of each other’s talk. Evidence about what a speaker’s intention is or whether an utterance is communicatively adequate, is sought by observing the reaction of the other participant in the conversation. How does this participant interpret the utterance and does he or she consider it to be adequate? Thus, the conversation analyst avoids making his or her own interpretations or judgements about the data. This approach is different from analyst-driven methods prevalent in the investigation of pragmatics in aphasia, where the clinician is commonly required to make interpretations about (1) how an utterance should be coded, using a theoretically driven categorization scheme, and (2) whether it is pragmatically acceptable. 3.1.2.

Principle 2: conversation is orderly

A CA methodology treats conversation as the orderly and organized product of speakers’ systematic methods of talking to and making sense of each other, on an utterance-by-utterance basis. The analysis therefore pays attention to intricate details of conversation such as pausing, overlapping talk, eye gaze, gesture, and the repair of interactional “trouble”, in addition to addressing relevant features of language such as phonetics, prosody, lexis and grammar. 3.1.3.

Principle 3: sequential context is important

Analysis of conversation has shown that, for interactants, one of the most important aspects of any utterance is the point in the conversation at which it occurs, that is, its sequential context. Each turn at talk is constructed within a sequential context of

350

Suzanne Beeke

prior turns, and itself creates the sequential context for upcoming turns. This contextualization is something that hearers use as a resource for interpreting conversational contributions, and speakers use when designing turns to make them understood. 3.1.4.

Principle 4: there is wariness of quantification

A CA methodology makes use of two main forms of analysis (Hutchby and Wooffitt 1998). One involves the investigation of a particular conversational phenomenon, e.g. word search behaviour, across various parts of the data in an attempt to uncover patterns of use (called “making a collection”). Another takes the form of a “single case analysis”, which provides an explication of conversational practices in a single extended sequence of talk. Both methods use the findings of published CA studies to illuminate what is happening in a given piece of talk. Because of the importance of sequential context in shaping how contributions to conversation are designed and understood, there is wariness of removing a conversational phenomenon from its context in order to count it. The concern is that this may lead to phenomena with a superficial similarity being grouped together, when in fact they may be performing quite separate conversational functions in different contexts. Quantitative analysis is possible within CA (Schegloff 1993), but should always be secondary to a qualitative analysis grounded in context and never a stand-alone count of linguistic characteristics, such as grammatical structures, for example.

4.

Conversation analysis and aphasia

Many early CA investigations of aphasia focused on the issue of repair (e.g. Milroy and Perkins 1992; Lesser and Milroy 1993; Wilkinson 1995, 1999b; Laakso 1997, 2003; Lindsay and Wilkinson 1999; Perkins 2003). The considerable CA literature on repair in typical (non-language disordered) conversation has revealed an organised system by which interactants deal with problems in speaking, hearing and understanding. The problem, referred to as a trouble source, is not necessarily an independently verifiable error, hence the preference for the term ‘repair’ instead of ‘correction’. The mechanism for repair involves two steps, whereby repair initiation (which highlights a trouble source) is followed by repair completion. The number of turns over which repair is carried out is referred to as the trajectory. Repair can be initiated or completed by self (the speaker of the trouble source) or by other (the recipient of the trouble source). Thus there are four varieties of repair: self-initiated self repair; other-initiated self repair; self-initiated other repair; and other-initiated other repair. The seminal work of Schegloff, Jefferson, and Sacks (1977) revealed an overwhelming “preference”3 in everyday conversation for selfinitiated self repair, followed by other-initiated self repair. Aphasic repair mechanisms have been shown to differ in several distinctive ways:

Aphasia: The pragmatics of everyday conversation

(i) (ii)

(ii) (iii) (iv)

351

aphasic repair sequences have a much longer trajectory than the one to four turn sequence that is typical of non-aphasic conversations; the preference for self-initiated self-repair is maintained but followed by self-initiated other-repair (not other-initiated self-repair as in non-aphasic talk); repair attempts are often abandoned unresolved; a sequence of confirmatory remarks occurs after the repair to indicate successful resolution; repair often becomes the overt focus of conversational activity.

Specific repair behaviours associated with the common symptom of aphasic word finding difficulty have also formed the focus of much research in the field (see e.g. Lubinski, Duchan, and Weitzner-Lin 1980; Laakso and Klippi 1999; Oelschlaeger 1999; Oelschlaeger and Damico 2003; Helasvuo, Laakso, and Sorjonen 2004). A key finding concerns the presence of a “hint and guess” sequence (Lubinski, Duchan, and Weitzner-Lin 1980). Laakso and Klippi (1999) describe the four stages of the sequence thus: (i) the aphasic speaker signals the presence of a problem (often by making an overt comment about having a word finding difficulty), (ii) the aphasic speaker appeals to their co-participant for help, (iii) the aphasic speaker offers a “hint” and the co-participant offers a “guess” or so-called candidate resolution, and (iv) a confirmation phase occurs. Stage (iii) can often become prolonged, with multiple hints and guesses occurring before the search ends (or is abandoned). Other key findings revealed by this work include a preference for the speaker with aphasia to attempt self-repair before launching a hint and guess sequence, and the use of gaze to indicate when word search is a self-directed activity (via gaze to the middle distance) versus an activity in which the conversational partner can participate (via direct gaze at the partner). Other studies have investigated the process of achieving mutual understanding (e.g. Goodwin 1995, 2003; Goodwin, Goodwin, and Olsher 2002; Klippi 2003; Lind 2002a). Goodwin’s (1995) seminal work in this area remains extremely influential for CA investigations of aphasia. He investigated the conversation of Rob4 (called Chil in later works), whose severe aphasia reduced his output to three words: yes, no and and. Goodwin explored how Rob and his interlocutors, his wife and nurse, collaborated in order to understand each other. This so-called “co-construction” of meaning was made possible by Rob’s ability to understand talk and to produce a competent reply at such a place where it was relevant to do so. Goodwin described the co-construction sequence as a “specialized language game” that involved Rob and his interactants in extensive work designed to uncover what he was trying to say. In it, the interactants offered guesses for acceptance or rejection, using what Goodwin calls “interpretive frameworks” to permit the selection of new guesses. For example, in Extract 1, below, guesses (marked by an arrow in the transcript) are alternatives from the category ‘set of foods that are spread on English muffins’.

352

Suzanne Beeke

Extract 1. From Goodwin (1995: 237, extract 2) Nurse

English muffin? (3.4)

Rob

ye:s. (0.4)

Nurse

a:

⎡nd what would you like on it. ⎣just one.

Wife (0.8) Nurse

A

jelly? (1.0)

Rob

no: (0.8)

Wife

A

Nurse

A

butt

⎡er? ⎣butter?

(0.3) Rob

yes. (0.6)

Nurse

okay.

For Rob, collaborating in this way requires him to attend to the sequential organisation of talk, and to tie his utterances to the talk of others, producing answers to questions at relevant points in the exchange. Goodwin found that, in addition to merely responding to guesses, Rob could also exit from a sequence of guesses that were not within an appropriate interpretive framework, or indicate that his co-participant should continue with guesses where something was “in the ballpark” but not entirely correct. To do these things, he systematically deployed verbal and nonverbal resources. For example, he produced no no whilst averting his gaze from the guesser in order to exit a sequence where guesses were not from an appropriate category set. Conversely, he said yes with rising intonation to indicate that his interactant was “onto something” but should continue to guess further within a category set. In conclusion, Goodwin (1995: 254) highlights how the process of co-constructing meaning requires others to treat Rob as a pragmatically competent interactant, whose behaviours constitute “an effort to say something meaningful, rather than the random movements of a man whose brain has been massively damaged.” More recently, publications in the field have begun to reveal the wide range of strategies used by people with aphasia and their everyday conversation partners, and the interactional mechanisms that govern their use (see e.g. Simmons-Mackie

Aphasia: The pragmatics of everyday conversation

353

and Damico 1996; Oelschlaeger and Damico 1998a, 1998b; Heeschen and Schegloff 1999, 2003; Lind 2002b; Madden, Oelschlaeger, and Damico 2002; Beeke 2003; Wilkinson, Beeke, and Maxim 2003; Beeke, Wilkinson, and Maxim 2003a, 2003b, 2007a, 2007b; Wilkinson et al. 2007; Bloch and Beeke 2008). This work is prompting a rethink of our understanding of some of the symptoms classically associated with aphasia, such as e.g. agrammatism, automatic speech and the overuse of pronouns. There is growing evidence to suggest that some such aspects of aphasic language may be strategic behaviours motivated by the drive to achieve pragmatic competence, rather than symptoms that spontaneously arise as a result of brain damage. By taking a CA approach, behaviours that are considered symptoms may be recast as strategic adaptations to aphasia. Such an approach is also illuminating the issue of whether pragmatic impairment in aphasia, when observed, is a primary disorder or a secondary consequence of having a severely restricted language system. Sections 5 and 6 below will present an overview of the current CA findings concerning conversational strategies and pragmatic impairments, respectively.

5.

A CA perspective: aphasic symptom or conversational strategy?

Recently, CA studies of aphasia have concluded that some aphasic language behaviours, far from arising randomly as a result of brain damage, may be motivated by the drive to promote pragmatic competence. The evidence will be reviewed below, with respect to two broad sub-types of aphasia, non-fluent and fluent. 5.1.

Symptoms and strategies in non-fluent aphasia

The disruptions to grammar that commonly affect speakers with so-called Broca’s aphasia have long been of interest to researchers and clinicians alike. The disorder of agrammatism is typically defined with reference to distinctive grammatical output characteristics: reduced complexity of syntactic structure; omission of morphological elements; word order problems; and reduced ability to produce verbs (see e.g. Berndt 1998; Goodglass, Kaplan, and Barresi, 2001; Marshall 2002). The output of people with agrammatism is often referred to as “telegraphic speech” (Goodglass 1976), and it has for the most part been investigated by analysing data elicited via picture description and storytelling (see Prins and Bastiaanse 2004). One of the first case studies of agrammatism in conversation was of an English speaker (Wilkinson 1995). It revealed that, whereas utterances elicited via picture description tended to be subject-verb-object constructions, utterances in conversation rarely contained verbs other than in stereotyped phrases such as I think, I mean and you know. Crucially, this study highlighted for the first time a potential mismatch between concepts of agrammatism construed via elicited data and via

354

Suzanne Beeke

conversational data. Klippi (1996) investigated group conversations amongst Finnish agrammatic speakers and found that all participants, as well as their speech and language therapist, made use of single-word turns to answer questions, give opinions and acknowledge the talk of others. Klippi concluded that a single-word turn was in fact the “adequate” form in certain conversational contexts. Her findings suggest that, whereas an impairment-focused approach may view single-word utterances as symptoms of grammatical impairment (i.e. examples of telegraphic speech), an interactional approach reveals that at least some single-word utterances are perfectly functional forms for the sequential context in which they occur, even if produced by a speaker with agrammatism. Subsequently, Heeschen and Schegloff (1999, 2003) provided two detailed and important CA accounts of telegraphic speech in the conversations of two Germanspeaking women with agrammatism. Their analyses revealed how both women deployed telegraphic speech as a way of encouraging the non-impaired interactant (a friend/family member) to help out by verbally “unpacking” and thus articulating a full version of what they (the women with aphasia) meant to say. Thus, the authors concluded that there was evidence of an “interactional” motivation for some of the telegraphic features of both womens’ conversational utterances. According to this approach, telegraphic speech represents a strategy to promote pragmatic competence, not a direct symptom of brain damage. In addition, Heeschen and Schegloff (1999, 2003) reported, as did Klippi (1996), that some utterances, which on the surface appear to be straightforwardly telegrammatic (as defined by aphasiological theory), can be shown to be appropriate grammatical forms for the interactional context in which they are produced. Beeke and colleagues were the first to systematically study agrammatism in the conversation of English speakers. They found that the interactional drive to take a turn at talk influenced the characteristics of telegraphic output. Beeke et al. (2003a, 2003b, 2007a, 2007b) discussed how two speakers (of differing aphasia severity) produced telegraphic turns that consisted of sequences of nouns, adjectives, discourse markers and conjunctions, packaged by a distinct pattern of prosody, characterized by level intonation on each non-final word in an utterance and falling intonation on the utterance-final word.5 A CA examination of such turns revealed a systematic pairing of structure with social action,6 such that different forms of telegraphic utterance were used to convey distinct pragmatic functions. For example, for one of the speakers studied (called Roy, see Beeke et al. 2007a, 2007b), conversation contained the following repertoire of telegraphic utterances: (1) a nouninitial turn, in which a noun was followed by a word or words that served to comment, e.g. racing, Newmarket, Epsom, anywhere, but me, Ascot no., in order to convey a new topic (akin to topic-comment structure); (2) an adjective-initial turn, in which an adjective was followed by because and a reason e.g. amazing because two years or three years, in order to offer a personal view on a current topic; and (3) a turn that combined talk and mime, designed to convey an event in the absence

Aphasia: The pragmatics of everyday conversation

355

of a verb (see Beeke et al. 2007b for examples). Each of these turn construction formats was designed to achieve a different conversational action and was shown to recur in the data; i.e., the authors found a collection of individual examples that were all constructed in the same way using the same resources. Beeke and colleagues concluded that, for these speakers, conversational turns were built using systematic methods which resulted in recognisable telegraphic turn formats, and the motivation for their structure was interactional. Another classical feature of Broca’s aphasia is so-called automatic speech, often associated with rigid formulas such as expletives, serial speech and song lyrics, but actually a continuum of behaviours including utterances such as I don’t know and all the time (Simmons-Mackie and Damico 1997)7. Prior research has often cast these formulaic or stereotyped utterances as automatic and involuntary, with little communicative value (see e.g. Goodglass and Kaplan 1983; Lum and Ellis 1994; Code 2005). However, a CA approach has begun to reveal ways in which formulaic expressions can be used within the context of aphasic conversation, and treated by non-aphasic conversation partners as examples of pragmatic competence, that convey referential meaning and perform social action (Beeke 2003; Johnson 2008; Soar 2008). Beeke (2003) presented an investigation of the conversation of Roy, an English speaker with agrammatism, and found that, in conversation with his adult daughter, Roy recurrently made use of the formula I suppose within a turn designed to do the action of assessing8 an event or state of affairs. For example, in response to his daughter’s comment that her sister in Australia is homesick, Roy says amazing because two years or three years but I suppose different, thus conveying his opinion on the matter9. Roy has very restricted output which includes yes and no, exactly, adjectives such as brilliant, the conjunctions and, but, and because, and formulaic language such as I don’t know and I think. His use of verbs in propositional speech is very limited – he produces only three in 23 minutes and 9 seconds of conversation (working, waiting, and fly). Beeke argued that I suppose represented an adaptation to the demands of constructing a conversational turn in the face of severe agrammatism. This pragmatic strategy accorded benefits, in that it created islands of fluency in Roy’s otherwise halting output, it gave the appearance of grammatical structure, and it projected at a grammatical and pragmatic level not only that there was more to come (performing a turn holding function), but also conveyed something about the as yet unspoken meaning (providing clarification for his listener). Soar (2008) extended the analysis of Roy’s formulaic language to other constructions, and concluded that formulas served two key functions in Roy’s conversation, turn holding (including you know, I’ll tell you what and I mean) and assessment (including who knows! and you say that). Johnson (2008) explored the pragmatic functions of formulaic language in the conversations of a 68-year-old man with severe agrammatic aphasia, called Donald. By analysing Donald’s conversations with his wife and family friends,

356

Suzanne Beeke

Johnson demonstrated that formulas performed two key pragmatic functions. Firstly, they requested a response from a conversation partner. Examples were: (i) requesting an object, e.g. please may I? in the context of his wife showing a bottle of wine to his friend, the result being that the wife gave Donald the bottle; and (ii) asking a question, e.g. what’s your? directed towards his friend in the context of a discussion about Donald’s birthday, with the result that the friend told Donald his (the friend’s) date of birth. Secondly, for Donald, formulas functioned to assess an event or a state of affairs (see note 6 on assessment), in order to make a joke or to reminisce. An example of a joke was alright for some!, produced in the context of a story by Donald’s wife about his wish to always check the percentage alcohol content of the wine they buy, and a specific instance of him rejecting a bottle that was only 12.5 %. An example of reminiscing was were the days, produced in the context of a discussion with a friend about their past sporting and drinking exploits. According to Johnson, the strategic use of formulas permitted Donald to produce turns that were not agrammatic (in contrast, she showed that his propositional speech consisted mostly of single word utterances), to influence the activity of his conversational partners, and to perform complex and subtle social actions such as joking and reminiscing. Johnson concluded that, for Donald, formulas acted as a resource for negotiating the social and linguistic barriers imposed by aphasia, and, furthermore, allowed him to establish himself as a competent conversationalist and social actor, despite severe impairments to his linguistic system. 5.2.

Symptoms and strategies in fluent aphasia

For those people whose aphasia is termed fluent, and thus predominantly causes word finding difficulties, much of the aphasiology literature reports on the overuse of pronouns, light verbs such as do and get, and general nouns such as thing and it (Bates, Hamby, and Zurif 1983; Edwards and Garman 1989). Wilkinson, Beeke and Maxim (2003) examined the conversations of two English speakers with fluent aphasia, and found a common pattern of turns characterised by a large number of such semantically “light” lexical forms. Several forms often occurred within a single turn, resulting in utterances such as I can do it, and most were accompanied by gesture or mime that carried the semantic content. According to Wilkinson, Beeke and Maxim this method of building an utterance facilitated the smooth progression of a turn to its completion, by minimising the word retrieval difficulties associated with semantically rich noun phrases and predicate argument structure. The authors argued that such utterances allowed speakers to produce relatively complex contributions to conversation without repair and linguistic non-competence becoming the focus of the conversational activity. They concluded that the overuse of such light lexical items could be seen as a strategic response to the demands of constructing a conversational turn given a speaker’s linguistic limitations, rather than as a direct symptom of his or her brain damage.

Aphasia: The pragmatics of everyday conversation

357

Yates (2009) examined the conversations of a man with Wernicke’s aphasia (Keith) and his wife (Annie), which are marked by frequent and multiple trouble sources relating to Keith’s output difficulties, in particular problems with word finding, but also a more widespread difficulty with framing his thoughts using language, or “thought-finding difficulties” as Yates termed them. Yates found that Keith repeatedly engaged in self-initiated self repair when conversing with his wife, although his repair attempts were very often unsuccessful. In addition, he sometimes made a direct comment about his word finding difficulty (I can’t get out; was that right?) or engaged in explicit discussion of the impact of his aphasia on his wife and others. Furthermore, Yates found that Keith used reported speech and semantically light lexical items such as thing and got as a means of adapting to the demands of producing a complete conversational turn, as previously noted by Wilkinson, Beeke and Maxim (2003), for other speakers with fluent aphasia. According to Yates, the existence of such phenomena in Keith’s conversation is highly note-worthy, given his aphasia type. Lack of self-monitoring and decreased insight are considered within the aphasiological literature to be defining symptoms of Wernicke’s aphasia (Goodglass and Kaplan 1983). However, these data show that Keith can self-monitor, and has some level of insight into his difficulties and their impact. Yates’ work extends to English the findings of Laakso (1997) for a group of Finnish speakers with Wernicke’s aphasia, and Roenfeldt (1999) for a German speaker. Both found evidence of successful self-monitoring of errors in conversation data, contrary to aphasiological expectation. Clearly, further research into the pragmatic abilities of people with Wernicke’s aphasia using a CA methodology is warranted. As these studies demonstrate, a CA approach permits us to rethink our understanding of aphasic behaviours which on the surface appear to be direct symptoms of brain damage, by revealing that at least some can function as strategies to promote pragmatic competence.

6.

CA perspectives on pragmatic impairment in aphasia

As has already been mentioned, although most people with aphasia have intact pragmatic abilities, it is generally accepted that aphasic language impairments can have secondary pragmatic consequences, in so far as they disrupt conversational timing, turn taking or the ability to clarify talk. Two examples will now be explored in data from an ongoing project that aims to design and evaluate a novel conversation-based speech and language therapy programme for agrammatism (see Beeke et al. 2011). The first extract presents an example of a pattern of disrupted conversational timing in the pre-therapy conversations of Kate, a woman in her early 40s who has agrammatism, and her twin sister, Shelly, who she lives with. An analysis of eight

358

Suzanne Beeke

pre-therapy conversations recorded at weekly intervals has revealed a common sequence of turns whereby Kate delivers an agrammatic turn and Shelley responds with a yes/no question intending to clarify the meaning. Kate subsequently answers the question extremely quickly (often in overlap), but then goes on to correct her response after a delay (answers yes then corrects her reply to no, or vice versa). However, during this delay, Shelly often continues the conversation by building on the first, incorrect answer, and misunderstandings ensue. This pattern is illustrated in Extract 2. In Extract 2, at line 05, Kate introduces the topic of a DVD, and indicates that it has something to do with herself, saying DVD um me, with stress on the word me (indicated by underlining in the transcript). The turn final uh suggests she may not be finished, but Shelley takes a turn to clarify the topic (line 06). In line 07, Kate provides an acknowledgement before repeating the element that Shelley has yet to pick up on (yes yes me). Kate’s mid level intonation on me (marked with a comma in the transcript) indicates that again, she has not finished her turn, but Shelley promptly launches a guess at Kate’s meaning (line 08, what you wanna send him a DVD). Kate’s response is an emphatic NO (capitals indicate an increase in loudness), and she continues by again introducing a topic, Larry, and commenting on it, fantastic (lines 09–10). For a third time, Kate’s turn appears unfinished as Shelley begins to speak; Kate’s turn final uh (line 10) is immediately followed by Shelley’s understanding check, oh he’s done – he’s cut his own DVD! (line 11). The lines that follow this understanding check are of particular interest here; Kate responds twice with yes in overlap (line 12), which gives the impression of quick and unequivocal understanding and acceptance of Shelley’s turn, and she goes on to give one final acknowledgement before attempting to continue (line 13). After Shelley requests more information at line 14 however (what with the band or), Kate delivers an emphatic no no! accompanied by a head shake, and goes on to say ME! once more, whilst pointing to herself (line 15).10 Lines 16 to 20 clarify the cause of the misunderstanding. The DVD is of Kate and the jazz group, Gene has not in fact cut his own DVD. In retrospect then, it appears that Kate’s acceptance at line 12 of Shelley’s suggestion that Gene had cut his own DVD was erroneous; the answer to that question should have been no. However Kate’s self-repair was delayed; it was not launched until after the intervening turn from Shelley at line 14. This sequence of a very prompt but erroneous yes/no answer to a question from Shelley, followed by a delay (often filled by another turn from Shelley) before the launch of a self-repair, is common to all of Kate and Shelley’s pre-therapy conversations, and as here, it causes much misunderstanding, reflected in the need for repair that backtracks over prior, and seemingly agreed-upon meanings. Kate’s behaviour appears to resonate strongly with what we know about the symptoms of aphasia; some people are left with an unreliable yes/no response.11 This extract clearly demonstrates the real life consequences of a symptom such as this, which can be characterised here as a disruption to the timing of conversation,

Aphasia: The pragmatics of everyday conversation

359

Extract 2. D1C4 (pre-therapy)–12:48–DVD me Kate and Shelley are in the kitchen, and Shelley is cooking dinner. Prior to line 01, Shelley has been asking what Kate’s partner Gene said when he telephoned earlier that day. Kate, Gene and a friend Larry are all members of a jazz band, referred to here as the Green Music Trio (not its real name). Kate is the singer. 01

Shelley

so what else did Gene say

02

Kate

awwwawww ⎡hhhhhhhhhhhh

03

Shelley

awwwawww ⎣he must have had something else to ⎡say

04

Kate

⎣hhh thank youste- um DVD ⎡(0.6) um (3.0) me=uh ⎣((extends index finger then holds all fingers out, spaced apart …

05 06

Shelley

DVD?

07

Kate

⎡ye::s:: yes me, ⎣… gesture held frozen …

08

Shelley

what you wanna send him a DVD

09

Kate

⎡NO! ⎣((shakes head))

⎡ah fantastic=uh ⎣… moves extended index finger in circular motion))

10 11

Shelley

12

A

13

A

14 15

⎤ ⎡(0.4) tuh (0.8) um (4.0) ps- (0.9) e- ur- (0.5) Larry ⎦ ⎣((moves extended fingers …

Kate

⎡ne- he’s

⎡cut his own DVD!

⎣ye-

⎣yea:::h

yes (0.6) um tuh Shelley

A

oh he’s do

what with the band

⎡or ⎣no no! YOU! ⎡(0.3) tuh ME! ⎤ um ⎣((shakes head, points to self)) ⎦

Kate Shelley

17

Kate

18

Shelley

⎡with ⎤ the Green Music Trio

19

Kate

⎣yes

20

oh your de

⎡mo!

16

⎣yea:::h:::

yes

21

Shelley

ahh

22

Kate

yeah

23 24

⎦

(2.0) Shelley

oh lovely!

360

Suzanne Beeke

such that misunderstandings are common and the need for repair is rife. Although pragmatic in nature, such consequences have their origins in the language disorder, and thus appear to be secondary to the aphasia rather than primary symptoms of pragmatic impairment. However, a CA approach reveals that there may be more to the so-called symptom of unreliable yes/no than factors internal to the person with aphasia, at least for Kate and Shelley. As this example demonstrates, Kate’s behaviour appears to be heavily influenced by Shelley’s turn taking habits, as well as by her own aphasia. Four times we observe Shelley taking the conversational floor in order to clarify Kate’s meaning at points where there are clear signs that Kate has not yet finished speaking (lines 06, 08, 11 and 14). This results in a high tempo of turn taking, and it is possible that Kate’s early incoming turns (her overlapping and unreliable acknowledgements) are a product of this time pressure. Thus, it may be the combined effects of having an unreliable yes/no response and attempting to initiate and complete a turn at talk quickly despite severe agrammatic aphasia that result in a secondary pragmatic impairment. In this way, pragmatic impairment can be said to emerge from the interactions of, and thus be the responsibility of, both the speaker with aphasia and the non-impaired conversational partner (Perkins 2007). A second extract, also taken from data collected for Beeke et al.’s (2011) conversation therapy project, provides an example of a pattern of disrupted turn taking in the conversations of Simon, who is agrammatic following a stroke in his mid 30s, and his wife Cath. An analysis of eight pre-therapy weekly conversation samples has revealed that Simon’s turn taking is disrupted in three key ways: (i) turns remain incomplete; (ii) turn design results in the action of the turn being unclear; and (iii) turns contain long and frequent pauses during which it is not clear if Simon intends to continue or not. Each of these is illustrated in Extract 3. In Extract 3, we see an example at line 02 of a pause where it is not clear if Simon will continue to speak or not. These occur frequently in his conversation. After saying yes in response to Cath’s topic-setting our bedroom, he directs his gaze towards the ceiling, but he is otherwise motionless. There is no non-verbal or verbal evidence to suggest that he is seeking the words to continue his turn. In line 03, we see evidence of Cath’s uncertainty about whether Simon will continue; she starts a turn with w – we wanted –, but Simon then comes in with the bed (line 04). As soon as Cath realises that Simon is continuing, she abandons her turn and at line 05 takes up his topic (our bed yeah?), in order to show that she is listening. Another long pause follows this (1.5 seconds, see line 06), during which it is again unclear if Simon will continue. At lines 06 to 09 the meaning and the action of Simon’s turn is unclear, and it appears to remain incomplete. These are also common features of his talk. It is not apparent what meaning the combination of gesture and the phrase inside and out is designed to convey. In addition, the action is unclear; is he commenting on the current state of the bedroom or making a suggestion for change? We see direct evi-

Aphasia: The pragmatics of everyday conversation

361

Extract 3. D2C7 (pre-therapy)–22:48–bed inside and out Simon and Cath are sitting on the sofa. At line 01, Cath returns to a prior topic of talk, their plans for redesigning their bedroom. 01 02

A

Cath

our bedroom.

Simon

yes

⎡(1.6) ⎤ ⎣((taps his chest three times)) ⎦

⎡(2.1) ⎣((looks up at ceiling))

03

Cath

w- we wanted-=

04

Simon

=the bed

05

Cath

our bed yeah?

Simon

(1.5)

⎤ ⎦

06

A A

07

A A

⎡(2.0) ⎤ ermm ⎣… index and middle finger in ‘V’ shape .. ⎦

08

A A

⎡(3.0) ⎣… moves hand left to right, rotates hand inwards))

⎤ ⎦

09

A A

inside >°and out° ¬S(x), where ‘→’ stands for the entailment relationship and ‘+>’ for the implicature relationship. The theory of Horn’s scale is very powerful: it adequately describes one part of the lexicon. (22) illustrates some such quantitative scales (see also Gazdar 1979: 56; Levinson 1983: 134): (22)





Second, the asymmetry between semantic entailments (truth-conditional meanings) and scalar implicatures (non-truth-conditional meanings) explains why lexical items are not genuinely ambiguous, but are simply constrained in their meaning by general pragmatic principles. This provides a good illustration of Grice’s recommendation, as formulated in his Modified Occam’s Razor (M.O.R.) (Grice 1979: 118–119): “Senses are not to be multiplied beyond necessity” [italics in Grice’s text].13 The case of or provides a spectacular example of how scalar implicature and the M.O.R. principle can produce new answers to the question of lexical meaning. The logical meaning of or is its inclusive one, which allows both disjuncts to be true together. If or had an inclusive meaning in use, (23) would be ambiguous as compared to (24) and (25): (23) Peter or Mary will come tonight. (24) Peter will come tonight and Mary will come too. (25) Peter will come tonight and Mary will not, or Peter will not come tonight and Mary will.

418

Jacques Moeschler

(24) is an illustration of the inclusive reading, while (25) is an illustration of its exclusive one. How can we explain that a speaker who says (23) generally intends to communicate (25) rather than (24)? One explanation is that by choosing or with the intention of communicating and, the speaker simply made a mistake in his lexical selection. Another is that the speaker is a follower of Grice, and cannot say and because he knows that only one person is coming, but he does not know who: in this case, he uses the word that is most compatible with the first maxim of quality. A third might be that the use of or triggers a default scalar implicature, implicating that (24) is false, explicitly reformulated in (26):14 (26) It is not the case that Peter and Mary will come. The scalar implicature reading explains two things: that or is semantically connected with a stronger one (and); and that the specific pragmatic exclusive reading of or is the result of the conjunction of its logical inclusive meaning and its scalar implicature. In other words, the scalar implicature of or is given in (27), and its pragmatic exclusive meaning is given in (28):15 (27) (p or q) +> not (p and q) (28) (p or q) and (not (p and q)) Horn’s theory of scalar implicatures also elegantly describes why certain logical expressions are not lexically realized in natural languages. Horn observed that the lexicons of natural languages do not contain words for negative particulars. Although languages do have words for positive universals (all) and particulars (some) as well as for negative universals (none), there are no words for negative particulars. Table 2 demonstrates this discrepancy for English: Table 1.

Lexical realizations of the corners of the logical square

A

I

E

O

all

some

no

*nall

always

sometimes

never

*nalways

both

one (of them)

neither

*noth

and

or

nor

*nand

The designations A, I, E and O stand for the four corners of the logical square (AffIrmo, nEgO), given in Figure 3:

Conversational and conventional implicatures

Figure 3.

419

The logical square (Horn 2004: 11)

Negative particulars are not lexicalized in natural language according to Horn because they have complex values. It is possible to express this property through what I refer to as Horn’s conjecture (Horn 2004: 11): Horn’s Conjecture16 Natural languages tend not to lexicalize complex values, since these need not be lexicalized. Now, what is interesting with connecting scalar implicatures and the logical square is that positive particulars appear as weak terms in the scale. The logical square illustrates the following positive scales: , , , . Some, for instance, implicates the negation of all, as shown in (29), meaning (30): (29) Some of Jacques’ students passed their pragmatics exam. (30) Not all of Jacques’ students passed their pragmatics exam. Horn was unable to create a scale of negatives such as because nall is not a word in English.17 He instead proposed that the negative particular that corresponds to no is a complex one: some … not, in which case would be the negative counterpart of . If this is the case, some … not implicates the negation of no, which cannot be rendered in a simple sentence, but only in a complex one. (31), therefore, quantitatively implicates (32): (31) Some of Jacques’ students did not pass their pragmatics exam. (32) It is not the case that none of Jacques’ students passed their pragmatics exam. What does (31) mean, and conversely, what does (29) mean? The logical square gives a simple answer: the negation of E is I, and the negation of A is O. By asserting (29), the speaker implicates (31), and by asserting (31), she implicates (29). This lead Horn (2004: 11) to claim that “While what is said in Some men are bald and Some men are not bald is distinct, what is communicated is typically identical: Some men are bald and some men aren’t”.18

420

Jacques Moeschler

What, exactly, is the status of scalar implicatures? Where do they come from? Different solutions have been proposed, but the classical neo-Gricean approach (based mainly on Horn’s and Levinson’s research) is predicated on making the first maxim of quantity and the sub-maxims of manner ‘Avoid ambiguity’ and ‘Avoid obscurity’ into a principle, the Q-Principle (Horn 1984: 13):19 The Q-Principle Make your contribution sufficient. Say as much as you can (given R). For Horn, the Q-Principle is Hearer-based, and is opposed to the R-principle, which is Speaker-based (Horn 1984: 13): The R-Principle Make your contribution necessary. Say no more than you must (given Q). The R-Principle is based on the maxims of Relation, the second maxim of Quantity and the maxims of Manner (“Be brief” and “Be orderly”). While the Q-Principle is illustrated by scalar implicatures, the R-Principle is typically exemplified by indirect speech acts. In (33), the speaker is not saying any more than necessary in conveying his request: a question is a more economical way of conveying a request than is the explicit performative given in (34): (33) Can you pass me the salt? (34) I request that you pass me the salt. The main advantage of the neo-Gricean approach to implicature (mainly scalar implicatures)20 is that it reduces the nine Gricean maxims to two principles, which are supposed to balance each other. The principle that corresponds to Horn’s R-Principle is Levinson I-Principle, or Principle of Informativeness:21 I-Principle Speaker’s maxim: the maxim of Minimization. “Say as little as necessary.” Recipient’s corollary: the Enrichment Rule. Amplify the informational content of the speaker’s utterance, by finding the most specific interpretation. Levinson (2000: 117–118) gives a series of examples illustrating the informational amplification of utterances, which are arranged in categories observed elsewhere: (35) Conditional perfection (Geis and Zwicky 1971) If you mow the lawn, I’ll give you five dollars. +> If you don’t mow the lawn, I won’t give you five dollars. (36) Conjunction buttressing (Atlas and Levinson 1981) John turned the key and the engine started. +> John turned the key and then the engine started. (37) Bridging (Clark and Haviland 1977) John unpacked the picnic. The beer was warm. +> The beer was part of the picnic.

Conversational and conventional implicatures

421

(38) Inference on a stereotype (Atlas and Levinson 1981) John said ‘Hello’ to the secretary and then he smiled. +> John said ‘Hello’ to the female secretary and then he smiled. (39) Negative strengthening (Horn 1989) I don’t like Alice. +> I positively dislike Alice. (40) Mirror maxim (Harnish 1976) Harry and Sue bought a piano. +> They bought it together. How are such implicatures triggered? The I-Principle obviously does not offer a sufficient explanation. In example (35), a causal reading is one possible more specific reading: John turned the key and, because of that, the engine started. Indeed, the I-Principle must be completed by reference to information that makes up background knowledge. Levinson (1987) makes an explicit reference to the maxim of relativity and the convention of non-disputability. Thus the principle of informativeness simply states that the best interpretation for an utterance is the most informative interpretation consistent with what is not disputable. Gazdar’s theory of potential implicature, which he calls im-plicatures, presents a slightly different and formally most convincing approach. A potential implicature, or im-plicature, is an implicature “which the sentence could possibly have prior to contextual cancellation” (Gazdar 1979: 55). In other words, the im-plicature, in order to be promoted from a potential to an actual implicature, should be consistent with propositions defining the context. According to this perspective the context is given and any implicature is a new contribution to it. Relevance Theory has challenged this classic view of implicature. The next section explains how and why this occurred (cf. Carston and Hall, this volume; Huang, this volume). 8.

Explicatures and weak implicatures

Implicatures have a very different status in Relevance Theory: according to this theory, implicatures can be false. If this is the case, it is because one of the required contextual premises is false. In other words, implicatures are defined as contextual implications; that is, as implications based on a contextualization (Sperber and Wilson 1986: 107–108): Contextual implication A set of assumptions {P} contextually implies an assumption Q in the context C if and only if (i) the union of {P} and {C} non-trivially implies Q, (ii) {P} does not non-trivially imply Q, and (iii) {C} does not non-trivially imply Q.22

422

Jacques Moeschler

This very general definition of contextual implication is a result of a strong reductionist version of Grice’s theory of implicature. Whereas neo-Griceans reduced the nine Gricean maxims of conversation to two principles, i.e. the Q-Principle and the R(I)-Principle, post-Griceans such as Sperber and Wilson reduced all maxims to one principle, the principle of relevance, which states that the speaker has produced the most relevant utterance in the circumstances. More precisely, the Principle of Relevance can be stated as follows (Sperber and Wilson 1986: 158): Principle of relevance Every act of ostensive communication communicates the presumption of its optimal relevance.23 A contextual implication is one way for an utterance to be relevant, relevance being defined as a balance between positive cognitive effects (addition of a new information, strengthening of an old information, suppression of an old information) and cognitive efforts, due to the length of the utterance, the deductive rules implied in non-trivial implication, the access to the concept forming the logical form of the sentence among others. Implicatures are thus defined in Relevance Theory as a type of cognitive effect. More precisely, they are the results of non-demonstrative inference, whose premises are based on an utterance (more precisely, on the logical form of the sentence uttered) and on a set of contextual assumptions. The result of interpreting a sentence in a context yields an implicated conclusion, which is arrived at by deduction. This approach contrasts with the classic Gricean procedure of working out an implicature, as well as with the neo-Gricean heuristics based on the Q- and the R(I)-Principles. In the recent version of Relevance Theory, implicatures are the result of a general procedure of comprehension, stated as follows (Wilson and Sperber 2004: 613): Relevance-theoretic comprehension procedure a. Follow a path of least effort in computing cognitive effects: Test interpretive hypotheses (…) in order of accessibility. b. Stop when your expectations of relevance are satisfied (or abandoned). In order to get a contextual implication, the hearer must access contextual assumptions. How is this possible? In Relevance Theory, the context is constructed, utterance after utterance, rather than given at the outset. Consider the following example: (41) Paul: How was the party? Peter: People left late at night. Peter’s answer must be interpreted against the contextual assumption (42a), which yields the implicated conclusion (42c) when combined with Peter’s utterance:

Conversational and conventional implicatures

423

(42) a. If people leave a party late, then the party is a success. b. People left Peter’s party late. c. Peter’s party was a success. It is obvious that context is not based on a set of background assumptions, but rather on a restricted set of propositions accessible in real time as the utterance is being interpreted. Some premises must be constructed during the interpretation of the utterance, even if they do not belong to the set of shared information. For example, suppose I invite Ahmed to dinner, and do not realize that Muslims do not drink alcohol. In this case, the implicature conveyed by Ahmed’s utterance in (43) – that is (44) – is not impossible to understand, because I can, in the setting of conversation, construe the missing contextual premise (45). Sperber and Wilson call this type of premise implicated premise:24 (43) Jacques: Ahmed, would you like a glass of wine? Ahmed: I am Muslim. (44) Muslims don’t drink alcohol. (45) Ahmed does not want a glass of wine. According to Relevance Theory, implicatures therefore belong to the Gricean category of particularized conversational implicatures. Almost all the research carried out on implicatures by neo-Griceans,25 however, has explored generalized conversational implicatures.26 One of Relevance Theory’s most important contributions to pragmatics, as well as to a general approach to implicature, was to elaborate the fifth Gricean criterion defining implicatures: determinacy. Sperber and Wilson (1986: 217–224, and 231–237) developed a very interesting theory of implicature. According to them, implicatures are not different in nature to what is said – or in Relevance Theoretic terms, to explicatures – but differ mainly in the strength through which they are entertained. Sperber and Wilson distinguish between two types of (nonce) implicatures: strong and weak implicatures. Strong implicatures are characterized by the strength through which they are conveyed, i.e. are under the responsibility of the speaker. In most cases, they are thus determinate in content. Some typical cases of strong implicatures follow: (46) Jacques: Axel, please go and brush your teeth. Axel: Dad, I’m not sleepy. (47) Anne, looking at Nat’s room: Your room is a pigsty. (48) Peter: Jacques, where do you live? Jacques: I live in Cluny.

424

Jacques Moeschler

In (46), Axel implicates that he does not want to brush his teeth and then go to bed; in (47) Anne implicates that Nat’s room is dirty and should be cleaned; finally, in (48), Jacques, who lives in a small village near Cluny, implicates that he is living close to a famous medieval town. Weak implicatures, on the contrary, are less determinate, and left to the responsibility of the hearer. In these cases, the utterance gives rise to a number of weak implicatures. Creative metaphors are typical cases of weak implicatures, as in these well-known metaphors in English and French: (49) No man is an island. (John Donne) (50) Juliet is the sun. (Shakespeare) (51) La femme est l’avenir de l’homme. (Aragon) ‘Woman is the future of man’. (52) L’homme est un roseau pensant. (Pascal) ‘Man is a thinking reed’. One of the main issues raised by any theory of implicature now arises. Are implicatures in fact non-truth-conditional aspects of meaning? The classical Gricean answer is yes, but the neo- and post-Griceans are very cautious about their answer. Scalar implicatures and informative implicatures will be used to illustrate this point. In scalar implicature, the pragmatic meaning of some, as well as for the conjunction or, is a restricted meaning, that is, a more specific meaning than its logical one: logically, some is compatible with all, and or can be read as inclusive. The following question must be asked: Which part of the meaning of the sentence is truthconditional? Is it the logical meaning or the implicature? If the restricted meaning determines the truth conditions of the utterance, it can no longer be interpreted as an implicature. So the question is now to what extent pragmatic meaning can be considered to be truth-conditional. Several scholars have given positive answers to this issue, using different labels: explicatures (Sperber and Wilson 1986), pragmatic primary processes (Recanati 2004) and impliciture (Bach 2004). In this context I will simply cite the argument given by Wilson and Sperber (1998; cf. Cohen 1971) as an explanation for the I-implicature of the temporal and. In the following examples, the complex proposition p and q is not truth-conditionally equivalent to q and p, because if it were true, (53) would be a tautology and (54) a contradiction. The temporal meaning of and cannot, therefore, be an implicature: (53) It’s always the same at parties: either I get drunk and no one will talk to me or no one will talk to me and I get drunk. (54) What happened was not that Peter left and Mary got angry but that Mary got angry and Peter left.

Conversational and conventional implicatures

425

Thus, the explanation is as follows: temporal enrichment, as well as all cases of I-implicatures (see examples 35 to 40) are cases of pragmatic enrichment arising at the level of explicatures. Pragmatic meaning is therefore truth-conditional, when it deals with the development of a full proposition.

9.

Conclusion: the role of implicatures in comprehension and communication

In this chapter, we have seen how a theory of implicature has become possible through Grice’s seminal work on non-natural meaning and the logic of conversation. For the last thirty years or so pragmatics has developed in many fields including the philosophy of language, logic, linguistics, psycholinguistics, and, in the past few years, in neuroscience. As will be shown, a variety of directions have been taken in explaining what it is to understand an utterance. These have yielded a variety of answers. The first move, which is represented by approaches that mainly explore generalized conversational implicatures, defines the understanding of an utterance as a process implying automatic and default reasoning. Generalized conversational implicatures and conventional implicatures are therefore defined as being part of the lexicon, and are not the result of any particular contextual device. Another development, which is currently inciting research, tries to include aspects of non-explicit meaning (primarily presuppositions and implicatures) in a very general, layered theory of meaning. For instance, Potts (2005: 23) has developed a theory of meaning that distinguishes between context-dependent meanings, including conversational implicatures and pragmatic presuppositions, and entailments, including at-issue entailments, conventional presuppositions and conventional implicatures. This development represents a strong intrusion of formal semantic techniques and of the semantic agenda into the classic domain of pragmatics. Finally, Relevance Theory presents a layered picture of meaning: pragmatic inferences intervene at the level of explicatures and implicatures. First, the explicature level contains three different meanings: basic explicature (propositional form), propositional attitude and illocutionary force, the last two being higherorder explicatures. Second, the implicature level contains two types of implicit meaning: the implicated premises and the implicated conclusions. Finally, an implicated conclusion can be, as we saw in section 8, strong or weak, depending on the strength of the entertained implicated premises. Even if the general picture of a theory of implicature is far from being homogeneous now, it is interesting to connect implicature and non-natural meaning. As defined in section 2, non-natural meaning (Grice) is what corresponds to the speaker’s informative intention (Sperber and Wilson). Now, what relation, if any, exists between non-natural meaning and implicature? The picture is not as clear-cut as it

426

Jacques Moeschler

could be drawn from a classical Gricean perspective. For Grice, there is a strong identity between non-natural meaning and implicature. However, in the neoGricean perspective, generalized conversational implicatures are default inferences, and since the speaker can deny having intended to implicate such and such proposition, it is not clear whether implicatures equate informative intention. Finally, in Relevance Theory, successful communication is not an absolute concept, but is more precisely associated to a continuum between the intended meaning and the interpretation of the speaker’s utterance. In that case, it is not surprising that the hearer can be in a situation where he cannot grasp some or most of the speaker’s implicatures. In other words, he should at least be capable of grasping the explicatures of an utterance, which seems to be a minimal condition to insure inferential communication.27

Notes 1. “A linguistic sign does not unite a thing and a noun, but a concept and an acoustic image”. 2. The concepts of informative and communicative intentions are mentioned in Sperber and Wilson (1986: 29): “Informative intention: to inform the audience of something; Communicative intention: to inform the audience of one’s informative intention”. 3. Gazdar (1979), Levinson (2000) and Sperber and Wilson (1986) respectively. 4. Huang (2007 and 2010) calls this type implicatures implicatureO, that is, implicatures arising “from strictly observing the maxims” (Huang 2010: 207). 5. The first case – the fulfillment or use of a maxim – is not developed by Grice, only implicitly mentioned throughout his first example, mainly implying the maxim of relation (relevance). This has lead to some radical interpretations of Grice’s framework, in which conversational maxims would be the results of a conversational maxim violation. See Borg (2006), for instance. Huang (2010) calls this type implicatureF (“generated by way of the speaker ostentatiously flouting the maxims”). 6. This reading of Grice’s example is somewhat different from that of Robyn Carston in her article on quantitative implicature (Carston 1998), in which the implicature is caused by the violation of the first maxim of quantity. In this case, the speaker knows where C lives but does not want to tell her audience. In Moeschler (2010), I explain both readings as examples of a pre-eminence of one maxim over the other: quantity > quality (Carston) vs. quality > quantity (Grice). 7. The corresponding proverb in French is à la guerre comme à la guerre. 8. Cf. note 4. 9. These implicatures are not conversational because they are not triggered by any conversational maxims. They are not conventional, either, because they are not part of the meaning of the word. 10. In Relevance Theory (Carston 2002), these implicatures are called nonce implicatures, and correspond to what Relevance Theory has defined as contextual implications or implicated conclusions (Sperber and Wilson 1986, 1995; Wilson and Sperber 2004). 11. Certain difficulties occur when making a lexical contrast between the act of saying and the result of such an act – that is, an utterance – in English. The word utterance refers

Conversational and conventional implicatures

12.

13. 14. 15.

427

both to the act of uttering and to its result, whereas French makes a conventional and technical difference between the act of uttering (énonciation) and its result (énoncé), cf. Benveniste (1966, 1974). This point is less obvious for GCIs (generalized conversational implicatures) because the implicature is carried by a specific expression. The restricted meaning ‘only some’ for some therefore is more determinate than indeterminate. This criterion is in fact restricted to particularized conversational implicatures. Horn (1985) provides substantial illustrations of the M.O.R. principle in relation to negation, modals and connectives. This reading is triggered by the Q-Principle (cf. infra). The following truth table is a demonstration of (28) (Moeschler and Reboul 1994: 198), where ‘v¯’ stands for the exclusive disjunction, ‘↔’ for the logical equivalence, ‘∨’ for the inclusive disjunction, ‘∧’ for the logical conjunction and ‘¬’ for the logical negation. In other words, the pragmatic meaning of or is equivalent to the meaning of the exclusive or (V¯): p

q

p∨q

p∧q

¬(p ∧ q)

(p ∨ q) ∧ ¬(p ∧ q)

p v¯ q

(p v¯ q) ↔ (p ∨ q) ∧¬ (p ∧ q)

1

1

1

1

0

0

0

1

1

0

1

0

1

1

1

1

0

1

1

0

1

1

1

1

0

0

0

0

1

0

0

1

16. This label is pointing to one of the most intriguing properties of natural language, described in detail in Horn (2011), that is, the general lack of lexical items for negatives particulars (O) in natural languages. 17. The same holds for French, where nitous is not a word for a negative particular: (i) *Nitous les étudiants de Jacques ont réussi leur examen de pragmatique. [Nall Jacques’ students passed their pragmatics exam.] 18. In Moeschler (2007a), I give a detailed analysis of Horn’s conjecture and propose a different solution, based on the Relevance Theoretical notion of explicature. 19. Another classic Neo-Gricean formulation appears in Levinson (2000: 76): “Q-principle Speaker’s maxim: Do not provide a statement that is informationally weaker than your knowledge of the world allows (…). Recipient’s corollary: Take it that the speaker made the strongest statement consistent with what he knows (…).” For Levinson, it is the recipient’s corollary rather the speaker’s maxim that yields scalar implicatures. 20. Different accounts of scalar implicatures have appeared in recent years. Danny Fox’s highly interesting account is based on Gricean reasoning and avoids Horn’s scale. This approach (Fox 2007) is based on the reformulation of the Maxim of quantity: “Maxim of Quantity (basic version): If S1 and S2 are both relevant to the topic of conversation and S1 is more informative than S2, if the speaker believes that both are true, the speaker should utter S1 rather than S2” (Fox 2007: 73). Fox reasons that “If we, the people who

428

21.

22. 23.

24. 25.

26. 27.

Jacques Moeschler

interpret the utterance [Sue talked to John or Fred], assume that s obeys the Maxim of Quantity, we conclude, for each disjunct, p, that it is false to claim that s believes that p is true, or if we keep to our convention of using the verb know instead of believe, we can state this as a conclusion that s does not know that p is true” (Fox, 2007: 73). Atlas and Levinson (1981: 40–41) give an explicit and formal definition (which I will not discuss here). I present a simpler version, from Levinson (2000: 114). In a more precise way, Horn’s principle subsumes Grice’s second sub-maxim of quantity and some of his sub-maxims of manner. But Levinson’s I-principle, as Horn’s R-principle, is the informative counterpart of his Q-principle. A trivial implication is an implication that requires only one premise as input. In the recent version of Relevance (Sperber and Wilson 1995; Wilson and Sperber 2004), the principle of relevance has been split into two principles, the cognitive principle of relevance and the communicative principle of relevance: Cognitive Principle of Relevance Human cognition tends to be geared to the maximization of relevance. (Wilson and Sperber 2004: 610) Communicative Principle of Relevance Every ostensive stimulus conveys a presumption of its optimal relevance. (Wilson and Sperber 2004: 612) Grice’s example John is an Englishman, therefore he is brave can be analyzed in a similar way. It is common to distinguish between two types of approaches that refer back to Grice: neo-Gricean approaches, represented mainly by Gazdar (1979), Horn (2007) and Levinson (2000), who attempted to reduce the nine conversational maxims to two principles; and post-Gricean approaches, whose main concern is to adjust the border between semantics and pragmatics, and to attribute truth-conditional properties to a pragmatic content. Post-Griceans are mainly represented by Sperber and Wilson (Sperber and Wilson 1986, 1995; Wilson and Sperber 1993, 2000, 2004), Carston (2002, 2004), Bach (2004, 2006) and Recanati (2004). The subtitle of Levinson’s book is The Theory of Generalized Conversational Implicatures. See Moeschler (2007b) for an argument based on intercultural communication and misunderstandings.

References Atlas, Jay and Stephen C. Levinson 1981 It-clefts, informativeness, and logical form: radical pragmatics (revised standard version). In: Peter Cole (ed.), Radical Pragmatics, 1–61. New York: Academic Press. Austin, John L. 1962 How to Do Things with Words. Cambridge, MA: Harvard University Press. Bach, Kent 2004 Pragmatics and the philosophy of language. In: Laurence R. Horn and George Ward (eds.), The Handbook of Pragmatics, 463–487. Oxford: Basil Blackwell.

Conversational and conventional implicatures

429

Bach, Kent 2006 What does it take to refer? In: Ernest Lepore and Barry C. Smith (eds.), The Oxford Handbook of Philosophy of Language, 516–554. Oxford: Oxford University Press. Benveniste, Emile 1966 Problèmes de linguistique générale. Paris: Gallimard. Benveniste, Emile 1974 Problèmes de linguistique générale II. Paris: Gallimard. Borg, Emma 2006 Intention-based semantics. In: Ernest Lepore and Barry C. Smith (eds.), The Oxford Handbook of Philosophy of Language, 250–266. Oxford:Oxford University Press. Carston, Robyn 1998 Informativeness, Relevance and Scalar Implicature. In: Robyn Carston and Seiji Uchida (eds.), Relevance Theory: Applications and implications, 179–236. Amsterdam: John Benjamins. Carston, Robyn 2002 Thoughts and Utterances. The Pragmatics of Explicit Communication. Oxford: Basil Blackwell. Carston, Robyn 2004 Relevance Theory and the Saying/Implicating distinction. In: Laurence R. Horn and George Ward (eds.), The Handbook of Pragmatics, 633–656. Oxford: Basil Blackwell. Clark, Herbert H. and Susan E. Haviland 1977 Comprehension and the given-new contrast. In: Roy O. Freedle (ed.), Discourse Production and Comprehension, 1–40. Hillsdale: Lawrence Erlbaum. Cohen, L. Jonathan 1971 Some remarks on Grice’s view about the logical particles of natural language. In: Yoshua Bar-Hillel (ed.), Pragmatics of Natural Language, 50–68. Dordrecht: Reidel. Cole, Peter (ed.) 1978 Syntax and Semantic 9: Pragmatics. New York: Academic Press. Cole, Peter (ed.) 1981 Radical Pragmatics. New York: Academic Press. Cole, Peter and Jerry L. Morgan (eds.) 1975 Syntax and Semantics 3: Speech Acts. New York: Academic Press. Ducrot, Oswald 1972 Dire et ne pas dire. Principes de sémantique linguistique. Paris: Hermann. Ducrot, Oswald 1973 Le structuralisme en linguistique. Paris: Ed. du Seuil. Ducrot, Oswald 1989 Linguistique et formalisation. In: Logique, structure, énonciation, 66–94. Paris: Ed. de Minuit. Fox, Danny 2007 Free choice disjunction and the theory of Scalar Implicatures. In: Uli Sauerland and Penka Stateva (eds.), Presupposition and Implicature in Compositional Semantics, 71–120. Basingstoke: Palgrave Macmillan.

430

Jacques Moeschler

Frege, Gottlob 1892 Sinn und Bedeutung. Zeitschrift für Philosophie und philosophische Kritik 100: 25–50. Frege, Gottlob 1952 On sense and reference. In: Paul T. Geach and Max Black (eds.), Translations from the Philosophical Writings of Gottlob Frege, 56–78. Oxford: Basil Blackwell. Gazdar, Gerald 1979 Pragmatics. Implicature, Presupposition, and Logical Form. New York: Academic Press. Geis, Michaels L. and Arnold Zwicky 1971 On invited inference. Linguistic Inquiry 2: 561–566. Cambridge, MA: The MIT Press. Grice, H. Paul [1957]1989 Meaning. In: H. Paul Grice, Studies in the Ways of Words, 213–223. Cambridge, MA: Harvard University Press. Grice, H. Paul 1975 Logic and Conversation. In: Peter Cole and Jerry L. Morgan (eds.), Syntax and Semantics 3: Speech Acts, 41–58. New York: Academic Press. Grice, H. Paul 1978 Further notes on logic and conversation. In: Peter Cole (ed.), Syntax and Semantics 9: Pragmatics, 113–127. New York: Academic Press. Grice, H. Paul 1981 Presupposition and conversational implicature. In: Peter Cole (ed.), Radical Pragmatics, 167–181. New York: Academic Press. Grice, H. Paul 1989 Studies in the Ways of Words. Cambridge, MA: Harvard University Press. Harnish, Robert 1991 Logical form and implicature. In: Steven Davis (ed.), Pragmatics, 316–364. Oxford: Oxford University Press. Hauser, Marc D., Noam Chomsky and W. Tecumseh Fitch 2002 The faculty of language: What is it, who has it, and how did it evolve? Science 298: 1559–1579. Horn, Laurence R. 1972 On the Semantic Properties of Logical Operators in English. Bloomington: IULC. Horn, Laurence R. 1984 Toward a new taxonomy for pragmatic inference: Q-based and R-based implicature. In: Deborah Schiffrin, (ed.), Meaning, Form, and Use in Context, 11–42. Washington: Georgetown University Press. Horn, Laurence R. 1985 Metalinguistic negation and pragmatic ambiguity. Language 61(1): 121– 174. Horn, Laurence R. 1988 Pragmatic theory. In: Frederick Newmeyer (ed.), Linguistics: The Cambridge Survey, Vol. I, 113–145. Cambridge: Cambridge University Press.

Conversational and conventional implicatures

431

Horn, Laurence R. 1989 A Natural History of Negation. Chicago: The University of Chicago Press. Horn, Laurence R. 2004 Implicature. In: Laurence R. Horn and George Ward (eds.), The Handbook of Pragmatics, 3–28. Oxford: Basil Blackwell. Horn, Laurence R. 2007 Neo-Gricean pragmatics: a manichean manifesto. In: Noel Burton-Roberts (ed.), Pragmatics, 158–183. Basingstoke: Palgrave Macmillan. Horn, Laurence R. 2011 histoire d’*O: Lexical pragmatics and the geometry of opposition. In: JeanYves Béziau and Gilbert Payette (eds.), New Perspectives on the Square of Opposition, 393–426. Bern: Peter Lang. Huang, Yan 2007 Pragmatics. Oxford: Oxford University Press. Huang, Yan 2010 Implicature. In: Louise Cummings (ed.), The Pragmatics Encyclopedia, 205–210. Abington: Routledge. Karttunen, Lauri and Stanley Peters 1979 Conventional implicatures. In: Choon-Kyu Oh and David A. Dinneen (eds.), Syntax and Semantics 11: Presupposition, 1–56. New York: Academic Press. Kempson, Ruth 1975 Presupposition and the Delimitation of Semantics. Cambridge: Cambridge University Press. Levinson, Stephen C. 1983 Pragmatics. Cambridge: Cambridge University Press. Levinson, Stephen C. 1987 Minimization and conversational inference. In: Jeff Verschueren and Marcella Bertucelli-Papi (eds.), The Pragmatic Perspective, 61–129. Amsterdam: John Benjamins. Levinson, Stephen C. 2000 Presumptive Meanings. The Theory of Generalized Conversational Implicature. Cambridge, MA: The MIT Press. Moeschler, Jacques 2007a Why are there no negative particulars? Horn’s conjecture revisited. Generative Grammar at Geneva 5: 1–13. Moeschler, Jacques 2007b The role of explicature in communication and in intercultural communication. In: Istvan Kecskes and Laurence R. Horn (eds.), Explorations in Pragmatics. Linguistic, Cognitive and Intercultural Aspects, 73–94. Berlin/New York: Mouton de Gruyter. Moeschler, Jacques 2010 Is discourse pragmatics possible? In: Alessandro Capone (ed.), Perspectives on Language, Use, Pragmatics, 217–241. Munich: Lincom Europa. Moeschler, Jacques and Anne Reboul 1994 Dictionnaire encyclopédique de pragmatique. Paris: Ed. du Seuil.

432

Jacques Moeschler

Morris, Charles W. 1938 Foundations of the Theory of Signs. In: Otto Neurath, Rudolf Carnap and Charles W. Morris (eds.), International Encyclopedia of Unified Science, 77–138. Chicago: The University of Chicago Press. Noveck, Ira and Dan Sperber (eds.) 2004 Towards Experimental Pragmatics. Basingstoke/NY: Palgrave Macmillan. Potts, Christopher 2005 The Logic of Conventional Implicatures. Oxford: Oxford University Press. Recanati, François 1994 Contextualism and anti-contextualism in the philosophy of language. In: Savas L. Tsohatzidis (ed.), Foundations of Speech Acts Theory, 156–166. London: Routledge. Recanati, François 2004 Pragmatics and semantics. In: Laurence R. Horn and Gregory Ward (eds), The Handbook of Pragmatics, 442–462. Oxford: Basil Blackwell. Russell, Bertrand 1905 On denoting. Mind 14: 479–493. Sadock, Jerrold M. 1978 On testing for conversational implicature. In: Peter Cole (ed.), Syntax and Semantic 9: Pragmatics, 281–297. New York: Academic Press. Saussure, Ferdinand de [1916]1968 Cours de linguisique générale. Paris: Payot. Searle, John R. 1969 Speech Acts. An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Searle, John R. 1979 Expression and Meaning. Studies in the Theory of Speech Acts. Cambridge: Cambridge University Press. Sperber, Dan and Deirdre Wilson 1986 Relevance. Communication and Cognition. Oxford: Basil Blackwell. Sperber, Dan and Deirdre Wilson 1995 Relevance. Communication and Cognition. Oxford: Blackwell, 2nd augmented edition. Stalnaker, Robert C. 1977 Pragmatic presupposition. In: Andy Rogers, Bob Wall and John P. Murphy (eds.), Proceedings of the Texas Conference on Performatives, Presuppopsition and Implicatures, 135–147. Arlington: Center for Applied Linguistics. Strawson, Peter F. [1950]1971 On referring. In: Jay F. Rosenberg and Charles Travis (eds.), Readings in the Philosophy of Language, 175–195. Englewood Cliffs: Prentice-Hall. Wilson, Deirdre 1975 Presupposition and Non-Truth-Conditional Meaning. New York: Academic Press. Wilson, Deirdre and Dan Sperber 1993 Linguistic form and relevance. Lingua 90(1/2): 1–25. Wilson, Deirdre and Dan Sperber 1998 Pragmatics and time. In: Robyn Carston and Seiji Uchida (eds.), Relevance Theory: Applications and Implications, 1–22. Amsterdam: John Benjamins.

Conversational and conventional implicatures

433

Wilson, Deirdre and Dan Sperber 2000 Truthfulness and relevance. UCL Working Papers in Linguistics 12: 215–254. Wilson, Deirdre and Dan Sperber 2004 Relevance theory. In: Laurence R. Horn and Gregory Ward (eds), The Handbook of Pragmatics, 607–632. Oxford: Basil Blackwell.

The construal of non-literal meaning-in-context

16.

Figurative language in discourse Alice Deignan

1.

Introduction

1.1.

An example

An assessment of climate change published in 2007 contained an error about the rate at which the glaciers in the Himalayas are melting. The publication of the mistake generated a number of texts, including an article by Rajendra Pachauri in The Guardian, a left-of-centre British newspaper, published on 26 March 2010. Pachuari argued strongly that climate change is happening rapidly and that the error about the Himalayan glaciers was immaterial to the wider issue. The 750-word text is typical of much written and spoken discourse in that it contains many expressions that are not used with what most speakers would regard as their basic, literal sense. These include the following metaphors (underlined): sound body [of science], distort, mammoth task, transparent, persecution and pay a [terrible] price. This is an illustration of the now well-established point (for example, Lakoff and Johnson 1980; Gibbs 1994; Kövecses 2002) that metaphors are ubiquitous in naturally-occurring text. The expressions were identified as metaphors using a procedure explained by the Pragglejaz Group (2007). For the first expression, a minimal outline of the procedure is as follows: an examination of body in this text shows that its contextual meaning is something like ‘large amount of, collection’. This meaning seems to be semantically related to what most speakers would regard as a more basic meaning of body, ‘physical form of a human or animal’. The semantic relation seems to be one of comparison, therefore the contextual meaning is metaphorically related to the basic meaning. Metaphor is the most widely-researched form of figurative language, but increasingly metonymy is also recognized as important and frequent in language. Metonymy is usually understood as the relationship between an entity and something that is a part or aspect of it, or that is associated with it. Pachauri’s article contains some examples of metonymy, including agencies, in the phrase “official investigations by intelligence agencies”. In this context, agencies refers to people who work for agencies, an example of the metonymy described as INSTITUTION FOR PEOPLE RESPONSIBLE by Lakoff and Johnson (1980: 39). The term ‘figurative language’ is used here to include metaphorical and metonymical expressions, as well as less well-researched phenomena such as personification, irony and hyperbole. An analysis of the figurative language used in Pachauri’s text raises some interesting issues. Firstly, his use of metaphor seems heavily evaluative. The posi-

438

Alice Deignan

tively evaluating terms: sound, mammoth and transparent refer to the work of climate scientists, while the negative terms: distort, persecution, and pay a terrible price refer to criticisms of scientists and the consequences of this criticism. This raises the question of whether figurative language has a special role in evaluation, and, more widely, what are its typical functions in discourse. The second issue concerns the density of figurative language at different points in the text. The headline of the article was: “Don’t hound the climate scientists”, and the final sentence reads: “Thankfully, with inputs from thousands of respected scientists, world governments and now the IAC, the panel is in a better position than ever to provide a robust and reliable scientific basis for tackling the growing challenge of climate change.” The headline has the metaphor hound, and in the final sentence there are several metaphors, including inputs, position, robust, basis, tackling, growing, and challenge, as well as the metonymies world governments, IAC and panel. This seems a very high density of figurative language for a single sentence, and raises the question of whether figurative expressions are frequently found in text-initial and final position, and if so, for what reason. To answer these questions, the text itself, and others, both similar in genre and different, need to be studied in detail. The analyst also needs to consider the figurative expressions found in this text against their uses in other contexts. The position and role of this text in the wider argument about climate change are also relevant. In other words, the figurative expressions found in this text cannot be understood without a consideration of their context at several levels: within the text, within the language system, and within this and related genres. Studying context in this way is known as taking a ‘discourse’ approach. A discourse approach to figurative language will help the analyst to understand how writers use it to further their goals, and can also shed light on the nature of figurative language itself. I now explain what I mean here by ‘discourse’ before discussing in more depth a discourse approach to figurative language. 1.2.

Discourse

The term ‘discourse’ is generally understood within language description to refer to language above the level of the sentence. The early studies of discourse, in the 1970s and 1980s, represented a departure from an approach to language analysis that had taken the grammatically well-formed sentence as its central unit. However, there is more to discourse analysis than a linguistic description of spoken and written texts; the study of discourse is also characterized by its concern with purpose (Thornbury 2005). It follows that discourse analysis “cannot be restricted to the description of linguistic forms independent of the purposes or functions which these forms are designed to serve in human affairs” (Brown and Yule 1983: xiii). A central strand of discourse analysis is the detailed linguistic examination of the organization of both spoken and written texts, and the role of language devices

Figurative language in discourse

439

in signalling this organization. Within the analysis of spoken discourse, attention has focused on issues such as topic development and change, turn taking, and opening and closing sequences (for example, Nunan 1993). In interactions where participants have specific roles and functions, researchers have considered how participants manage the ongoing talk. The context that has probably been most widely researched is the classroom, following seminal work on the structure of teacher and pupil talk by Sinclair and Coulthard (1975). Discourse in the workplace has also been researched (e.g. work in the collection edited by Sarangi and Roberts, 1999). It will be shown below that figurative language seems to play an important organizational role in both spoken and written discourse. The analysis of discourse, therefore, takes account of the personal and social goals of participants in the discourse. This dimension has been extended within the Critical Discourse Analysis movement from the concern with the participant as an individual to the study of wider social, ideological and political aspects of discourse. Within this movement, Fairclough (1989) points out that discourse needs to be understood in three ways, as the text itself, as how it is produced and interpreted within an interaction, and as how it is produced and interpreted within the wider social context. He also importantly notes that any single discourse is situated against the backdrop of the range of discourse types that are sanctioned by the group, and at a higher level, by society. Chilton (2009: 42) writes that “viewing discourse in this way means that it may be closely linked with social and political institutions, their characteristic linguistic forms being bound up with both conceptualisation and interaction among individuals and groups.” Later in this chapter, I review some of the important and useful work on the ideological use of figurative language in discourse. The analysis of discourse can also be understood in terms of different types of texts, or, seen in terms of language users, the discourses of different groups. This approach is taken by Semino (2008) in her study of metaphor in discourse. In this approach, researchers analyze texts produced by specific groups of language users, and try to determine how these differ from other kinds of texts. Some research of this kind aims to develop more detailed and accurate descriptions of genres and registers for pedagogical purposes. Other researchers have goals similar to those of the Critical Discourse Analysis school, and seek to uncover ideological positions underlying specific genres and registers, and to describe how these are encoded in language. This is an important aspect of figurative language studies: Gibbs (2010: 3) writes that metaphor researchers with a commitment to authentic data “have begun to study metaphor in virtually every type of discourse, in fields as diverse as politics, the law, music or food”.

440 1.3.

Alice Deignan

Importance of discourse studies of metaphor

Much recent work on figurative language has been carried out within the cognitive school, in the Conceptual Metaphor Theory tradition (Lakoff and Johnson 1980; Kövecses 2002). This work has tended to focus on issues of concern to that field, such as mental mapping and processes, rather than the issues related to discoursal use, noted above. Caballero (2003a: 145) writes that “the interest in unearthing the cognitive motivations and processes at work in metaphor has led to neglect of its linguistic realization and discourse role”. The cognitive approach has nonetheless provided important theoretical input for applied linguists studying figurative language, and most current discourse work uses the theoretical framework developed by cognitive linguists, albeit as a construct to be questioned, supplemented and refined. The discourse approach insists on the importance of studying language in its context of use (for example, Cameron 1999). In this approach, it is argued that studying linguistic expressions in isolation may miss central aspects of their meaning and use; for instance, Gibbs and Lonergan write that “metaphors are not decontextualised entities, to be studied in isolation” (2009: 251), and it has been shown that much metonymic meaning is only retrievable with a detailed understanding of context (Deignan, Littlemore and Semino, forthcoming). Research findings are still far from exhaustive, but they suggest that figurative language plays a central role in discourse. Such findings importantly complement the cognitive view, which has tended to focus on the information-bearing properties of figurative language. 1.4.

Goals of this chapter

This chapter considers the role that figurative language plays in spoken and written discourse. It then considers what the findings from this investigation can tell us about the nature of figurative language. The chapter begins by discussing research into the functions of metaphor in discourse; a brief account of the critical discourse approach to metaphor analysis is included in this section. I then discuss figurative language use within different discourse types, and review some studies into the form that figurative language takes within discourse. Finally, I consider the implications of the work reviewed for our understanding of figurative language.

2.

The functions of figurative language in discourse

Following Moon (1998), in her work on fixed expressions and idioms, and Caballero (2003a), in her discussion of metaphors of architecture, I use Halliday’s (1994) three-way classification of the functional components of meaning: the ideational, interpersonal and textual.

Figurative language in discourse

2.1.

441

The ideational function of figurative language in discourse

Clearly, like almost all language, most figurative language has some semantic content and therefore an ideational function. As Caballero (2003a) points out, a number of studies of metaphor use in specialized genres focus on ideational functions, because they are concerned with meanings specific to particular groups of language users; this type of study is discussed in section 3. It is widely accepted that the special ideational contribution of metaphor may be describing and referring to abstract topics: Lakoff and Johnson (1980) claim that without figurative language, little could be said about non-concrete aspects of our lives. Deignan, Littlemore and Semino (forthcoming) show that metonymy is also a central resource for expressing genre-specific meanings, perhaps more important than metaphor in some genres. Figurative language has been shown to play a role in the communication and development of academic and scientific knowledge. Its contribution to scientific theory has been discussed; for instance, a number of writers discuss the analogy of electricity with water as a tool for understanding and theorizing (for example, Gentner and Woolf 2000). Boyd (1993) argues that metaphors can play a role in developing new theory, particularly in young sciences. He lists metaphors drawn from computer science used in cognitive psychology, although the data do not seem to have been taken directly from naturally-occurring discourse. Experimental evidence (for example, Gentner and Gentner 1983) suggests that analogies certainly play a very important role in the conceptual structuring of theoretical and academic knowledge. Gülich (2003) also found that metaphor is used to package and convey knowledge. She analyzed a corpus of spoken German, consisting of transcribed medical seminars in rehabilitation centres, and interviews with chronically ill patients. Gülich was interested in the strategies used by experts to communicate with non-experts, and found that metaphor (including comparison and analogy) was one of the four key procedures they used to illustrate the functioning of various organs. In her data, the heart was sometimes described as an engine, and the circulatory system as a traffic system. Deignan, Littlemore and Semino (forthcoming) show that metonymical expressions are common in specialized genres such as the language of football, and conversations between workers in a children’s nursery. For example, in football commentary, ball frequently stands metonymically for actions involving the ball, generating expressions such as a through ball. Cameron’s (2003) analysis of classroom talk showed that teachers sometimes use metaphor and analogy to explain new topics to their pupils. For instance, she analyzes textbooks used with primary school children, in which the heart is described metaphorically as a pump, and the ozone layer is described as a blanket. It should be noted though that Cameron found that such pedagogical uses of metaphor were not the most frequent uses in the classrooms she observed; other uses, particularly interpersonal ones (discussed below), were more frequent. Interest-

442

Alice Deignan

ingly, her data also raise questions about the effectiveness of metaphor as a pedagogical tool. When she discussed the HEART IS A PUMP metaphor with children who had read a text based around this, she found that the metaphor was misunderstood by some of the children. For these children, the most salient kind of pump was a bicycle pump, which, they said, prevents bicycle tyres from going floppy, by introducing air under pressure into the tyre. By analogy, some of them reasoned, the heart prevents the body from going floppy – that is, from dying. The children seemed to have completely missed the pedagogical point of the metaphor, that the heart enables blood to circulate around the body, a metaphor which depends on understanding that pumps often push substances round a system. The ideational function of metaphor is also seen in talk about another abstract area: emotions. In another discourse study, Cameron (2008a, 2008b) analyzed in depth several conversations between a former member of the Irish Republican Army (IRA) convicted of planting a bomb in Brighton, UK, in 1984, and the daughter of a man killed by the bomb, in which each tries to understand the other’s feelings at the time of the attack and later. Cameron finds a fascinating pattern of metaphor development, and metaphor shifting and sharing, with concentrations of figurative language use at key points in the talk. One of these is where each speaker acknowledges the other’s pain, verbalized in their sharing the figurative use of a healing metaphor. Another is where the ex-IRA member explores the emotions felt by himself and other IRA members, and rejects hatred as their primary motivation (Cameron and Stelma 2004). At first glance, her data suggest that metaphor is used as a resource to express strong emotion, but Cameron shows that the picture is more complex than this. At one point in the transcripts, the victim’s daughter describes the moment that she heard that her father had been killed, and her feelings then and later. This passage is almost completely free of metaphorical use, and yet is perhaps the point where the strongest emotions are described. Cameron’s series of analyses suggests that metaphor is used less for directly reporting feelings, but rather to organize and explore feelings and relationships, to create a coherent structure from them. Many discourse studies, such as Gülich’s (2003), begin with the functions of explaining and illustrating, which they then seek to explore. They are not attempts to classify all other functions of figurative language. Figurative language, especially metaphor, certainly seems to be an important resource for communicating, explaining and illustrating knowledge. However, the small number of studies that have analyzed figurative language in whole texts as a starting point, rather than starting with stretches of text in which it is known that knowledge is being explained and illustrated, suggest that explaining and illustrating are not the most frequent functions of figurative language. While findings about the ideational function of figurative language are valuable, in the picture beginning to emerge from the currently small number of discourse studies, this function is perhaps its least important; interpersonal and textual meanings seem to be much more significant. These are now discussed.

Figurative language in discourse

2.2.

443

The interpersonal function of metaphor in discourse

In this section, I describe some studies showing how figurative language is used in the management of relationships between speakers. I discuss the evaluative function of figurative language in discourse, and then widen this to the ways in which figurative language carries ideological meaning in institutional and political discourse. In her study of educational and classroom discourse, Cameron (2003) found that one of the most frequent uses of figurative language was interpersonal: it is used to manage behaviour and relationships between teachers and learners. For instance, teachers use figurative language in what Cameron calls ‘agenda management’: setting goals for activities, opening and closing them, and urging pupils to work harder. Figurative language is also used in evaluative feedback and ‘control’ sequences, where teachers issue instructions about behaviour. Cameron claims that the function of metaphor is to help avoid threats to face (2003: 136). Metaphor can achieve this because of its off-record nature; other kinds of figurative language, such as irony and humorous hyperbole, can also reduce threat to face. Cameron and Stelma (2004) investigated the relative density of figurative language in stretches of text. Like Cameron’s work on educational discourse, they examined texts in detail, by hand, considering each word as a potentially figurative use. They are particularly concerned with the distribution of figurative language and how this coincides with the topics under discussion. They note that figurative language seems to be distributed unevenly in many texts, and designate as ‘clusters’ the areas where figurative use is dense. Several clusters were identified in the conversations described above between the former IRA bomber and the daughter of one of his victims. These clusters were found where a good deal of interpersonal work is being carried out, specifically, where each participant is struggling to explain him/herself to the other, and to understand the other. This can be seen as an ideational function, but it seems significant that while most of the text contains ideational uses of language, it is where the most interpersonal work is required, in terms of bridging gaps in understanding, that clusters of figurative language are found. Pollio and Barlow (1975) examined psychotherapeutic discourse, and also found that there was a concentration of figurative language at the points where, according to expert informants, substantial therapeutic insight occurred. Moon (1998) conducted one of the earliest large-scale corpus investigations into idiomatic and fixed expressions in naturally occurring texts. A large number of the expressions she studied were figurative. She discusses the metaphorical expression the worm has turned in some detail, pointing out that the referential meaning is only a small part of the information that we presumably all have about how the expression is used in text. Importantly, the expression “evaluates positively the action or behaviour of a person who has [previously in the text, or elsewhere] been negatively evaluated, typically as weak, dull, passive or submissive” (1998: 216).

444

Alice Deignan

Moon built a database of 6776 fixed expressions and idioms, starting from expressions that were marked as ‘phrases’ in the first edition of the Cobuild English Language Dictionary (of which she was an editor), supplemented by further corpus research. She analyzed citations from a subset of the Oxford Pilot Corpus, developed as part of the British National Corpus initiative (Moon 1998: 44–48). In her database, 89 % of all expressions that have some figurative content also have an evaluative function. One reason for this may be interpersonal; in these data, negatively evaluating figurative expressions were much more frequent than positively evaluating ones. This may be because figurative language can be off-record, and is therefore a face-saving way of criticizing: criticism is often expressed through figurative language. Semino (2008) came to a similar conclusion in her analysis of the former British Prime Minister Tony Blair’s use of metaphors. In 2003, addressing the Labour Party conference, Blair stated “Forward or back. I can only go one way. I’ve not got a reverse gear” (cited in Semino, 2008: 81). Semino argues that the use of the figurative expressions forward, back and reverse gear enables Blair to criticize the section of his own party that adheres to what are seen as traditional socialist values, in opposition to his new vision for the Labour Party. She claims that for Blair to have used literal language to make this point would have involved on-record negative evaluation of some of his parliamentary colleagues, which would have been very unwise politically. Deignan (2010) explored the evaluative function of figurative language using corpus data. Like Moon, she found that evaluation is a very important feature of figurative expressions. She investigated four mechanisms through which evaluation may be performed, each of which has been discussed in the metaphor literature. These are: – through the exploitation of entailments, as discussed by Lakoff in, among other work, his discussion of the use of metaphor to frame the wars in the Persian Gulf (2003); – through the exploitation of connotations (Deignan 2000); – through the choice of a source domain that resonates with listeners or readers; and – through creating metaphorical scenarios (Musolff 2004), that is, fragments of stories, which contain implicit evaluations. Deignan’s data were naturally-occurring, taken from both spoken and written discourse. She searched them using corpus techniques, which have the advantage of allowing the analyst to study many citations of the same expression. The disadvantage of this approach, compared to Cameron and Stelma’s approach of handsearching a smaller amount of data (2004, discussed above), is that it is far more difficult for the analyst to locate figurative expressions within the unfolding discourse. This means that it is difficult to study their organizational role within the text as a whole, and most corpus analysts do not attempt this. Further, within a cor-

Figurative language in discourse

445

pus approach, the analyst does not “know” the speakers or writers in the same way that an analyst conducting a more detailed search would, and ascribing interpersonal meaning is therefore undertaken with much more caution. However, despite the more limited context available, for the study of the discourse meaning and function of a specific expression or expressions, the richness of data gathered through corpus searches is invaluable. In her study, Deignan found a predominance of multi-word figurative expressions that appeared to be evaluative in nature. By studying several hundred citations of each, she was able to match these to very specific evaluative and pragmatic functions. In many cases these were not clear-cut examples of either metaphor or metonymy, but had elements of both, as well as, in many cases, hyperbole and irony. Examples are turning point, on thin ice, [vanish] into thin air, and down at heel. Of the four evaluative mechanisms she considered, most of these expressions most closely resemble Musolff’s (2004) scenarios, though they are often even more specific and local in meaning and function than the expressions he describes, which are mainly taken from political journalism. For instance, corpus examples of turning point include: (1) This round of peace talks marks a turning point, from squabbling over procedure to negotiating the issues. (2) That appearance taught me not to be afraid of TV and I think I would put it down as the turning point in my career. In each citation, turning point seems to evaluate positively. An analysis of further corpus citations confirms this; it refers to a significant moment when a situation changed for the better. [Vanish] into thin air combines metaphor with hyperbole to stress the unexpectedness of an event, in citations such as: (3) He saw as much as 65 per cent of his net worth vanish into thin air. Although these scenario-type expressions were the most frequent way in which figurative language evaluated in her data, Deignan found evidence of all four of the evaluative mechanisms listed above. She speculates that scenario-type expressions may be the most frequent form of evaluation through figurative language in unplanned text, but that other forms may predominate in planned speaking or writing. For instance, developing a story with entailments, as illustrated in Lakoff’s Gulf War data (2003), may be more usual in carefully planned discourse, such as political speeches or advertising. The exploitation of connotations seems to be common in advertising, particularly in the choice of brand names and slogans (Cook 1989). The potential of figurative language to evaluate and persuade is of interest to Critical Discourse Analysts, because it is one way in which ideology may be encoded in a text. As a persuasive technique, using figurative language may be a particularly powerful tool because it is off-record: evaluation is generally implied rather than stated. In some cases, the reader or listener may be invited to supply the evaluation themselves rather than being offered it, as will be shown below.

446

Alice Deignan

An early writer to demonstrate these points was Schön (1993, originally published in 1979), who analyzed social policy documentation written in the 1960s, which he claimed, “shaped public consciousness about the issues of housing” (1993: 146). He showed that many people’s understanding of inner city housing had been constructed “through a complementary process of naming and framing” (1993: 146), largely through a disease metaphor. He argues that once planners and the general public accept the metaphor, the solution becomes obvious: “once we are able to see a slum as a blighted area, we know that blight must be removed” (1993: 147). It can be seen from Schön’s example that once an entity – a social phenomenon, a group of people, a situation – has been metaphorically cast in a particular way, it is a small step to map other aspects of the metaphor onto it. In the case of casting housing as a disease, the solution of a “cure” in the form of metaphorical surgery can seem natural, even inevitable. The writer may not even present the figurative solution explicitly; the readers can supply it themselves, making the persuasion both implicit and powerful. There have been a large number of studies that have taken this basic insight and applied it to the analysis of collections of texts, with the twin aims of making explicit the underlying world-views of the writers, and showing how these have been packaged persuasively. The great majority of these studies use naturally-occurring corpora, usually of written texts, because of the relative ease of collecting written discourse data, although a few studies have focused on political speeches (for instance, Charteris-Black 2005; Semino 2008). Both hand-searching and corpus techniques have been used, often in combination. One of the themes that has been examined in a number of studies is that of race. In an analysis of fiction on the topic of Middle Eastern political history, Van Teeffellen (1994) examined how Israelis and Arabs were figuratively described, finding that there was a tendency to describe Arab people using natural metaphors, such as fire and water, implying lack of conscious volition and control, and collectively, using animal metaphors, again downplaying their individuality and humanity. Santa Ana (1999) analyzed figurative language used in the Los Angeles Times in articles on immigration, and found that this often described immigrants using the lexis of animals. Coffin and O’Halloran (2006) analyzed a corpus of The Sun, a British tabloid newspaper, examining the depiction of asylum seekers and immigrants from the countries that joined the EU in 2004. They did not specifically focus on figurative language, but many of the examples they cited, such as flood and floodgates, are metaphorical. They claim that for regular readers of this newspaper and similar texts, the frequent exposure to such terms would build up to create or reinforce racist attitudes, even though an individual text would, on the surface, appear relatively harmless. Other topics that have been the object of this kind of analysis include foreign policy, in particular the recent wars in the Persian Gulf, and the discourse of business (for example, Koller 2004). Such studies are almost invariably based on the analysis of collections of texts of a particular genre or related subgenres, tending to take a corpus approach to

Figurative language in discourse

447

analysis. As discussed above, this means a focus on word meaning and function across multiple contexts, rather than an analysis of detailed interpersonal meaning as it unfolds in a particular discourse event. A difficulty for all such work is sampling; large corpora can rarely be hand-searched, and the analyst therefore needs to ensure in some way that they are not cherry-picking examples that confirm their hypothesis and ignoring other evidence from the corpus. Charteris-Black (2004; 2005) has attempted to systematize the methodology to deal with this sampling problem. He hand-searches samples of his corpora, and uses his findings as the basis for automatic corpus searches. In his studies he has attempted to combine the value of corpus searches with the insight into meaning given by more detailed study of individual texts. The relatively small corpora that he uses allow for a comparison of citations for the more frequent figurative expressions across several texts, as well as close study of individual examples in their extended discourse context. Charteris-Black names his approach Critical Metaphor Analysis, and he has used his methodology over a number of collections of naturally-occurring texts (2004; 2005). He has carried out detailed studies of the metaphors used by recent political leaders in Britain and the US (2005), arguing that “critical analysis of metaphors provides particular insight into why the rhetoric of political leaders is successful” (2005: 197). He claims that figurative language contributes to the four key rhetorical strategies that he has identified in political speaking: establishing ethos, heightening emotional impact, communicating argument and communicating myth. Interestingly, he argues that the most successful political discourses use more than one interlinked metaphor, multiplying the number of entailments. One of his studies was an analysis of 14 speeches by the former British Prime Minister Tony Blair, totalling 50,000 words. He finds significant use of the JOURNEY conceptual metaphor both to describe policy and to describe Blair’s drawn-out attempts to move the Labour Party away from its traditional socialist roots. Blair also uses creation and destruction metaphors. Charteris-Black argues that the skilful use of these metaphorical networks help Blair to position himself as a politician of conviction. At the linguistic level, Charteris-Black shows that Blair often chooses colloquial, familiar figurative expressions such as going down the drain, gathering dust and getting in through the back door, and argues that this is an attempt to show himself as a man of the people, as someone who can “draw on the informal register of pub conversation” (2005: 163). This function of figurative language, creating intimacy though the use of colloquial tropes which allude to shared culture and world views, was noted by Gibbs and Gerrig (1989). It has also been found in the speeches of another successful political orator, Silvio Berlusconi, by Semino and Masci (1996). The studies discussed in this section show that figurative language serves several functions that can be classified as interpersonal. Firstly, it can be used in the

448

Alice Deignan

management of face in informal or institutional conversation, and seems to be chosen especially frequently when speakers are negotiating differences in experience and understanding. It is frequently found in difficult interpersonal moves such as criticism, and negative evaluation more generally. A second interpersonal function of figurative language is conveying evaluation, through several mechanisms but especially using detailed figurative scenarios. This function permeates everyday, unplanned discourse. Figurative language can also be exploited to present ideological viewpoints, and is used very skilfully in the carefully planned discourse of politicians. This also seems to occur in journalism, another genre produced by highly skilled language users: in the article on global warming, discussed at the beginning of this chapter, there were a number of evaluative uses of figurative language which seem to be of this kind. Finally, figurative language is used to allude to shared experience and shared ways of framing that experience, in an attempt to stress bonds between speaker and hearer. 2.3.

The textual function of metaphor in discourse

Moon’s (1998) corpus analysis of fixed expressions and idioms and their evaluative function was briefly described above, and her example of the worm has turned was reported. The following examples of the same expression are from the Oxford English Corpus: (4) The government thought it could get away with bullying people and its thinking has looked more and more out of touch and arrogant as more people have been persuaded of the protest’s legitimacy. And that is where the worm turns. The British public will never start any trouble but when there’s no escape, it turns into a snarling dog … (5) It comes in every parent/child relationship – when the worm turns. Those big, bright eyes that once looked at you with adoring love begin to roll upwards, and the slight clucking of the tongue can be heard … (6) The [cricket] runs came along at a rate of about one a minute, which very few teams have managed against such a parsimonious attack. The worm had turned. Moon notes that as well as its evaluative function, the worm turns seems to have a textual function, acting “as a bridge between statements of the status quo and of a new state of things” (1998:216), and thus typically marking a change of direction within the discourse. This can be seen in (1) above; the text up to the worm turns describes a government bullying the British public, while the text after the expression describes the public no longer tolerating this, and the expression, as Moon puts it, acts as a bridge. In (2), the expression seems to predict and encapsulate what is about to be said; the contextual meaning of the worm turns is filled out in

Figurative language in discourse

449

the following sentence. In (3), the phrase summarizes what has been said in the preceding sentences. Moon cites other figurative expressions which seemed to have a text organizational function, including: with hindsight, in a nutshell, at the end of the day, more to the point, far be it from me, and I’m bound to say. Corts and Pollio (1999) studied the occurrence and frequency of metaphor in college lectures, analyzing each stretch of spoken text by hand and in detail, in contrast to the cross-sectional corpus approach taken by Moon. They were particularly interested in what they termed ‘bursts’ of metaphor. These are similar to the clusters which Cameron and Stelma (2004) regard as significant, that is, short stretches within the discourse where there was a much greater frequency of metaphor than in the text as a whole. In their data, bursts of figurative language coincided with sections in the lecture where a new topic was introduced, the lecturer apparently anticipating that students would not be familiar with the material. Bursts also occurred when a familiar topic was presented from a new perspective. In a later study, Corts (2006) arrived at similar findings. He studied two lectures each from two disciplines, Geology and Greek Mythology. Each lecture contained between two and five bursts, and on average one third of all figurative language was produced in the bursts. Figurative language bursts also coincided with greater output of gesture. In some, but not all, bursts, a single, novel metaphor was used to present a new cognitive area, and this was elaborated on to develop the new ideas the lecturer was attempting to convey. This suggests that the figurative language had an ideational function, but also that it signalled new directions in the discourse. Corts and Meyers (2002) studied figurative language use in three Baptist sermons, like Corts and Pollio, concentrating on bursts and the apparent motivation for these. In these data, bursts seemed to coincide with the development of a topic using a single metaphor. Perhaps surprisingly, the topics framed figuratively in this way were rarely novel, but more likely to be the main topic of the sermon. If the main topic has been framed figuratively – often the case, as it may be a concept difficult to express in literal terms – then it will be developed using metaphorical and metonymic extensions of the same mapping, resulting in a high density of figurative language. This is clearly an ideational use of figurative language, but the writers also argue that it contributes importantly to coherence, an aspect of textual function. In one of the sermons, for example, the metaphorical mapping “take up one’s cross” is developed and expounded using related “burden” metaphors. Corts and Pollio (1999) also found that in all the lectures they studied, there was use of conventionalized realizations of a systematic metaphor that could be paraphrased as A LECTURE IS A JOURNEY. They write: “the language used suggested a tour lead (sic) by the lecturer where ideas are pointed out and seen” (1999: 95). These seemed to occur at transition points, dividing the lecture up into components or stages. The figurative expressions served to “identify subject changes, direct the audience’s focus and, finally, stress the interplay between lecture topics” (1999: 105). Expressions found included we’ll come back to that and from this point of

450

Alice Deignan

view. Corts and Meyers (2002) also found evidence of a journey metaphor “in which the act of speaking is spoken of in terms of moving forward and other metaphors” (2002: 406). These journey metaphors did not appear in bursts, but were fairly evenly distributed throughout the sermons, apparently serving to signal the stages of the argument. Cameron and Low (2004) also found that figurative language has a textual function. They reported that in academic textbook data, metaphors were often used to open and to summarize. For instance, one of their texts began with “Astronomers have discovered a huge fountain of antimatter at the heart of the galaxy” and finished with “‘It is like finding a new room in the house we have lived in since childhood,’ said Charles Dermer of the US Naval Research Laboratory” (2004: 362). Subsections within the texts in their data were also sometimes summarized with a figurative expression. Green and Kupferberg (2000) analyzed spoken discourse data: radio discussions, in which members of the public are invited to present personal narratives. In their corpus, the narratives were on the topic of addiction. Green and Kupferberg argue that speakers often use figurative language both to evaluate and to summarize their experience, after the telling of a personal story. Here again, as in Moon’s data, the interpersonal and textual functions seem closely entwined. The text discussed at the beginning of this chapter contained the metaphor hound in its headline and a very dense cluster of metaphors in the final sentence. The studies discussed here suggest that this is not unusual, and that these metaphors have both a textual and an evaluative function. In this section, a number of studies of the function of figurative language in discourse have been summarized, using Halliday’s three-way model of functional meaning. Although the picture is not yet complete, some aspects of figurative language having been more fully investigated than others, two observations can be made. Firstly, the ideational function of metaphor does not seem to be the most important of the three; indeed, there are grounds for considering it as the least important. Metaphor clusters or bursts seem to coincide with interpersonal and textual functions, rather less so than with ideational ones, though there is inevitably some ideational content. Secondly, figurative language often seems to carry out more than one of the three functions simultaneously: Moon’s (1998) example of the worm has turned shows interpersonal and textual functions combined. The interpersonal and ideational are also combined in almost all evaluative uses of figurative language. Figurative expressions normally refer to an entity or an attribute while indicating how this is evaluated. For instance, one of Deignan’s (2010) examples is the expression down at heel; this means ‘in need of repair’ and also connotes squalor and dishonesty. It is relatively rare for a figurative expression to mean simply ‘bad’ or ‘good’, that is, to evaluate alone. Despite this overlapping of functions, Halliday’s model usefully frames these functions and provides important insights into figurative language use in discourse.

Figurative language in discourse

3.

451

Figurative language and discourse type

The use of figurative language in specific genres and registers has been the object of a number of studies. Cameron (1999) points out that certain discourse communities make extensive use of systematic metaphors that are less widely used in the language generally. She refers to this phenomenon as “discourse systematicity”. She places discourse systematicity between “local systematicity”, which refers to the development of a metaphor within a single text and “global systematicity”: networks of uses of related metaphors that are widespread in the language as a whole, such as journey metaphors. Cameron’s example of discourse systematicity is the use of lexis from speech being used to talk about writing, in school activities, found in her data in examples such as “what is the writer telling us” (1999: 16). Deignan, Littlemore and Semino (forthcoming) review and carry out a number of studies of the use of figurative language in different genres and registers, arguing that figurative language use is highly bound to culture and context. They claim that a detailed model, encompassing the notions of discourse community, genre and register, is needed to handle this aspect of figurative language use rigorously. In the absence of such a model, findings from the various studies that have been conducted are not directly comparable. Nonetheless some very valuable research has been done, covering a range of discourse types. This research clearly demonstrates that different discourse communities use different conceptual metaphors and metonymies for a range of specialized purposes. In this chapter, findings of studies of discourse types are described within the Hallidayan framework of register analysis (Halliday and Hasan 1985). Register is a way of describing language variation due to context, using three variables: field, tenor and mode. These variables have correspondences with the ideational, interpersonal and textual functions (described above) respectively. Studies of figurative language use suggest that it is highly context-sensitive and that it is affected by each of these three variables. 3.1.

Field

Field “refers to what is happening, to the nature of the social action that is taking place” (Halliday and Hasan 1985: 12) and is usually interpreted as the subject matter of the text. Field is an important factor in determining lexical choice, of which figuratively used words are a component. A number of writers have pointed out that specialist registers have their own metaphors and metonyms; for instance, Caballero (2003) notes the use of textile metaphors in architecture, while White (2003) explores growth metaphors in economics. Goatly (1993, 1997) is one of the few analysts to have attempted comparative analyses using a register framework. He makes the point that field is significant to the interpretation of polysemous expressions as being either literally or metaphorically used in context. The point can be illustrated using corpus examples. The

452

Alice Deignan

words nightmare and limp each have current literal and metaphorical meanings. Around two thirds of citations of nightmare are metaphorical, and one quarter of citations for limped are metaphorical, in 100-citation samples from the Oxford English Corpus – that is, a reader could reasonably expect either a literal or metaphorical meaning for either word. However, in the following two sentences, the field is football, and this is likely to lead the reader to interpret nightmare metaphorically and limped literally. (7) England were facing their biggest World Cup nightmare last night as Wayne Rooney suffered a major injury scare. The nation’s most important striker limped off in agony. (The Sun newspaper, 31 March 2010) The following sentences are from a text about the writer’s son; the field is likely to lead the reader to interpret nightmare literally. (8) I like being with him in the morning and cleaning up the kitchen with him. He had a nightmare last night so he came into our bedroom. The word possession has several figurative senses, including one meaning ‘having control of the ball in a game’. This only accounts for around 5 citations in a random sample of one hundred from the Oxford English Corpus, but given that the topic of the following extract is again football, this seems to be the one that the reader is likely to access first. (9) Right from minute one the possession wasn’t up to standard. (The Sun newspaper, 31 March 2010) Field is not the only clue in interpreting expressions as literal or figurative; Deignan (2005, 2009) argues that there are clues to intended meaning in the linguistic co-text, which probably work at a subconscious level. 3.2.

Tenor

Tenor “refers to who is taking part, to the nature of the participants, their statuses and roles: what kinds of role relationship obtain” (Halliday and Hasan 1985: 12). Skorzcynska and Deignan (2006) compared the figurative language used in collections of text that had the same field: topics such as technology, pricing, competition and regulation within economics, and the same mode: both collections of texts were formal, written, publicly available articles. The central difference between their collections of texts was the tenor, that is, the relationship between the writer and the intended audience. One set of texts was written for fellow researchers, peers, while the other set consisted of articles from popular magazines such as The Economist, intended for well-educated but not specialist readers. Skorzcynska and Deignan found that while both kinds of text used figurative language, very few figurative expressions were common to both sets of texts. The researchers analyzed

Figurative language in discourse

453

the functions of the figurative expressions and found that these were different, the popular texts using metaphor as an explanatory tool, and the research texts using it to model theory. These functions seem to be in part a product of the tenor of each register. Skorzcynska (2010) analyzed metaphors from a business textbook and found that there was very little overlap between these and texts from popular and research articles. Again, the significant difference between her two sets of data seems to have been tenor. These studies suggest that tenor is an important factor in the choice of figurative language. 3.3.

Mode

Mode usually defines whether a text is written or spoken, and within this, a description of mode includes details such as whether it is face to face, if spoken, and planned or unplanned, if written. There has been less research into this variable, and what has been done concerns metaphor rather than other kinds of figurative language. There has been some research into metaphors in talk, and a good deal into metaphors in writing, but very little work has compared the two modes, and no significant studies have been carried out that have controlled for field and tenor in order to study how mode affects figurative language choice. Cameron (2008a) provides an account of some key characteristics of metaphor in talk. Perhaps the most significant of these is that she has found that figurative meanings are often negotiated and developed by speakers; it seems very unlikely that this occurs in writing. Speakers sometimes take on each other’s metaphors, as a way of aligning themselves with the affective meanings expressed through them. It was noted above that figurative language is used in talk between experts and non-experts as a resource for tackling gaps in each speaker’s knowledge. Because talk gives the opportunity for ongoing feedback, speakers adjust the way that they present information according to how well they think they have been understood. Cameron (2008: 206–207) cites a doctor explaining to a patient how a bladder problem will be treated, using a number of metaphors including recalibrate, messages, re-acclimatise and drill: actually what’s required is a behavioural approach to try and tackle that problem and recalibrate the messages in a way so that actually your brain re-acclimatises to only passing only getting the message pass urine (…) we call it bladder drill. Gülich’s (2003) study of expert and non-expert medical talk was discussed above; the conversations that she cites show similar developments of figurative language as an explanatory device. In her data this is not a one-way process: both medical staff and patients use figurative language to convey their expertise to the other. While not usually a medical expert, the patient is an expert in how he/she feels, and patients de-

454

Alice Deignan

velop figurative language to convey their experiences. In one of her extracts, a patient describes a heart attack as a band around the chest that tightened progressively until suddenly it burst and was left hanging. The metaphor is developed over several utterances in response to feedback and encouragement from the other speaker, a doctor. As was the case for Cameron’s examples, this level of negotiation and finetuning to the discourse needs of the other would not be possible in writing. 3.4.

Register and type of figurative language

In some cases, it seems that different discourse registers show different types of figurative language. Caballero’s (2003a, 2003b) study of figurative language in architecture reviews shows that figurative expressions, including metaphor, simile and personifications, are used mainly to provide “the lexical means to refer to certain building elements or to the way buildings are spatially located” (2003a: 158). As metaphor is a well-known mechanism for extending the lexicon, this is to be expected in technical language, and is probably a product of the variable of field. Less expectedly perhaps, Caballero observes that specific types of metaphors are chosen. In her corpus, she finds a number of conceptual metaphors, including BUILDINGS ARE CLOTH and BUILDINGS ARE LIVING ORGANISMS . She also finds many image metaphors, that is, metaphors whose grounds are visual similarity; 46 % of metaphors in her corpus are of this kind. The following citation from her data contains three sequences of image metaphors: Fixed glass skylights span each of the side arches; beneath those, a bonelike cage of steel ribs, controlled by pneumatic struts, can open and close like an eyelid, revealing the sunken, tile-encrusted orb of the IMAX to the outdoors. (Cabarello 2003b: 182)

Caballero notes that image metaphors are generally considered within the cognitive school to be relatively unimportant to reasoning, but that within her corpus, they were significant. This is clearly related to the function of architectural discourse, to describe the visual, and, as she notes, to move from two-dimensional to three-dimensional impressions. Her work shows that register, the context of discourse, can influence not only the choice of figurative language, but the type of figurative mapping that best serves the purposes of the discourse. While some very useful descriptive work has been done on figurative language use in different registers, studies have been inconsistent in the types of figurative language investigated, many only investigating metaphor. Few studies use a tightly defined procedure for identifying the types of figurative language that they discuss. Further, many researchers appear to have compiled their databases by choosing specific publications, rather than describing a register or genre – the register or genre that these may represent then needs to be inferred. In the case of spoken data, data are often collected from a particular context or professional field, but within this there may be different registers or genres. For instance, Gülich’s (2003) dataset

Figurative language in discourse

455

contains conversations between professionals and patients as well as seminar discussions, two text types that are probably different registers. This means that it is very difficult to use findings comparatively to analyze how figurative language use varies according to discourse community, genre or register. It seems very likely that there are systematic variations in figurative language according to discourse type, but more research, framed within a theory of register and genre, is needed.

4.

The form of figurative language

The form taken by figurative language is of interest to discourse scholars, both for its own sake and because, they argue, form has implications for our understanding of figurative language, processing and production. Research into two aspects of form, lexical collocation and lexico-grammar, is now discussed. 4.1.

Collocation

Collocation refers to the tendency for words to co-occur regularly, and is a feature of all natural discourse. Collocational patterns are usually recognized by fluent speakers of a language once they are demonstrated through the analysis of data from language in use, but they cannot generally be predicted accurately by intuition alone (Deignan 2005). Analysts generally use automatic techniques over large amounts of corpus data to investigate them. Deignan (for example, 2005, 2009) has carried out a number of investigations into the collocational patterns found in figurative language. Recently she has analyzed data from the Oxford English Corpus to show that for some words, the more fixed a collocation is grammatically and lexically, the more likely it is to have a figurative meaning (2008). For instance, regular collocates of direction are words such as opposite, step, heading, moving, and right. These are found in corpus citations such as the following: (10) … a pose in which one part of the body is turned in the opposite direction from another. (11) … the final ironic plot twist in Seligmann’s novel takes the opposite direction to that in Roth’s. For collocations consisting of two words, such as the co-occurrence of opposite and direction in the above citations, Deignan found no obvious patterns. In some cases, the collocations tended to be mostly literal uses, as is the case for opposite and direction, while in others they are mostly figurative. Other collocations are almost evenly split between literal and figurative meanings; when the verb head occurs in collocation with direction, it is equally likely to have a figurative meaning as a literal one. However, she found that for collocations of three or more words, especially where the words occur in an uninterrupted sequence, there is an over-

456

Alice Deignan

whelming tendency for the meaning to be figurative. The following citations are typical of this tendency: (12) It’s a first step in the right direction towards healthy, safe and better tasting lunches. (13) I’ve decided that it may be time for me to start moving in a new direction with my professional writing career. It was noted that in citations where head co-occurs with direction, the meaning is as likely to be about literal movement as about abstract tendencies. When the citations are filtered to include only those that contain right, in, for example, head in the right direction, more than 96 % show the figurative meaning of the expression. Researchers such as Sinclair (for example, 2004) and Wray (2002) have suggested that regular, fixed collocations, often termed ‘chunks’, constitute a single choice by the language user. It seems likely that such chunks are stored, produced and processed as wholes. Natural discourse seems to be made up of a large number of such chunks, the most fixed of which, according to these findings, may be figurative in meaning. 4.2.

Grammatical form

Cameron (2003) observes that the metaphors that are often discussed in the literature are nominal, but that in her data from classrooms metaphorical meaning is more frequently found in other parts of speech, especially verbs. This finding is consistent with early work on metaphor in poetry by Brooke-Rose (1958). Deignan (2006) carried out some detailed comparisons between the literal and figurative uses of the same words, and found that there is almost always some variation in the form. In many cases, a word that is a noun in its literal sense has a figurative meaning that is another part of speech. This is the case for many animal metaphors. For instance, nominal metaphorical uses of dog, pig, rat, monkey and wolf can be found if a sufficiently large corpus is searched, but they are infrequent. In comparison, the verbs to dog, to pig out/ pig oneself (eat a lot), to rat on somebody (give away their secrets), to monkey with something (fool) and to wolf (food) down are much more frequent. For some researchers (for example, Steen et al. 2010), change in a part of speech means that a use is no longer figurative, because they would view different parts of speech as derivationally related, but not the same word. This seems to deny the psychological reality for language users of the relationship between for instance, pig (noun) and pig out (verb). Deignan (2005, 2006) also found that even where there was no change of part of speech, figurative and literal uses of the same word were usually found in different lexico-grammatical patterns. For instance, face is used figuratively to refer to a person’s public image; in this sense it is almost invariably an uncount noun, in expressions such as lose face and save face, whereas its usual literal use is count (Deignan 2005).

Figurative language in discourse

457

The grammar of figurative language seems to affect processing. Steen (2004) investigated whether different metaphors were more likely to be recognized by participants depending on features of their discourse, and found that they were. Metaphors were more easily recognized by his informants when they co-occurred with other metaphors, and when they were in post-verbal position in the sentence as opposed to preverbal or verbal. He also found that his informants were more likely to recognize nominal metaphors than other types, perhaps partly explaining the predominance of nominal metaphors in the literature.

5.

Conclusion: Discourse and the nature of figurative language

This chapter has described studies of figurative language use in discourse, and has considered its function, register and form. In this section, I consider what the findings from these studies can tell us about the nature of figurative meaning. Firstly, figurative language is found to express all of the three functions of language described by Halliday (1994). While much research has focused on the ideational function – what metaphors and metonyms refer to or describe in the real world –, figurative language in discourse seems to express the interpersonal and textual functions at least as frequently. Given that studies of figurative language have tended to be especially concerned with its semantic content, this finding suggests that a shift in direction is indicated. It was also argued here that figurative language often expresses more than one function simultaneously. This capacity to compress several kinds of meaning into single expressions suggests that much figurative language is highly economical. Cameron and Deignan (2006) describe such expressions as “metaphoremes” to convey their multi-functionality, and speculate that they are the result of a process of emergence from the dynamics of discourse activity. Secondly, it seems from these studies that figurative meaning and use are extremely context-sensitive. Figurative language use appears to vary according to function, and according to the variables of register. Comparative studies are in their infancy, but it seems almost certain that there are significant intra-cultural and inter-cultural variations as well. Cameron’s various studies of spoken data (for instance, 2008) have shown that figurative meaning is often renegotiated between speakers as discourse unfolds. This has important implications for the cognitive school notion of metaphorical mappings, which can thus be seen as a resource that speakers readily manipulate and extend, rather than a fixed semantic framework. Thirdly, it seems that figurative language use is both more fixed and more flexible than we have hitherto appreciated. Deignan’s (2008) work on collocations suggests that much figurative language use has stabilized into fairly fixed three- or four-word chunks, which have very specific evaluative and pragmatic uses. Deig-

458

Alice Deignan

nan’s (2006) and Cameron’s (2003) work on grammatical patterns of figurative language suggests that as new figurative meanings evolve, they separate themselves formally from their literal counterparts. This can perhaps be understood in terms of language as a communicative system with little tolerance of ambiguity. Language users probably disambiguate words as literal or figurative using both context and register, and the details of lexico-grammatical patterns. For reasons of space, this chapter has discussed figurative language as if it were a single entity. The literature on metaphor is vastly more extensive than that on other tropes, though there is a sizeable and growing body of research into metonymy. Much less has been written about hyperbole, irony and other types of figurative meaning, and this has inevitably been reflected in the discussion in this chapter. As has been pointed out in several places in this chapter, there are many gaps remaining to be researched before we can have a full understanding of how figurative language operates in natural discourse.

References Boyd, Richard 1993 Metaphor and theory change: What is a “metaphor” a metaphor for? In: Andrew Ortony (ed.), Metaphor and Thought, 481–532. Cambridge: Cambridge University Press. Brooke-Rose, Cristina 1958 A Grammar of Metaphor. London: Mercury. Brown, Gillian and George Yule 1983 Discourse Analysis. Cambridge: Cambridge University Press. Caballero, Rosario 2003a. Metaphor and genre: The presence and role of metaphor in the building review. Applied Linguistics 24(2): 145–167. Caballero, Rosario 2003b How to talk shop through metaphor: Bringing metaphor research to the ESP classroom. English for Specific Purposes, 22(2), 177–194. Cameron, Lynne 1999 Operationalising ‘metaphor’ for applied linguistic research. In: Lynne Cameron and Graham Low (eds.) Researching and Applying Metaphor, 3–26. Cambridge: Cambridge University Press. Cameron, Lynne 2003 Metaphor and Educational Discourse. London; Continuum. Cameron, Lynne 2008a Metaphor and talk. In: Raymond Gibbs (ed.), The Cambridge Handbook of Metaphor and Thought, 197–211. Cambridge: Cambridge University Press. Cameron, Lynne 2008b Metaphor shifting in the dynamics of talk. In: Mara Sophia Zanotto, Lynne Cameron and Marilda Cavalcanti (eds.), Confronting Metaphor in Use: An Applied Linguistic Perspective, 45–62. Amsterdam: John Benjamins.

Figurative language in discourse

459

Cameron, Lynne and Graham Low 2004 Figurative variation in episodes of educational talk and text. European Journal of English Studies, 8(3): 355–373. Cameron, Lynne and Juurd Stelma 2004 Metaphor clusters in discourse. Journal of Applied Linguistics 1: 7–36. Cameron, Lynne and Alice Deignan 2006 The emergence of metaphor in discourse. Applied Linguistics. 27(4): 671–690. Charteris-Black, Jonathan 2004 Corpus Approaches to Critical Metaphor Analysis. London: Palgrave Macmillan. Charteris-Black, Jonathan 2005 Politicians and Rhetoric: The Persuasive Power of Metaphor. London: Palgrave Macmillan. Chilton, Paul 2009 Reading Sonnet 30: Discourse, Metaphor and Blending. In: Andreas Musolff and Jorg Zinken (eds.), Metaphor and Discourse, 40–58. Basingstoke: Palgrave Macmillan. Coffin, Caroline and Kieran O’Halloran 2006 The role of appraisal and corpora in detecting covert evaluation. Functions of Language, 13(1): 77–110. Cook, Guy 1989 Discourse. Oxford: Oxford University Press. Corts, Daniel P. 2006 Factors characterizing bursts of figurative language and gestures in college lectures. Discourse Studies 8(2): 211–233. Corts, Daniel P. and Howard R. Pollio 1999 Spontaneous production of figurative language and gesture in college lectures. Metaphor and Symbol 14(2): 81–100 Corts, Daniel P. and Kristina Meyers 2002 Conceptual clusters in figurative language production. Journal of Psycholinguistic Research 31(4): 391–408. Deignan, Alice 2000 Persuasive use of metaphor in discourse about business and the economy. In: Chris Heffer and Helen Stauntson (eds.), Words in Context: A Tribute to John Sinclair on his Retirement, 156–168. Birmingham: English Language Discourse Research Monographs. Deignan, Alice 2005 Metaphor and Corpus Linguistics. Amsterdam: John Benjamins. Deignan, Alice 2006 The grammar of linguistic metaphors. In: Anatol Stefanowitsch and Stefan Th. Gries (eds.) Corpus-Based Approaches to Metaphor and Metonymy, 106–122. Berlin: Mouton de Gruyter. Deignan, Alice 2008 Collocation and metaphorical meaning. Paper given at the 7th International Conference on Researching and Applying Metaphor, University of Extremadura, Caceres, Spain.

460

Alice Deignan

Deignan, Alice 2009 Searching for metaphorical patterns in corpora. In: Paul Baker (ed.), Contemporary Corpus Linguistics, 9–31. London: Continuum. Deignan, Alice 2010 The evaluative properties of metaphors. In: Graham Low, Zazie Todd, Alice Deignan and Lynne Cameron (eds.) Researching and Applying Metaphor in the Real World, 357–373. Amsterdam: John Benjamins. Deignan, Alice, Jeannette Littlemore and Elena Semino forthcoming Figurative Language, Genre and Register. Cambridge: Cambridge University Press. Fairclough, Norman 1989 Language and Power. London: Longman. Gentner, Dedre and Donald R. Gentner 1983 Flowing waters or teeming crowds: Mental models of electricity. In: Dedre Gentner and Albert L. Stevens (eds.), Mental Models, 99–129. Hillsdale, New Jersey: Erlbaum. Gentner, Dedre and Philip Woolf 2000 Metaphor and knowledge change. In: Eric Dietrich and Arthur B. Markman (eds.), Cognitive Dynamics: Conceptual Change in Humans and Machines, 295–342. Mahwah, NJ: Lawrence Erlbaum Associates. Gibbs, Raymond 1994 The Poetics of Mind: Figurative Thought, Language and Understanding. Cambridge: Cambridge University Press. Gibbs, Raymond 2010 The wonderful chaotic, creative, heroic, challenging world of Researching and Applying Metaphor: A celebration of the past and some peeks into the future. In: Graham Low, Zazie Todd, Alice Deignan and Lynne Cameron (eds.) Researching and Applying Metaphor in the Real World, 1–18. Amsterdam: John Benjamins. Gibbs, Raymond and R. Gerrig 1989 How context makes metaphor comprehension seem special. Metaphor and Symbolic Activity 4: 145–158. Gibbs, Raymond, and Julia Lonergan 2009 Studying metaphor in discourse: Some lessons, challenges and new data. In: Andreas Musolff and Jörg Zinken (eds.), Metaphor and Discourse, 251–261. London: Palgrave Macmillan. Goatly, Andrew 1993 Species of metaphor in written and spoken varieties. In: Mohsen Ghadessy (ed.), Register Analysis: Theory and Practice, 110–148. London: Pinter. Goatly, Andrew 1997 The Language of Metaphors. London: Routledge. Green, David, and Kupferberg, Irit 2000 Detailed and succinct self-portraits of addicts in broadcast stories. Discourse Studies 2(3): 305–322. Gülich, Elisabeth 2003 Conversational techniques used in transferring knowledge between medical experts and non-experts. Discourse Studies 5(2): 235–263.

Figurative language in discourse

461

Halliday, Michael.A.K. 1994 An Introduction to Functional Grammar, 2nd edn. London: Edward Arnold. Halliday, Michael.A.K. and Ruqyia Hasan 1985 Language, Context and Text: A Social Semiotic Perspective. Deakin: Deakin University Press. Lakoff, George 2003 Metaphor and War: The metaphor system used to justify war in the gulf. Available at: http://ftp.funet.fi/pub/sci/neural/neuroprose/lakoff.war.p s.Z. accessed: 23 April 2012. Lakoff, George and Mark Johnson 1980 Metaphors We Live By. Chicago: University of Chicago Press. Koller, Veronika 2004 Metaphor and Gender in Business Media Discourse: A Critical Cognitive Study. Basingstoke: Palgrave Macmillan. Kövecses, Zoltån 2002 Metaphor: A Practical Introduction. Oxford: Oxford University Press. Moon, Rosamund E. 1998 Fixed Expressions and Idioms in English: A Corpus-based Approach. Oxford: Clarendon. Musolff, Andreas 2004 Metaphor and Political Discourse: Analogical Reasoning in Debates about Europe. London: Palgrave Macmillan. Nunan, David 1993 Introducing Discourse Analysis. Harmondsworth: Penguin. Pollio, Howard R. and Jack M. Barlow 1975 A behavioural analysis of figurative language in psychotherapy: One session in a single case study. Language and Speech 18: 236–254. Pragglejaz Group 2007 MIP: A method for identifying metaphorically used words in discourse, Metaphor and Symbol, 22(1): 1–39. Santa Ana, Otto 1999 “Like an animal I was treated”: Anti-immigrant metaphor in US public discourse. Discourse and Society 10: 191–224. Sarangi, Srikant and Celia Roberts (eds.) 1999 Talk, Work and Institutional Order: Discourse in Medical, Mediation and Management Settings. Berlin: Mouton de Gruyter Schön, Donald 1993 Generative metaphor: A perspective on problem-solving in social policy. In: Andrew Ortony (ed.), Metaphor and Thought, 2nd edn., 137–163. Cambridge: Cambridge University Press. Semino, Elena 2008 Metaphor and Discourse. Cambridge: Cambridge University Press. Semino, Elena and Michela Masci 1996 “Politics is football”: Metaphor in the discourse of Silvio Berlusconi in Italy. Discourse and Society 7: 243–269. Skorczynska, Hanna and Alice Deignan 2006 Readership and purpose in the choice of economics metaphors. Metaphor and Symbol, 21(2): 105–120.

462

Alice Deignan

Skorzcynska, Hanna 2010 A corpus-based evaluation of metaphors in a business English textbook. English for Specific Purposes 29: 30–42. Sinclair, John and Malcolm Coulthard 1975 Towards an Analysis of Discourse: The English used by Teachers and Pupils. Oxford: Oxford University Press. Sinclair, John 2004 Trust the Text: Language, Corpus and Discourse. London: Routledge. Steen, Gerard 2004 Can discourse properties of metaphor affect metaphor recognition? Journal of Pragmatics, 36(7): 1295–1313. Steen, Gerard, Aletta Dorst, J. Berenicke Herrmann, Anna A. Kaal, Tina Krennmary and Trijntje Pasma 2010 A Method for Linguistic Metaphor Identification. Amsterdam: John Benjamins. Thornbury, Scott 2005 Beyond the Sentence: Introducing Discourse Analysis. Oxford: Macmillan. Van Teeffellen, Toine 1994 Racism and metaphor: the Palestinian-Israeli conflict in popular literature. Discourse and Society 5: 381–405. White, Michael 2003 Metaphor and economics: the case of growth. English for Specific Purposes 22: 131–151. Wray, Alison 2002 Formulaic Language and the Lexicon. Cambridge: Cambridge University Press.

17.

Humour and irony in cognitive pragmatics Geert Brône

1.

Introduction

Ever since the publication of Arthur Koestler’s early cognitive approach to humour in The Act of Creation (1964/1989), in which he inquires into the common cognitive grounds of highly disparate phenomena such as humour, artistic creativity and scientific discovery, cognitive psychologists and linguists have explored the cognitive mechanisms underlying humour and irony. The motivations for inquiring into the cognitive and inferential build-up of these two interrelated phenomena are manifold: (a) humour and irony are prominent types of linguistic creativity and for that reason alone deserve systematic scrutiny (the ecological validity argument, Graesser, Long, and Mio 1989; Tannen 1989; Glenn 2003), (b) they are complex cognitive phenomena on the interface between linguistic, cognitive, social, cultural, emotional and bodily factors, making them prime cases for interdisciplinary research (Goldstein 1990; Coulson 2000), and (c) they often build the playground for the ‘fluid’ conceptual system (Hofstadter 1995). Humour and irony have aroused interest in a variety of fields, ranging from philosophy to sociology and neuroscience. The multitude of theories and views that have emanated from the attempts to get to the essence of the phenomena are generally divided into three major categories: the cognitive theories that give a central role to incongruity and its resolution, social theories that highlight the importance of aggression, disparagement and the confirmation of superiority in humour, and the tension-release models that find their origin in Sigmund Freud’s early theory of the joke. Linguistic studies generally adopt an incongruity(resolution) perspective. Ever since the publication of Victor Raskin’s seminal work on the Semantic Mechanisms of Humor (1985), linguistic humour research has had a decidedly cognitive orientation, and the cognitive psychological roots of the Semantic Script Theory of Humor (SSTH, cf. 2.1.) have been adopted in a large number of studies. The specific case of irony has remained on the agenda of (cognitive) pragmatics ever since Grice’s discussion of the phenomenon as a specific case of the violation of conversational maxims. This has generated a range of theories that have been tested extensively in psycholinguistic and neurolinguistic experiments. Over the past two decades, a number of overviews have been published that discuss the main findings in linguistic and cognitive-psychological humour and irony research (Attardo 1994, 2001a, 2005; Giora 2003; Raskin 2008; Gibbs and Colston 2007). Given these excellent surveys, this chapter takes a more restricted focus in the sense that it presents an overview of the contribution of Cognitive Lin-

464

Geert Brône

guistics (henceforth CL) to the study of humour and irony. Cognitive Linguistics, with its focus on the conceptual aspects of language use, provides an adequate framework for the analysis of the interplay of quantitative and qualitative aspects in humour and irony. However, for a long time both phenomena have been a largely unexamined topic in the cognitive linguistic paradigm, in contrast to the study of other instances of ‘creative’ and non-literal language use (like metaphor, metonymy, conceptual blending, etc.). Only since the turn of the century has CL developed an interest in humour as an additional research topic. The present chapter aims to present an exhaustive overview of this programme in cognitive pragmatics, with an explicit focus on the contribution of a CL approach to the larger field of linguistic humour and irony research. While irony and humour are not generally considered to be strictly distinguishable categories, the dichotomy will serve to structure the chapter. The first part deals with cognitive and pragmatic mechanisms of meaning construction in humour. I present the basic architecture of the two most influential linguistic humour theories, the Semantic Script Theory of Humor (SSTH, 2.1.) and the General Theory of Verbal Humor (GTVH, 2.2.). In 2.3., I discuss the contribution of Cognitive Linguistics to linguistic humour research, both from the perspective of its main proponents Rachel Giora (2.3.1.) and Seana Coulson (2.3.2) and in terms of key topics (2.3.3.). The second part of the chapter provides a concise overview of cognitive pragmatic theories of irony. Following Attardo’s (2000) excellent survey, I first present an outline of the two major groups of theories, viz. those that view irony as a figure of speech (3.1.), and those that take indirect echoing and mentioning as key features (3.2.). In 3.3., the few existing cognitive linguistic accounts of irony and sarcasm, situated within the framework of mental spaces theory, are discussed against the background of the main theories in the field. The chapter is closed off with some suggestions for future research on humour and irony in cognitive pragmatics.

2.

Cognitive approaches to verbal humour

2.1.

Semantic Script Theory of Humor (SSTH)

The publication of Victor Raskin’s Semantic Mechanisms of Humor (1985) marks the beginning of a strong linguistic tradition in humour research, with a decidedly cognitive orientation. Although this seminal work sets out to develop a competence theory of humour in the generative-linguistic tradition, hence pursuing a formal linguistic model that specifies the necessary and sufficient conditions for a text to be categorized as humorous, Raskin’s actual proposal shares a number of basic assumptions with CL. In specifying his conception of a viable semantic theory, Raskin rejects a formal semantic view like that of Katz and Fodor (1963) in favour of a stronger contextual semantics that hinges on the notion of semantic scripts

Humour and irony in cognitive pragmatics

465

(Minsky 1975; Schank and Abelson 1977; and many others, cf. Kecskes, this volume). A script is defined as “a large chunk of semantic information surrounding the word or evoked by it. The script is a cognitive structure internalized by the native speaker and it represents the native speaker’s knowledge of a small part of the world” (Raskin 1985: 81). Scripts thus are broadly compatible with frames (Fillmore 1982, 1985), domains (Lakoff and Johnson 1980; Langacker 1999) and Idealized Cognitive Models (ICMs, Lakoff 1987), which are more generally used in CL.1 In comparison to the latter, however, Raskin stresses that scripts are strongly tied to lexical elements, so that “every script is a graph with lexical nodes and semantic links between the nodes” (1985: 81). Cognitive linguists generally assume a less structurally bound perspective on frames (Coulson 2000, 2006; cf. 2.3.2.). Raskin’s competence theory of verbal humour is dubbed the Semantic Script Theory of Humor (SSTH). The main hypothesis of the theory is described in terms of two conditions which must be satisfied for a text to be characterized as a ‘singlejoke-carrying text’: (i) The text is compatible, fully or in part, with two different scripts (ii) The two scripts with which the text is compatible are opposite (Raskin 1985: 99) It should be noted that the theory takes the standard joke type with a punch line as the prototypical instance of verbal humour. For analytical reasons, Raskin restricts the scope of his model to that particular subcategory (Raskin 1985: 45–46) but stresses that in principle his theory is applicable to longer text types (see however Attardo 2001a and 2.2.2. below). The basic tenets of the theory can be illustrated using the stock example in (1), analyzed in great detail and from different perspectives over the last two decades: (1) “Is the doctor at home?” the patient asked in his bronchial whisper. “No,” the doctor’s young and pretty wife whispered in reply. “Come right in.” When conceived in a linear fashion, the text up to the punch line contains a number of lexical elements (doctor, patient, bronchial) that jointly activate the salient script of a (visit to the) DOCTOR. The punch line, however, seems at odds with the initial script-based interpretation, which leaves the reader or hearer with a semantic-pragmatic puzzle. According to the SSTH, this incongruity forces a pragmatic reorientation: the reader/hearer concludes that the normal bona fide pragmatic mode of communication, based on Grice’s principle of cooperation, does not hold for this particular piece of discourse. Rather, on the basis of the violation of (some of the) Gricean maxims, the reader/hearer assumes an alternative non-bonafide mode of communication, which is guided by conventions and knowledge of humorous texts (Raskin 1985: 100). Part of this knowledge is the overlap and opposition of semantic scripts as defined in the competence theory of the SSTH. Once the alternative communicative mode has been adopted, the hearer/reader can

466

Geert Brône

search for an alternative script-based interpretation that is compatible with the entire text. In the case of (1), elements like young, pretty and whispered, in combination with the information from the punch line of the joke, allow for the activation of the script of LOVE , and more specifically ADULTERY. The possibility of adopting an alternative script refers to the first aspect of Raskin’s definition quoted above: the text is partly or fully compatible with two different scripts. The reference to ‘different’ scripts pertains to the second aspect of the definition, viz. the relation between the two scripts. According to the SSTH, the two scripts should be opposite within the context of the joke. Oppositeness is defined in terms of a local antonymy: “two linguistic entities whose meanings are opposites only within a particular discourse and solely for the purposes of that discourse” (Raskin 1985: 108). In the scene described in (1), there is an obvious opposition between the scripts DOCTOR and ADULTERY (with their different event structure and finality). The dynamic unfolding of script-based interpretations as language users process jokes is aptly termed script-switching (Raskin 1985) or frame-shifting (Coulson 2000): speakers/hearers switch from an initial salient script interpretation to a secondary construal that is compatible with the punch line and which is in a relationship of (local) opposition to the initial script. The switch is typically elicited by a specific textual element or segment, referred to as the semantic script-switch trigger (or disjunctor, Attardo 1994: 95). In the case of (1) the trigger coincides with the wife’s reply (“come right in”). Script-switch triggers need to be distinguished from so-called connectors (Attardo 1994: 96), which are defined as the elements in the set-up of the text that are compatible with multiple script interpretations. Connectors thus are ambiguous elements that – in retrospect – put the reader/hearer on the wrong interpretational track before reaching the punch line. In (1), most notably the patient’s question (“Is the doctor at home?”) serves as a connector, because it is ambiguous as to what its pragmatic force is. Apart from pragmatically ambiguous text segments, jokes can obviously also revolve around (lexical) semantic ambiguity, as in the case of the polysemous strike in (2) and the idiomatic expression in (3): (2) The first thing which strikes a stranger in New York is a big car. (Raskin 1985: 26) (3) I decided to start saving for a rainy day, so I went to a savings and loan and deposited my umbrella. (Giora 2003: 173) It is generally assumed that the SSTH is the first systematic linguistic attempt at a formal theory with clearly defined and empirically testable hypotheses. As a consequence, virtually all studies on linguistic aspects of verbal humour since 1985 more or less explicitly relate to Raskin’s foundational work. From the perspective of CL, the epistemological basis of the SSTH constitutes a direct link with the contextual semantics advocated in CL. Through its script-theoretic roots the SSTH

Humour and irony in cognitive pragmatics

467

adopts a non-reductionist view that is essential in order to account for complex semantic-pragmatic reinterpretation processes involved in joke processing. Raskin’s basic theoretical claim on script-switching has been picked up in CL, most notably so in the empirical work by Coulson (2.3.2). Vaid et al. (2003) focus on the temporal discourse dynamics of the (re)interpretation process and provide evidence for a sudden shift in the priming of different script-based meanings. The results of these empirical studies are generally in support of the script-switching hypothesis, at least for the circumscribed subphenomenon of one-liner punch line jokes. The selection of the specific genre of a single-joke-carrying text allowed Raskin to formalize his theory, in the hope that its insights would be scalable to the larger phenomenon of verbal humour (cf. Brône, Feyaerts, and Veale 2006: 224–226). As a consequence, the focus of the SSTH is restricted to a single semantic phenomenon while other relevant linguistic parameters, including phonetic patterning, genre knowledge, sociolinguistic variables, etc., are excluded from the analysis. Hence the granularity of an SSTH analysis is necessarily coarse, as it consists of listing lexically activated scripts and how they are combined to build a coherent interpretation. The output is a highly schematic, almost “semic dichotomy” (Vandaele 2002: 223) between abstract scripts, which was often criticized as reducing the complexity of the phenomenon (Dolitsky 1986; Mulkay 1988; Giora 1991; Martin 2007). These criticisms of reductionism led to the development of a revised theory on the epistemological foundations of the SSTH, the so-called General Theory of Verbal Humor (GTVH, Attardo and Raskin 1991; Attardo 1994, 2001a), which is introduced in the following section. 2.2.

General Theory of Verbal Humor (GTVH)

The General Theory of Verbal Humor is generally considered to be an offshoot of the SSTH, developed by Salvatore Attardo and Raskin. This attempt at an extension of the theory has occurred in two major movements. The first revision, presented in Attardo and Raskin (1991), essentially extends the breadth of the analysis by introducing a series of knowledge resources that are relevant next to the central script opposition (2.2.1.). This is primarily a methodological revision, as it proposes a more encompassing linguistic (rather than purely semantic) analysis, while the empirical basis remains the standard punch line joke. In the second revision phase, the SSTH/GTVH is tailored to become a tool for the analysis of other humour types, including longer narrative texts and conversational humour (2.2.2.). 2.2.1.

The first extension: extending the depth of analysis

A first substantial revision of the SSTH involved an extension of the scope of analysis to include a variety of potentially relevant parameters and areas of linguistics. This first incarnation of the General Theory of Verbal Humor (Attardo and

468

Geert Brône

Raskin 1991) adds a number of so-called knowledge resources (KRs) to the already established script opposition. The combination of these resources, which are used when construing a joke, allows for further subclassifications of humorous phenomena. The other KRs are: (i)

(ii)

(iii)

(iv)

(v)

LANGUAGE :

This KR provides all information pertaining to the formulation or verbalization of the joke text. At this level, the GTVH situates wording and positioning of the punch line. It is argued that this KR only plays a secondary role, with the exception of the case of verbal jokes (puns, wordplay, as in I wondered why the baseball was getting bigger. Then it hit me). NARRATIVE STRATEGY : This KR refers to the macrostructure of the joke (Attardo and Raskin 1991: 335) or the question in what type of textual organization the joke is cast. For instance, jokes can take the shape of a narrative, a mini-dialogue, a riddle, etc. (e.g. the well-known knock knock jokes). TARGET : Jokes tend to have a target (the butt of the joke), which can be an individual, a social group or category, a stereotype or even an ideological structure (like marriage, as in Man is incomplete until he is married. Then he is finished). SITUATION : In the GTVH, the knowledge resource SITUATION applies to the specific build-up of the fictional world: “The situation of a joke can be thought of as the ‘props’ of the joke: the objects, participants, instruments, activities, etc.” (Attardo 1994: 225). Situation hence does not refer to the communicative situation in which a joke is produced (Kotthoff 1998; Brock 2004: 76). The latter dimension, and its influence on the semantic construal of a humorous text, has not received systematic scrutiny in the GTVH. LOGICAL MECHANISM : Incongruity-resolution theories of humour assume that a cognitive mechanism is needed to achieve a (partial) resolution of the incongruity. In the GTVH, this cognitive rule that establishes a local logic is labelled the logical mechanism and describes how opposed scripts in a joke are partially connected in a dynamic way (Attardo 1997). Attardo, Hempelmann, and Di Maio (2002) present an attempt at (a) a taxonomy of logical mechanisms (including figure/ground reversal, false analogy, role reversals, etc.) and (b) a formalization of the resolution process using mathematical set and graph theory. Logical mechanisms in this account are viewed as mappings between elements in different sets, based on a relationship of similarity. The type of analysis that is presented using graph theory and mappings bears strong similarities to cognitive linguistic work on conceptual mappings between domains, frames or mental spaces (e.g., Lakoff 1987; Fauconnier 1994, 1997, 2004; Fauconnier and Turner 1998, 2002; Coulson 2000).

The six knowledge resources build the core of the analytical tool of the first GTVH incarnation. Attardo and Raskin (1991: 313–316) propose a hierarchical relationship between the separate KRs, with higher-order KRs determining the possibil-

Humour and irony in cognitive pragmatics

469

ities for lower-level ones. This hypothesis is tested empirically in Ruch, Attardo, and Raskin (1993) using a large-scale rating study. The results partially support the proposed hierarchy of KRs. In sum, the first extension of the SSTH to the GTVH renders a multidimensional linguistic apparatus for the analysis of jokes. 2.2.2.

The second extension: broadening the empirical scope of analysis

The introduction of knowledge resources such as NARRATIVE STRATEGY and SITUATION in the GTVH paved the way for the analysis of text types other than the prototypical punch line joke. In Attardo (1998, 2001a) the GTVH is extended further to account for genre-specific structural features of various text types, including longer narrative texts and conversations. One of the basic observations that fuelled the second extension is that longer texts as well as conversational humour do not adhere to the classical joke structure of a set-up phase followed by a punch line. Longer narrative texts, for instance, derive their humorous potential at least partly from the fact that a specific topic or textual choices recur throughout a text. Aspects of the surface structure may be stored in long-term memory (Attardo 2001a: 58–59, following Kintsch 1998, and Ericsson and Kintsch 1995) and serve as material anchors for textual coherence. For instance, when over the course of a longer text a writer uses a surface form a number of times, it can be retained, especially if it is expressed in a pragmatically marked way (e.g. an alliterative sequence like let love lead the way may be recalled by a reader who reads the following sequence 10 pages further on in the text: don’t let life live you, on the basis of the surface structure analogy). The availability of this type of discourse representation allows for the construction of structural links between various parts of longer text sequences. Attardo (2001a) introduces a series of concepts for the analysis of humorous coherence at the text level, including so-called jab lines and strands. Jab lines are the equivalent of punch lines in texts that do not have the structure of a single-jokecarrying text. Semantically jab lines are fully identical to punch lines, the only difference being that they do not occur in final position and that they may function as part of the narrative plot of longer texts (e.g. in episodes of the British television series Fawlty Towers the sequence of events is typically guided by a series of humorous verbal mishaps by Basil Fawlty). Strands emerge when multiple jab lines in a longer stretch of discourse are structurally or semantically related. The connecting element in a strand can be any of the knowledge resources described above. The ensuing theory of humorous texts consists of an extrapolation of insights taken from the SSTH and the first version of the GTVH. The case studies presented by Attardo for narrative texts (1998, 2001a, 2002) and by Archakis and Tsakona (2005) for conversational data reveal that the actual analysis of humorous instances is highly similar to the earlier work, with the exception that the textual po-

470

Geert Brône

sitioning of humorous lines and their mutual relations in the larger text structure are accounted for. As a consequence, some of the strengths and weaknesses of the SSTH also apply to the GTVH. From a cognitive linguistic perspective, two major concerns remain. A first problem arises when reviewing the role of the KR LANGUAGE and its function in the analytical practice of the GTVH. As a low-level parameter in the hierarchy of knowledge resources, it is usually not treated systematically in GTVH analyses (or only as a symbolic label, cf. Ritchie 2004: 76–77).2 The interaction between language and the other knowledge resources is not developed in the GTVH beyond the traditional script-semantic approach. One parameter that could be developed further in order to attain a more fine-grained linguistic analysis is the LOGICAL MECHANISM . This is the second major concern in the current development of the GTVH. The logical mechanism was introduced to describe the local connection or manipulation of the central scripts in a humorous text, which provides the semantic motivation for a partial resolution of the incongruity. Although this parameter could potentially generate interesting insights into specific semantic constellations that occur in humorous texts, the current taxonomic efforts presented in the GTVH (Attardo, Hempelmann, and Di Maio 2002) only offer vague descriptions like ignoring the obvious, exaggeration, faulty reasoning, etc. It cannot be plausibly argued that these labels have the status of ‘logical’ or ‘cognitive’ mechanisms. Here lies a potentially important contribution of CL with its focus on the procedural aspects of meaning construction in terms of semantic construal operations (cf. Fabiszak, this volume, and Taylor, this volume). The following section presents an overview of research in CL that has addressed the issue of humorous meaning construction and its effects on cognitive processing. 2.3.

Cognitive Linguistics and humour research

Recent years have witnessed a significant interest in the phenomenon of humour in Cognitive Linguistics (see Brône and Feyaerts 2004; Brône, Feyaerts, and Veale 2006 and Brône 2010 for an overview). This development is not surprising, given the epistemological basis of CL. A linguistic framework that claims a focus on the conceptual aspects of language use should be able to naturally cover cases of linguistic expressivity that are motivated through the marked, non-prototypical use of the everyday cognitive mechanisms described in that paradigm. After all, humour is interpreted on the fly in everyday discourse, just like the other cases of “figurative language” CL has given centre stage to, such as metaphor, metonymy and conceptual integration. And just like these other cases, the study of humour can yield interesting insights into some of the specifics of cognitive processing, insights that extend beyond the perspective of linguistic humour theories alone. Fauconnier (1997: 125) summarises this potential of ‘deviant’ phenomena for the general cognitive linguistic enterprise:

Humour and irony in cognitive pragmatics

471

Errors, jokes, literary effects, and atypical expressions use the same cognitive operations as everyday language, but in ways that actually highlight them and can make them more salient. As data, they have a status comparable to laboratory experiments in physics: things that may not be readily observable in ordinary circumstances, which for that reason shed light on ordinary principles.

In accordance with this general idea, the development of a cognitive linguistic line of humour research can be attributed to two factors. On the one hand, a series of studies have been published that critically assess existing humour theories from a cognitive perspective (e.g. Giora 1991; Veale 2004; Brône and Feyaerts 2004; Ritchie 2006; Hamrick 2007; Antonopoulou and Nikiforidou 2009). These studies focus on one or more cognitive linguistic constructs and address the question to what extent these may contribute to the development of humour theories. On the other hand, quite a number of cognitive linguists have resorted to humour as an illustration of specific cognitive linguistic concepts, independent of their relevance to the development of linguistic humour theories.3 In these contributions, humour essentially serves as a test case for the explanation of a semantic phenomenon. What follows is an overview of the major lines in cognitive linguistic humour research. Inevitably, this overview will be dominated by two researchers who have to date developed the most systematic accounts, viz. Rachel Giora (2.3.1.) and Seana Coulson (2.3.2.). Next to a review of their seminal work, the following sections also present a survey of the studies exploring the usefulness of cognitive linguistic constructs for the analysis of verbal humour (2.3.3.). 2.3.1.

Marked informativeness and optimal innovation

Rachel Giora’s work in the field of lexical semantics and psycholinguistics has had a strong influence on CL and linguistic humour research. More specifically, her research programme on information structure and salience as part of the mental lexicon (Giora 1997, 2003; cf. also Giora, this volume) has proven highly relevant for the analysis of different manifestations of creativity. In what follows, two specific hypotheses are discussed: the marked informativeness hypothesis (Giora 1991), and the so-called optimal innovation hypothesis (Giora 2002, 2003; Giora et al. 2004). Both hypotheses are developed as cognitive models with specific predictions on cognitive processing and subjective appreciation, which can be tested using standard techniques from psycholinguistics. 2.3.1.1. The marked informativeness hypothesis In Giora (1991), the typical effect of surprise that is associated with punch line jokes is analyzed in terms of textual information structure. According to the graded informativeness requirement (Giora 1985, 1988), a well-structured text starts with the least informative element and gradually adds more informative el-

472

Geert Brône

ements, with the final element corresponding to the most informative one (Giora 1991: 467). This general principle of information value is operationalized using prototype theory from psychology: elements with a high information value exhibit properties that differ from the prototype of the category. Those properties that distinguish one member from another category member (distinctiveness) enhance its information value. As the most informative members of a category differ most strongly from the category prototype, they generally are considered to be least likely and hence yield effects of surprise on the part of readers or hearers. By linking categorization theory to information structure, Giora offers an account of the marked semantic structure of narrative jokes. In discourse linguistic terms, the more a text segment differs from the previously established discourse representation, the more informative it is. These considerations lead Giora to formulate the marked informativeness requirement for jokes. According to the hypothesis, a joke is well-formed if and only if it obeys the Relevance Requirement and violates the Graded Informativeness Requirement […] in that it ends in a markedly informative message and causes the reader to perform a linear shift: the reader is made to cancel the first unmarked interpretation upon processing the second marked interpretation. (Giora 1991: 470)4

The first part of the requirement specifies that although the marked constituent in the punch line differs significantly from the information in the established discourse representation, it is compatible with the rest of the text (cf. section 2.2.1 above on script overlap in the SSTH). The second part of the requirement stipulates the markedness of the final element in comparison to the prototypical text category. The transition from the initial discourse representation to the forced new one is abrupt (large cognitive distance, Giora 1991: 471). In the case of stock example (1) the initial interpretation of the element whisper is linked to the context of a doctorpatient interaction. In terms of information value and categorization, the bronchial whisper is prototypical (and thus unmarked and with a low degree of informativeness) in the setting of clinical discourse, whereas the interpretation in terms of the low voice of secret lovers, which is foregrounded at the punch line of the text, is informatively marked. According to the marked informativeness requirement, an important distinction between metaphor and wordplay on the one hand, and jokes on the other, resides in the cognitive activation levels of different interpretations. Whereas in jokes the unmarked (or salient) interpretation is suppressed in favour of the marked one, the ambiguous nature of metaphors and wordplay is argued to hinge on the simultaneous activation of multiple meanings. Empirical support of the suppression hypothesis is presented in Giora (2003: 170–175) and indirectly also in Vaid et al. (2003) and Coulson and Kutas (1998, 2001).

Humour and irony in cognitive pragmatics

473

2.3.1.2. The optimal innovation hypothesis In a series of more recent studies, Giora and her colleagues have broadened the perspective developed in the marked informativeness requirement and addressed the question as to what principles guide the processes of creativity and innovation. According to the optimal innovation hypothesis (Giora 2002, 2003; Giora et al. 2004), aesthetic pleasure as an effect of creativity resides in a delicate balance between conventionality on the one hand and innovation on the other. Central to the optimal innovation hypothesis is the concept of salience, developed in great detail by Giora as a full-fledged research programme (Giora 1997, 1999, 2003). According to the graded salience hypothesis, meanings are organized hierarchically in the mental lexicon in such a way that salient meanings are more easily accessible than less salient ones (hence the label graded). Differences in salience are determined by a series of parameters, including conventionality, frequency, familiarity and prototypicality (Giora 2003: 15–22). Although it may seem paradoxical at first glance, the salience hypothesis stipulates that aesthetic innovation essentially relies on salience for its effect. The optimal innovation hypothesis explicitly builds on a long-standing research tradition on expressivity and innovation, and formulates a cognitive hypothesis that situates creativity in the tension between expressivity and referential accuracy. Aesthetic pleasure, it is argued, is not a function of innovation alone, but rather of innovation that allows for the recoverability of the familiar or salient: (1) The Optimal Innovation Hypothesis – Pleasurability is sensitive to optimal innovation. (2) Optimal innovation – A stimulus would be optimally innovative if it involves (a) a novel – less or nonsalient – response to a given stimulus, which differs not only quantitatively but primarily qualitatively from the salient response(s) associated with this stimulus and (b) at the same time, allows for the automatic recoverability of a salient response related to that stimulus so that both responses make sense (Giora et al. 2004: 116) The envisaged balance between innovation and familiarity can be illustrated using the examples in (4) (taken from Giora 2003: 180). (4) a. b. c. d.

Body and soul (salient, fixed expression) Bodies and souls (variant version) Body and sole (optimal innovation) Bobby and Saul (pure innovation)

The fixed expression in (4a) has a conventional salient meaning that is stored in the mental lexicon. Aesthetic creativity emerges as an effect of variation or deautomat-

474

Geert Brône

ization of a salient stimulus (Mukarˇovsky´ 1964; Miall and Kuiken 1994; Langlotz 2006). This variation is principled in the sense that the novel response should differ both qualitatively and quantitatively from the salient meaning associated with the familiar stimulus. For instance, although (4b) differs quantitatively from the entrenched expression, it does not trigger a conceptually different meaning. The example in (4c), however, varies both quantitatively and qualitatively from the salient basis (part 2a of the hypothesis), in that it evokes a completely different meaning (this is a slogan for a shoe manufacturer) but in such a way that the new form allows for the recovery of the salient form-meaning pair on which it builds (part 2b). The notion of recoverability is essential to the optimal innovation hypothesis: “[A]esthetic creativity is, at least to a certain extent, a matter of graded innovativeness” (Giora 2003: 178). That innovation needs to be ‘constrained’ by salience becomes apparent in example (4d). The coordinated nominal phrase has a syntactic and rhythmic structure that is similar to the fixed expression in (4a), but the new phrase does not (clearly) allow for the recoverability of the salient (requirement 2b of the hypothesis). In a series of experiments, Giora and her collaborators empirically tested the optimal innovation hypothesis (cf. the overview in Giora et al. 2004). For instance, a large-scale rating experiment shows that optimally innovative stimuli like (4c) are rated as more pleasurable than the other three variants.5 In order to test the cognitive processing cost of optimal innovation, a series of online experiments was conducted (including self-paced reading), which strongly suggest that optimally innovative stimuli involve the activation of both the novel and salient response. The previously discussed marked informativeness requirement is only one means to achieve optimal innovation: The structure of most jokes is such that it keeps us attending to the salient response until the punch line point where a reversal is enforced allowing for the recognition of the novel. The pleasure derivable from this joke hinges on recognizing the innovative in the salient. (Giora 2002: 15).

In jokes, like in the type of wordplay in (4c), the effect crucially depends on a salience imbalance (Brône and Feyaerts 2003). By focusing on the tension between convention and novelty or the deautomatization of routine thinking, Giora’s work adds a relevant dimension to script-based theories of humour, viz. creativity as a mode of subversiveness (Giora 2003: 182–183). By focusing on the humorous, and more generally the aesthetic, potential of deautomatization patterns at different levels of linguistic organization, Giora’s work reflects a central interest in both pragmatics and Cognitive Linguistics. In pragmatic accounts, a strong plea is made for the incorporation of norm violation as a central aspect of humour (Yamaguchi 1988; Graesser, Long, and Mio 1989; Curcó 1998; Brock 2004; Kotthoff 2006; see Attardo 2005, 2008 for an overview). In comparison to the SSTH and GTVH, Giora gives centre stage to the inherent

Humour and irony in cognitive pragmatics

475

tension between expectation and reality, or between conventionality and innovation. The majority of cognitive linguists engaged in humour research take this tension as a starting point as well. As pointed out above, for CL verbal humour may serve simultaneously as a test case and a show case for the flexibility of some of the constructs developed in that paradigm. In its subversion of linguistic norms or habitual patterns of speech, humour highlights the working of semantic-conceptual mechanisms. Phenomena that at first glance appear to parasitize on the linguistic system may in fact provide a unique perspective on it. 2.3.2.

Humorous phenomena in the space structuring model

A second major contribution to the study of humour from a cognitive linguistic perspective is Seana Coulson’s theoretical and empirical work within the so-called space structuring model. Coulson explicitly situates her work in the framework of CL by combining insights from mental spaces theory (Fauconnier 1994, 1997), conceptual blending theory (Fauconnier and Turner 1998, 2002) and Cognitive Grammar (Langacker 1987, 2008) in developing a cognitive linguistic discourse model. In a series of empirical studies, Coulson illustrates the potential of the space structuring model as an analytical tool for different humorous phenomena, including irony and sarcasm (Coulson 2005a, cf. 3.3.), political cartoons (Coulson 2005b), punch line jokes (Coulson 1996, 2000; Coulson, Urbach, and Kutas 2006; Bergen and Coulson 2006) and wordplay (Coulson and Oakley 2000, 2003; Brône and Coulson 2010). In line with Giora and the optimal innovation hypothesis, one of the main strengths of this line of research is that the hypotheses generated by the model are tested empirically using a variety of experimental techniques. According to the space structuring model, linguistic and non-linguistic elements in the discourse context selectively activate structures from background knowledge in the form of frames, which are distributed in different mental spaces and connected for the purpose of local meaning construction. The role of knowledge in long-term memory in this sense is primarily “one of constraining the topology of the mappings and completing unfilled slots in frames at different levels” (Coulson 2000: 159).6 Coulson, Urbach, and Kutas (2006: 231) list the following basic assumptions of the space structuring model: (i)

the embodiment assumption: the structure of language at least partially reflects bodily constraints on perception and action; (ii) the immediacy assumption: the integration of linguistic and non-linguistic information occurs rapidly, and does not (necessarily) require the prior construction of a propositional representation of sentence meaning; (iii) the elaboration assumption: language comprehension involves activating the cognitive models constructed by the listener.

476

Geert Brône

The first and third assumptions correspond to basic tenets of CL (viz. the embodiment and construal hypotheses) and hence are not exclusive to the space structuring model. The second assumption, the immediacy assumption, highlights the strong situatedness of meaning construction, which is essential to the analysis of creative discourse phenomena. In this account, lexical elements bear on the immediate discourse context, in the sense that the current discourse representation facilitates the integration of elements that are compatible with the representation. At the same time, however, lexical input contributes to the construction of the discourse representation, as they activate frames in long-term memory and construe the conceptual content in a specific way. Coulson draws on a range of creative and humorous phenomena for the exploration of the basic assumptions of the space structuring model. The aim of these studies is not primarily to define a higher-order teleological principle (like Giora’s optimal innovation hypothesis) or a general humour theory (like the SSTH and GTVH). Rather, the studies can be viewed as a reaction against a long-standing praxis in linguistics and cognitive science to build hypotheses on the basis of construed and simplified examples in minimal discourse contexts, rather than looking at the reality of discourse, where language and context interact in sometimes exotic and atypical ways (Coulson 2000: 278). One such ‘exotic’ phenomenon is the punch line joke. In examples like (5a), the reader construes a cognitive model of the referential situation on the basis of the linguistic input. These models direct discourse expectations in one particular direction, which becomes apparent in jokes like (5b). The final nominal phrase forces the construction of a cognitive model that is radically different from the one initially set up. This process is labelled frame-shifting in Coulson (2000) and generally corresponds to Raskin’s scriptswitching. (5) a. When I asked the bartender for something cold and full of rum, he recommended his daiquiri. b. When I asked the bartender for something cold and full of rum, he recommended his wife. (6) He told me he just bought a set of balloon tires and I told him that I didn’t even know he owned a balloon. In contrast to Raskin, Coulson focuses more explicitly on the semantic creativity of the construal mechanisms involved in the frame-shifting process. Using conceptual blending theory as a general framework (Fauconnier and Turner 1998, 2002), Coulson shows that frames as internalized cognitive structures or conventionalized slot-filler structures alone cannot account for the flexibility of meaning construction, specifically in cases that are considered to be marked or atypical (Coulson, Urbach, and Kutas 2006: 232). The space structuring model emphasizes the creative construction and integration of cognitive models, rather than merely their accessing.7 In the specific case of (5b), a local opposition between the scripts of REC-

Humour and irony in cognitive pragmatics

477

and INSULT can be distinguished, but part of the effect can be attributed to the connection of these contrasting frames. By linking the phrase full of rum, which is conventionally associated with drinks, to a person, the phrase acquires a new and local meaning (“alcoholic”) that only appears meaningful in that specific context (i.e. a nonce sense meaning, Clark 1983). In (6) the conventional meaning of the nominal group balloon tires is deconstructed in favour of a novel and marked, but still analyzable meaning (Coulson 2000: 63). According to Coulson, examples like these, which are reminiscent of Giora’s optimal innovation hypothesis (2.3.1.2.), cannot be explained in terms of the activation of an alternative script or frame stored in long-term memory. Rather, they hinge on the creative frame-based combination to yield novel gestalts. Various aspects of the space structuring model have been tested empirically. Coulson and her colleagues have used a battery of experimental techniques to verify the psychological reality of frame-shifting. The complex relation between word meaning and contextual information in punch line jokes produces the hypothesis that frame-shifting requires a higher cognitive processing load than cases that do not force such a reconfiguration. This hypothesis was tested using self-paced reading (Coulson and Kutas 1998), the analysis of event-related brain potentials (ERP, Coulson and Kutas 2001) and eye-tracking (Coulson et al. 2006). The tests using these three experimental paradigms all confirm the basic hypothesis: reading times, brain wave patterns and gaze behaviour of test subjects are all indicative of the processing complexity of frame-shifting in one-line jokes. In addition the ERP experiment shows neural indices of the frame-shifting process after the final word of the joke has been processed, and the eye-movement data reveal that the higher processing cost is related to higher-level processing. The fact that jokes elicit more regressive eye movements upon encountering the punch word (in comparison to straight controls) provides additional evidence for the psychological reality of a reanalysis process like frame-shifting. In general, Coulson’s work on the space structuring model is relevant for cognitive humour research for two reasons. First, its focus on the creative construction and mapping of frames in ongoing discourse provides a more dynamic account of meaning construction than advocated in the SSTH. The hypotheses emerging from the model are tested systematically using a variety of empirical methods from psycholinguistics and neurolinguistics. Second, the specific studies presented in the framework of the space structuring model are generally understood as test cases for the suitability of general cognitive principles, even for the analysis of ‘exotic’ language phenomena like jokes, wordplay and innovative expressions. In this sense, Coulson’s work can be conceived of as a strong plea in favour of a genuinely usage-based CL and cognitive science, where conventional and creative forms of language use are treated on a par.

OMMENDATION

478 2.3.3.

Geert Brône

Topics in cognitive linguistic humour research

In recent years, the systematic research programmes of Coulson and Giora have sparked an interest in creative-humorous phenomena in Cognitive Linguistics. In the steadily growing body of publications in this field, four major topics can be distinguished, viz. metaphor, metonymy, mental spaces theory and construction grammar. This section presents a brief overview of the major contributions in each of these fields. 2.3.3.1. Metaphor Metaphor is beyond any doubt the conceptual mechanism that has been studied most extensively in CL (Lakoff and Johnson 1980; see Kövecses 2002 or Grady 2007 for an overview). In contrast, linguistic humour research has paid astonishingly little attention to the relation between metaphor and humour. This is particularly surprising given the resemblance between the cognitive-linguistic definition of metaphor and the concept of script opposition and overlap in the SSTH/GTVH. Cognitive definitions of both metaphor and humour typically revolve around concepts such as domain mapping, opposition and similarity, but obviously, not all metaphorical construals are humorous, and not all humour involves a degree of metaphoricity. Conceptual metaphor theory defines metaphors as systematic mappings between two domains, one of which is a target domain that is conceptualized by mapping structure from the source domain. The metaphorical cross-domain mapping is motivated by a perceived similarity between the two concepts (Rudzka-Ostyn 1994: 436). This highly schematic definition is reminiscent of script-based humour theories, which assume the (sequential) activation of two opposed scripts that are connected to achieve a local resolution of incongruity (Attardo, Hempelmann, and Di Maio 2002). Although several scholars have referred to the theoretically interesting (and vague) conceptual boundaries between metaphor and humour (Koestler 1964/1989; Hofstadter and Gabora 1989; Giora 1991, 2003), linguistic humour research has largely marginalized this topic by only addressing the potential ambiguity of metaphorical expressions (Attardo 1994; Alexander 1997). Only Pollio (1996) explores the boundary issue somewhat more systematically, using insights from, among others, the Lakovian approach. It is argued that humorous bisociation (to use Koestler’s 1964 term) differs from metaphorical conceptualization in the profiling of domain boundaries. Whereas metaphors essentially fuse the source and target to form a single entity, suppressing the obvious domain boundaries, humorous stimuli (un)intentionally emphasize the dissimilarities between domains. Veale, Feyaerts, and Brône (2006) follow a similar argument, in that they illustrate that in many cases of interactional humour, utterances with an underlying metaphorical structure are used as a cue for a trumping strategy. In (7), this strategy

Humour and irony in cognitive pragmatics

479

consists of a speaker B subverting a speaker A’s utterance by distorting the source/ target structure of the initial metaphorical utterance. (7) A: [consoling tone] B: [angry tone]

Perhaps it is a blessing in disguise, dear. Well, it must be a bloody good disguise then.

By extending the stock metaphorical phrase used by A (a blessing in disguise) with the adjectival phrase bloody good, B performs a figure/ground-reversal on the semantics of the idiomatic expression. The metaphorical extension foregrounds and exploits the initially backgrounded component parts of the expression and their relative contribution to the idiomatic meaning. Veale, Feyaerts, and Brône (2006) and Brône (2008, 2010) argue that figure/ground-reversal is a prominent mechanism of verbal humour, which applies not only to metaphorical conceptualizations but to a range of humorous subphenomena (e.g. the relation between the component parts in a compound like waterpolo in 8). (8) Tim hasn’t played waterpolo since that tragic day when his horse drowned. 2.3.3.2. Metonymy Metonymy, in a cognitive linguistic approach, is generally viewed as a cognitive mechanism enabling the selection of a salient reference point in a frame to refer to a different concept in the same frame or to the frame as a whole (Langacker 1993; Panther and Radden 1999). This conceptual approach to metonymy covers phenomena that were previously treated in pragmatics as inferences or conversational implicatures (Gibbs 1999; Panther and Thornburg 2003). Given the general agreement that humour interpretation involves complex inferential activity, metonymy can be argued to play a substantial role in this process. Barcelona (2003) claims that the inferential work in joke interpretation is facilitated by preexisting metonymic connections in a cognitive frame. Metonymic connections, in his account, are the driving force behind the script switching process (Raskin 1985). Example (9), taken from Barcelona (2003: 93–97), illustrates the inference guiding role of metonymy. (9) Opposition MP (referring to the Prime Minister): “But what can we expect, after all, of a man who wears silk underpants” The Prime Minister (rising calmly): “Oh, I would never have thought the Right Honorable’s wife could be so indiscreet!” The indirect insult by the MP hinges on the stereotypical cultural model that only women and homosexuals wear silk underwear. This metonymic reference point (WEARING SILK UNDERPANTS stands for PERSONALITY TRAIT ) is exploited by the prime minister by linking the MP’s supposed knowledge to the indiscretion of the latter’s wife (‘you know that I wear silk underwear because your wife, with whom I

480

Geert Brône

have a sexual relationship, has confessed this to you’). In this and other cases, metonymic patterns guide the reorientation process that is central to the humorous effect. Brône and Feyaerts (2004) and Feyaerts and Brône (2005) illustrate, using various types of verbal as well as non-verbal humour, that the interpretation process is often complicated through the use of marked metonymic reference-point structures. In the examples in (10), the target concepts UGLINESS (10a) and STUPIDITY (10b) are expressed metonymically through the use of an unconventional effect (10a) or cause (10b). The authors argue that the use of marked metonymic reference-point structures yields a balanced processing difficulty that is comparable to Giora’s optimal innovation (cf. 2.3.1.2.). (10) a. Your father surely shot the stork after he saw you. b. Did you get a blow on the head when you were a child? 2.3.3.3. Mental spaces theory and conceptual blending Mental spaces theory (Fauconnier 1994, 1997, 2004) was introduced to linguistics and cognitive science to coherently deal with a wide range of problematic semantic phenomena that all involve a partitioning of information. Mental spaces, in Fauconnier’s account, are small conceptual structures “that proliferate when we think and talk, allowing a fine-grained partitioning of our discourse and knowledge structures” (Fauconnier 1997: 11). These local mental constructs are structured by frames in long-term memory but are flexible in that they are constructed and modified in the course of ongoing text and thought. Apart from the systematic work of Coulson in the space structuring model, humour research has only very recently touched upon the relevance of mental spaces. Ritchie (2006) uses mental spaces theory to account for jokes that revolve around a misinterpretation by a story character, and proposes an account of viewpoint shifts/ reinterpretations. The starting point is the observation that the level at which the reinterpretation process occurs is not as unambiguous as most humour theories observe. The initial misinterpretation can be situated on the level of either the reader/ hearer (as is claimed by most theories) or a character in the joke. In order to capture this stratification, Ritchie presents an account in terms of viewpoints or belief sets, formalized by means of mental spaces. Ritchie argues that the reinterpretation process does not necessarily have to be directly in the mental space of the audience, but rather, it suffices that it happens in a viewpoint that is accessible to the audience. Similar to Ritchie’s approach, Brône (2008) presents an account of interactional humour in terms of complex constellations of viewpoint mental spaces. Among the phenomena singled out for analysis is a form of humorous one-upmanship referred to as hyper-understanding (cf. example 11).

Humour and irony in cognitive pragmatics

481

(11) Ken: and I’m cheap, – – – Margaret: I’ve always felt that about you, Ken: oh shut up, (– – laughs) fifteen bob a lesson at home, – (Clark 1996: 353) Hyper-understanding revolves around a speaker’s ability to exploit potential weak spots in a previous speaker’s utterance by playfully echoing that utterance while simultaneously reversing the initially intended interpretation. Central to the phenomenon is the differentiation of viewpoints, yielding a layered discourse representation. In (11), the teasing reaction by Margaret simultaneously echoes the previous turn (through the anaphoric element that) and sets up a pretence space with an alternative reading of that turn (the metaphorical rather than the intended literal reading of cheap). Mental spaces theory provides a useful analytical tool to account for the complex interplay of different communicative layers of action that is typical of interactional humour. Mental spaces theory builds the basis for a more recent subbranch of cognitive semantics, viz. conceptual blending theory (Fauconnier and Turner 1998, 2002; Coulson 2000). This framework accounts for hybrid cognitive models that emerge through processes of conceptual integration in online cognition and language use. Conceptual blending theory posits a set of partially compositional processes that guide the integration of conceptual content in a range of cognitive phenomena, including some of the classical mechanisms described in CL (metaphor, metonymy, counterfactual reasoning, etc.). Much like Koestler’s (1964/1989) notion of bisociation, which is generally regarded as an early forerunner of the theory, conceptual blending theory aims to uncover an underlying mental operation in various manifestations of creative or imaginative cognition. Different ‘acts of creation’, from wit and humour to the scientific and artistic genius, involve the integration of input from different sources to yield novel conceptualizations. Conceptual blending theory has been used to account for the phenomenon of frame-shifting (cf. 2.3.2.), irony and sarcasm (cf. 3.3.), and a range of other phenomena in which the local integration of conceptual structure yields a witty, expressive or humorous effect. Fauconnier and Turner (2002), Coulson and Oakley (2003), Feyaerts and Brône (2005), Brône and Feyaerts (2005), Lundmark (2005), Brône and Coulson (2010) provide blending analyses of local ambiguities that revolve around the source-target resonance in metaphorical construal. In an example like (12), the literal basis of a salient metaphorical-idiomatic expression (take the froth off) is reactivated via a metonymically motivated link (in this case a PRODUCER - PRODUCT relation between Carlsberg and beer). (12) Falling sales take the froth off Carlsberg results (Financial Times, 27th February 2003).

482

Geert Brône

The opportunistic recruitment of encyclopaedic structure (Fauconnier and Turner 2002: 282) for the construction of a secondary interpretation has been labelled double grounding (Feyaerts and Brône 2002). The witty ambiguity in double grounding emerges as an effect of one key element being projected onto two input spaces in a complex network of integrated mental spaces. Brône and Coulson (2010) present empirical evidence for the psychological reality of double grounding. The results of a combined reading-time and rating study suggest that double grounded metaphors (e.g. US slowdown punctures Bridgestone’s profits) require a higher processing cost (reflected in longer reading times) and produce higher wittiness ratings than single grounded metaphors (like e.g. US slowdown punctures Cold Stone’s profits). Apart from a range of verbal phenomena, conceptual blending theory has been applied to multimodal phenomena as well, including cartoons (Hofstadter and Gabora 1989; Hünig 2002; Bergen 2004; Marín-Arrese 2003, 2006; Feyaerts and Brône 2002; Brône and Feyaerts 2005; Coulson 2005b), comics (Veale, Feyaerts, and Forceville in press), and multimodal advertisements (Fauconnier and Turner 2002; Lundmark 2005). Multimodal representations are often highly compressed cues that need to be unpacked in multiple knowledge structures in order to be understood (Coulson 2005c). Conceptual blending theory provides an analytical model for the complex relation between a formal blend (a material anchor, Hutchins 2005) and the integrated conceptualization it triggers. 2.3.3.4. Construction grammar The majority of cognitive linguistic humour studies focus on the exploitation of basic construal mechanisms like metaphor, metonymy and conceptual blending. Only a few studies present an account of the grammatical particularities of humour, based on insights from Cognitive Grammar and construction grammar. Construction grammar (Fillmore, Kay, and O’Connor 1988, Fillmore et al. 2003; Goldberg 1995, 2006) posits that the basic units of language are not atomic syntactic structures and their combinatory rules, but rather form-meaning pairings in the sense of Cognitive Grammar (Langacker 1987, 2008). These pairs are referred to as grammatical constructions, which build the basic symbolic units at different levels of linguistic organization. In this view, conventional syntactic structures have a semantic value that is partly independent of the lexical elements that instantiate the construction. As a consequence, construction grammar dismisses the classical distinction between lexicon and syntax in favour of a continuum from lexicon to constructicon (or the repertoire of conventional schematic constructions). Antonopoulou (2002) and Antonopoulou and Nikiforidou (2009) use insights from construction grammar to account for the potential discrepancies between constructional and lexical semantics in verbal humour. The example in (13), taken

Humour and irony in cognitive pragmatics

483

from a story by Raymond Chandler, crucially hinges on the caused motion construction described in great detail in construction grammar (Goldberg 1995). (13) Keith squats forward and fights his mother’s thigh up into the car, while Frank leans forward … (Antonopoulou and Nikiforidou 2009: 298) This constructional pattern ([SUBJ [V OBJ OBL]]) is prototypically instantiated by verbs that have the causing of motion as part of their semantics, like e.g. to throw (Jack threw the napkin off the table), which licenses two thematic roles (agent, patient), corresponding to the subject and object position of the construction. However, the construction can also be instantiated with other verb types, like they laughed the poor guy out of the room, where the valence of the intransitive verb laugh is ‘augmented up’ to that of a transitive verb because the construction in which it is embedded requires a transitive relationship. This phenomenon is referred to as coercion or type-shifting (Michaelis 2005). In (13) the integration of fight, which does not conventionally have a directional complement, into the caused motion construction triggers a humour-relevant incongruity: the construction profiles an external causal force for an action (lifting a thigh) that under normal circumstances does not require such an external force. As a consequence, the constructional semantics triggers the construal of the object of movement (mother) as an indiscrete mass. Apart from schematic constructional patterns, humour frequently draws on more or less lexically fixed idiomatic constructions, referred to as (semi-)substantive constructions. In such cases, which we have briefly discussed in the sections on Giora’s optimal innovation hypothesis (2.3.1.2.) and metaphor (2.3.3.1.), the transparency or analyzability of idiomatic expressions is exploited for expressive purposes. Antonopoulou and Nikiforidou argue that construction grammar provides an adequate model for a uniform analysis of these different manifestations, with the additional advantage that construction grammar can naturally integrate insights from frame semantics (on which it builds): “A significant contribution of CxG [Construction Grammar, GB] to humour is that it fills in a noticeable gap in the GTVH and in any theory which focuses on content rather than form, thereby underrating the role of language” (Antonopoulou and Nikiforidou 2009: 310). A somewhat different perspective on the applicability of construction grammar to the study of verbal humour is presented in Bergen and Binsted (2004, in press). The focus of this study is on the phenomenon of scalar humour that hinges on the so-called XYZ-construction (‘X is so Y that Z’). This schematic construction, which is productive in non-humorous language use (14a), profiles a pragmatic relation between the first and second phrase of the construction: Z illustrates that X has a high degree of Y. This scalar inference inherently evoked by the XYZ-construction can be exploited in multiple ways in humour.

484

Geert Brône

(14) a. The end of the movie was so shocking it physically hurt. b. Yo’ mama’s so old, she was a waitress at the Last Supper. c. Yo mama’s so fat, a picture of her fell off the wall. (Bergen and Binsted 2004) The example in (14b) is hyperbolic because the scene that is evoked in Z falls outside the normal scope of Y. (14c) shows that such a hyperbole can be based on a false presupposition (viz. that representations have a weight that is proportional to the object represented). The study by Bergen and Binsted presents an overview of the different levels at which the conventional pragmatic inference pattern associated with the XYZ-construction can be exploited.

3.

Cognitive approaches to irony

In comparison to the previous section on cognitive(-linguistic) humour research, the overview of cognitive irony research is necessarily brief and fragmentary for two reasons.8 First, the perspective taken in the present chapter is that of cognitive pragmatics and of the more narrowly defined field of Cognitive Linguistics, rather than the broad field of pragmatics and psycholinguistics. In comparison to the burgeoning field of cognitive linguistic humour studies, irony thus far has not played a central role in discourse-oriented CL, with a few notable exceptions (cf. 3.3.). Second, in the last few years, a number of excellent overviews have been published that compare and critically assess the various theories of irony presented in pragmatics and psycholinguistics (e.g. Attardo 2000; Giora 2003; Gibbs and Colston 2007). The existing overviews categorize theories of irony on the basis of different typological parameters, including the processing hypothesis endorsed (one-stage vs. two-stage processing, Giora 2003) and the major pragmatic principle that is assumed to underlie the phenomenon (Attardo 2000). For the purpose of this chapter, we take the latter perspective and focus on cognitive-pragmatic approaches to irony. The brief sketch presented here draws heavily on the excellent theoretical state of the art presented by Attardo (2000, 2001a). In general we can distinguish between two major groups of theories on the basis of the fundamental pragmatic principle they propose for irony, viz. irony as a figure of speech (3.1.) and irony as a form of mention or pretence (3.2.). This rough distinction will serve as a basis for classifying the multitude of (sometimes subtly) different theories of irony developed in linguistics and cognitive science.

Humour and irony in cognitive pragmatics

3.1.

Irony as a figure of speech

3.1.1.

The standard pragmatic model

485

A series of accounts in pragmatics take the common-sense notion of irony as “communicating the opposite of (or at least something different than) what is expressed literally” as a starting point for a definition of irony as a figure of speech. The bestknown proposal undoubtedly is Grice’s treatment of irony as an example of implicature (Grice 1975, 1989). Irony, like other cases of ‘indirect’ speech acts, revolves around a violation of one of the four Gricean maxims, in this case the maxim of quality (“do not say what you believe to be false”) and thus flouts the central cooperative principle. Deliberately flouting maxims does not result in a breakdown of communication, however, as can be shown in example (15). (15) [A and B are walking on an icy pavement. B almost slips.] A: Graceful as ever, I see. B: Oh yes. With his ironic remark “graceful as ever”, speaker A flouts the maxim of quality, but he intends for his co-participant to understand the implicature (viz. that the remark is not intended as a compliment). B’s reaction suggests awareness of this deliberate manipulation for pragmatic purposes. In terms of its cognitive processing hypothesis, the standard pragmatic account can be categorized as a two-stage model. According to Grice’s model, the interpretation of irony – like other types of non-literal language use – presupposes the activation of a literal compositional meaning first. In the case of (15), the addressee B would first process a literal reading of A’s remark, before rejecting it in favour of a reading that is contextually appropriate. The psychological reality of such a twostage model is contested, as there is experimental psycholinguistic evidence in support of both a two-stage process (Giora, Fein, and Schwartz 1998) and a direct access of the intended ironic meaning (Gibbs 1986, 1994; see Giora 2003 for an overview). Whereas Grice restricts his definition of irony to the violation of the maxim of quality, Sperber and Wilson (1981), Kaufer (1981) and Attardo (2000) argue that a violation of any of the maxims can trigger irony. Example (16), for instance, does not hinge on a speaker saying something which he or she believes to be false. Rather, the speaker violates the maxim of relevance by stating an all too obvious fact. Grice’s restricted definition hence seems to cover only a subset of the phenomenon. (16) [Two people caught in a downpour] It seems to be raining. (Sperber and Wilson 1981: 300, cited in Attardo 2000: 799)

486 3.1.2.

Geert Brône

Relevant inappropriateness hypothesis

Partly as a reaction to the above-mentioned issues of scope with the traditional Gricean account of irony, Attardo (2000) proposes an alternative view that hinges on the notions of disruption and appropriateness. This essentially Neo-Gricean approach (cf. Huang, this volume) views irony as a strictly pragmatic phenomenon, the meaning of which is arrived at inferentially on the basis of shared presuppositions and contextual cues. With the introduction of specific pragmatic principles that account for the momentary violation of the Cooperative Principle without losing communicative relevance, Attardo’s proposal has a significantly broader scope than the traditional Gricean account. The starting point for the relevant inappropriateness hypothesis is the basic fact that a hearer will assume that an utterance like the one in (17) both temporarily violates the Cooperative Principle and is relevant in the given context, thus triggering a rejection of the literal interpretation in favour of an ironic reading. (17) [A and B are standing in pouring rain] A: Lovely weather. According to Attardo’s account, this process is guided by a specific principle that helps to explain why communication does not break down despite the manifest violation. He labels this the principle of least disruption: violations of the Cooperative Principle are tolerated as long as they are kept to a minimum, i.e. they are restricted to a single conversational unit (Attardo 2000: 815). This principle serves as a basis for providing a functional explanation of irony within a pragmatic framework. On the assumption that hearers recognize the minimal disruption of the normal Cooperative Principle, Attardo claims that two factors guide the processing of the intended ironic meaning, viz. the maxim of relevance, and the hearer’s assumption that irony involves an antonymic or antiphrastic meaning typically linked with a (negative) evaluative stance (2000: 816; cf. Sperber and Wilson’s relevance theoretic account of irony in section 3.2. below). In other words, in its momentary violation of the Cooperative Principle, an ironic utterance still cooperatively refers to a specific element in the context that is being evaluated. The assumption of a momentary violation of the Cooperative Principle raises the question of the contextual appropriateness or inappropriateness of ironic utterances. For instance, an utterance like the one in (17) is contextually inappropriate when produced in pouring rain because it contains a literal non-truth. An example like (18), which is equally ironic, does not involve the expression of something that is literally false (the farmers may in fact love spring rains) but is inappropriate given the contextual circumstances. (18) Two farmers in a drought-stricken area are talking and farmer A says: “Don’t you just love a nice spring rain?” (Attardo 2000: 816)

Humour and irony in cognitive pragmatics

487

Importantly, however, although both examples violate the conditions for contextual appropriateness, they maintain a high degree of relevance. The discussion of examples like these leads Attardo to a revision of the traditional Gricean definition of irony: An utterance u is ironical if 1. u is contextually inappropriate, 2. u is (at the same time) relevant, 3. u is construed as having been uttered intentionally and with awareness of the contextual inappropriateness by S, and 4. S intends that (part of) his/her audience recognize points 1–3, 5. unless H construes u as being unintentional irony, in which case 3–4 do not apply (Attardo 2000: 817) The key concept of (in)appropriateness is explicitly linked to Clark’s notion of common ground (Clark 1996; cf. Horton, this volume). Attardo proposes an operational definition of appropriateness: “an utterance u is contextually appropriate iff all presuppositions of u are identical to or compatible with all the presuppositions of the context C in which u is uttered […], except for any feature explicitly thematized or denied in u” (Attardo 2000: 818). This definition of appropriateness differs, among others, from Sperber and Wilson’s concept of relevance in that it is truthsensitive: the truth value of the presuppositions underlying an utterance has implications for its appropriateness. To illustrate the difference, Attardo discusses examples such as (19), which are contextually inappropriate, as the presuppositions of the utterance do not match the contextual circumstances of the utterance, but at the same time highly relevant for the exchange. (19) How is this night treating you? (uttered in daylight) By incorporating the notion of appropriateness into the (Neo-)Gricean theory of irony, Attardo considerably extends the scope of the original proposal, as it not only covers cases that hinge on the violation of maxims and but also those that violate Austin’s felicity conditions. As for its take on the cognitive processing of irony, the relevant inappropriateness theory belongs to the category of two-stage models, which hold that the processing of irony involves the rejection of a first (literal) interpretation in favour of the ironical alternative reading. In comparison to most psycholinguistic theories, Attardo’s model does not make explicit predictions concerning the processing order and activation. This is in contrast to Giora’s work on irony within the framework of the graded salience hypothesis (cf. 2.3.1.), which will be introduced in the following section (cf. also Giora, this volume).

488 3.1.3.

Geert Brône

Irony and the graded salience hypothesis

One approach that views irony as a figure of speech, but proposes a clear processing hypothesis that differs from the standard pragmatic account, is to be found in Giora’s work on graded salience (cf. 2.3.1.). As mentioned in the overview of Giora’s work on verbal humour, the graded salience hypothesis states that salient meanings have priority in cognitive processing. Importantly, the concept of salience does not coincide with the traditional notion of literal meaning. Rather, salience is defined as a function of a range of parameters, including conventionality, familiarity and frequency. According to the graded salience hypothesis, novel meanings can only be processed after initial activation of a salient meaning. For the specific case of irony, the graded salience hypothesis posits a two-stage process (in contrast to the direct access view introduced by among others Gibbs 1986, 1994; see Giora 2003: 62–72 for an overview) that revolves around the notion of ‘indirect negation’ (Giora, Fein, and Schwartz 1998; Giora and Fein 1999; Giora 2003). More specifically, irony is argued to involve two distinct meanings, a literal and an implied ironical meaning, with a relationship of indirect or non-explicit negation between both: [I]rony is a form of negation that does not use an explicit negation marker. Often an affirmative (What a lovely day for a picnic said on a stormy day) rather than a negative (What a lousy day for a picnic said on a sunny day) expression is used to implicate that a specific state of affairs is different or far from the state of affairs that is taken for granted, expected or more desirable and that is made explicit by the expression. Such a view assumes that irony comprehension involves activating the salient, often literal meaning automatically. (Giora 2003: 72)

The most important difference between the traditional Gricean account and Giora’s salience-based approach is to be situated in the cognitive processing assumption. Whereas the traditional pragmatic theory postulates that the literal meaning is suppressed and the newly activated ironic reading becomes the only one available, Giora convincingly argues that both senses are in fact retained. A series of experiments (summarized in Giora 2003 and Giora, this volume) support the hypothesis that literal meanings are still available after activation of the non-literal reading and that ironical interpretations are generally less salient than literal ones and are thus – along the lines of the graded salience hypothesis – harder to process. In a series of additional experiments, Giora and her collaborators tested the influence of the familiarity of ironical utterances as well as contextual influences on the processing cost, but discussing the impact of these studies (mainly as arguments against competing models of irony) is well beyond the scope of this chapter (see Giora 2003 and Giora, this volume, for excellent overviews).

Humour and irony in cognitive pragmatics

3.2.

489

Irony as mention or pretence

A second major group of theories of irony situate the core of the phenomenon in its reference to a previous (speech) event in combination with the expression of an attitude towards that event. The best-known representative of this strand is Sperber and Wilson’s echoic mention theory (Sperber and Wilson 1981, 1986/1995; Wilson and Sperber 1992). Rooted in Relevance Theory, the mentioning account views irony as a special case of echoic utterances, which attain their relevance by virtue of the fact that the speaker informs the hearer that he/she has a particular previously encountered utterance in mind and has a particular attitude towards that utterance (e.g. in reported speech). The specific case of irony is argued to involve “echoic interpretive use in which the communicator dissociates herself from the opinion echoed with accompanying ridicule or scorn” (Wilson and Sperber 1992: 75).9 A concise definition is given in Sperber and Wilson (1986/1995: 239): Irony invariably involves the implicit expression of an attitude, and that the relevance of an ironical utterance invariably depends, at least in part, on the information it conveys about the speaker’s attitude to the opinion echoed.

The most important difference between Sperber and Wilson’s approach and the traditional Gricean account is exactly the question of relevance: the reason why speakers may opt for an indirect expression rather than a direct one resides in its potential to be critical. In addition, the echoic mention theory does not posit a twostage processing of irony, but rather states that the context may guide a direct access to the relevant (i.e. ironic) interpretation. In the case of examples such as (20) (Sperber and Wilson 1986/1995: 239, cited in Giora 2003: 64), which both involve an echoic mention, the theory assumes no significant processing difference between a non-ironic (20.b1) and an ironic/sarcastic (20.b2) reading. Sperber and Wilson argue that, given the right contextual input, no shift from an initial to an alternative reading is required. (20) a. He: It’s a lovely day for a picnic. b1. (The sun shines) She (happily): It’s a lovely day for a picnic, indeed. b2. (It starts to rain) She (sarcastically): It’s a lovely day for a picnic, indeed. Although there is some empirical evidence that supports the mention account (see Gibbs 1994 for an overview), the theory has been criticized on several grounds, including the question whether mentioning is a necessary and sufficient condition for irony to occur (Attardo 2000: 805–806) and the fact that the evidence put forward in favour of a direct processing of irony does not in fact exclude a two-stage model (Giora 2003: 64–65). Apart from Sperber and Wilson’s widely discussed account, a number of alternative views have been proposed that all hinge on a similar idea). One of these proposals is Herbert Clark’s view on irony as (joint) pretence (Clark and Gerrig 1984;

490

Geert Brône

Clark 1996). According to this view, irony involves a speaker setting up a pretence layer in discourse, in which the pretend speaker addresses the pretend hearer (who may or may not have a counterpart in the actual communicative situation). Similar to the mention theory of irony, Clark argues that the ironist’s purpose in setting up a pretence layer is to express a critical attitude towards what the pretend speaker is saying. In the case of example (20.b2) above, the ironist performs a staged communicative act in which the assumed speaker addresses a pretend audience that is somewhat naive. The function of this pretence is – both for the ironist and the intended audience that recognizes the irony – to ridicule the pretend speaker and hearer. The idea of a double-layered pretence seems to suggest a cognitive processing model that assumes a simultaneous double activation (both literal and ironic), rather than a linear two-stage model (see however Giora 2003: 67–68 for a more detailed discussion). A final important contribution that needs to be mentioned in the overview of mention theories is Kumon-Nakamura, Glucksberg, and Brown’s (1995) allusional pretence view of irony, an adaptation of Kreuz and Glucksberg’s (1989) earlier reminder theory. As is clear from the chosen label, the theory claims that irony crucially involves an allusion to an expectation or norm that is violated in the specific context. For instance, in the case of (21), the speaker alludes to the social norm that drivers should signal before making a turn in traffic. (21) [A driver is confronted with a car ahead that makes a sudden turn without signalling] Driver: I just love people who signal when turning. As in Clark’s model, the ironist sets up a pretence in complementing the other driver, and in doing so is pragmatically insincere and highly critical of the current situation. In a series of experiments, Kreuz and Glucksberg (1989) test a number of assumptions that are consistent with the reminder theory of irony (see, however, Attardo 2000: 808–809 for a critical discussion). This concludes the overview of the most important accounts of irony in a broad pragmatic and psycholinguistic perspective. In what follows, we briefly discuss the few approaches to irony within the more narrowly defined field of CL, with specific reference to the above-mentioned theories. 3.3.

Cognitive Linguistic approaches to irony and sarcasm

The only strand of research in Cognitive Linguistics that has turned its attention to the phenomenon of irony from a theoretical perspective is mental spaces theory. In two independently developed accounts, Coulson (2005a) and Kihara (2005) propose a mental space approach to irony and sarcasm that is reminiscent of the mention theories (and more specifically Kreuz and Glucksberg’s reminder theory discussed above). The starting point for both accounts is the basic observation that the

Humour and irony in cognitive pragmatics

491

traditional account of irony, which states that irony involves saying the opposite of what is meant, cannot account for examples such as (21) above. Strictly speaking, the driver agrees with the literal meaning of the utterance, but is nevertheless highly ironic or sarcastic (cf. Sperber and Wilson 1998: 284). The ironic effect, according to Coulson and Kihara, can be attributed to the contrast between the expected reaction in the given circumstances (chastise the reckless driver) and the apparent compliment in the actual utterance or “the discrepancy between the contextual scenario in which an utterance occurs and the verbiage of the sarcastic utterance” (Coulson 2005a: 132). Similar to Kreuz and Gluckberg’s and Clark’s account, the mental space approaches assume that an ironist sets up a pretence and essentially points at the violation of norms and expectations. The merit of the mental space approach to irony is that it links the phenomenon to a series of other discourse phenomena that have been shown to revolve around complex constellations of connected mental spaces. In the specific case of irony, Coulson argues that a conceptual integration of structure from two distinct mental spaces takes place, viz. one space corresponding to the actual (communicative) situation and the corresponding verbal behaviour (the so-called expected reaction space) and one counterfactual space that is set up for the purpose of the ironical utterance (counterfactual trigger space). A positive utterance such as the one in (21), uttered after a dangerous manoeuvre, essentially hints at a desired state of affairs (drivers signal before making a turn) and hence triggers a counterfactual scenario. Irony thus essentially involves the integration of an expression from the counterfactual trigger space with contextual information from the expected reaction space. Kihara’s (2005) mental space approach to irony differs only from Coulson’s account in that it does not assume a conceptual integration of structure from the two distinct mental spaces (along the lines of conceptual blending theory). The core of the ironical effect is attributed to a reference to a counterfactual expectation space. If a speaker shouts out Great! after having received bad news, he or she sets up a counterfactual space corresponding to an imaginary scenario in which such a reaction would be appropriate or expected. The view is summarized as follows: Here is a neck-cube-like flipping of the view on irony. Whereas all the major theories on irony so far have dealt with ironical remarks as referring critically to something in reality by means of a kind of lie, I propose that ironical remarks always refer to something in the counterfactual expectation space that is linked by a connector to its ironical counterpart in the base reality space. (Kihara 2005: 519–520)10

For the example in (21), Kihara argues that the ironist construes a counterfactual expectation space, in which an act (i.e. signalling) has a specific effect (i.e. appeals to the speaker). The elements from this mental space have counterparts in the base reality space (act ‘non-signalling’ and effect ‘speaker’s anger’). In comparison to Coulson’s approach the bulk of the inferential work is not situated in a so-called integrated space or blend, but rather the expectation space acquires central relevance.

492

Geert Brône

Apart from this technical difference between the specific mental space constellation assumed by both accounts, they seem to be identical in their general take on the phenomenon and their coverage. A third study that uses mental spaces theory to account for specific aspects of verbal irony is Attardo (2001b). Although situated in a (Neo-)Gricean framework (cf. section 3.1.2. for Attardo’s relevant inappropriateness hypothesis for irony), this study proposes a mental space account of ironical mode adoption. Mode adoption is to be understood as an acceptance on H’s [hearer’s] part of a possible world, as defined by S, which differs from Wr, i.e., the world that S and H mutually know as “reality” in some respect. By acceptance, I mean the agreement to operate, at least momentarily, within that possible world (Attardo 2001b: 10–11).

Mode factive utterances (i.e. utterances allowing mode adoption), such as irony, metaphorical utterances, fiction, etc., trigger the construction of a new mental space (next to the base reality space) so as to allow the hearer to avoid having to reject the utterance as ill-formed. In accordance with Fauconnier’s presupposition float principle (1994: 61), it is argued that the construction of e.g. an ironical mental space avoids a clash between the presuppositions of an utterance and those of the speaker and hearer.

4.

Concluding remarks: perspectives for future research

Despite the steadily growing body of studies exploring a cognitive perspective on humour and irony, there is still some considerable ground to be covered. First of all, many concepts and constructs proposed in linguistic humour research, and not in the least in cognitive accounts, are in need of empirical validation in the form of corpus studies, psycholinguistic and neurolinguistic testing (see Attardo 2001a: 208 for a similar argument). The phenomena discussed in this chapter, including logical mechanisms, script opposition, optimal innovation, double grounding, ironical mode adoption, hyper-understanding and many more, all raise specific cognitive production and processing hypotheses that need to be tested for theories to hold ground. Coulson’s and Giora’s systematic empirical approach to phenomena such as frame-shifting and optimal innovation, in which a variety of empirical methods are combined, may serve as an example for future work in cognitive humour and irony research. Converging evidence from different paradigms and disciplines is a goal to be pursued in a field that is inherently interdisciplinary in nature. Next to the much-needed empirical work, there is also still a considerable theoretical gap between the dominant linguistic humour theories on the one hand and the more narrowly defined field of Cognitive Linguistics on the other. Although, as explained in more detail in 2.1. and 2.2. above, the SSTH and GTVH share a

Humour and irony in cognitive pragmatics

493

number of theoretical premises with CL, they do not have identical objectives and employ a different conceptual apparatus. This at times has led to a degree of mutual suspicion or lack of information (see Brône and Feyaerts 2004 and Attardo 2006 for a discussion of this problem). One of the key challenges for future research is to explore more fully to what extent both traditions are compatible and where precisely the contribution of CL to the field of linguistic humour research can be situated. To avoid that newcomers fall victim to the common problem of ‘reinventing the wheel’, to which humour research has proven to be particularly liable, CL needs to carefully chart the common ground, state of affairs and unexplored territory. An example such as the construction-grammar approach to verbal humour by Antonopoulou and Nikiforidou (2009, cf. supra 2.3.3.4.) serves as an illustration that cognitive linguistic constructs may provide the basis for an analysis of aspects that have been largely ignored in recent linguistic humour research, and that a careful positioning of such a contribution is welcomed by the representatives of the field (Attardo 2009). A final issue that needs to be raised more thoroughly in future research is that of the boundaries or interaction between related phenomena. The classic example is the relation between conceptual metaphor and humour, both of which are argued to involve a mapping between two distinct knowledge structures (cf. 2.3.3.1.). Especially within the framework of CL, which views meaning construction in discourse as an inherently creative construal process, the question needs to be raised as to what separates humour (as cases of ‘full creativity’, Bergen and Binsted 2004) from the ‘normal’ cases of meaning construction involving structure mapping, conceptual integration, coercion, etc. To date, this issue has been largely ignored in the literature, although the output of such a discussion may have important implications for the theoretical foundations of a truly cognitive approach to humour and irony.

Notes 1. See however Cienki (2007) for a more detailed overview of the commonalities and differences between the various concepts that have been introduced to account for structured cognitive content. 2. For instance, the large majority of punch lines and jab lines analyzed in Attardo (2001a) get the value irrelevant for the KR LANGUAGE . In the few cases where the KR is labelled, two distinct categories can be distinguished: in some cases, the parameter is used to label a specific humour type in which language plays a central role (e.g. pun, register humour), whereas in other cases, the value refers to specific linguistic characteristics of the humorous utterance (e.g. idiom, paronymy, co-ordinating conjunction). The KR hence seems to function primarily as a label to isolate instances of verbal humour that centrally revolve around linguistic form, rather than as a real knowledge resource that is tapped into in various types of humour.

494

Geert Brône

3. Obviously, both motivations for exploring humorous phenomena from a cognitive linguistic perspective cannot always be clearly distinguished. Nevertheless, the focus is different: in the first case, a cognitive linguistic niche is worked out in the broader context of linguistic humour research. In the second case, studies only need to account for their relevance within the research paradigm of Cognitive Linguistics. 4. A similar account of joke structure, which takes accessibility of meaning as a starting point, can be found in Weiner & De Palma (1993) and Weiner (1996). In these studies, like in the marked informativeness hypothesis, humorous effects are explained in terms of prototype theory and meaning shifts. 5. In addition and in contrast to traditional accounts of creativity, the optimal innovation hypothesis attaches importance to familiar stimuli. According to the hypothesis, salient stimuli elicit a higher degree of aesthetic pleasure than pure innovations of the type in (4d). This hypothesis is confirmed in the rating tasks presented in Giora et al. (2004). 6. Following Barsalou (1983, 1999) and Glenberg, Kruley and Langston (1994), Coulson has a different perspective on the role of frames as knowledge structures than traditional script-based theories (Schank and Abelson 1977, but also Raskin 1985). Whereas the latter aim to account for semantic phenomena by resorting to conventionalized and experiential gestalts (frames, scripts, schemata), Coulson argues that the static nature of these knowledge structures does not suffice to explain the flexibility of meaning construction in online language use. 7. It should be noted that Attardo (2001a: 6–7) endorses a more flexible view of script semantics than in classical script theories. In more recent incarnations of the GTVH, scripts are described not as static knowledge structures, but rather as dynamic gestalts that can be construed for local purposes. 8. It should be noted that we do not assume a principled difference between irony and sarcasm. Sarcasm is generally seen as an aggressive form of irony, with the distinction between the two being gradual rather than strictly categorical. For an argument in favour of a more strict distinction, see Haiman (1990, 1998). 9. Curcó (1998) extends the notion of indirect echo to the broader category of verbal humour. 10. The base reality space corresponds to the speaker’s and hearer’s conception of reality at a specific point in discourse.

References Alexander, Richard J. 1997 Aspects of Verbal Humour in English. Tübingen: Gunter Narr Verlag. Antonopoulou, Eleni 2002 A cognitive approach to literary humour devices: Translating Raymond Chandler. The Translator 8: 195–220. Antonopoulou, Eleni and Kiki Nikiforidou 2009 Deconstructing verbal humour with contruction grammar. In: Geert Brône and Jeroen Vandaele (eds.), Cognitive Poetics. Goals, Gains and Gaps, 289–314. (Applications of Cognitive Linguistics 10.) Berlin/New York: Mouton de Gruyter.

Humour and irony in cognitive pragmatics

495

Archakis, Argiris and Villy Tsakona 2005 Analyzing conversational data in GTVH terms: A new approach to the issue of identity construction via humor. Humor. International Journal of Humor Research 18: 41–68. Attardo, Salvatore 1994 Linguistic Theories of Humor. (Humor Research 1.) Berlin/New York: Mouton de Gruyter. Attardo, Salvatore 1997 The semantic foundations of cognitive theories of humor. Humor. International Journal of Humor Research 10(4): 395–420. Attardo, Salvatore 1998 The analysis of humorous narratives. Humor. International Journal of Humor Research 11(3): 231–260. Attardo, Salvatore 2000 Irony as relevant inappropriateness. Journal of Pragmatics 32(6): 793–826. Attardo, Salvatore 2001a Humorous Texts: A Semantic and Pragmatic Analysis. (Humor Research 6). Berlin/New York: Mouton de Gruyter. Attardo, Salvatore 2001b Humor and irony in interaction: From mode adoption to failure of detection. In: Luigi Anolli, Rita Ciceri and Giuseppe Riva (eds.), Say not to say: New Perspectives on Miscommunication, 165–185. Amsterdam: IOS Press. Attardo, Salvatore 2002 Cognitive stylistics of humorous texts. In: Elena Semino and Jonathan Culpeper (eds.), Cognitive Stylistics. Language and Cognition in Text Analysis, 231–250. Amsterdam/Philadelphia: John Benjamins. Attardo, Salvatore 2005 Humor. In: Jan-Ola Östman and Jef Verschueren (eds.), Handbook of Pragmatics 2003–2005, 1–25. Amsterdam/Philadelphia: John Benjamins. Attardo, Salvatore 2006 Cognitive linguistics and humor. Humor. International Journal of Humor Research 19(3): 341–362. Attardo, Salvatore 2008 A primer for the linguistics of humor. In: Victor Raskin (ed.), A Primer for Humor Research, 101–156. Berlin/New York: Mouton de Gruyter. Attardo, Salvatore 2009 A commentary on Antonopoulou and Nikiforidou. In: Geert Brône and Jeroen Vandaele (eds.), Cognitive Poetics. Goals, Gains and Gaps, 315–318. (Applications of Cognitive Linguistics 10.) Berlin/New York: Mouton de Gruyter. Attardo, Salvatore and Victor Raskin 1991 Script theory revis(it)ed: joke similarity and joke representational model. Humor. International Journal of Humor Research 4(3): 293–347. Attardo, Salvatore, Christian F. Hempelmann and Sara Di Maio 2002 Script oppositions and logical mechanisms: Modeling incongruities and their resolutions. Humor. International Journal of Humor Research 15(1): 3–46. Barcelona, Antonio 2003 The case for a metonymic basis of pragmatic inferencing: Evidence from jokes and funny anecdotes. In: Klaus-Uwe Panther and Linda Thornburg (eds.), Me-

496

Geert Brône

tonymy and Pragmatic Inferencing, 81–102. Amsterdam/Philadelphia: John Benjamins. Barsalou, Lawrence W. 1983 Ad hoc categories. Memory and Cognition 11(3): 211–227. Barsalou, Lawrence W. 1999 Perceptual symbol systems. Behavioral and Brain Sciences 22: 577–609. Bergen, Benjamin K. 2004 To awaken a sleeping giant. Cognition and culture in September 11 political cartoons. In: Michel Achard and Susanne Kemmer (eds.), Language, Culture, and Mind, 23–35. Stanford: CSLI Publications. Bergen, Benjamin K. and Kim Binsted 2004 The cognitive linguistics of scalar humor. In: Michel Achard, and Susanne Kemmer (eds.), 79–92. Language, Culture and Mind. Stanford: CSLI Publications. Bergen, Benjamin K. and Kim Binsted In press Embodied grammar and humor. In: Geert Brône, Kurt Feyaerts and Tony Veale (eds.), Cognitive Linguistics meets Humor Research. Current Trends and New Developments. Berlin/New York: Mouton de Gruyter. Bergen, Benjamin K. and Seana Coulson 2006 Frame-shifting humor in simulation-based language understanding. IEEE Intelligent Systems 21(2): 59–62. Brock, Alexander 2004 Blackadder, Monty Python und Red Dwarf. Eine linguistische Untersuchung britischer Fernsehkomödien. (Blackadder, Monty Pyhton and Red Dwarf. A linguistic analysis of British television comedy). Tübingen: Stauffenburg Verlag. Brône, Geert 2008 Hyper- and misunderstanding in interactional humor. Journal of Pragmatics 40(12): 2027–2061. Brône, Geert 2010 Bedeutungskonstitution in verbalem Humor. Ein kognitiv-linguistischer und diskurssemantischer Ansatz. (Meaning construction in verbal humour. A cognitive linguistic and discourse semantic approach). Frankfurt a.M./Bern: Peter Lang. Brône, Geert and Seana Coulson 2010 On the cognitive processing of deliberate ambiguity in newspaper headlines: the case of double grounding. Discourse Processes 47: 212–236. Brône, Geert and Kurt Feyaerts 2003 The cognitive linguistics of incongruity resolution: Marked reference-point structures in humor. University of Leuven, Department of Linguistics Preprint 205. 58 pages. Brône, Geert and Kurt Feyaerts 2004 Assessing the SSTH and GTVH: A view from cognitive linguistics. Humor. International Journal of Humor Research 17(4): 361–372. Brône, Geert and Kurt Feyaerts 2005 Headlines and cartoons in the economic press: Double grounding as a discourse supportive strategy. In: Guido Erreygers and Geert Jacobs (eds.), Lan-

Humour and irony in cognitive pragmatics

497

guage, Communication and the Economy, 73–99. Amsterdam/Philadelphia: John Benjamins. Brône, Geert, Kurt Feyaerts and Tony Veale 2006 Introduction: Cognitive linguistic approaches to humor. Humor. International Journal of Humor Research 19(3): 203–228. Cienki, Alan 2007 Frames, Idealized Cognitive Models, and Domains. In: Dirk Geeraerts and Hubert Cuyckens (eds.), Handbook of Cognitive Linguistics, 170–187. Oxford: Oxford University Press. Clark, Herbert H. 1983 Making sense of nonce sense. In: Giovanni B. Flores D’Arcais and Robert Jarvella (eds.), The Process of Language Understanding, 297–332. Chichester: John Wiley and Sons. Clark, Herbert H. 1996 Using Language. Cambridge/New York: Cambridge University Press. Clark, Herbert H. and Richard Gerrig 1984 On the pretense theory of irony. Journal of Experimental Psychology: General 113: 121–126. Coulson, Seana 1996 The Menendez brothers virus: Blended spaces and internet humor. In: Adele Goldberg (ed.), Conceptual Structure, Discourse, and Language, 67–81. Palo Alto, CA: CSLI. Coulson, Seana 2000 Semantic Leaps. Frame-Shifting and Conceptual Blending in Meaning Construction. Cambridge: Cambridge University Press. Coulson, Seana 2005a Sarcasm and the Space Structuring Model. In: Seana Coulson and Barbara Lewandowska-Tomaszczyk (eds.), The Literal and the Nonliteral in Language and Thought, 129–144. Berlin: Peter Lang. Coulson, Seana 2005b Extemporaneous blending: Conceptual integration in humorous discourse from talk radio. Style 39(2), 107–122. Coulson, Seana 2005c What’s so funny? Cognitive semantics and jokes. Cognitive Psychopathology/ Psicopatologia cognitive 2(3): 67–78. Coulson, Seana 2006 Constructing meaning. Metaphor and Symbol 21–4: 245–266. Coulson, Seana and Marta Kutas 1998 Frame-shifting and sentential integration. UCSD Cognitive Science Technical Report 98–03. Coulson, Seana and Marta Kutas 2001 Getting it: Human event-related brain response to jokes in good and poor comprehenders. Neuroscience Letters 316: 71–74. Coulson, Seana and Teenie Matlock 2001 Metaphor and the space structuring model. Metaphor and Symbol 16: 295– 316.

498

Geert Brône

Coulson, Seana and Todd Oakley 2000 Blending Basics. Cognitive Linguistics 13(4): 175–196. Coulson, Seana and Todd Oakley 2003 Metonymy and conceptual blending. In: Klaus-Uwe Panther and Linda Thornburg (eds.), Metonymy and Pragmatic Inferencing, 51–79. Amsterdam/Philadelphia: John Benjamins. Coulson, Seana, Thomas P. Urbach and Marta Kutas 2006 Looking back: Joke comprehension and the space structuring model. Humor. International Journal of Humor Research 19–3: 229–250. Croft, William and D. Alan Cruse 2004 Cognitive Linguistics. Cambridge: Cambridge University Press. Curcó, Carmen 1998 Indirect echoes and verbal humour. In: Villy Rouchota and Andreas Jucker (eds.), Current Issues in Relevance Theory, 305–326. Amsterdam/Philadelphia: John Benjamins. Dolitsky, Marlene 1986 Review of Semantic Mechanisms of Humor (V. Raskin). Journal of Pragmatics 10(2): 269–273. Ericsson, K. Anders and Walter Kintsch 1995 Long-term working memory. Psychological Review 102, 211–245. Evans, Vyvyan 2009 How Words Mean: Lexical Concepts, Cognitive Models, and Meaning Construction. Oxford: Oxford University Press. Fauconnier, Gilles 1994 Mental Spaces: Aspects of Meaning Construction in Natural Language. Cambridge, MA: MIT Press. Fauconnier, Gilles 1997 Mappings in Thought and Language. Cambridge: Cambridge University Press. Fauconnier, Gilles 2004 Pragmatics and Cognitive Linguistics. In: Laurence Horn and Gregory Ward (eds.), The Handbook of Pragmatics, 657–674. Oxford: Blackwell. Fauconnier, Gilles and Mark Turner 1998 Conceptual integration networks. Cognitive Science 22(2): 133–187. Fauconnier, Gilles and Mark Turner 2002 The Way We Think. Conceptual Blending and the Mind’s Hidden Complexities. New York: Basic Books. Feyaerts, Kurt and Geert Brône 2002 Humor through ‘double grounding’: Structural interaction of optimality principles. In: Anders Hougaard and Steffen Nordhal Lund (eds.), The Way We Think (Odense Working Papers in Language and Communication 23), 313–336. Odense: Syddansk Universitets Trykkeri. Feyaerts, Kurt and Geert Brône 2005 Expressivity and metonymic inferencing: stylistic variation in non-literary language use. Style 39(1): 12–36. Fillmore, Charles J. 1982 Frame semantics. In: The Linguistic Society of Korea (ed.), Linguistics in the Morning Calm, 111–137. Seoul: Hanshin.

Humour and irony in cognitive pragmatics

499

Fillmore, Charles J. 1985 Frames and the semantics of understanding. Quaderni di Semantica 6(2): 222–254. Fillmore, Charles J., Paul Kay and Catherine O’Connor 1988 Let alone: Regularity and idiomaticity in grammatical constructions. Language 64: 501–539. Fillmore, Charles J., Paul Kay, Laura A. Michaelis and Ivan A. Sag 2003 Construction Grammar. Stanford, CA: CSLI Publications. Gibbs, Raymond W. Jr. 1986 On the psycholinguistics of sarcasm. Journal of Experimental Psychology: General 115: 3–15. Gibbs, Raymond W. Jr. 1994 The Poetics of Mind: Figurative Thought, Language, and Understanding. Cambridge/New York: Cambridge University Press. Gibbs, Raymond W. Jr. 1999 Speaking and thinking with metonymy. In: Klaus-Uwe Panther and Günter Radden (eds.), Metonymy in Language and Thought, 61–76. Amsterdam/Philadelphia: John Benjamins. Gibbs, Raymond W. Jr. and Herbert L. Colston (eds.) 2007 Irony in Language and Thought. A Cognitive Science Reader. London: Routledge. Giora, Rachel 1985 Towards a theory of coherence. Poetics Today 6: 699–716. Giora, Rachel 1988 On the informativeness requirement. Journal of Pragmatics 12: 547–565. Giora, Rachel 1991 On the cognitive aspects of the joke. Journal of Pragmatics 16(5): 465–486. Giora, Rachel 1997 Understanding figurative and literal language: The graded salience hypothesis. Cognitive Linguistics 7: 183–206. Giora, Rachel 1998 Irony. In: Jef Verschueren, Jan-Ola Östman, Jan Blommaert and Chris Bulcaen (eds.), Handbook of Pragmatics, 1–21. Amsterdam: John Benjamins. Giora, Rachel 1999 On the priority of salient meanings: Studies of literal and figurative language. Journal of Pragmatics 31: 919–929. Giora, Rachel 2002 Optimal innovation and pleasure. In: Oliviero Stock (ed.), The April Fools’ Day Workshop on Computational Humour: Proceedings of the Twentieth Twente Workshop on Language Technology, 11–28. Enschede: UT Service Centrum. Giora, Rachel 2003 On our Mind: Salience, Context and Figurative Language. New York: Oxford University Press. Giora, Rachel and Ofer Fein 1999 Irony comprehension: The graded salience hypothesis. Humor. International Journal of Humor Research 12(4): 425–436.

500

Geert Brône

Giora, Rachel, Ofer Fein and Tamir Schwartz 1998 Irony: Graded salience and indirect negation. Metaphor and Symbol 13: 83–101. Giora, Rachel, Ofer Fein, Ann Kronrod, Idit Elnatan, Noa Shuval and Adi Zur 2004 Weapons of mass distraction: Optimal innovation and pleasure ratings. Metaphor and Symbol 19: 115–141. Glenberg, Arthur M., Peter Kruley and William E. Langston 1994 Analogical processes in comprehension: Simulation of a mental model. In: Morton Anne Gernsbacher (ed.), Handbook of Psycholinguistics, 609–640. San Diego: Academic Press. Glenn, Phillip J. 2003 Laughter in Interaction. Cambridge: Cambridge University Press. Goldberg, Adele E. 1995 Constructions: A Construction Grammar Approach to Argument Structure. Chicago/London: The University of Chicago Press. Goldberg, Adele E. 2006 Constructions at Work: The Nature of Generalization in Language. Oxford: Oxford University Press. Goldstein, Laurence 1990 The linguistic interest of verbal humor. Humor. International Journal of Humor Research 3(1): 37–52. Grady, Joseph 2007 Metaphor. In: Dirk Geeraerts and Hubert Cuyckens (eds.), The Oxford Handbook of Cognitive Linguistics, 188–213. Oxford: Oxford University Press. Graesser, Arthur C., Debra Long and Jeffrey S. Mio 1989 What are the cognitive and conceptual components of humorous text? Poetics 18: 143–163. Grice, H. Paul 1975 Logic and conversation. In: Peter Cole and Jerry L. Morgan (eds.), Syntax and Semantics: Speech Acts, 41–58. New York: Academic Press. Grice, H. Paul 1989 Studies in the Way of Words. Cambridge, MA: Harvard University Press. Haiman, John 1990 Sarcasm as theater. Cognitive Linguistics 1(2): 181–205. Haiman, John 1998 Talk is Cheap: Sarcasm, Alienation, and the Evolution of Language. Oxford: Oxford University Press. Hamrick, Phillip 2007 Notes on some cognitive mechanisms of humor. In: Diana Popa and Salvatore Attardo (eds.), New Approaches to the Linguistics of Humor, 140–150. Galatai: Editura Academica. Hofstadter, Douglas 1995 Fluid Concepts and Creative Analogies. New York: Basic Books. Hofstadter, Douglas and Liane Gabora 1989 Synopsis of the workshop on humor and cognition. Humor. International Journal of Humor Research 2(4): 417–440.

Humour and irony in cognitive pragmatics

501

Hünig, Wolfgang K. 2002 British and German Cartoons as Weapons in World War I. Invectives and Ideology of Political Cartoons, a Cognitive Linguistics Approach. Bern/Frankfurt a.M.: Peter Lang. Hutchins, Edwin 2005 Material anchors for conceptual blends. Journal of Pragmatics 37(10): 1555–1577. Katz, Jerrold J. and Jerry A. Fodor 1963 The structure of a semantic theory. Language 39: 170–210. Kaufer, David S. 1981 Understanding ironic communication. Journal of Pragmatics 5: 495–510. Kihara, Yoshihiko 2005 The mental space structure of verbal irony. Cognitive Linguistics 16: 513–530. Kintsch, Walter 1998 Comprehension: A Paradigm for Cognition. Cambridge: Cambridge University Press. Koestler, Arthur 1964/89 The Act of Creation. London: Hutchison. Kotthoff, Helga 1998 Spaß verstehen: zur Pragmatik von konversationellem Humor. (Understanding verbal play: the pragmatics of conversational humour). Tübingen: Niemeyer. Kotthoff, Helga 2006 Pragmatics of performance and the analysis of conversational humor. Humor. International Journal of Humor Research 19(3): 271–304. Kövecses, Zoltán 2002 Metaphor. A Practical Introduction. Oxford: Oxford University Press. Kreuz, Roger J. and Sam Glucksberg 1989 How to be sarcastic: The reminder theory of verbal irony. Journal of Experimental Psychology: General 118: 347–386. Kumon-Nakamuro, Sachi, Sam Glucksberg and Mary Brown 1995 How about another piece of pie: the allusional pretense theory of discourse irony. Journal of Experimental Psychology: General 124(1): 3–21. Lakoff, George 1987 Women, Fire and Dangerous Things. What Categories Reveal about the Mind. Chicago: Chicago University Press. Lakoff, George and Mark Johnson 1980 Metaphors We Live By. Chicago: Chicago University Press. Langacker, Ronald W. 1987 Foundations of Cognitive Grammar. Volume 1: Theoretical Prerequisites. Stanford, CA: Stanford University Press. Langacker, Ronald W. 1993 Reference-point constructions. Cognitive Linguistics 4(1): 1–38. Langacker, Ronald W. 1999 Grammar and Conceptualization. Berlin/New York: Mouton de Gruyter. Langacker, Ronald W. 2008 Cognitive Grammar. A Basic Introduction. New York: Oxford University Press.

502

Geert Brône

Langlotz, Andreas 2006 Idiomatic Creativity. A Cognitive-Linguistic Model of Idiom-Representation and Idiom-Variation in English. Amsterdam/Philadelphia: John Benjamins. Lundmark, Carita 2005 Metaphor and Creativity in British Magazine Advertising. Luleå: Luleå University of Technology. Marín-Arrese, Juana 2003 Humour in language revisited: The view from cognitive linguistics. In: José L. Camarés, Carmen Escobedo, Daniel García and Natalia Menéndez (eds.), El Humor en Todas las Épocas y Culturas. Oviedo: Universidad de Oviedo. CD. Marín-Arrese, Juana 2006 Cognition and cultural in political cartoons. LAUD Papers in Linguistics 666. Duisburg: Universität Duisburg-Essen. Martin, Rod A. 2007 The Psychology of Humor. An Integrative Approach. New York: Academic Press. Miall, David. S. and Don Kuiken 1994 Foregrounding, defamiliarization, and affect response to literary stories. Poetics 22: 389–407. Michaelis, Laura A. 2005 Entity and event coercion in a symbolic theory of syntax. In: Jan-Ola Östman and Miriam Fried (eds.), Construction Grammar(s): Cognitive and Cross-Language Dimensions, 45–88. Amsterdam/Philadelphia: John Benjamins. Minsky, Marvin 1975 A framework for representing knowledge. In: Patrick H. Winston (ed.), The Psychology of Computer Vision, 211–277. New York: McGraw Hill. Mukarˇovsky´, Jan 1964 Standard language and poetic language. In: Paul L. Garvin (ed.), A Prague School Reader on Esthetics, Literary Structure and Style, 17–30. Washington: Georgetown University Press. Mulkay, Michael 1988 On Humor. Oxford: Basil Blackwell. Panther, Klaus-Uwe and Günter Radden (eds.) 1999 Metonymy in Language and Thought. Amsterdam/ Philadelphia: John Benjamins. Panther, Klaus-Uwe and Linda Thornburg (eds.) 2003 Metonymy and Pragmatic Inferencing (Pragmatics and Beyond 113). Amsterdam/Philadelphia: John Benjamins. Pollio, Howard R. 1996 Boundaries in humor and metaphor. In: Jeffrey S. Mio and Albert N. Katz (eds.), Metaphor: Implications and Applications. Mahwah, NJ: Lawrence Erlbaum, 231–253. Raskin, Victor 1985 Semantic Mechanisms of Humor. Dordrecht/Boston/ Lancaster: D. Reidel. Raskin, Victor (ed.) 2008 The Primer of Humor Research. Berlin/New York: Mouton de Gruyter.

Humour and irony in cognitive pragmatics

503

Ritchie, Graeme 2004 The Linguistic Analysis of Jokes. London/New York: Routledge. Ritchie, Graeme 2006 Reinterpretation and viewpoints. Humor. International Journal of Humor Research 19(3): 251–270. Ruch, Willibald, Salvatore Attardo and Victor Raskin 1993 Toward an empirical verification of the General Theory of Verbal Humor. Humor. International Journal of Humor Research 6(2): 123–136. Rudzka-Ostyn, Brygida 1994 Metaphor, schema, invariance: the case of verbs of answering. In: Keith Carlon, Kristin Davidse and Brygida Rudzka-Ostyn (eds.), Perspectives on English. Studies in Honour of Professor Emma Vorlat, 408–447. Leuven: Peeters. Schank, Roger C. and Robert P. Abelson 1977 Scripts, Plans, Goals and Understanding. New York: Wiley. Sperber, Dan and Deirdre Wilson 1981 Irony and the use-mention distinction. In: Peter Cole (ed.), Radical Pragmatics, 295–318. New York: Academic Press. Sperber, Dan and Deirdre Wilson 1986/95 Relevance: Communication and Cognition. Oxford: Blackwell. [Second edition with Postface 1995] Sperber, Dan and Deirdre Wilson 1998 Irony and relevance: A reply to Drs. Seto, Hamamoto and Yamanashi. In: Robyn Carston and Seiji Uchida (eds.), Relevance Theory: Applications and Implications, 283–293. Amsterdam: John Benjamins. Tannen, Deborah 1989 Talking Voices: Repetition, Dialogue, and Imagery in Conversational Discourse. Cambridge: Cambridge University Press. Vaid, Jyotsna, Rachel Hull, Roberto Heredia, David Gerkens and Francisco Martinez 2003 Getting a joke: The time course of meaning activation in verbal humor. Journal of Pragmatics 35: 1431–1449. Vandaele, Jeroen 2002 Humor mechanisms in film comedy: Incongruity and superiority. Poetics Today 23(2): 221–249. Veale, Tony 2004 Incongruity in humor. Root cause or epiphenomenon? Humor. International Journal of Humor Research 17(4): 419–428. Veale, Tony, Kurt Feyaerts and Geert Brône 2006 The cognitive mechanisms of adversarial humor. Humor. International Journal of Humor Research 19(3): 305–340. Veale, Tony, Kurt Feyaerts and Charles Forceville In press E unis pluribum. The art of creative duality in words, images and sounds. In: Tony Veale and Kurt Feyaerts (eds.), The Agile Mind. Weiner, E. Judith 1996 Why is a riddle not like a metaphor? In: Jan Hulstijn and Anton Nijholt (eds.), Automatic Interpretation and Generation of Verbal Humor. Proceedings of the 12th Twente Workshop on Language Technology (TWLT12), 111–119. Enschede: UT Service Centrum.

504

Geert Brône

Weiner, E. Judith and Paul De Palma 1993 Some pragmatic features of lexical ambiguity and simple riddles. Language and Communication 13(3): 182–193. Wilson, Deirdre and Dan Sperber 1992 On verbal irony. Lingua 87: 53–76. Yamaguchi, Haruhiko 1988 How to pull strings with words. Deceptive violations in the garden-path joke. Journal of Pragmatics 12: 323–337. Ziv, Avner 1984 Personality and Sense of Humor. New York: Springer.

Part V: The emergence of linguistic structures from meaning-in-context

18.

Emergent and usage-based models of grammar Peter Harder

1.

Introduction: From Platonic structures to linguistic practices

Although the cognitive revolution in linguistics is essentially a movement away from a narrow focus on grammatical structure, its task would be incomplete if it did not also include a rethinking of grammar. From the perspective of cognitive pragmatics, this is reflected in the fact that the term pragmatics historically arose as part of a trajectory away from a presupposed core area that was understood as non-pragmatic. In order to move beyond the wastebasket understanding of pragmatics (cf. Bar-Hillel 1971), it is necessary to include that area, of which grammar is the central part, in a pragmatically-based account of language. In a cognitive context, the obvious point of departure is Langacker’s cognitive grammar (cf. Langacker 1987, 2008a). Cognitive grammar is inherently a usagebased model (cf. Langacker 1988): grammar consists of cognitive structures emerging from human experience. The details of the relations between grammar and the pragmatics of usage have, however, only gradually become the subject of focal attention. The development was highlighted in Barlow and Kemmer (2000), which also laid out a list of points on the agenda for exploring this dimension, including the importance of frequency and entrenchment; the role of comprehension and production as integral, rather than peripheral, to the linguistic system; linguistic representations as emergent, rather than stored or fixed entities; the importance of usage data in theory construction and the crucial role of context in the operation of the linguistic system. As in other academic fields, the turn towards usage must be understood as part of a general change in perspective, which occurred during the latter half of the 20th century. Wittgenstein (1953), pointing to language use and forms of life as the basis of meaning, is usually seen as the seminal figure. The basic pattern of understanding since Plato had been that true insight reflects an underlying order that is accessible to the mind (cf. for instance Philosophy and the mirror of nature, Rorty 1980), but not directly reflected in appearances or sensory evidence, and takes the form of stable structures rather than processes occurring in and changing over time. In order to provide a valid account of an object of description, changing and superficial properties have therefore to be set aside and its unchanging essence identified by a process that depends on the mind rather than what meets the eye. Although I think it is fair to say that this way of thinking is losing ground in favour of models that take process rather than structure as the point of departure, it takes time to work out the full implications of abandoning a way of thinking that

508

Peter Harder

has held sway for more than 2000 years. Moreover, the basic pattern manifested itself in different forms depending on the object of investigation and academic traditions in the field. Within linguistics, mainstream Chomskyan generative grammar provided a focal point of departure for a usage-based and emergentist countermovement, in the sense that it aggressively and explicitly stood for everything that a usage-based approach must be against. Generative grammar saw itself as the first truly scientific approach to language; it looked for a source of explanation in the form of an innate structure; it regarded language use as performance, the debased actualization of ideal underlying competence; it aligned itself with a Platonic mode of inquiry by understanding language acquisition as presenting scientists with “Plato’s problem”: just as the slave in the dialogue Meno was shown to be preequipped with knowledge, the child could be shown to be pre-equipped with language (otherwise there would be no way to account for their actual performance). Although generative grammar was not the point of departure for all usage-oriented approaches to grammar (the British tradition has existed since before the rise of generative grammar), its hegemonic status means that most recent discussions of the issue have been conducted against the background of generative assumptions.

2.

The sources of usage-based and emergent approaches to grammar

The development that made language use a gradually more central issue in linguistics involved several contributing sub-developments within macro-linguistics. The rise of the discipline of discourse analysis was one step towards recognizing actual usage as basic (cf. Harris 1952). Although the object of description was units above the sentence, Harris’ initial approach was entirely structural and distribution-based. Because such larger discourse units resisted analysis based on grammar-like distributional patterns, however, it developed into a type of linguistics that rejected langue in favour of parole as the object of description. For the same reason it did not give rise to models of grammar: by defining its territory as above and beyond sentence grammar (see e.g. Shiffrin 2006: 170), it established a division of labour that preserved grammar from the full implications of including discourse into linguistics (other usage-based approaches similarly defined themselves as lying outside grammatical territory, cf. below). As discourse analysis became a going concern, however, it contributed to marking the issue of how exactly grammar entered into the picture of language as usage. In this it was parallel with sociolinguistics, which acquired a pervasive impact on linguistics as a whole with Labov (1966, 1972). Two factors were responsible for the impact: First, Labov’s findings showed that sociolinguistic variation was a property of the individual speaker and not just of the geographical extension of language(s). This was significant because it undermined the Chomskyan assumption of homogeneity as an adequate basis for linguistic description. Secondly, variation

Emergent and usage-based models of grammar

509

in the individual could be described by systematic patterns (and was thus not just a matter of random performance errors). The rigid dichotomy between absolute homogeneity on the one hand and randomness on the other, which had more or less forced scientists to seek refuge in homogeneity in order not to be left with chaos, was no longer tenable. In contrast to discourse analysis, sociolinguistics did not define itself as a marginal or merely hyphenated discipline: social variation was integral to language and had implications for linguistics as such (Labov’s chair was in linguistics rather than sociolinguistics). The sociolinguistic revolution was, however, strictly limited to the expression side of language. The essential methodology was to look for semantically equivalent forms in order to identify the precise role of strictly social factors as distinct from semantic factors. Although it broke down the formal homogeneity of grammar, it thus created a new methodologically-based separation between structural and semantic relations on the one hand and relations between linguistic expressions and the social context on the other. Although the other keyword of the title of this chapter, emergent, shares with usage-based the contrast with an eternal underlying structure that antedates use, its focus is slightly different. Emergence is a feature of complex systems, in which the interaction of elements may yield effects of two different kinds. The standard type of effects are directly predictable from the immediate input to the process, while emergent effects arise when immediate effects conspire with features of the larger system to yield properties that cannot be predicted from the input factors themselves. Emergence is not an occult phenomenon – it enters into basic physical structure: when the gases oxygen and hydrogen combine, they turn into water, which has the property of being liquid in spite of the fact that it is not found in the atoms that are combined (because the relations between the atoms in a water molecule give rise to this behavioural property in the combination of the two). There is no “underlying” wetness anywhere in the elements – it comes about as a side effect of their combination, due to properties of the larger system. The more complex the phenomenon, the greater the scope for emergent effects. The preface to The Emergence of Language (MacWhinney 1999) begins with a series of examples, including supermarket queues, which have the property of being (roughly) equally long, although there exists no built-in mechanism to ensure this – it is the interaction of elements in the larger context that works so as to bring this result about. In all such cases, theories postulating an inherent mechanism would be the legitimate prey of Occam’s razor. The properties of being usagebased and emergent thus combine into a compelling alternative to traditional views that were based on a source of explanation that is at the same time pre-existing, stable, and distinct from actual use. MacWhinney (1999: xi) mentions a number of forms of emergence, depending on the time scale: evolutionary, diachronic, embryological, developmental and online emergence. Although MacWhinney leaves evolutionary emergence out of his

510

Peter Harder

own account, there is a point in using elements of evolutionary theory to understand the process of emergence (as opposed to a “magic bullet” theory of the evolution of language), cf. the discussion below, section 7, of Croft’s evolutionary theory of language change. The parallel is illustrated by Chomsky’s parallel distaste for evolutionary theory and for performance-based accounts. Through the developments enumerated above, the Platonic approach to language gradually acquired more and more competitors that took their point of departure in actual events rather than immutable structures. There is no necessary contradiction involved, however, as long as language as such is recognized as standing apart from all its uses. This remained true for most authors focusing on language use, cf. above. The conflict arises, however, when usage is unleashed on grammar understood as the central fortress of Platonic-flavoured structuralism – and that is what makes the topic of usage-based models of grammar the most controversial topic in the area of usage-based approaches to language.

3.

The British tradition

Above I have referred mostly to the mainstream American tradition. Modern usage-based linguistics, however, includes a separate tradition that goes back to the anthropological inspiration from Malinowski (1923) and includes Firth, Halliday, Sinclair and Stubbs, among others. Malinowski’s (1923) key legacy to this tradition was the dictum that language is a mode of action rather than a countersign of thought. Via his collaboration with J.R. Firth, this approach was imported into British linguistics in the form of a number of inspirational principles that include the insistence on attested data as the object of linguistics, the contextual theory of meaning and the importance of collocations as an intermediate level between context of situation and language (cf. Lyons 1977 II: 609–612; Stubbs 1996: 3, 22–23). It gave rise to a tradition involving discourse analysis (with Sinclair and Coulthard’s [1975] analysis of teaching discourse as a widely influential example) and corpus linguistics in close alliance (cf. Sinclair 1991; Stubbs 1996). Sinclair summed up the position of the collocation-based tradition vis-à-vis Chomskyan linguistics by setting up the “idiom principle” as opposed to the “open choice” principle (Sinclair 1991: 110): rather than always using abstract rules for how to combine words, speakers mostly combined words on the principle of what words go together in actual usage. Wray (2001) is a development of this tradition. When it comes to actual models of grammar, the most influential manifestation of the British tradition is systemic-functional linguistics with Halliday as the central figure, cf. Halliday (1985/1994). There are a number of similarities with usagebased linguistics as associated with cognitive linguistics. First of all, language is meaning-based. The social process that is the primary manifestation of language in

Emergent and usage-based models of grammar

511

Hallidayan thinking is a process of creating, handling and exchanging meanings. In contrast to most of the American tradition, the issue of structuring is not primarily associated with the formal side, but also and essentially with the side of meaning. As in cognitive linguistics, meaning is described as a continuum rather than a division between lexicon and grammar, cf. the term lexicogrammar. Fourthly, the experiential dimension is embedded in a larger, functional conception of what semantics is that also includes an explicitly recognized interactive (“interpersonal”) dimension, as well as a textual dimension. Some of Halliday’s theories about links between sentence structure and interpersonal and textual features have gained very wide recognition both inside and outside linguistics. Thus Halliday (1970), in which standard notions of subjecthood are reanalysed into actor, subject and theme, with roles defined in terms of the different major subcomponents of the semantic system, has remained a classic. Hallidayan achievements, however, are perhaps more impressive in terms of the diverse practices in which they have played a role rather than when viewed from the perspective of models of grammar. As pointed out by Butler (2003: 202–203), there has been relatively little effort to develop a consistent model based on the wide range of implementations. Instead, diverse actual uses of systemicfunctional linguistics are viewed as the main manifestation of systemic-functional linguistics. As illustrated by the sarcastic remark from Hasan (as quoted in Butler 2003: 203), on whether it would “be useful if Halliday put out a flag every time he considered an alternative position”, it has not been a priority to re-evaluate the central claims and concepts in the light of the different findings and developments (cf. also the critique in Bache 2008). Hallidayan theory has perhaps been more influential outside than inside linguistics (for a fuller discussion, cf. Harder 2010).

4.

Emergent grammar

In the American context, the reorientation towards actual usage within linguistics is closely associated with the influence from sociology. Ethnomethodology developed into Conversation Analysis, cf. Sacks, Jefferson, and Schegloff (1974), working out a descriptive practice that applied the basic commitments of ethnomethodology to the fine-grained analysis of language as human activity, viewing it as unfolding, and being structured by participants, as online flow. The name interactional linguistics covers the field that spans conversation analysis and conversation-analytically inspired description of linguistic structures (relative clauses, transitivity, anaphora etc). This area also gave rise to a theoretical position that constitutes a usage-based model of grammar, emergent grammar (cf. Hopper 1987). This model contradicts in every way the ideal, static, structure-oriented, formal position associated with generative grammar. In Hopper’s foundational formu-

512

Peter Harder

lation, emergence as a property of grammar was linked with the anthropological position of Clifford (1986: 19), in which culture is said to be “temporal, emergent and disputed” as well as with the element of deferral that is one aspect of the différance concept of Derrida (discussed with reference to Culler 1982). After introducing these sources of inspiration, Hopper goes on to say “I believe the same is true of grammar, which like speech itself must be viewed as a real-time, social phenomenon, and therefore temporal; its structure is always deferred, always in a process but never arriving” (Hopper 1987: 141). The theoretical implications of this position for the status of grammar will be taken up in section 7 below. The actual descriptive practice had been in existence before the theoretical position was articulated explicitly, in articles that describe familiar categories such as noun, verb and transitivity as they function in actual discourse contexts (cf. Hopper and Thompson 1980, 1984, 1985), and had introduced into linguistics a range of usage facts that had previously been outside the purview of mainstream linguistic description. A central aspect was the existence of a basic level of “probabilistic tendencies within continua” in actual language use, underlying apparently clear-cut structural divisions. For instance in the case of transitive verbs, hard core transitive situations in discourse constituted a cluster of properties (physical force, affected objects, perfective aspect, etc.) that did not characterize all instantiations of transitive verbs and were, in fact, very rare in everyday conversation (Hopper and Thompson 1980, Thompson and Hopper 2001). Similarly, underlying the linguistic categories of noun and verb, there were discourse tasks such as denoting participants as opposed to processes, and these were cross-linguistically more robust and significant than the distributional properties in individual languages. Similarly, Du Bois (1987) demonstrated that ergativity as a syntactic pattern can be understood by seeing it against the background of discourse patterns. A range of such studies made the basis of supposedly pure structural patterns in patterns of actual language use transparent. Including the pathway of emergence into the grammar itself created new descriptive opportunities. For example, Fleischman (1990) describes tense as it functioned in narrative discourse in the framework of emergent grammar (cf. Fleischman 1990: 12), introducing pragmatic patterns as found in narrative texts into grammar (Fleischman 1990: 67). In order to make room for variational patterns associated with narrative tense shifts, she introduces additional (Halliday-inspired) levels, including a textual level, in addition to the traditional referential level that is traditionally associated with grammar. There is thus a continuous pathway from discourse uses to grammatical subcategories, and the lack of clear-cut structural definitions of grammatical categories is consistent with the basic assumption that all categories are emergent rather than clear-cut. Usage-based description in the emergentist sense is inherently associated with a directionality that goes bottom up: the actual tokens are primary, and the formation of types comes later. Also, larger structures are based on more local pat-

Emergent and usage-based models of grammar

513

terns. The point of departure is patterns based on individual items, cf. also Verhagen (2002). This approach has a direct empirical link with language acquisition (cf. MacWhinney 1975; Tomasello 2003: 117). Patterns arise, in usage as well as in acquisition, because each element carries information about what it can combine with, and complexity is built up gradually through the rise of more complex constructions that in turn combine with each other. Perhaps the most central factor, which is therefore also a centrepiece of this approach, is the role of frequency (as announced already by Hopper 1987). Frequency is a test case in that if structure exists entirely in itself, frequency in use would be entirely irrelevant to it. As documented in Bybee and Hopper (2001), however, a mounting body of evidence has shown that frequency has a long range of interesting correlations with structural phenomena. Most profoundly, as suggested in the title Frequency and the emergence of linguistic structure, it has a role as a key criterion for what structural patterns arise in a given language: it is the most frequent constellations of linguistic features that graduate into linguistic structure. Citing Zipf as a key precursor, Bybee and Hopper argue that frequency and repetition serve to epitomize a number of phenomena that show how usage events are channelled into structure. A very basic strand is the question of the origin of the units that enter into a language, where the suggestion is that there is a very direct relation with ongoing usage: the more frequently a given combination of features recurs, the greater its chance of becoming recognized as a unit in its own right. Frequency as such may not be a notion that captivates the imagination. But as empirical work pursued this line of inquiry, the basic issue of frequency per se became differentiated and integrated with the question of what kind of linguistic units were being investigated. This process showed clearly that there was a differentiated set of relations between frequency and structure, for example, that type frequency and token frequency play different but interlocking roles (cf. Smith 2001 on the role of type frequency in the generalization of the have-perfect in English). Another differentiated observation is that highly frequent items are the first to be affected by sound change and the last to be affected by analogical change (Bybee and Hopper 2001: 17). Bybee (2006) gives a general overview of the differentiated roles of frequency as the bridge between usage and grammar. Increasing sophistication of frequency-based theory goes with a methodological development in the form of the vastly increased power of corpus linguistics, due to the compilation of electronic corpora and to rapidly developing methods for electronic data processing (including powerful statistical tools; cf. section 8 below). The enhanced power of corpus methods put a greater range of hypotheses within the purview of empirical investigation. The role of type frequency exemplifies the basic step from raw usage frequency into structurally differentiated frequency: type frequency presupposes the formation of types as something distinct from actual tokens. The interplay between

514

Peter Harder

theoretical power and empirical requirements can be illustrated with cases where the formation of predictive hypotheses forces clarification of issues that cannot be resolved by appeal to intuition alone. One of the phenomena that became empirically validated was a theoretical argument (put forward by Langacker 1987: 42), criticising the radical separation between rules and lists (the rule-list fallacy). Hare, Ford, and Marslen-Wilsen (2001) show that it has to be assumed that past tense forms, also of regular verbs, are stored individually, with frequency information attached to the individual past-tense form – otherwise there would be nothing to which this (attested) frequency effect could apply. Regardless of more specific claims, emergence constitutes a centrepiece in any usage-based theory, because it shows how non-conventional meaning – in other words meaning associated with actual usage events – constitutes the input to a process that has linguistic units at the output end. Frequency-based accounts, however, need to link up with cognitive mechanisms, most directly with entrenchment or routinization. Unless frequent occurrence goes with frequency effects inside speakers’ cognitive systems, it does not matter for the emergence of features of the language. Further, however, the concept of routinization in itself is still not enough: there is also a differentiated issue involving exactly what comes to be routinized. If we assume that a chunk of meaning starts off as something due to the actual situation and subsequently becomes understood as encoded by one of the linguistic expressions, it involves abductive change, cf. Andersen (1973): because the meaning is there and the expression is also there, the meaning is (in deductive terms, fallaciously) realigned as being associated with the expression. This requires a selection process, whereby part of a full situational interpretation becomes part of a linguistic convention. In the case of units below full utterance size, there is also an extra abstraction factor which is necessary in order for the process to produce grammar out of usage: In order for meaning in a whole usage event to become conventionalized as a property of a specific linguistic unit that is a constituent of the full utterance, there must be an act of linking a particular aspect of a whole situational meaning with a specific linguistic unit. Two abductive sub-processes have to take place: from situation to utterance and from utterance to a specific expression (cf. on since below). The gradual entrenchment of new units is a process that applies most obviously to lexical units, including collocations, a point that goes back to Firth but gained new currency following Pawley and Syder (1983) and Sinclair (1991). But it applied equally to frequency effects associated with syntactic patterns (cf. Bybee and Hopper 2001: 14). In mainstream linguistics, such patterns are viewed as separate from abstract general syntactic rules. However, the validity of this distinction was challenged by the rise of a type of model of grammar that relocates the core of grammar from abstract rules to usage-based units: construction grammar.

Emergent and usage-based models of grammar

5.

515

Construction grammar

Construction grammar took its point of departure in complex units intermediate between the word and the sentence levels, which play a previously underestimated role in understanding human languages. As illustrated by the let alone construction (cf. Fillmore, Kay, and O’Connor 1988) the abstract generalizations of sentence grammar cannot account for constructional properties such as the peculiar relation type between the two clauses indicated by the phrase let alone itself. The “construction” format of description acquired a seminal influence as a model of grammar when it was realized that it could be used to combine the “unit” format of description with internal syntactic complexity. From the point of view of cognitive linguistics, construction grammar is part of the basic foundations, since language is defined as a “structured inventory of conventional linguistic units” (cf. Langacker 1987: 73). Construction grammar as an independent theoretical development created a new focus for the understanding of structure because it made Chomskyan abstract generalizations superfluous and captured the facts they were designed to cover with the same constructional format that was used for less general phenomena. As described by Croft and Cruse (2004: 227, 247, 256) the development from a purely generative view of grammar via a generative view that includes constructions to a purely constructional model highlights the list dimension of language. The general format for language description, also for higher-level units, is an inventory of units, called the constructicon (cf. Goldberg 2006, citing Jurafsky 1996): a list of all units of form and meaning with special properties, such that a learner needs to acquire them as individual units. The constructional format reflects a bottom-up orientation in two ways: From a hierarchical point of view, it takes the units (of whatever size) as the basic elements, and like cognitive grammar looks on combinations as the work of the speaker rather than the grammar. It is bottom-up also in the usage-based sense of the term: as discussed above, constructional patterns arise out of usage in a very transparent fashion (cf. Bybee and Hopper 2001), with frequency as a key factor, and the idiomatic dimension highlighted by Sinclair is effortlessly subsumed in this approach: idioms are constructions, too. Fillmore went on to develop his own corpus-based constructional approach, the FrameNet project, building up a computationally implemented account of the relation between verbs and the semantic frames that they go with, cf. Ruppenhofer et al. (2006). If constructions take over the entire territory which used to be populated by abstract general categories, one may ask whether generalizations have been lost from sight. This question has been taken up by Goldberg (2006: 45–46; 2009), who emphasizes the fact that the focus on constructions should not be taken to imply a rejection of general laws in favour of idiosyncratic items. Rather, the advantage of constructions is that they can accommodate anything from the most general to the most local phenomena: “adjective-noun modification” is a construction just as

516

Peter Harder

much as the let alone construction. As we have seen, an emphasis on actual usage instances as the basis of all generalizations does not entail that unique instances is all we have: exemplars are stored, but so are high-level generalizations (Goldberg 2006: 50). Construction grammar has two major advantages: First, it provides a descriptive model that can accommodate intermediate-level patterns that were marginal in both the traditional word-based and the modern sentence-based approach to grammar. Secondly, it highlights the direct pathway from usage to syntax by focusing on pattern extraction – the fundamental mechanism for making structure possible, cf. the discussion of frequency above. Radical construction grammar (Croft 2001) takes the constructional approach to its logical limit. Where emergent grammar as well as earlier versions of construction grammar presupposed the existence of categories such as noun and verb (and looked for usage-based motivations for them), radical construction grammar does away with all basic atomic categories and takes individual, language-specific constructions as the basic terms of the theory. Utterances are viewed as instances of constructions, and all other linguistic categories arise as distributional abstractions based on similarities between utterances, which (like the smaller units they contain) are basically variable and idiosyncratic. The only shared properties are the kinds of things speakers can use language for: encoding things and processes, predicating properties, asking questions and making statements. Whatever shared linguistic features there may be – local, general or even universal – arise out of the practices languages are used to enact.

6.

Grammar, grammaticalization and the lexicon

One issue that is not settled with the rise of construction grammar is the question of what characterizes a usage-based model of grammar as distinct from a usage based approach to the lexicon. Although construction grammar is a form of grammar, the idea is not to separate, but rather integrate grammar and lexicon, to the point where it appears that “the distinction between grammar and the lexicon has disappeared altogether” (cf. Bates and Goodman 1999: 37). This aspect is also involved in the area of grammaticalization, an issue which is central to usage-based models of grammar, in that grammaticalization by definition constitutes the process whereby grammar emerges from usage. But as with synchronic grammar, the emergent perspective on grammaticalization focuses on the gradual process rather than on what emerges at the end. Discussing Meillet’s ([1912] 1958) seminal view, Hopper concludes with approval that Meillet […] backgrounded the project of accounting for “the grammar” of the language, and insisted instead on the study of the history of particular grammatical forms, in effect treating grammar as all and only a collection of such forms. His notion of grammaticaliz-

Emergent and usage-based models of grammar

517

ation […] says nothing about forms “entering the grammar” of the language, but suggests only that one and the same word is now lexical, now grammatical in nature. (Hopper 1991: 33)

Classic grammaticalization theory (cf. Bybee, Perkins, and Pagliuca 1994) focuses on pathways going from lexical to grammatical forms. To this Givón added an extra pathway, whereby today’s syntax is yesterday’s discourse pattern, as when topics become “syntacticized” into grammatical subjects (Givón 1979: 209) and a further stage of the same path is that “today’s morphology is yesterday’s syntax”, cf. Givón (1971). Although the topic of grammaticalization could not be studied without an idea of what grammar is, discussions have relied on an informal consensus based on what has always been understood as grammar, such as inflections, auxiliaries and constructions, rather than developing a new, usage-based theory of what constitutes grammar. The topics of grammaticalization and lexicalization are treated in chapter 19 of this handbook (see Trousdale, this volume), so below I will discuss only the specific issue of the emergence of grammar out of usage, based on a theory presented in Boye and Harder (2012). The key idea is that grammatical elements are ancillary and hence secondary to lexical elements – and so when a new grammatical element comes into being, it simultaneously creates an ancillary (“auxiliary”) relationship and a new, ancillary element. This proposal makes explicit the inherent link between grammaticalization and complexification (cf. Dahl 2004; Heine and Kuteva 2007: 24, assumption 10d) and involves a scenario for the step from pre-grammatical to grammatical language that brings the rise of structural relations into the picture in addition to the constructional focus on new units. The scenario involves two canonical phases, whose general features are compatible both with an evolutionary and an acquisitional pathway. The first step involves the transition from a holophrastic code, where each simplex expression constitutes a whole utterance, to the multi-word stage. At the holophrastic stage a speaker can invoke meanings in isolation from each other, such as mummy and up. Two utterances in succession would each have their separate meanings. The sequence Mummy! Up! could mean, for instance, that mummy summons the mother, while up signals that the child wants to be taken up. This changes with the rise of combinations of the “pivot-grammar” type (cf. Braine 1963), when mummy up begins to mean that ‘mummy’ is the trajector of ‘up’. Systematicity of this kind in the use and understanding of combinations, however embryonic, means that there is now something extra provided by the linguistic utterance, namely a relation between the two meanings. This extra element is ancillary and secondary to the item meanings (~ lexical meanings), and implies that a complex expression is now more than the sum of its parts. When this first step has occurred, the possibility arises of a second step, where separate grammatical expressions come into being – as happens when a lexical expression such as a verb acquires conventionalized “auxiliary” status in relation to

518

Peter Harder

other lexical expressions. A verb can only turn into a grammatical auxiliary if there is already a grammatical relation between that verb and the rest of the clause. This scenario allows linguists to speak of grammar as something that emerges from usage and as intimately linked up with the lexicon, while at the same time having grammar as an identifiable subdomain of the language. Grammar cannot exist except as ancillary to those lexical and semantically primary elements (cf. the term “primary information-bearing units” in Croft 2001), in relation to which grammatical elements are backgrounded and semantically secondary. At the same time grammar adds something identifiable and essential to human languages: although grammatical elements in themselves are secondary, the possibility of having complex wholes that include grammatical and relational elements constitutes an immense enhancement of the communicative potential of linguistic utterances. This scenario makes explicit an affinity between the pragmatics of language use and the emergence of structure that goes against the classic assumptions whereby structure and pragmatics belonged at opposite ends of the linguistic enterprise.

7.

System, usage and variation

Getting rid of the heritage of underlying eternal structure in linguistics is a complex process. A central part of it is to develop a new understanding of what constitutes the coherence of the field: if there is no stable underlying system where we can put all the durable results of linguistic description, how does it all hang together? One of the reasons why this may be an uncomfortable question is the classic identification between invariance and scientific reliability. The engineer’s professional pride is predicated on his confidence that if the calculations are solid, the bridge will be solid. But this identification is only valid for objects of description which inherently have that kind of invariance, prototypically solid physical objects. A language is not a physical object, however, and not only can it be present in several places at once, it also manifests itself in variable ways. Part of the motivation for traditional thinking is simply the need to invent a scientifically respectable object of description for linguistics, in which this erratic behaviour can be spirited away. The task that presents itself to usage-based linguistics is therefore to figure out how to rethink foundations so that they will allow for the kind of coherence that is appropriate for the description of language without slipping back into unwarranted idealizations. Suspicion of unwarranted hypostatization of language as some kind of quasi-object may suggest that it would be best to give up the idea of any stable construct as a successor concept to the language system and settle for unity at the level of processes alone, in harmony with an emergentist approach:

Emergent and usage-based models of grammar

519

I suggest that the real problem here is the underlying assumption that the unity of the field should somehow exist in the unity of the system itself. Liberating ourselves from this structuralist prejudice, we may see that the source of the unity of linguistic structure may well be external to it, that is, in the processes giving rise to all these bits of knowledge. (Verhagen 2002: 434)

However, I believe an account where the unity of linguistic structure is solely external would throw an important baby out with the bathwater. As argued in detail in Harder (2010), I think a pragmatically- oriented cognitive linguistics needs to recognize three different but intimately connected objects of description: the flow of usage (1) is the basic level, but in addition there is (2) language as a property of the speech community (cf. below on niche construction), and finally (3) language as a property of individual speakers (that which qualifies them to be members of the speech community). In order to see the reasons for this differentiation, it is necessary to consider the usage- fundamentalist position where there is assumed to be no distinction at all: there is really only usage. The thoroughgoing usage-fundamentalist position would be one which conflates not only system and usage, but also grammar and lexicon, as well as mind and interaction: the only thing that exists is situated utterance understanding. This position does not allow for a usage-based model of grammar, because such a position would introduce a distinction between the usage that constitutes the base and the grammar that is based on but distinct from it. As far as I know, no one supports this position, but in one respect emergent grammar is close enough to illustrate the problem. This becomes apparent when emergent grammar suggests that grammar is always emergent, cf. the passage from Hopper (1987) above. (For a thoroughgoing fundamentalist, even this would be allowing grammar too much ontological independence, of course). Hopper’s formulation expresses an important insight, namely that change is always happening, and ongoing usage is producing patterns all the time – but in saying that the only grammar that exists is what is now emerging, it suggests that grammatical patterns exist only as (frequency-based) generalizations from actual usage. One concrete instance where this has been argued is the discussion of the status of complement-taking predicates, cf. Thompson (2002), Boye and Harder (2007). Thompson argues based on corpus evidence that clauses such as I think and I believe in a statistically preponderant proportion of cases should not be understood as matrix clauses: I think he’s here most often means ‘he’s here’ with an epistemic reservation, not: “this is what I think: ‘he’s here’”. The assumption is that because they rarely constitute the main point of an utterance, it follows that they are not main clauses, grammatically speaking. The interest in this context is not in the specific issue of complement-taking predicates, but in the direct identification between system and (patterns found in) usage. But if grammar does not consist simply in statistically salient patterns found in actual real language, what is it then?

520

Peter Harder

The answer depends on the role of evolutionary dynamics, and the interplay between the individual and the social level that it involves. Evolutionary dynamics provides a panchronic frame around the immediate actual situation: as argued by Croft (2000), linguistic items constitute “lineages” of expressions that are reproduced across successive temporal states, analogous to animal species rather than individuals. In order to understand why an item-in-use has the properties it currently has, we need to see it as a point on a trajectory in time, influenced by two sets of factors that cannot be collapsed into one: (1) the factors associated with its (re)production in the immediate situation, including speaker intentions and (2) the social factors associated with the diachronic proliferation, continuation, or extinction of the form – the selection processes that determines whether a forms stays in the language or gets lost. The second set means that linguistic items are the products of factors that go beyond the immediate usage situation, including what is going on in the mind of the actual speaker. There is (for example) no specific usage event that constitutes the extinction of a linguistic expression, and if we can only talk about actual attested events we can therefore not talk about a key event type in historical linguistics. Synchronically, we may understand language at the social level as constituting environmental affordances for individual action (cf. Gibson 1966 and MacWhinney 1999: 218 for a discussion in relation to language). The word affordance was coined by Gibson to account for perceptual input from the environment to the senses, but the point it makes generalizes to other ways in which opportunities offered by the environment impinge on the individual. If other people speak French, for instance, this constitutes an environmental affordance for someone who speaks French and would like to interact with them. Viewing language as an affordance endows it with another form of existence than being part of the ongoing interactive flow of usage. The affordance exists (or does not exist, as the case may be), before the individual has said anything. In relation to the evolutionary dynamics briefly outlined above, it constitutes the source of selection pressure: those who speak in the ways that best tap existing affordances are more successful than those who speak in different ways. In evolutionary terms, the set of affordances in a particular place may be regarded as a niche, i.e. as a causal background for adaptation in organisms living in this environment (cf. Deacon 1997 on niche construction in relation to the evolution of language). If grammar is regarded as inherent only in actual usage, there can be no interface between usage and affordances, and no selection pressure. Within the three-way division (introduced above) of language as an object of description, the reconstruction I propose (cf. Harder 2010) is to say that langue, language as a system, exists primarily in the sociocultural niche, rather than in the individual mind or in the flow of usage. One advantage of seeing the language system as a set of sociocultural affordances is that the system thus conceived can accommodate variation of the kind that is sociolinguistically central.

Emergent and usage-based models of grammar

521

Just as the option of using auxiliary inversion to formulate a question is an affordance (in English), variant options including those on the cline between the prestige norm and the community norm constitute affordances. Choosing the wrong variant may subject speakers to selection pressures; an extreme example is described in Judges 12:1–6, according to which the use of the Ephraimite variant of the word shibboleth cost 42,000 people their lives. The kind of stability and coherence that such a concept of language system can have deviates from Platonic ways of thinking in the following respects: (1) it is not eternal, but under constant diachronic change; (2) the categories it contains are no more abstract than recognized in speaker practices (as criterial for adaptive pressures); (3) it is not purely formal, but based on conventions for conveying meaning that are in force in the community. Its status is close in spirit to the Durkheim-inspired views of Saussure on language as a social institution (cf. Saussure [1916] 2005: 33). Where Saussure had little to say about what linked language as a social institution with language in the mind (cf. Geeraerts 2003), the evolutionary framework offers a relation consisting in adaptation: if a community has a certain set of linguistic affordances, members get adaptive advantages out of using them, e.g. by calling a horse cheval or Pferd depending on where they are. The position I suggest thus opts out of the confrontation between synchrony and diachrony and the associated bias in favour of the diachronic perspective that is sometimes found in a usage-based perspective, as when Bybee, Pagliuca, and Perkins (1994: 213) argue for the advantages of viewing linguistic expressions essentially as links in a diachronic chain, while they view synchronic systems as epiphenomenal (1994:1). Instead, it adopts the panchronic approach (cf. Heine, Claudi, and Hünnemeier 1991: 248) that is built into the evolutionary perspective, cf. above. With respect to grammar in the narrow sense, viewing it as a feature of the niche also allows for the inclusion of a variable command of the system between individuals. As pointed out recently by Da˛browska (forthcoming), there is a much greater degree of individual grammatical variation than generally assumed. In the case of command of features like long-distance dependencies and participial clauses, there are significant differences (correlating with education). A niche approach allows us to say that in an English speech community, there are participial clauses out there in language as a social system, whether an individual speaker is fully adapted to them or not. This, again, does not run counter to the central claims of emergentist approaches to grammar. All grammatical affordances arise through a process of emergence, and also a panchronic approach must recognize all stages from fully emerged conventions to nonce formations without conventional force. The only point on which there is a difference is in the force that should be attributed to statistical usage data when it comes to making conclusions about the affordances that are available in the niche. When Thompson and Hopper (2001) present data show-

522

Peter Harder

ing the low transitivity typical of everyday interaction, they take care to emphasize that their results do not generalize to all forms of usage, thus maintaining a distinction between data distilled out of actual usage and grammatical patterns as such – but they come close to conflating the two again when they talk about “the very nature of clause grammar in itself”: “We hope to have shown that the degree of Transitivity of the utterances people use in everyday interaction, and indeed the very nature of clause grammar itself, is tightly related to what they are doing with their talk” (Thompson and Hopper 2001: 54). From the point of view I have argued, it is essential to distinguish between transitivity as a (comparatively rare) feature of what people in fact choose to say, and transitivity as a part of grammar. If we look strictly at “the very nature of clause grammar itself”, once the feature has emerged into grammar, frequency would only affect it if it became so low that transitivity was threatened with extinction. Failing that, transitivity is an affordance that you can choose or not choose, and it remains part of the system even if you do not choose it very frequently in everyday conversation. Finally, systemic-functional linguistics is an interesting case with regard to the question of the relation between system and usage, in that it may be said to be maximalist both when it comes to usage orientation and system orientation. With respect to the usage-based dimension, the overlap in interest between systemicfunctional linguists and American emergent-style functionalism is reflected e.g. in the collaboration between Matthiessen and Thompson (1988) on rhetorical relations between clauses. So much the greater is the contrast on the systemic side. While the whole apparatus of systemic-functional linguistics is open to continuing reinterpretation as dictated by specific texts and usage situations, all such developments are viewed as contributions to one overarching system. Halliday thus envisages a descriptive system in which all meaningful choices can be handled, including those that are now accounted for in terms of “inference, knowledge of the universe, and the like […]. What we have to do is extend and enrich our semantics to the point where we can handle these things as part of the system and process of language” (Halliday 2002: 11). If all choices ultimately go into the system, it becomes pretty big; and it is not clear exactly what is gained by operating with a system that is explicitly designed to include knowledge of the universe and other related matters (cf. Bache 2010 on the way systemic-functional linguistics has used the structural heritage from Hjelmslev). The system of conventional affordances that I have argued for is smaller than one that includes world knowledge and total statistical information about usage. This is convenient in that it limits the task of language learning to a manageable format. On the other hand, the system is much bigger than the maximally abstract and general format that was the ideal aim of classic structural description: It is not enough to know that, e.g., the English demonstratives this, that, these, those can be inserted in a neat grid according to abstract structural distinctions between (hor-

Emergent and usage-based models of grammar

523

izontally) proximal and distal forms, and (vertically) between singular and plural forms. As emphasized in the usage-based discussion, underneath this high level of abstraction there is a large number of more local regularities that cannot be inferred from this bare two-by-two contrast (cf. e.g., Biber et al. 1999: 350). The living usage-based system to which a speaker may find it convenient to adapt consists in all the choices that are conventionally associated with the choice of a demonstrative form, including, e.g., the use in which those is a peg for a postmodifier and does not contrast with a proximal form (those in favour, etc.). The mere contrast would not get a language user very far. As part of this awareness, the system also includes types of statistical information necessary to understand the current state of conventional entrenchment, including ongoing change and variation (cf. Hare, Ford, and Marslen-Wilson 2001).

8.

Empiricism, intuition and data: methodological implications of usage-based grammar

The implications of a usage-based approach involve a number of difficult issues from the foundational to the heuristic level of linguistics. The Platonic aspects of the tradition constitute an idealistic position in relation to which usage-based linguistics represents a counter-position; but much depends on how the contrast between the two is understood. One of the strands is the contrast between the humanist interpretive tradition and empirical science, a confrontation which went on throughout the twentieth century. Basically, empirical science is undoubtedly on the winning side, and usagebased linguistics is essentially an empirically committed development. However, a simple contrast between an empirical and an interpretive approach is not the key to an adequate usage-based alternative to Platonic structure. Such a head-on conflict has surfaced in a number of guises, beginning with behaviourism and Bloomfieldian linguistics, whose position is one form of insistence on basing linguistics on overt manifestations of language rather than Platonic ideas. As this example suggests, the issue is not simply a matter of choosing between intuitions and hard facts or the other, but to find out how intuitive human understanding is to interface with empirical findings. A classic confrontation on this point is Fillmore (1992), responding to a charge from Labov of “woolly minded introspectionism”. (Fillmore had claimed that the deictic phrase yea big could not occur in telephone conversations; hearing the phrase over the telephone immediately afterwards, Labov used it as an example of the merits of empirically based linguistics over intuition). As argued by Fillmore, however, the way to progress is to develop patterns of mutual enrichment: some relevant facts are provided by intuitive understanding (and could not be delivered by empirical findings alone) – other relevant facts can only be established empirically.

524

Peter Harder

The most significant advance on the empirical side is the vastly increased power of corpus linguistics, as discussed above. Information technology has made it possible to amass and process amounts of data that go beyond the wildest dreams of previous generations of linguists, whose principled commitment to the use of actual usage data sometimes went with a realization that for purely practical reasons it was hard to get at the phenomena they looked for. (As an example, investigating the expression of future time in English, Wekker (1976) found that his corpus material included no examples of will + present perfect). In addition to the increase in computational power, sophistication of statistical techniques has provided greatly enhanced powers of teasing out subtler aspects of distribution. One example of developments tapping these new possibilities is the “collostructional” enhancement of construction grammar (cf. Stefanowitsch and Gries 2003), whose aim is to bring out the degrees of collocational affinity between different elements in constructions: what lexemes ‘prefer’ what constructions and vice versa – and with what strength? This approach can follow the process of emergence more closely than previously possible, although it takes a large number of datapoints. Recently it has also been brought to include the variational dimension, showing differences of preference between lexemes and constructions in written and spoken language (Stefanowitsch and Gries 2008). There are still plenty of issues for which we need better data, but the options have expanded greatly. It is clear from the discussion above that there can be no doubt about the need to base linguistics much more unambiguously on usage data rather than purely armchair intuitions, and that the new power of corpus linguistics has a key role to play in living up to this commitment. Yet subtler techniques do not in themselves resolve the issue of what exactly is required to evaluate the significance of purely quantitative findings for linguistic description, as evident from discussions within corpus linguistics, cf. Tognini-Bonelli (2001). At the most radically empiricist and quantitative end we find the building of speech recognition models based on trigrams and Markovian transitional frequencies between elements (Tognini-Bonelli 2001: 82n). For reasons lucidly argued by Chomsky (1957), such an approach cannot hope explain the structure of human languages. The same issue is relevant in relation to another empirically-oriented new descriptive approach, the connectionist approach to simulation (cf. Rumelhart and McClelland 1986). The process whereby a neural net comes to give the desired output without needing recourse to a precise rule, merely through gradual adjustment of a hidden layer of “neurons”, could easily be understood as a computational implementation of behaviourism: there is no need to have any explicit account of what is going on, if an essentially unknowable process gets the right result. Such a process can be assumed to underlie purely behavioural learning such as a snail whose movements are sensitive to the level of moisture around it, and insisting that that is all that is going on in speech would be simplistic. In order to understand what should be understood by a usage-based approach, it is essential to be aware of

Emergent and usage-based models of grammar

525

this reductionist undercurrent in the usage and data-oriented movement. From that extreme, there is a cline of approaches which to various degrees use corpus data to throw light on issues defined in terms of intuition-based research questions, to the extreme intuitive end where corpora just furnish attested examples of preconceived, intuition-based structures. A natural question to ask is where should usagebased linguistics place itself on this cline? That question, however, is wrongly conceived. In order to avoid reductionism, usage-based linguistics must reject not only purely quantitative empiricism, but the whole idea of a cline from a more statistical to a more interpretive methodology. Cognitive Linguistics arose as a recontextualizing movement (cf. Geeraerts 2003), after the structuralist century of removing language from its anchoring in the rest of the world. The first barrier it broke down was between language and the rest of cognition; the usage-based agenda means that the time has come to break down the barrier between language as a mental fact and language as a social fact. The aim is therefore to include the quantitative and variational and the mental dimensions in one integrated field of inquiry, not to play down either intuitive understanding or attested events. Geeraerts and the Leuven group on Quantitative Lexicology and Variational Linguistics (QLVL) have demonstrated what this implies in a series of studies (mainly focusing on lexical issues). A clear example of the interplay between usage variation and interpretive understanding is the study investigating lexical choice among names of cancerous diseases (Grondelaers and Geeraerts 2003). The internal variational dimension they focus on is lexical specificity, i.e. the choice between superordinate terms like illness or disease, and more precise diagnoses such as cancer, or breast cancer, and this is plotted against the external dimension of personal vs. professional contexts. The data show (as predicted) that there is a significantly higher frequency of the less threatening and more general terms illness and disease in personal contexts (as compared with generic and professional contexts), reflecting an avoidance strategy when speaking of people close to you. Clearly, it would be arbitrary and unrevealing to separate this description into two sub-descriptions, one of which dealt with level of semantic specificity and another dealt with quantitative differences between different social contexts. A more grammatically flavoured example is the variational-and-semantic investigation of the use of the “preliminary subject/existential” er (roughly equivalent to “existential there”) in Belgian and Netherlandic Dutch. The issue is the question of what determines its presence or absence in construction types with initial locative (cf. Grondelaers, Speelman, and Geeraerts 2008), as in the illustration case translatable into English as in the ashtray (there) was a hailstone. Unaided introspection about conceptual content does not provide immediately striking ideas about how to account for the choice; but an interpretation that can be investigated with the help of corpus data is that er serves as an “inaccessibility

526

Peter Harder

marker” i.e., its occurrence correlates with the extent to which the object that comes after er is unpredictable. Thus after the concert came a reception would be without er, while after the concert there was an earthquake would tend to have er. In addition, there is regional variation. Belgian speakers are more prone to use er than speakers of Netherlandic Dutch. By logistical regression, it can be shown that one can account for a very large part of the variation (85 %) by careful corpus methods. This kind of study thus exemplifies some of the aspects that are going to be central for the future perspectives of the usage-based approach to grammar. Viewed from the social angle, it transcends the limitation of traditional sociolinguistics to the expression side, as discussed above, bringing variation and conceptualization together. Viewed from the grammatical angle, it studies the system as a feature of a multidimensional social universe (emerged as well as emergent features). Getting this complex act together is going to be a much more difficult task than either focusing on quantitative variation alone or on structural distinctions alone, as pointed out in Arppe et al. (2010). Continuing the discussion raised by Fillmore (1992), they take up the unresolved issues that face corpus linguistics in doing justice to both intuitions and quantitative data of different kinds, including experimental evidence and the nature of explanation in corpus linguistics. But difficult or not, taking this path towards bringing all relevant types of approaches to bear collectively on linguistic issues suggests that usage-based linguistics may be on the way to becoming simply linguistics, paving the way for a future where linguistic descriptions which have no basis in usage – including models of grammar – have been phased out.

References Andersen, Henning 1973 Abductive and deductive change. Language 49: 765–793. Arppe, Antti, Gaëtanelle Gilquin, Dylan Glynn, Martin Hilpert and Arne Zeschel 2010 Cognitive Corpus Linguistics: Five points of debate on current theory and methodology. Pre-publication version September 2009. Corpora 5(1): 1–27. Bache, Carl 2008 English Tense and Aspect in Halliday’s Systemic Functional Grammar. A Critical Appraisal and an Alternative. London: Equinox. Bache, Carl 2010 Hjelmselv’s Glossematics: A Source of Inspiration to Systemic-Functional Linguistics? Journal of Pragmatics 42(9): 2562–2578. Bar-Hillel, Yehoshua 1971 Out of the Pragmatic Wastebasket. Linguistic Inquiry 2(3): 401–407. Barlow, Michael and Suzanne Kemmer (eds.) 2000 Usage Based Models of Language. Stanford: CSLI Publications.

Emergent and usage-based models of grammar

527

Bates, Elizabeth and Judith C. Goodman 1999 On the emergence of grammar from the lexicon. In: Brian MacWhinney (ed.), The Emergence of Language, 29–80. Mahwah, New Jersey: Lawrence Erlbaum. Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad and Edward Finegan 1999 Longman Grammar of Spoken and Written English. Edinburgh: Pearson Education. Boye, Kasper and Peter Harder 2007 Complement-taking predicates: usage and linguistic structure. Studies in Language 31(3): 569–606. Boye, Kasper and Peter Harder 2012 A usage-based theory of grammatical status and grammaticalization. Language 88: 1–44. Braine, Martin D.S. 1963 The ontogeny of English phrase structure: The first phase. Language 39: 1–14. Butler, Christopher S. 2003 Structure and Function. A Guide to Three Major Structural-Functional Theories. (Studies in Language Companion Series 63.) Amsterdam/Philadelphia: John Benjamins. Bybee, Joan, Revere Perkins and William Pagliuca 1994 The Evolution of Grammar. Tense, Aspect, and Modality in the Languages of the World. Chicago: University of Chicago Press. Bybee, Joan and Paul Hopper (eds.) 2001 Frequency and the Emergence of Linguistic Structure. (Typological studies in language 45.) Amsterdam/Philadelphia: John Benjamins. Bybee, Joan 2006 From usage to grammar: The mind’s response to repetition. Language 82: 529–551. Chomsky, Noam 1957 Syntactic Structures. The Hague: Mouton. Clifford, James 1986 Introduction: Partial truths. In: James Clifford and George Marcus (eds.) Writing Culture. The Poetics and Politics of Ethnography, 1–26. Berkeley: University of California Press. Croft, William A. 2000 Explaining Language Change. An Evolutionary Approach. London: Longman. Croft, William A. 2001 Radical Construction Grammar. Syntactic Theory in Typological Perspective. Oxford: University Press. Croft, William A. and D. Alan Cruse 2004 Cognitive Linguistics. Cambridge: Cambridge University Press. Culler, Jonathan 1982 On Deconstruction: Theory and Criticism after Structuralism. Ithaca: Cornell University Press. Dabrowska, Ewa forthcoming Individual differences in native language attainment: A review article. Paper presented at the Conference on Cognitive and Functional Approaches to Language. Tartu, Estonia, May 2008.

528

Peter Harder

Dahl, Östen 2004 The Growth and Maintenance of Linguistic Complexity. Amsterdam/Philadelphia: John Benjamins. Deacon, Terrence 1997 The Symbolic Species. The Co-evolution of Language and the Human Brain. Harmondsworth: Penguin. Du Bois, John 1987 The Discourse Basis of Ergativity. Language 63(4): 805–855. Fillmore, Charles J., Paul Kay and Mary Catherine O’Connor 1988 Regularity and idiomaticity in grammatical constructions: the case of Let alone. Language 64 (3): 501–538. Fillmore, Charles J. 1992 “Corpus linguistics” or “Computer-aided armchair linguistics”. In: Jan Svartvik (ed.), Directions in Corpus Linguistics. Proceedings of Nobel Symposium 82. Stockholm, 4–8 August 1991, 35–60. (Trends in Linguistics 65.) Berlin/New York: Mouton de Gruyter. Fleischman, Suzanne 1990 Tense and Narrativity. Austin: University of Texas Press. Geeraerts, Dirk 2003 Decontextualizing and recontextualizing tendencies in 20th-century linguistics and literary theory. In: Ewald Mengel, Hans-Joerg Schmid, and Michael Steppat (eds.), Anglistentag 2002 Bayreuth, 369–379. Trier: Wissenschaftlicher Verlag. Gibson, James J. 1966 The Senses Considered as Perceptual Systems. Boston: Houghton Mifflin Company. Givón, Talmy 1971 Historical syntax and synchronic morphology. An archeologist’s field trip. Chicago Linguistic Society 7: 394–415. Givón, Talmy 1979 On Understanding Grammar. New York: Academic Press. Goldberg, Adele 2006 Constructions at work. The Nature of Generalization in Language. Oxford: Oxford University Press. Goldberg, Adele 2009 The nature of generalization in language. Cognitive Linguistics 20(1): 93–127. Grondelaers, Stefan and Dirk Geeraerts 2003 Towards a pragmatic model of cognitive onomasiology. In: Hubert Cuyckens, René Dirven and John Taylor (eds.), Cognitive Approaches to Lexical Semantics, 67–92. Berlin: Mouton de Gruyter. Grondelaers, Stefan, Dirk Speelman and Dirk Geeraerts 2008 National Variation in the use of er ‘there’. In: Gitte Kristiansen and René Dirven (eds.), Cognitive Sociolinguistics. Language Variation, Cultural Models, Social Systems, 153–203. Berlin: Mouton de Gruyter. Grondelaers, Stefan and Speelman, Dirk. 2008 Can careful design overcome the methodological drawbacks of grammaticality judgments? Evidence from existential sentences in Belgian Dutch. Paper pres-

Emergent and usage-based models of grammar

529

ented at the Conference on Cognitive and Functional Approaches to Language. Tartu, Estonia, May 2008. Haiman, John (ed.) 1985 Iconicity in Syntax. (Typological studies in language 6.) Amsterdam/Philadelphia: John Benjamins. Halliday, Michael A.K. 1970 Functional diversity in language, as seen from a consideration of modality and mood in English. Foundations of Language 6: 322–361. Halliday, Michael A.K. 1985/1994 An Introduction to Functional Grammar. London: Edward Arnold. Halliday, Michael A. K. 2002 On Grammar. In: The Collected Works of Michael. A. K. Halliday, Volume 1. London: Continuum. Harder, Peter 2010 Meaning in Mind and Society. A Functional Contribution to the Social Turn in Cognitive Linguistics. Berlin/New York: Mouton de Gruyter. Hare, Mary L., Michael Ford and William D. Marslen-Wilson 2001 Ambiguity and Frequency effects in regular verb inflection. In: Joan Bybee and Paul Hopper (eds.), Frequency and the Emergence of Linguistic Structure, 181–200. (Typological studies in language 45.) Amsterdam/Philadelphia: John Benjamins. Harris, Zellig S. 1952 Discourse Analysis. Language 28: 1–30. Heine, Bernd, Ulrike Claudi and Friederike Hünnemeier 1991 Grammaticalization: A Conceptual Framework. Chicago: University of Chicago Press. Heine, Bernd and Tania Kuteva 2007 The Genesis of Grammar. A Reconstruction. Oxford: Oxford University Press. Hopper, Paul 1987 Emergent Grammar. In: Jon Aske, Natasha Beery, Laura Michaelis and Hana Filip (eds.), Proceedings of the Thirteenth Annual Meeting of the Berkeley Linguistics Society, 139–157. Berkeley: Berkeley Linguistics Society. Hopper, Paul 1991 On some principles of grammaticization. In: Elizabeth C. Traugott and Bernd Heine (eds.), Approaches to Grammaticalization, 17–35. Amsterdam/Philadelphia: John Benjamins. Hopper, Paul and Sandra A. Thompson 1980 Transitivity in grammar and discourse. Language 56(2): 251–299. Hopper, Paul and Sandra A. Thompson 1984 The discourse basis for lexical categories in universal grammar. Language 60 (3): 703–752. Hopper, Paul and Sandra A. Thompson 1985 The iconicity of the universal categories “noun” and “verb”. In: John Haiman (ed.), Iconicity in Syntax, 151–183. (Typological studies in language 6.) Amsterdam/Philadelphia: John Benjamins. Kristiansen, Gitte and René Dirven (eds.) 2008 Cognitive Sociolinguistics. Language Variation, Cultural Models, Social Systems. (Cognitive Linguistics Research 39.) Berlin/New York: Mouton de Gruyter.

530

Peter Harder

Labov, William 1966 The Social Stratification of English in New York City. Washington, DC: Center for Applied Linguistics. Labov, William 1972 Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Langacker, Ronald W. 1987 Foundations of Cognitive Grammar, Volume 1: Theoretical Prerequisites. Stanford: Stanford University Press. Langacker, Ronald W. 1988 A usage-based model. In: Brygida Rudzka-Ostyn (ed.), Topics in Cognitive Linguistics, 127–61. (Current Issues in Linguistic Theory 50.) Amsterdam: John Benjamins. Langacker, Ronald W. 2008a Cognitive Grammar. A Basic Introduction. Oxford: Oxford University Press. Langacker, Ronald W. 2008b A Functional Account of the English Auxiliary. MS for book chapter, originally given as a plenary lecture at the conference on Cognitive and Functional Perspectives on Dynamic Tendencies in Languages, May 29–June 1 2008, Tartu, Estonia. Lyons, John 1977 Semantics. Volumes I–II. Cambridge: Cambridge University Press. MacWhinney, Brian 1975 Pragmatic patterns in child syntax. Papers and Reports on Child Language Development 10: 153–165. MacWhinney, Brian (ed.) 1999 The Emergence of Language. Mahwah, NJ: Lawrence Erlbaum. Malinowski, Bronislaw 1923 The problem of meaning in primitive languages. In: Charles K. Ogden and Ivor A. Richards: The Meaning of Meaning, 296–336. London: Routledge and Kegan Paul. Matthiessen, Christian M.I.M. and Sandra A. Thompson 1988 The structure of discourse and “subordination”. In: John Haiman and Sandra A. Thompson (eds.), Clause Combining in Grammar and Discourse, 275–329. (Typological studies in language 18.) Amsterdam/Philadelphia: John Benjamins. Meillet, Antoine 1912 L’évolution des formes grammaticales. Scientia 12, No. 26, 6. Reprinted in Antoine Meillet 1958 Linguistique historique et linguistique générale, 130– 148. Paris: Champion. Pawley, Andrew K. and Frances H. Syder 1983 Two puzzles for linguistic theory: nativelike selection and nativelike fluency. In: Jack C. Richards and Richard W. Schmidt (eds.), Language and Communication, 191–226. London and New York: Longman. Rorty, Richard 1980 Philosophy and the Mirror of Nature. Oxford: Basil Blackwell. Rumelhart, David E., James L. McClelland and The PDP Research Group 1986 Parallel Distributed Processing. Explorations in the Microstructure of Cognition. Cambridge, MA: MIT Press.

Emergent and usage-based models of grammar

531

Ruppenhofer, Josef, Michael Ellsworth, Miriam R.L. Petruck, Christopher R. Johnson and Jan Scheffczyk 2006 FrameNet II: Extended Theory and Practice. http://framenet.icsi.berkeley.edu/ index.php?option=com_wrapper&Itemid=126 (Accessed March 23, 2008). Sacks, Harvey, Emanuel Schegloff and Gail Jefferson 1974 A simplest systematics for the organization of turn-taking for conversation. Language 50(4): 696–735. Saussure, Ferdinand de [1916] 2005 Cours de linguistique générale. Paris: Payot. Shiffrin, Deborah 2006 Discourse. In: Ralph W. Fasold and Jeff Connor-Linton (eds.), An Introduction to Language and Linguistics, 169–203. Cambridge: Cambridge University Press. Sinclair, John and Malcolm Coulthard 1975 Towards an Analysis of Discourse: the English used by Teachers and Pupils. Oxford: Oxford University Press. Sinclair, John 1991 Corpus, Concordance, Collocation. Oxford: Oxford University Press. Smith, K. Aaron 2001 The role of frequency in the specialization of the English Anterior. In: Joan Bybee and Paul Hopper (eds.) Frequency and the Emergence of Linguistic Structure, 361–82. Amsterdam: Benjamins. Stefanowitsch, Anatol and Stefan Th. Gries 2003 Collostructions: investigating the interaction between words and constructions. International Journal of Corpus Linguistics 8(2): 209–243. Stefanowitsch, Anatol and Stefan Th. Gries 2008 Channel and constructional meaning: A collostructional case study. In: Gitte Kristiansen and René Dirven (eds.), Cognitive Sociolinguistics. Language Variation, Cultural Models, Social Systems, 129–152. (Cognitive Linguistics Research 39.) Berlin/New York: Mouton de Gruyter. Stubbs, Michael 1996 Text and Corpus Analysis. Oxford: Blackwell. Thompson, Sandra 2002 “Object complements” and conversation. Towards a realistic account. Studies in Language 26(1): 125–164. Thompson, Sandra and Paul Hopper 2001 Transitivity, clause structure, and argument structure: Evidence from conversation. In: Joan Bybee and Paul Hopper (eds.), Frequency and the Emergence of Linguistic Structure, 91–120. (Typological studies in language 45.) Amsterdam/Philadelphia: John Benjamins. Tognini-Bonelli, Elena 2001 Corpus Linguistics at Work. Amsterdam/Philadelphia: John Benjamins. Tomasello, Michael 2003 Constructing a Language. A Usage-Based Theory of Language Acquisition. Cambridge, M.A.: Harvard University Press. Traugott, Elisabeth C. and Bernd Heine (eds.) 1991 Approaches to Grammaticalization. Amsterdam/Philadelphia: John Benjamins.

532

Peter Harder

Verhagen, Arie 2002 From parts to wholes and back again. Cognitive Linguistics 13: 403–439. Wekker, Herman C. 1976 The Expression of Future Time in Contemporary British English. Amsterdam: North Holland. Wittgenstein, Ludwig 1953 Philosophical Investigations. Oxford: Basil Blackwell. Wray, Alison 2001 Formulaic Language and the Lexicon. Cambridge: Cambridge University Press.

19.

Grammaticalization, lexicalization and constructionalization from a cognitive-pragmatic perspective Graeme Trousdale

1.

Introduction

In this chapter, the role of cognitive pragmatics in aspects of language change is considered.1 Using the definition of cognitive pragmatics (as elsewhere in this handbook) as ‘the construal of meaning-in-context’, the chapter explores some of the ways in which morphosyntax and semantics often co-evolve (Bybee, Perkins and Pagliuca 1994). The co-evolution of both meaning and form is a feature of grammaticalization studies; in this chapter, drawing on some principles of cognitive (particularly constructional) approaches to language, I consider some of the ways in which an understanding of cognitive processes (such as construal and invited inferencing) enable us to account for particular aspects of language change, including, but not restricted to, grammaticalization. Particularly, the focus in later parts of this chapter is on constructional change (i.e. changes affecting conventionalized pairings of form and meaning), including the creation of new constructions. Since construal is at the heart of cognitive pragmatics, I will briefly examine how construal has been understood and employed in synchronic analyses of language; this leads to a consideration of how changes in construal are related to the gradual emergence of new constructions over time. Construal involves a combination of factors, such as viewpoint, profiling and scanning (Langacker 1987), all of which reflect different aspects of a language user’s conceptualization, and the manifestation of that mental process in linguistic form: the notion of construal is concerned with how language users depict some aspect of a mentally represented world. This depiction holds for representations of things and processes (whether states or events) and for the grammatical constructions used to associate events with processes: in both cases, “the mind is an active participant in the creation of semantic structure, and conceptualizes or construes the experiences of the speaker in the world in certain ways” (Clausner and Croft 1999: 2–3). It is this conceptualization of experience which is at the heart of usage-based approaches to language change (Bybee 2010). Furthermore, construal has typically been associated with synchronic analysis; part of the objective of this chapter is to consider some of the ways in which language change may be (at least in part) understood as correlating with changes in construal in successive generations of language users. While many aspects of construal might be considered in a chapter such as this, I will focus on

534

Graeme Trousdale

three particular aspects of construal of relevance to language change.2 The first of these concerns the degree of specificity at which a particular linguistic expression is coded; the second concerns the viewpoint encoded in the expression; and the third concerns the degree of subjectivity inherent in the expression. The related diachronic processes to be considered are schematization and subjectification. The chapter is structured as follows. Specificity, viewpoint and subjectivity are briefly introduced in section 2. Definitions and exemplifications of grammaticalization and lexicalization are provided in 3. In that section, particular consideration is given to the ways in which grammaticalization and lexicalization are similar, and the ways in which they are different. Building on the similarities between grammaticalization and lexicalization, the notion of constructionalization is introduced, as a way of considering the development of form-meaning pairings from an explicitly cognitive linguistic perspective (section 4). Using two examples of grammatical constructionalization in the history of English (namely, the marking of conditional perfection with as long as, and the evolution of a new degree modifier construction as good as), I explore in some detail the nature of construal, changes in construal over time, and the relationship between pragmatic and semantic change. The conclusion (section 5) returns to the issue of cognitive pragmatics as ‘construal of meaning-in-context’, and how this features in a constructional approach to language change.

2.

Three aspects of construal

This section is concerned with the three aspects of construal mentioned in the introduction above: specificity, viewpoint and subjectivity. Specificity in construal is concerned with the degree of detail at which things and processes are expressed. In cognitive linguistics (see e.g. Croft and Cruse 2004), concepts are often said to be organized in the mind in a taxonomy which reflects this variation in specificity, from more general to more specific (SQUIRREL is more specific than RODENT , RODENT more specific than MAMMAL , MAMMAL more specific than ANIMAL and so on). Such taxonomies are not restricted simply to lexical concepts, and may be extended to aspects of the clause, such as the degree of attention to detail at which a speaker encodes a process and the participants involved. Langacker has argued that “an expression’s meaning cannot be reduced to an objective characterization of the situation described; equally important for linguistic semantics is how the conceptualizer chooses to construe the situation and portray it for expressive purposes” (1990: 5; emphasis original). In this aspect of construal, Langacker means (among other things) the focus of the speaker’s attention, the awareness of participants in a process and the relationship between those participants, so different scenes may be construed with varying degrees of detail:

Grammaticalization, lexicalization and constructionalization

535

(1) A grey-haired, slim French woman was repeatedly hitting a young man with her handbag. (2) An older woman hit a younger man. (3) Someone did something. More crucially for our discussion of grammaticalization, it is often the case that a more semantically general expression is a more likely candidate for participation in grammaticalization: (4) John was sprinting to catch the bus. John was running to catch the bus. John was going to catch the bus. The more semantically specific forms sprint and run are generally less likely to serve as candidates for grammaticalization to future markers than the more general form go. But as we will see below (section 4), while such semantic generality may be a characteristic of many grammaticalizing items, there is often a significant degree of pragmatic enrichment which accompanies meaning change in grammaticalization (see, for example, Sweetser 1990; Traugott 1995), so changes at one level of meaning may be compensated for at other levels of meaning. Viewpoint is similarly relevant for understanding some of the ways in which new form-meaning pairings encoding grammatical relations may evolve. This is most clearly observable in the development of prepositions as instances of grammaticalization; such new constructions often encode the speaker’s viewpoint and focus of attention (e.g. John is in front of Mary, Mary is behind John (behind < Old English be ‘at’ + hindan ‘the back parts’). This is an instance of a more general feature of viewpoint. Langacker (1987: 122) argues: “In observing a complex scene with multiple participants, we can assume different positions in relation to them, with direct consequences for their perceived proximity and salience.” He distinguishes two aspects of viewpoint, vantage point and orientation. The latter is concerned with “alignment with respect to the axes of the visual field (or some comparable coordinate system)”, and is perhaps of less relevance to grammaticalization and lexicalization than vantage point, which Langacker defines simply as “the position from which a scene is viewed” (1987: 123). The importance of this for grammaticalization (and to a lesser extent, lexicalization) is significant, as it ties in with the third of our aspects of construal, subjectivity. The literature on subjectivity, and the associated diachronic process of subjectification, is vast: a thorough overview of recent research is provided by LópezCouso (2010). The discussion of subjectivity and subjectification in this chapter will be restricted to how this state and process respectively relate to construal and changes in construal only. Speaker perspective is at the heart of construal. Consider in this regard the two clauses in (5) and (6).

536

Graeme Trousdale

(5) John waved as he passed. (6) John waved as he passed me. In (6), a reference point is explicitly referred to; in Langacker’s terms, the reference point appears “onstage”, and is verbalized as the complement of a verb. This is an example of objective construal. In (5), the speaker is only implicitly present, but nonetheless functions as a cognitive reference point. We understand as hearers or readers of expressions such as (5), that John’s ‘passing’ movement takes as its reference point the speaker, even though this is not explicitly coded. This is an example of subjective construal. For Traugott, subjective expressions have a prime semantic or pragmatic meaning which indexes “speaker attitude or viewpoint” (2010a: 32). This is based on Lyons’ (1982: 102) observation that “subjectivity refers to the way in which natural languages, in their structure and their normal manner of operation, provide for the locutionary agent’s expression of himself and his own attitudes and belief”. As noted above, Traugott is and has been particularly interested in how linguistic expressions come to index speaker perspective (that is, on the diachronic process of subjectification which leads to synchronic subjectivity).3 A critical issue which must be established from the outset is that subjectification and grammaticalization, though they may co-occur in instances of language change, are substantially different (Traugott 2010a: 32). Subjectification is in essence a semantic-pragmatic change, while grammaticalization is primarily a morphosyntactic change. This may be shown in instances of semasiological shifts involving none of the morphosyntactic parameters associated with grammaticalization (such as decategorialization and bonding). For instance, the pejoration of Old English (OE) sælig ‘blessed, innocent’ to Modern English (ModE) silly ‘foolish, stupid’ involves an increase in subjectivity, but no concomitant increase in grammatical status: silly is still an adjective (and thus has not decategorialized), nor is it bound as a suffix to a root. Of course, some instances of subjectification may also show signs of grammaticalization – but simply because two processes are at work on the same form does not mean the processes themselves are the same. For example, OE prætig ‘crafty’ undergoes meaning changes associated with subjectification (‘crafty’ > ‘attractive’ > ‘rather’), but the latter shift also involves change from adjective to degree adverb (as in the non-oxymoronic use of pretty ugly ‘rather ugly’). This brief discussion of the synchronic notions of specificity/schematicity and subjectivity, and the allied diachronic processes of schematization and subjectification, will be elaborated on in section 4 of this chapter.

Grammaticalization, lexicalization and constructionalization

3.

537

Grammaticalization and lexicalization compared

Both grammaticalization and lexicalization have been defined in a number of ways; some of these definitions focus more on the similarities between the two processes, some on the differences. Grammaticalization is typically seen as the development of new grammatical markers, typically (though not exclusively) from erstwhile lexical items as they are used in particular contexts. Examples include the case marker ba in Hungarian (Old Hungarian vila béle ‘world core/guts: directional’ > vilagbele ‘world into’ > világba; Antilla 1989: 149, Lehmann 1995: 85), and the English indefinite article a(n), from the Old English numeral an ‘one’ (Hopper and Martin 1987). Lexicalization also involves the development of new forms, but here, the output construction has referential, or contentful, meaning. While grammaticalization typically produces forms that are bound, like tense affixes and case markers, lexicalization typically produces open class items like nouns and verbs; in terms of the formal changes involved, we often see the development of monomorphemes from compounds (e.g. [[cup][board]] ‘piece of wood on which cups are placed’ > [cupboard] ‘storage unit’), or the creation of referential idioms (e.g. up and coming, ‘emerging’). As Brinton and Traugott (2005) observe, many of the formal processes involved in grammaticalization also feature in lexicalization. For instance, fusion involves the bonding of two or more elements which form a new unit: in grammaticalization, this is seen in the development of new prepositions like beside (< by + side); in lexicalization, this is seen in the development of new phrasal verbs like GIVE up ‘surrender’. Sometimes the boundary between these bonded forms dissolves, resulting in coalescence (in grammaticalization, parts of the future construction BE going to coalesce to gonna; in lexicalization, the bonded form god spell ‘good news’ coalesces to gospel). Both grammaticalization and lexicalization involve reanalysis, defined by Langacker (1977: 58) as “change in the structure of an expression of class of expressions that does not involve any immediate or intrinsic modification of its surface manifestation”. Although both grammaticalization and lexicalization share a number of features, there has been sufficient research on both topics (especially the former) for each to warrant separate treatment. Grammaticalization is elaborated on in section 3.1, and lexicalization in section 3.2. 3.1.

Grammaticalization

Traugott (2010b) observes that grammaticalization has recently often been conceived in one of two ways: as involving reduction, or as involving expansion. The reduction model (for example, Lehmann 1995 [1982], Haspelmath 2004) focuses primarily on formal loss; the expansion model (for example, Himmelmann 2004) focuses primarily on functional gain. This does not, however, mean that the reduction model does not attend to changes in function, nor that the expansion model

538

Graeme Trousdale

does not attend to changes in form. The reduction model is manifest most clearly in clines of grammaticalization, particularly those clines which focus on changes in the morphosyntactic freedom of the developing gram. One such cline is given in (7): (7) content word > grammatical word > clitic > inflectional affix (Brinton and Traugott 2005: 85) The development of OE willan ‘to want’ to the ModE auxiliary will and clitic ’ll is an example of this cline. As a clitic, ’ll has no independent status and must attach to a host (a reduction in morphosyntactic flexibility); it has also undergone a loss of phonetic substance. The replication of such changes across languages leads to the first of our definitions of grammaticalization: (8) Grammaticalization of a linguistic sign is a process in which it loses in autonomy by becoming more subject to constraints of the linguistic system. (Lehmann 2004: 155) This definition of grammaticalization encapsulates the kinds of changes we see in the development of tense and aspect marking. An example which illustrates this kind of linguistic change is the shift from Latin dare habes ‘give.INF have2. S . PRES ’ to 7thc. Romance daras, the loss of autonomy indicated by a main verb (habes) becoming an inflectional suffix (-as). By contrast, an expansion model of grammaticalization has been invoked to account (among other things) for the development of particular kinds of discourse markers and connectives. Himmelmann (2004) discusses three kinds of expansion, of relevance to both grammaticalization and lexicalization (on which see further section 3.2 below): host-class expansion, syntactic-expansion and semantic-pragmatic expansion.4 The first of these is concerned with the increase in the range of categories a particular form can occur with. For instance, the grammaticalized degree modifier sort of is initially found with nouns (as in he’s a sort of teacher); in the early nineteenth-century, this restriction is relaxed as speakers use the grammaticalized form to modify adjectives (He’s sort of confused) and verbs or verb phrases (it sort of annoyed me). Syntactic expansion concerns the loosening of restrictions on syntactic position (for example, the early twentieth century use of sort of as a free adverb, as in It annoyed me, sort of). Finally, semantic-pragmatic expansion is concerned with the development of polysemies, or increase in appropriate pragmatic contexts of use. For instance, the development of quantifier uses of a bit of (= ‘some’), as in a bit of apple, from the partitive uses (e.g. a bit (= ‘bite’) of apple), represents a particular kind of semantic-pragmatic expansion. This view of grammaticalization is more aligned with the definition provided by Brinton and Traugott (2005: 99):

Grammaticalization, lexicalization and constructionalization

539

Grammaticalization is the change whereby in certain linguistic contexts speakers use parts of a construction5 with a grammatical function. Over time the resulting grammatical item may become more grammatical by acquiring more grammatical functions and expanding its host-classes.

Furthermore, like Brinton and Traugott (2005), Himmelmann (2004) is keen to explore ways in which grammaticalization and lexicalization are similar, and ways in which they are distinct. For instance, Himmelmann observes that grammaticalization and lexicalization share a “common point of origin, i.e., the spontaneous and productive combination of lexical items in discourse” (2004: 33). As we will see immediately below, some of the processes involved in grammaticalization are identical to those in lexicalization; however, the crucial difference concerns the function played by the newly-created linguistic item: in grammaticalization, the new item has a procedural or indexical function, while in lexicalization, the new item has a referential or conceptual function. As Terkourafi observes, procedural meaning is provided by linguistic expressions which “contribute information about how to manipulate and combine […] concepts into a conceptual representation” (2011: 358–359); referential meaning, by contrast, concerns that part of meaning that links signs in the linguistic system to referents in the experienced world. 3.2.

Lexicalization

Brinton and Traugott define lexicalization as: the change whereby in certain linguistic contexts speakers use a syntactic construction or word formation as a new contentful form with formal and semantic properties that are not completely derivable or predictable from the constituents of the construction or the word formation pattern. Over time there may be further loss of internal constituency and the item may become more lexical. (2005: 96)

Central to this definition is the notion of lexicalization as a kind of language change, specifically a change involving both form and meaning, where the meaning of the new form is contentful rather than procedural. When considering lexicalization as a kind of language change, typically, the macro-change involved is gradual, though the process may involve abrupt micro-changes (reanalyses), on which see further Traugott and Trousdale (2010). This is useful for distinguishing other types of change which have been described as lexicalization (see also Brinton 2002). One such change is the extragrammatical word formation process known as clipping (for example, the loss of the initial two syllables of the noun attitude to create the new noun tude ‘a bad, angry or disrespectful attitude’). There is no gradual change in form and meaning in this case (though of course the spread of the new lexical item in a social network or speech community may be gradual): a new form associated with a specific sense is created instantaneously. Examples of the kind of language change typically considered to be gradual lexicalization include the development of simplex forms from compounds, as in

540

Graeme Trousdale

Pre-OE *hlaf ‘loaf’ + *weard ‘guardian’ > OE hlaford > ME laferd > ModE lord, and the development of unanalyzable wholes from syntactic strings, as in nuts and bolts ‘practical details’. In fact, we see in the process of lexicalization a range of shifts along the syntax-lexicon continuum. Whatever the shift in position, we see a decrease in linguistic ‘flexibility’ (which characterizes the syntactic end of the continuum). As lexicalization develops, syntactic strings are no longer understood compositionally; erstwhile compounds become morphologically complex forms; complex forms become simplex. An expression involving non-idiomatic string nuts and bolts can be recoded in an expression where the two nouns are reversed, with little (if any) change in meaning: (9) In the toolbox was a selection of nuts and bolts. (10) In the toolbox was a selection of bolts and nuts. Such reversal is, of course, not possible when nuts and bolts is used idiomatically: (11) Let’s get right to the nuts and bolts, congressman. (Corpus of Contemporary American English; Davies 2008-) (12) *Let’s get right to the bolts and nuts, congressman. Lehmann treats lexicalization as “a process in which something becomes lexical” (2002: 14), where “lexical” is understood as a loss of internal structure. In comparing lexicalization with grammaticalization, Lehmann (2002) considers a structure of the type [X Y]Z, i.e. a sequence (Z) of two elements (X and Y). In lexicalization Z as a whole is affected, and the dependency relationship between X and Y is “renounced”. For instance, in a change from Latin ad ipsum ‘to itself-ACC ’ to Italian adesso ‘now’ (Giacalone Ramat 1998: 122), a prepositional phrase develops into an adverb through a process where the dependency relationship between the preposition and the pronoun is erased, and instead of being analyzed as compositional, the string is accessed as a whole, and a new unit is created. In grammaticalization, the dependency relationship between X and Y changes as one or the other becomes a grammatical formative (and thus more constrained): the process “involves analytic access to a unit” (Lehmann 2002: 13, emphasis added). For instance, in the change from Old Hungarian vila ‘world’ + bele ‘guts [+ directional]’ to Modern Hungarian vilagba ‘world [+ directional]’ (Antilla 1989 [1972]: 149), to which reference was made in section 3, bele develops to a case marker: the dependency relationship between the (still “free”) lexical item and the grammaticalized case marker has changed. In this account, no attention is paid to the meaning of the newly-coined linguistic item. For instance, adesso ‘now’, while belonging to the inventory, can hardly be said to have developed a more lexical, referential meaning as opposed to a more grammatical function. Examples from English of changes which involve increased holistic access as well as referential meaning include daisy (< OE dæges eage ‘day.GEN eye’), woman (< OE wifmann ‘wife.man’) and walnut (< OE

Grammaticalization, lexicalization and constructionalization

541

walhhnutu ‘Celtic/foreign.nut’). More crucially, erstwhile sequences of grammatical items may be analyzed holistically and develop a contentful meaning (e.g. shilly-shally ‘to vacillate’ < shall I, shall I). It is observable from the preceding discussion that in this view of grammaticalization and lexicalization, the processes of fusion and coalescence are operative in both changes. These processes are formal (i.e. they affect some or all of the syntactic, morphological and phonological properties of the input string). Brinton and Traugott (2005: 63–67) identify a series of ways in which fusion and coalescence are said to be properties of lexicalization. From a diachronic perspective, fusion entails morphological boundary loss. This may involve a shift from, among other possibilities, syntagm to compound (e.g. kinsman < OE cynnes mann ‘kin.GEN man’), or from complex to simplex structure (e.g. bailiwick < ME bailie ‘bailiff’ + -wick ‘area’). In each case, we see a development from something more compositional, more clearly assembled through a number of component parts, to something more holistic. However, fusion is not restricted to changes where the meaning of the more recent item is ‘contentful’; changes which produce more indexical or procedural constructions also involve fusion, as in Latin cantare habeo > French je chanterai ‘I will sing’, and English BE going to > gonna as a future marker. Fusion is therefore not a process which distinguishes grammaticalization from lexicalization, but rather is a process often observed in instances of both kinds of language change. Furthermore, form changes involving “the loss of phonological segments” (Brinton and Traugott 2005: 27), or coalescence, are also observable in cases of lexicalization (in the development of barn < OE bere ‘barley’ + ærn ‘house’) and grammaticalization (in the development of gonna < going to). Coalescence, like fusion, is not a process which distinguishes grammaticalization from lexicalization; rather, since the erasure of boundaries sometimes leads to morphophonological change, in both types of change we see a new, more atomic, less analyzable construction emerging. In a similar vein, Wischer (2000), for instance, notes that coalescence and syntactic reanalysis are sub-processes associated with both lexicalization and grammaticalization, but that these are formal, not functional changes. In distinguishing grammaticalization from lexicalization, Wischer observes that a new semantic component is added in lexicalization “so that the new lexical meaning differs from the former compositional meaning”; by contrast, grammaticalization involves semantic bleaching and “an implied categorical or operational meaning is foregrounded” (Wischer 2000: 364–365). For Himmelmann (2004), host-class, syntactic and semantic-pragmatic expansions are key features of grammaticalization, but contraction in all three domains is not a necessary feature of lexicalization. Lexicalization involves a historically compositional expression being “treated more and more as a single processing unit” (Himmelmann 2004: 34), and host-class contraction is said to occur because it is said to apply to only one member of a particular class. However, contraction in syntactic context and in sem-

542

Graeme Trousdale

antic-pragmatic context is not criterial: in fact, semantic-pragmatic context may expand, contract or remain unchanged. In summary, grammaticalization and lexicalization are processes of change which share a number of properties: both involve sequences of reanalysis, and both involve clines, and in many cases, both involve formal fusion and subsequent coalescence; similarly, both typically involve increase in idiomaticity. A significant distinguishing feature is the function of the output construction: in grammaticalization, the new form has a procedural or indexical function, while in lexicalization, the new form is contentful. Brinton and Traugott (2005: 110) provide a useful summary of ways in which grammaticalization and lexicalization may be seen as similar, and ways in which they may be seen as different.

4.

Constructionalization

The final part of this paper is an attempt to bring together some of the previous discussion, and to show how constructionalization (Traugott 2008a, 2008b; Trousdale 2008, 2010) is a helpful model for understanding how the “construal of meaningin-context” is relevant for the development of grammatical constructions. Specifically, the discussion focuses on the role of pragmatics and construal in change, and the link between construal and subjectification. We saw in section 3 that there were a number of similarities and differences between grammaticalization and lexicalization. What unites the two types of change is that in both cases, there is the gradual development of new form-meaning pairings, through a series of reanalyses; what distinguishes them is that in grammaticalization, the new form-meaning pairing has an indexical or procedural function, while in lexicalization, the new construction has a referential meaning. Since constructions are conventional symbolic units (Langacker 1987, Croft 2005), the creation of new form-meaning pairings may be referred to as constructionalization; grammatical constructionalization produces procedural/indexical constructions, while lexical constructionalization produces referential/conceptual constructions. Since the focus of the present chapter concerns construal, and semantic-pragmatic changes, and since (as Himmelmann 2004 observes) semantic-pragmatic context may increase, decrease or stay the same in lexicalization, the focus for the remainder of this chapter will be on grammatical constructionalization. Constructionalization has been defined as “the formation of new units (constructions) out of hitherto independent material” (Bergs and Diewald 2008: 5). This formation is typically the product of reanalysis, though analogy to other constructional types plays an important role (see Fischer 2007 for a discussion of the importance of analogy in grammaticalization). These reanalyses take place at different levels of the construction, which gives the appearance of a gradual change, though the reanalyses themselves are necessarily abrupt (Traugott and Trousdale

Grammaticalization, lexicalization and constructionalization

543

2010). While there is no predetermined order for reanalyses at different constructional levels, the hypothesis is that pragmatic changes precede semantic changes; and these meaning changes precede formal changes. How are the three aspects of construal identified in section 1 relevant to grammatical constructionalization? Specificity is relevant because in grammatical constructionalization, the new construction is more schematic and general than the input constructions; a more schematic construction emerges as new polysemies develop when speakers abstract across instances of language use. This may be illustrated by the development of BE going to as a future marker. A new schema develops, linked to host-class expansion. Prior to constructionalization, where go is interpreted as a motion verb, the following verb had to be non-stative; now, when BE going to marks futurity, stative verbs are acceptable (e.g. You’re not going to like this). Similarly, prior to constructionalization, the morphosyntactic subject of BE had to be capable of motion; after constructionalization, dummy subjects were possible, in raising constructions (e.g. There’s going to be trouble).Viewpoint and subjectivity are involved in the diachronic process of subjectification, an important part of grammatical constructionalization. As observed above, constructionalization involves form-meaning changes; subjectification, by contrast, is crucially a meaning change. Subjectification perforce applies only to the meaning part of a sign. In the changes discussed below, we see the importance of both reanalysis and analogy – analogy is relevant for invited inferencing (Traugott and Dasher 2002), reanalysis is relevant for semanticization of generalized invited inferences. Invited inferencing involves speakers (or writers) using lexical items in particular contexts through which pragmatic implicatures arise, and to which hearers (or readers) may become attuned: in other words, invited inferencing must involve interaction. Should such implicatures become conventionalized in a speech community, with a historically pragmatic implicature becoming semanticized, then change has occurred as a result of this interactive process. As part of the exploration of the role of construal in grammatical constructionalization, we can identify different constructional levels, as well as constructs (Traugott 2007; Fried 2008). Constructs are attested instances of use, and are the locus of language change in a usage-based model. Micro-constructions are individual construction types (for example, the individual degree modifiers a lot vs. a bit vs. sort of vs. kind of); meso-constructions are groupings of constructions which behave in similar ways (e.g. groups of degree modifiers such as a lot/a bit vs. sort of/kind of); and macro-constructions are the most schematic construction which nevertheless sanctions all the relevant meso-constructions (e.g. a Degree Modifier Construction). These issues are explored in relation to the development in English of two constructions with the form [as x as], illustrated by the following (from the British National Corpus, henceforth BNC-BYU (Davies 2004-):

544

Graeme Trousdale

(13) by Christmas 1942 the situation was as good as hopeless (14) change is fine as long as it’s done sympathetically and actually has some purpose. The constructions to be discussed originate as instances of the Comparative Construction, which is still productive in contemporary English, and illustrated by the following examples (again, all from BNC-BYU): (15) there’s a mirror in the sitting room as big as that wall (16) Pop round as early as you can (17) only Scotland could resort to something quite as sad as turning a pub into a museum of dispirited sport In the rest of this section, I detail the development of as good as as a degree modifier (section 4.2) and as long as as a marker of conditional perfection (section 4.3). The purpose of these sections is simply to provide some data and a brief account of the changes involved. A more detailed analysis in terms of construal is provided in section 5. 4.1

As good as

Evidence from contemporary English of as good as as an approximator (i.e. a kind of downtoner in the category of intensifiers defined by Quirk et al. 1985: 589) can be found in the following examples from the BNC-BYU: (18a) [=(13) above] by Christmas 1942 the situation was as good as hopeless [modifies A] (18b) real short-term interest rates are six and a half per cent (they are as good as zero in the United States) [modifies N] (18c) Mr Kinnock (apparently incited by Labour’s teaching hospitals) as good as came out for it [modifies finite V] (18d) seeing fine teams commit one horrendous play to throw away a crucial championship game as good as won [modifies Past Participle] (18e) They are horses, or as good as [free adjunct] In addition to showing a meaning change compared with the meaning of expressions like John is as good as Bob, the examples in (18) show host class expansion (18a-d) and syntactic expansion (18e). Examples of a degree modifier use can be found from early Middle English on, first as a modifier of an adjective (ded ‘dead’): (19) A slepyng mon to men is tolde As good as ded for pe tyme, Bi niht or day (c. 1260 S. Bernard, (Legenda Aurea) 321–6 [LION Poetry])

Grammaticalization, lexicalization and constructionalization

545

‘A sleeping man to men is told/As good as dead for the time/By night or day’ Later, it can be found as modifier of participles (both past and present), then infinitive forms of verbs, as illustrated by (20), (21) and (22) respectively (20) For the batayle was as good as ent, Sithen thei hadde sclayn the knyght vigorous, Sir Ector, and Dephebus (c. 1400 Laud Troy Book, 13232–4 [LION Poetry]) ‘For the battle was as good as ended, since they had slain the vigorous knight, Sir Hector, and Dephebus’ (21) And with that woord he thrusting in his head, Euen then did turne him towards her as good as being dead, And wretchedly did totter on the poste with strangled throte. (1567 Golding Metamorphosis 851–853 [LION Poetry]) ‘And with that word, he, thrusting in his head, even then turned himself toward her, nearly being dead, and in a wretched way tottered on the post with a strangled throat’ (22) You doe as good as passe by … which is greater Jeofaile then our Maximilian. [OED s.v. jeofail, n.] (1641 Smectymnuus Vind. Answ. xi. 111 The Acts of Dioclesian Maxim.) ‘You do as good as pass by, which is a greater jeofail [= mistake, typically by a lawyer in court GT] than our Maxmilian’ The use of as good as as a modifier of a tensed verb seems to be a nineteenth century phenomenon, illustrated in (23): (23) the first time I ever saw him he as good as told me that he had set fire to the place – it surprised me rather – I have never been engaged in any of these sort of things (Trial of Charles Wilson Bond and Thomas Nye, 10th June 1867 [BAILEY]), example provided by Elizabeth Traugott, pc. The gradual alignment of the construction as an instance of the more schematic degree modifier construction is manifest by the fact that the former is a peripheral, non-prototypical member of the latter: (24) *It was an as good as hopeless situation. (cf. (18a)) (25) *?He played the piano as good as inaudibly. (cf. He played the piano practically inaudibly.) We may account for this change as follows. First, comparative constructions of equality (such as as near as, as fine as and as good as) highlight similarities be-

546

Graeme Trousdale

tween two things x and y, but maintain the distinction between the two (i.e. that the two things x and y are equal with regard to some property, but are not identical); if something is as good as dead (and the most general meaning of good, historically, is ‘Of things: Having in adequate degree those properties which a thing of the kind ought to have’, OED s.v. good, a.), then the implicature is that it approximates as closely as possible to being dead, without actually being so. For as good as, this is established by the ME period. Then, this implicature is semanticized to mean simply ‘practically’, ‘nearly’: this late ME development is a generalization, and would count as an instance of semantic/pragmatic expansion in the sense of Himmelmann (2004). Host-class expansion, as an approximator used to modify adjective-like forms (participles, 15th/16th centuries), then verbal forms, non-finite (17th century), then finite (19th century) is a later development. 4.2

As long as

Consider the present-day English example in (26), (26) I’ll look after them as long as you pay me. This example, a slightly modified version of example (36i) in Huddleston and Pullum (2002: 1134), has two potential meanings: ‘I’ll look after them for the period of time during which you pay me’ (the historically older meaning), and ‘I’ll look after them if and only if you pay me’ (the historically newer meaning). The newer pattern is “one where as long as has lost its comparative meaning and been reanalysed as a compound preposition meaning ‘provided’” (Huddleston and Pullum 2002: 1134); so here, a pattern associated with the Comparative Construction has been reanalyzed as an instance of the Conditional Construction. Although this is undoubtedly an instance of the Conditional Construction, through a series of invited inferences, it has a more specific meaning of conditional perfection (Geis and Zwicky 1971): not just ‘if’, but ‘if and only if’. The construction has some overlap with as far as (another instance of grammatical constructionalization, whereby an older comparative comes to have a topic restricting function), but there is a difference relating to profiling (Langacker 1987, 1991): with as far as, what is profiled is the endpoint, but with as long as, what is profiled is the distance between the start and endpoint. Like as far as, the temporal sense of as long as is likely to have been a metaphorical extension from the spatial sense, but this was a pre-OE development, since in both OE and ME, the spatial and temporal meanings co-existed (see also Traugott and Dasher 2002: 36–37), as illustrated by (27) and (28): wæron full neah tu (27) þa these were.pl full nigh two hæfdon lx ara, sume mara had.pl. 60 oars, some more

swa lange swa þa oðru sume as long as the others some

Grammaticalization, lexicalization and constructionalization

547

‘these [boats] were nearly twice as long as the others: some had 60 oars, some more’ (896/7 Chron A. (Plummer)) (28) ne mihte nan mann naht pisum folce swa not had-power no man nought this.dat people.dat. as lange swa hi heoldon heora god on riht long as they held.pl their god in right ‘and no man could do anything to these people, as long as they valued their god in the right way’ (1000 Ælfric Judith) There is an issue as to whether (28) has the invited inference of a “conditional”, given its formal properties: the -on marking on the verb (i.e. heoldon) is not conclusive either way, since not all verbs in conditional clauses were marked with subjunctive in OE (Mitchell 1985: §§ 3560–3591, Traugott 1992: 256–257), and by later OE, the changes in the unstressed vowels meant that a distinction between indicative and subjunctive in past plurals was neutralized (Campbell 1959: § 735, 752; Hogg 1992: § 6.62). But thinking about the “construal of meaning-in-context”, there is a suggestion that it probably does have such an inference: the textual context establishes independence of the people for the period that, but also on the understanding that, they maintain loyalty to God. Indeed, this example provides a good instance of a critical context in the sense of Diewald (2002). In Diewald’s model, there are three context stages. The first is the ‘untypical’ context. In this context, there are implicatures which “show clusters of contextual features which had not been customary before” (Diewald 2002: 109). The second is the “critical” context, where ambiguities (in both form and meaning) invite “several different interpretations, among them the new grammatical meaning” (Diewald 2002: 103). Finally, there is the “isolating” context, where one interpretation (and not the other) is possible. As Traugott (2010c) has shown, critical contexts (and bridging contexts in the sense of Heine 2002) are not obligatory for grammaticalization, though of course there may be such contexts for morphosyntactic reanalyses. By the late ME period, generalization of the invited inference is clear, as in example (29): (29) A, dere hosbond, thynk on me, And save yowersylfe as long as e may, For trewly itt wyll no otherwyse be! (?1490 Saint Mary Magdalen (Bodleian Digby MS 1333, 1749–1755) [LION Drama]) ‘Ah, dear husband, think about me (as long as you can), and save yourself (if it’s possible), for truly, it can be no other way’

548

Graeme Trousdale

The temporal reading is still there, but backgrounded (‘Keep yourself protected for as long as you can’ ~ ‘Escape if you can’), and the example is suggestive of the generalized invited inference discussed by Traugott and Dasher (2002: 37). Semanticization of this inference is evident by the mid-nineteenth century: (30) You is good to a hundred yard, you say, and they aint much furder off nor that; but as long as they don’t git skeered, we’ll let ’em kim up powerful nigh. (1857 Bennett, The Border Rover, Ch. 19 [LION Prose]) These data provide good evidence to link Traugott and Dasher’s theory of semantic change with Himmelmann’s expansion views on grammaticalization, and the more general issue of constructionalization. The data suggest that constructionalization may begin with the negotiation of invited inferences associated construct meaning (or utterance-token meaning in the sense of Traugott and Dasher 2002); conventionalization of invited inferences as generalized invited inferences applies to micro-constructions, not constructs; this stage correlates with the semantic/pragmatic expansion of Himmelmann (2004), and it is here where we can clearly see the role of cognitive pragmatics in this aspect of constructionalization. Here, the ‘construal of meaning-in-context’ gives rise to new conventionalized meanings, a critical part of any constructionalization process (given that, in constructionalization, both form and meaning must be new). The cycle may repeat as speakers and hearers continue to exploit the newly available structures, reanalyzing other similar micro-constructions in the same way. The creation of such meso-constructions means that the generalization of invited inferences may apply to even more schematic types; but the reanalysis of micro-constructions as instances of “new” mesoconstructions involves changes in form as well as meaning, correlating with the host class expansion of Himmelmann (2004), and therefore cannot be solely linked to construal of meaning-in-context. Here we see this in the spread of the conditional meaning of as long as from verbs denoting temporary states to “event structures involving patterns of reasoning and cognition” (Traugott and Dasher 2002: 37). This is part of the important issue of renewal, recursion and resetting (Traugott and Dasher 2002: 39). When the cycle repeats at the meso-constructional level, it is not the case that recursion stops at the micro-constructional level; speakers may still use older constructional types to invite further inferences, and the older constructional types may still undergo changes. Constructionalization involves stepwise realignments within the constructional network: semantic-pragmatic expansion seems to be the precursor for other kinds of expansion (Himmelmann 2004); constructionalization involves changes at formal levels (syntax, morphology, phonology), not just semantic-pragmatic change, for reasons to be explained immediately below.

Grammaticalization, lexicalization and constructionalization

4.3.

549

Constructionalization and cognitive pragmatics

In this section, I return to the three types of construal identified in section 1 and, having provided an outline of two different instances of grammatical constructionalization in English, show how cognitive pragmatics helps to elucidate some of the key issues. 4.3.1.

Specificity and schematicity

It has been suggested that in grammatical constructionalization, constructions become more schematic (Traugott 2007, 2008a, 2008b; Trousdale 2008, 2010). This may be interpreted in more than one way. In one interpretation, increased schematicity means a decrease in semantic detail associated with a new construction. In another interpretation, increased schematicity means a more abstract construction which is capable of sanctioning new, more specific constructions (Langacker 2008: 215). How can these two interpretations be brought together? One way is to consider the relation between semantic bleaching and pragmatic strengthening which has often been attested in work on grammaticalization (Traugott and Dasher 2002; Hopper and Traugott 2003). On the one hand, the semantic bleaching associated with the development of quantifier a bit of from OE bita ‘bite’, or future BE going to from motion go, clearly involves speakers, over time, construing new meanings in particular pragmatic contexts which do not involve the historically more specific meaning of a given item. Thus the motion sense of go may be backgrounded (he’s going to visit his sister) or absent (there’s going to be trouble) as the new construction BE going to develops. But the critical issue here is that we are comparing the ‘lexical’ semantics of an item (go) at one stage with the constructional semantics of another item (BE going to) at a later stage – in constructionalization, the meaning of BE going to (and not just go) is crucial, once the reanalyses have produced this new form-meaning pairing. The future meaning is not simply associated with go, but with the reanalysis of the combination go, the progressive (BE -ing) and the purposive (to), and that the future meaning comes to be distributed over the sequence BE going to. As the new construction becomes entrenched, speakers attend less to the individual contributions of the historically composite and specific parts, and more to the general and schematic meaning of the new “whole”. This is facilitated by the second development of pragmatic strengthening (Brinton 1988; Traugott 1989, 1995; Sweetser 1990, among others). In the early stages of grammaticalization, invited inferences may be strengthened. These contextually bound interpretations (in at least some cases of grammaticalization) involve polysemies, and these polysemies are associated with the kind of semanticpragmatic expansions noted by Himmelmann (2004) and discussed above. Not only does this illustrate the interplay between reduction and expansion processes in grammaticalization, it also shows the interplay between different aspects of con-

550

Graeme Trousdale

strual in this kind of language change. This type of schematization – whereby more abstract constructions sanction a wider range of types (as, for example, BE going to collocates with statives (he’s going to like this) as well as non-statives (he’s going to lick this), as noted above at the start of section 4 – is closely allied to another aspect of pragmatic strengthening, namely subjectification, or the strengthening of “the subjective stance of the speaker” (Traugott 1995: 49), particularly – in the case of examples such as this – that there has been “a pragmatic reanalysis such that the experiencer of an (abstract) sense of motion has been identified with the speaking subject” (Traugott 1995: 36). Similar claims can be made for the development of as good as and as long as. In the case of the latter, for example, the development of a (solely) conditional perfection reading arises as a result of an initial pragmatic strengthening (i.e. the initial invited inference which develops in particular instances of use). This pragmatic strengthening is in itself related to the semantic bleaching, and development of polysemies associated with schematization. In the constructionalization of as long as, the bleaching is associated not only with the adjective long, but with the comparative meaning invoked by the discontinuous sequence as … as; the “loss” is shared by all members of the newly developing construction. At the same time, the pragmatic strengthening associated with conditional readings of as long as is a product of a pragmatic reanalysis which is a concurrent part of the incipient constructionalization of as long as as a unit. As noted above, this constructionalization cannot be said to have taken place until both meaning and form have changed; nonetheless, the early stages of constructionalization often involve pragmatic reanalyses that strengthen particular readings in particular instances of use (i.e. meanings associated with constructs, rather than constructions). Subsequent polysemies, and the development of purely conditional readings of as long as, are associated with a later stage of schematization, because it is at this stage that a wider range of types is sanctioned (by virtue, in part, of the other kinds of expansion noted by Himmelmann 2004). 4.3.2.

Viewpoint and subjectification

In their analysis of subjectification, De Smet and Verstraete define the process as one “whereby an element or a construction develops new senses that require speaker reference” (2006: 365). In this final section, I explore how some aspects of changes in viewpoint associated with subjectification are implicit in the cases of grammatical constructionalization discussed above, and therefore how cognitive pragmatics, understood as the “construal of meaning-in-context” is implicated too. Since the construct is the locus of all linguistic change, semantic-pragmatic changes associated with subjectification must also be located there. The beginning of the change is hypothesized to be the Speaker/Writer’s production of a construct of the source construction (in the present cases, the Comparative Construction)

Grammaticalization, lexicalization and constructionalization

551

which invites inferences that ‘require speaker reference’; inferences may arise as properties of the source macro-construction, or in collocations of particular microconstructions with other elements (e.g. as good as followed by adjectives such as dead). The next stage is that the Hearer/Reader accepts implications that ‘require speaker reference’ by a process of constructional (re)alignment. This is associated with the notion of schematization discussed in the previous section. Acceptance of the invited inferences – for example, that the Speaker/Writer is using the form associated with a micro-construction of the comparative macro-construction to imply a meaning associated with a different macro-construction (a conditional, and specifically, conditional perfection) – is associated not only with a new polysemy for as long as, but also for the beginning of a new micro-construction to mark conditional perfection. At this stage, however, there has been no constructionalization, since it is not the case that both form and meaning have changed. Only when a new coded meaning has been established can it be suggested that semanticization has taken place (Traugott and Dasher 2002: 38). In constructional terms, semanticization of generalized invited inferences may lead to the development of a new micro-construction as a conventional symbolic unit, but as noted above, morphosyntactic change is a necessary feature of constructionalization: semantic change alone will not constitute constructionalization, since constructionalization involves a sequence of reanalyses leading to new form-meaning pairings (i.e. where both form and meaning need to be new). Once this new construction has been entrenched, there may be further schematization, in that the more abstract construction type may sanction new specific constructions as a consequence of host class expansion (Himmelmann 2004). For instance, as good as as a micro-construction, once established, develops in the ME period by becoming more schematic and sanctioning a wider range of forms (see Figure 1):

Figure 1. The relationship between increased schematicity and host-class expansion

As host-class expansion takes place, the productivity of the micro-construction increases, but the effects on the constructional network need not stop there. The micro-construction may develop syntactic patterns associated with meso-construc-

552

Graeme Trousdale

tions. For instance, in the case of as long as, movement to the left periphery aligns the micro-construction with other English conditionals. This position is not available for all instances of the old string associated with the comparative construction: (31a) He’ll buy it as long as it’s cheap. As long as it’s cheap, he’ll buy it. (conditional perfection) (31b) My left leg is as long as my right leg is. *As long as my right leg is, my left leg is. (comparative) This syntactic expansion (Himmelmann 2004) is part of the formal changes associated with grammatical constructionalization, rather than the meaning changes associated with subjectification. Therefore, constructionalization goes beyond the development of new senses (cf. De Smet and Verstraete 2006); subjectification may be implicated in grammatical constructionalization, but without changes in both form and meaning, constructionalization has not taken place. Subjectification is more likely in early stages of grammatical constructionalization; this is to be expected, given the association between subjectification and primary grammaticalization (Traugott 2010a: 40–41). In summary, the relationship between subjectification and grammatical constructionalization concerns the way in which semantic-pragmatic changes relate to morphosyntactic changes. Discussing the relationship between invited inferences and generalized invited inferences, Traugott and Dasher state: If the morphosyntactic contexts in which the IINs [invited inferences GT] arise are replicated often enough and become salient enough in the community, GIINs [generalized invited inferences GT] will arise. These widely used and understood GIINs enable the innovating speaker to recruit constructions and lexical items to functional categories. (2002: 35)

Traugott and Dasher (2002) are using ‘construction’ here in the sense of ‘string’, not a conventional form-meaning pairing (see note 5); but it is true that generalized invited inferences clearly facilitate grammatical constructionalization. In both cases examined in detail in this section, we see the “enriching of the form-meaning pair in question with the speaker’s perspective” (Traugott 1999: 188). The changes involve new micro-constructions, which arise as a result of speaker-hearer negotiation of: (a) polysemies of the existing construction (created in part by schematization) (b) invited inferences (and accepted implications) of utterances in particular contexts and instances of use (where construal plays an important role) (c) realignments with ‘neighbouring’ constructions in the network and formal reanalysis as reconfiguration in a new part of the constructional network, when invited inferences give rise to generalized invited inferences.

Grammaticalization, lexicalization and constructionalization

5.

553

Conclusions

This chapter has discussed the importance of aspects of cognitive pragmatics (understood as the construal of meaning-in-use) for particular kinds of language change. Having established three particular aspects of construal of particular relevance to semantic-pragmatic change, I looked at features of grammaticalization and lexicalization, particularly in terms of similarities and differences between the two types of change. Building on those similarities and differences, the notion of constructionalization was introduced. Exemplifying the discussion using two (related) case studies of grammatical constructionalization, I showed how observations on changes in construal – particularly, schematization, and subjectification – help to elucidate some aspects of this cognitive approach to the emergence of grammar. It need hardly be said that context has always been important in grammaticalization – nearly twenty years ago, Bybee, Perkins and Pagliuca wrote that “[e]verything that happens to the meaning of a gram happens because of the contexts in which it is used” (1994: 297). This context includes pragmatics. An expansion model of grammaticalization (one which fits well with the principles of constructionalization) considers semantic/pragmatic expansion in some detail. This expansion involves growth in polysemies, and conventionalizations/semanticizations of pragmatic implicatures. These have consequences for the morphosyntactic changes which are an essential part of grammatical constructionalization. One benefit of the constructionalization approach is that it links up the semantic/pragmatic changes associated with subjectification to the morphosyntactic changes associated with grammaticalization; by doing so, it reinforces the claim that subjectification is a semantic-pragmatic change, and grammaticalization is a morphosyntactic change (Traugott 1995, 2010a). An explicitly cognitive linguistic/constructional model is not the only one which links semantic-pragmatic change to formal changes – this is also addressed by e.g. Himmelmann (2004) in his discussion of host-class and syntactic expansion. Nonetheless, this application of a particular model of cognitive linguistics demonstrates some of the ways in which cognitive pragmatics may be implicated in more widespread and systemic changes. Furthermore, it lends support to the more general constructional claim regarding the importance of context (see the contributions to Bergs and Diewald 2009). Other kinds of context which are relevant include for example diachronic collostructional analysis (Hilpert 2008) but this is not explicitly linked to semantic/pragmatic expansion, being more concerned with host class expansion. For semantic/pragmatic expansion, aspects of construal appear to be central.

554

Graeme Trousdale

Notes 1. I am grateful to Hans-Jörg Schmid and to Maria José Lopez-Cóuso for their perceptive and helpful comments on an earlier draft of this chapter. All shortcomings are my own. 2. Other aspects of construal (e.g. metaphor and metonymy) are also relevant, but limitations of space mean that only a particular set of construal operations can be considered in detail. 3. For a detailed discussion of the different positions taken by Langacker and Traugott on subjectivity and subjectification, see De Smet and Verstraete (2006) and López-Couso (2010). 4. See Traugott (2008a) for further discussion of these expansions in relation to the development of quantifiers and degree modifiers from binominal sequences in English. 5. In this quotation, ‘construction’ is used to refer to a syntactic string or syntagm, rather than a form-meaning pairing as the term is understood in construction grammars of various kinds (for instance Langacker 1987; Goldberg 1995, 2006; Croft 2001).

References Antilla, Raimo 1989 (1972) Historical and Comparative Linguistics, 2nd edition. Amsterdam: John Benjamins. Bergs, Alexander and Gabriele Diewald 2008 Introduction: Constructions and language change. In: Alexander Bergs and Gabriele Diewald (eds.), Constructions and Language Change, 1–21. Berlin: Mouton de Gruyter. Bergs, Alexander and Gabriele Diewald (eds.) 2009 Contexts and Constructions. Amsterdam: John Benjamins. Brinton, Laurel J. 1988 The Development of English Aspectual Systems. Cambridge: Cambridge University Press. Brinton, Laurel J. 2002 Grammaticalization versus lexicalization reconsidered: On the late use of temporal adverbs. In: Teresa Fanego, María José López-Couso and Javier PérezGuerra (eds.), English Historical Syntax and Morphology, 67–97. Amsterdam: John Benjamins. Brinton, Laurel J. and Elizabeth Closs Traugott 2005 Lexicalization and Language Change. Cambridge: Cambridge University Press. Bybee, Joan 2010 Language, Usage and Cognition. Cambridge: Cambridge University Press. Bybee, Joan, Revere Perkins and William Pagliuca 1994 The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago: University of Chicago Press. Clausner, Timothy C. and William Croft 1999 Domains and image schemas. Cognitive Linguistics 10: 1–31.

Grammaticalization, lexicalization and constructionalization

555

Croft, William and D. Alan Cruse 2004 Cognitive Linguistics. Cambridge: Cambridge University Press. Croft, William 2001 Radical Construction Grammar. Oxford: Oxford University Press. Croft, William 2005 Logical and typological arguments for Radical Construction Grammar. In: JanOla Östman and Mirjam Fried (eds.), Construction Grammars: Cognitive Grounding and Theoretical Extensions, 273–314. Amsterdam: John Benjamins. Davies, Mark 2004BYU-BNC. (Based on the British National Corpus from Oxford University Press). Available online at http://corpus.byu.edu/bnc/. Davies, Mark 2008The Corpus of Contemporary American English: 425 million words, 1990present. Available online at http://corpus.byu.edu/coca/. De Smet, Hendrik and Jean-Christophe Verstraete 2006 Coming to terms with subjectivity. Cognitive Linguistics 17: 365–92. Diewald, Gabriele 2002 A model for relevant types of contexts in grammaticalization. In: Ilse Wischer and Gabriele Diewald (eds.), New Reflections on Grammaticalization, 103– 120. Amsterdam: John Benjamins. Fischer, Olga 2007 Morphosyntactic Change: Functional and Formal Issues. Oxford: Oxford University Press. Fried, Mirjam 2008 Constructions and constructs: mapping a diachronic process. In: Alexander Bergs and Gabriele Diewald (eds.), Constructions and Language Change, 47–79. Berlin: Mouton de Gruyter. Geis, Michael and Arnold Zwicky 1971 On invited inferences. Linguistic Inquiry 2: 561–566. Giacalone Ramat, Anna 1998 Testing the boundaries of grammaticalization. In: Anna Giacalone Ramat and Paul J. Hopper (eds.), The Limits of Grammaticalization, 107–127. Amsterdam: John Benjamins. Goldberg, Adele E. 1995 Constructions: A Construction Grammar Account of Argument Structure. Chicago: University of Chicago Press. Goldberg, Adele E. 2006 Constructions at Work: The Nature of Generalization in Language. Oxford: Oxford University Press. Haspelmath, Martin 2004 On directionality in language change with particular reference to grammaticalization. In: Olga Fischer, Muriel Norde and Harry Peridon (eds.), Up and down the Cline – the Nature of Grammaticalization, 17–44. Amsterdam: John Benjamins. Heine, Bernd 2002 On the role of context in grammaticalization. In: Ilse Wischer and Gabriele Diewald (eds.), New Reflections on Grammaticalization, 83–101. Amsterdam: John Benjamins.

556

Graeme Trousdale

Hilpert, Martin 2008 Germanic Future Constructions: A Usage-based Approach to Language Change. Amsterdam: John Benjamins. Himmelmann, Nikolaus P. 2004 Lexicalization and grammaticization: opposite or orthogonal? In: Walter Bisang, Nikolaus Himmelmann and Björn Wiemer (eds.), What Makes Grammaticalization – A Look from its Fringes and its Components, 19–40. Berlin: Mouton de Gruyter. Hopper, Paul J. and Janice Martin 1987 Structuralism and diachrony: the development of the definite article in English. In: Anna Giacalone Ramat, Onofrio Carrubo and Giuliano Bernini (eds.), Papers from the 7th International Conference on Historical Linguistics, 295– 304. Amsterdam: John Benjamins. Hopper, Paul J. and Elizabeth Closs Traugott 2003 Grammaticalization, 2nd edition. Cambridge: Cambridge University Press. Huddleston, Rodney and Geoffrey K. Pullum 2002 The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press. Langacker, Ronald W. 1977 Syntactic reanalysis. In: Charles N. Li (ed.), Mechanisms of Syntactic Change, 57–139. Austin: University of Texas Press. Langacker, Ronald W. 1987 Foundations of Cognitive Grammar, Volume I: Theoretical Prerequisites. Stanford: Stanford University Press. Langacker, Ronald W. 1990 Subjectification. Cognitive Linguistics 1: 5–38. Langacker, Ronald W. 1991 Foundations of Cognitive Grammar, Volume II: Descriptive Application. Stanford: Stanford University Press. Langacker, Ronald W. 2008 Cognitive Grammar: An Introduction. Oxford: Oxford University Press. Lehmann, Christian 1995 (1982) Thoughts on Grammaticalization. Munich/Newcastle upon Tyne: LINCOM Europa. Lehmann, Christian 2002 New reflections on grammaticalization and lexicalization. In: Ilse Wischer and Gabriele Diewald (eds.), New Reflections on Grammaticalization, 1–18. Amsterdam: John Benjamins. Lehmann, Christian 2004 Theory and method in grammaticalization. In: Gabriele Diewald (ed.), Grammatikalisierung. Special issue of Zeitschrift für Germanistische Linguistik 32: 152–187. Lopez-Cóuso, María José 2010 Subjectification and intersubjectification. In: Andreas H. Jucker and Irma Taavitsainen (eds.), Historical Pragmatics, 127–163. Berlin: Mouton de Gruyter.

Grammaticalization, lexicalization and constructionalization

557

Lyons, John 1982 Deixis and subjectivity: Loquor, ergo sum? In: Robert J. Jarvella and Wolfgang Klein (eds.), Speech, Place and Action: Studies in Deixis and Related Topics, 101–124. New York: Wiley. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech and Jan Svartvik 1985 A Comprehensive Grammar of the English Language. London: Longman. Sweetser, Eve 1990 From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure. Cambridge: Cambridge University Press. Terkourafi, Marina 2011 The pragmatic variable: toward a procedural interpretation. Language in Society 40: 343–72. Traugott, Elizabeth Closs 1989 On the rise of epistemic meanings in English: an example of subjectification in semantic change. Language 65: 31–55. Traugott, Elizabeth Closs 1995 Subjectification in grammaticalization. In: Dieter Stein and Susan Wright (eds.), Subjectivity and Subjectivisation: Linguistic Perspectives, 31–54. Cambridge: Cambridge University Press. Traugott, Elizabeth Closs 1999 The rhetoric of counter-expectation in semantic change: a study in subjectification. In: Andreas Blank and Peter Koch (eds.), Historical Semantics and Cognition, 177–196. Berlin: Mouton de Gruyter. Traugott, Elizabeth Closs 2007 The concepts of constructional mismatch and type-shifting from the perspective of grammaticalization. Cognitive Linguistics 18: 523–557. Traugott, Elizabeth Closs 2008a Grammaticalization, constructions and the incremental development of language: suggestions from the development of degree modifiers in English. In: Regine Eckhardt, Gerhard Jäger and Tonjes Veenstra (eds.), Variation, Selection, Development: Probing the Evolutionary Model of Language Change, 219–250. Berlin: Mouton de Gruyter. Traugott, Elizabeth Closs 2008b The grammaticalization of NP of NP constructions. In: Alexander Bergs and Gabriele Diewald (eds.), Constructions and Language Change, 21–43. Berlin: Mouton de Gruyter. Traugott, Elizabeth Closs 2010a (Inter)subjectivity and (inter)subjectification – a reassessment. In: Kristin Davidse, Lieven Vandelanotte and Hubert Cuyckens (eds.), Subjectification, Intersubjectification and Grammaticalization, 29–71. Berlin: Mouton de Gruyter. Traugott, Elizabeth Closs 2010b Grammaticalization. In: Silvia Luraghi and Vit Bubenik (eds.), The Continuum Companion to Historical Linguistics, 269–283. London: Continuum Press. Traugott, Elizabeth Closs 2010c Contexts and constructionalization. Plenary paper, International Conference on Construction Grammar 6, Prague, September 2010.

558

Graeme Trousdale

Traugott, Elizabeth Closs and Richard B. Dasher 2002 Regularity in Semantic Change. Cambridge: Cambridge University Press. Traugott, Elizabeth Closs and Graeme Trousdale 2010 Gradience, gradualness and grammaticalization – how do they intersect? In: Elizabeth Closs Traugott and Graeme Trousdale (eds.), Gradience, Gradualness and Grammaticalization, 19–44. Amsterdam: John Benjamins. Trousdale, Graeme 2008 Constructions in grammaticalization and lexicalization: evidence from the history of a composite predicate construction in English. In: Graeme Trousdale and Nikolas Gisborne (eds.), Constructional Approaches to English Grammar, 33–67. Berlin: Mouton de Gruyter. Trousdale, Graeme 2010 Issues in constructional approaches to grammaticalization in English. In: Katerina Stathi, Elke Gehweiler and Ekkehard König (eds.), Grammaticalization: Current Views and Issues, 51–72. Amsterdam: John Benjamins. Wischer, Ilse 2000 Grammaticalization versus lexicalization – “methinks” there is some confusion. In: Olga Fischer, Anette Rosenbach and Dieter Stein (eds.), Pathways of Change: Grammaticalization in English, 355–370. Amsterdam: John Benjamins.

20.

Sociopragmatics of language change Terttu Nevalainen

1.

Introduction

In the Notes and Queries section of The Guardian Weekly (9 April, 2010), a reader raised a simple but intriguing question: “[w]hy do some expats retain the accent of their place of origin, while others adopt the accents of their new surroundings”. The three replies to this query, excerpted in (1) to (3), refer to some of the basic issues to be addressed in this chapter. (1) Some would argue that people hang on to accents from their native country deliberately as a kind of identity statement, and that may be so in many cases. (D.W., Queensland, Australia) (2) I often find myself mimicking whomever I work at communicating with: for example, I have New York vowels for tourists in my city, Strine on Skype for my Aussie friends […] and current slang with my son’s friends. Surely such is simply an attempt to meet people half-way? (D.S., Toronto, Canada) (3) As Eliza Doolittle found out, it may improve your acceptability among your neighbours in some circumstances to adopt their speech patterns. In Canada there seems to be no great social pressure to change. Most people quite unconsciously imitate the speech of those around them to some extent … (J.D., Halifax, NS, Canada) Judging by these replies, some of the processes involved in accent preservation and change are directly observable and even subject to a degree of self-monitoring. The first response suggests that people may consciously reject new influences because they want to preserve their own linguistic identity. As discovered by the second writer, being able to adopt new speech patterns is based on interaction, speakers quasi-automatically accommodating their usage to that of their interlocutors. The third reply makes the point that the motives for processes of accommodation are complex, ranging from conscious attempts to improve one’s social acceptability to unconscious reactions to others in an environment with a minimum of social pressure to conformity. There has been a good deal of scholarly discussion about the above issues and their relation to language variation and change. Over the last forty years, research has produced a number of “subtheories” (Hudson 1996: 228) or “socially relevant mini-theories” (Coupland 2001: 6) which address these particular areas. They include the accommodation, acts of identity, and social network theories of linguistic

560

Terttu Nevalainen

choices, and the politeness or ‘face’ theory of interaction. These highlight the cross-disciplinary nature of sociolinguistics and pragmatics in that they draw on various neighbouring disciplines, ranging from social psychology (accommodation theory, social identity theory) to sociology (social network theory). They all relate to various aspects of language use, and are of relevance to any sociopragmatic account of language change. In this chapter I will specifically concentrate on factors that have been associated with language change and its diffusion in the community, and the theoretical models that have been proposed to account for these processes.

2.

Approaches to language change

Linguists approach the motives of language change from a variety of different angles. One of the traditional arguments concerns the extent to which language change can be accounted for in terms of language-internal factors, or purely endogenous growth, as opposed to language-external factors arising from the agency of language users. The proponents of endogeny aim to provide maximally economical accounts of language change, typically in terms of its end-products rather than the processes leading to them. The actuation of language change is accounted for by language-internal argumentation and idealized linguistic data (cf. Milroy 2003: 148–150). Those willing to debate the issue always find it safer not to assume external influence – unless there is clear evidence to the contrary – and criticize sociolinguists for their eagerness to correlate social and linguistic factors. Lass (1997: 373–374) argues that the fact that something, such as a pattern of variation or a change, is manifested in society does not necessarily make it ‘social’. The observed correlation could simply be a contingent fact. He also protests against the view that the propagation of a change or maintenance of a norm should in some sense be based on constant processes of speakers negotiating consensus in face-toface interaction. Contingent facts apart, considering language-internal developments in the light of external factors provides more comprehensive accounts of language change, to which various usage-based models of language contribute. Those who analyse the role of social factors emphasize the agency of speakers in language change: language does not change on its own; it is always speakers who change their language (e.g. Milroy 1992, 2003). A basic distinction can be made between speaker innovation and language change. Speaker innovation is an act of the individual speaker, whereas language change takes place within the language system. For an innovation to become a language change it has to be adopted by other speakers. This requirement makes language change ultimately a social phenomenon. A related distinction is made between the actuation of linguistic change and its propagation across the community. The latter has occupied the centre stage in so-

Sociopragmatics of language change

561

ciolinguistic research, in variationist studies of both the present and the past. By contrast, major challenges are posed by actuation, the fact that a linguistic feature changes in a particular language at a given time, but not in other languages with the same feature, or in the same language at other times (Weinreich, Labov, and Herzog 1968: 102). Unlike the propagation of a change, its actuation may not be accessible with the standard methods of variationist sociolinguistics. Depending on the level of abstraction of the discussion, the approaches adopted to shed light on the actuation problem may range from contact linguistics to typology (Herold 1997; De Vogelaer 2006). The actuation problem can also be recast in terms of speaker innovation and language change (Milroy and Milroy 1985). Ever since Weinreich, Labov, and Herzog’s classic article, observing a linguistic innovation take place has been recognized as one of the hardest areas for social models of language change to come to grips with. As Milroy (1992: 169–170) points out, investigators may be able to observe something close to speaker innovation, but they have no way of predicting whether it will become a language change. In quantitative studies, the initial stages of an incipient change may be too easily dismissed as ‘noise’. A major issue for research is, therefore, to develop models of social interaction that will help us better understand the conditions in which an innovation may be adopted and become a change. The study of accommodation in social networks of different kinds, including coalitions and communities of practice, has become a common cross-disciplinary way of modelling some of these conditions. In addition, other models, particularly those of politeness and social identity, have provided answers to questions like why innovations are sometimes adopted as changes and sometimes rejected (Milroy 1992: 221; Milroy and Gordon 2003).1 There are many conceptualizations of sociopragmatics, both speaker- and message-oriented, and those which highlight interaction. These rarely address the issue of language change directly in structural terms. However, specializations like historical sociopragmatics have a diachronic dimension by definition, although they are not restricted to diachrony and their focus need not be on language change (e.g. Archer 2005: 4–8). Culpeper (2009: 182) writes that “historical sociopragmatics concerns itself with any interaction between specific aspects of social context and particular historical language use that leads to pragmatic meanings”. This approach stresses the contextualization of the social dimensions of language use rather than the language-internal changes that the emerging pragmatic meanings may give rise to. The theoretical notions that inform current historical sociopragmatic work include (1) activity types, frames and genres, (2) situated roles and identities, (3) relational notions such as “face”, “face-work”, power relations, social distance, affect, rights and obligations, and (4) attitudes and opinions (Culpeper 2009: 181). These notions cover part of the same ground as the theoretical frameworks listed by Hudson (1996) and Coupland (2001) and the models outlined by Milroy (1992, 2003).

562

Terttu Nevalainen

Since the focus of this chapter is on language change, I will discuss frameworks that have been developed for modelling the circumstances in which speaker innovation may become language change. My focus will be on language change from the perspectives of language users and the contexts and interactional settings of language use on both the micro- and macro-levels. Most of the research to be discussed has been carried out on two languages which differ from each other in many respects: English, the current lingua franca with a long written record, and Finnish, a non-Indo-European language with some 5.5 million speakers and a much shorter literary history. After considering the fundamental sociolinguistic issue of who, in principle, can change language (section 3), I introduce the theory of accommodation, which provides a basis for understanding how an innovation can be adopted and propagated through social networks (section 4). In the subsequent sections the discussion is focused on social evaluation and identity projection (section 5), and a comparison of these competing approaches, combined with the cognitive issue of conscious awareness in language change (section 6).

3.

Who can change language?

The pivotal role that children and adolescents play in language change is usually taken for granted (e.g. Labov 2001), but they are in fact much less frequently the object of study in sociolinguistic research than adults.2 Similarly, while the sociolinguistic, pragmatic, and cognitive theories developed to account for the motivations of linguistic variation and change rarely address child language use, sophisticated models are constructed to account for the sociopragmatics of interadult communication. This means an implicit acknowledgement of the role of adults in language variation and change. The current understanding of language change readily admits both generational and lifespan changes, in which “individual speakers change over their lifespans in the direction of a change in progress in the rest of the community” (Sankoff 2005:1011). This approach accounts for ongoing communal changes but is neutral with respect to the origins of linguistic innovations. In this chapter we are interested in the conditions under which speaker innovation may become language change. A good linguistic starting point for the discussion is offered by Kerswill (1996), who is interested in the types of linguistic feature an individual can acquire at different ages and, consequently, the amount of influence people at different ages exert on the speech of others. He comes to the conclusion that: “[N]either children nor adults can be seen as the sole originators of change: the picture is complex, as it turns out that people of all ages can (and do) modify and restructure their language – though exactly what they can change is to some extent age-related” (Kerswill 1996: 178). The wealth of sociolinguistic literature makes it possible to chart the kinds of change that can be expected to take place not only in childhood or adolescence but

Sociopragmatics of language change

563

also in adulthood. Table 1, based on Kerswill (1996: 200), shows a hierarchy of the difficulty of acquiring second-dialect features. The table suggests that, except for certain phonological and grammatical rules and systems, most features can be acquired throughout the lifespan, albeit they can be expected to vary in difficulty and hence likelihood of adoption. If changes can also take place throughout the lifespan as suggested, we can realistically discuss them in terms of sociopragmatic models designed on the basis of adult communication. Table 1.

Difficulty/ease of acquiring second-dialect features (modified from Kerswill 1996: 200)

Difficulty (1 most, 9 least difficult)

Feature

1–2

lexically unpredictable phonological rules grammatical change: parameters new phonological oppositions, prosody

3–4

5 6–7 8–9

Age acquired

by 3 ( ? ) by 8 ( ? ) by 3–15

grammatical change: new morphological classes morphologically conditione changes

adolescence, lifespan? not before 4–7; then lifespan

recategorialization of words and lexical sets

lifespan

phonological mergers exceptionless phonological shifts

lifespan

lexical diffusion of phonological changes lexical productivity, new vocabulary

4.

Accommodation

4.1.

General distinctions

lifespan

One of the interpersonal models that studies of language change commonly refer to is the Communication Accommodation Theory (CAT). It is associated with the work of the social psychologist Howard Giles and his co-workers (cf. e.g. Giles, Coupland, and Coupland 1991). Accommodation processes are found to take place with respect to all aspects of face-to-face interaction, including gaze, eye contact, proximity, body movement and speech tempo. As far as language is concerned, no essential difference is observed between grammatical and phonological variables. Accommodation is often associated with convergence – the interlocutors’ manner

564

Terttu Nevalainen

of communication becoming more similar – but divergence is equally natural. Studies on accommodation ask questions such as who accommodates to whom, and how. Linguists find the why-question particularly intriguing, i.e. “why people modify their language in the presence of others in the way and to the extent that they do” (Trudgill 1986: 2). Giles, Coupland, and Coupland (1991: 11) present an overview of the distinctions that have been found to characterize linguistic and language-related processes of convergence and divergence. These range from varietal choices to the speakers’ awareness of the extent to which they modify their speech patterns. Accommodation can take place upwards, towards a consensual prestige norm, or downwards, towards stigmatized varieties. Convergence is interpreted in terms of a speaker’s or speaker group’s need for greater social integration or identification and hence a means of increasing their attractiveness and gaining approval. Everyday convergence in social networks emerges as the breeding-ground for long(er)term shifts in individual and group-level linguistic usage (see 4.3.). Interactants need not accommodate to the same extent but a distinction can be made between symmetrical as opposed to asymmetrical accommodation. Giles, Coupland, and Coupland (1991: 20) point out that group accommodation is often asymmetrical and unilateral towards a power source. The less powerful tend to adjust to those perceived as more powerful. By contrast, divergence may reflect the strategy of the speaker identifying with the communicative norms of some reference group outside the communicative situation. Accommodation is, however, usually partial: there is no full convergence on all levels and features. Interlocutors can converge on some features but diverge on others. The degree of accommodation can also vary quantitatively. For example, convergence on speech rate and lexical diversity can be quantified and characterized as large or moderate. Accommodation can be unimodal or multimodal, involving, besides the language code, voice qualifications such as laughter. The interactants’ own subjective evaluations may vary as to the degree to which they modify their speech patterns. These subjective evaluations may differ from what can in actual fact be observed to take place. However, Giles, Coupland, and Coupland (1991: 24) point out that it is noteworthy that much communication is both produced and received at low levels of awareness. Even code-switching can take place below the level of the interactants’ conscious awareness. Language change from above and below the level of social awareness is one of the key notions found in sociolinguistic literature on language change; it is discussed in section 6.3. 4.2.

Short- and long-term accommodation

As the survey in 4.1 indicates, accommodation is characterized by complex and variable patterns of communicative interaction, which basically enable the actu-

Sociopragmatics of language change

565

ation of language change. Many of these characterizations are the subject of other frameworks relevant to the sociopragmatic motivations for language change. Before moving on to those, the distinction between long-term as opposed to shortterm accommodation needs to be established. At the level of the individual, the difference was illustrated by examples (1) and (2) in section 1. The assumption is that it is continued, everyday accommodation processes that form the context for language change, rather than short-term contacts in brief interactional episodes. The interaction of individuals in social networks provides the immediate social context in which language change can be actuated. Table 2, based on Auer and Hinskens (2005), summarizes the different aspects of the process. Table 2.

From short-term accommodation to language change (modified from Auer and Hinskens 2005: 336)

Micro-level:

Individual (interactional episode): short-term accommodation ↓ Individual (the individual): long-term accommodation ↓

Meso-level:

Networks: incipient language change ↓

Macro-level:

Speech community: language change

Migration provides a natural context for testing the effects of long-term accommodation. Population movements provide some of the most striking laboratory cases for establishing the conditions in which language change is likely to take place, including comprehensive dialect shifts. I will illustrate the outcomes of internal migration by discussing four studies, three of which are examples of (partial) linguistic convergence, and one of divergence. They illustrate both the short- and long-term consequences of catastrophic events such as the Second World War, which can be seen as catalysts for the demographic conditions behind these linguistic changes. Migration in Texas and Oklahoma Bailey et al. (1996) and Bailey (2002) discuss the effects of the rapid acceleration of urban growth in Texas and Oklahoma as a result of World War II in the light of three large-scale population surveys, which represent a cross-section of the population. Mobilization for the war involved the expansion of the industrial base and dramatically altered the patterns of migration in this part of the United States. The result was a large-scale urbanization and redistribution of the Southern population. These demographic developments were accompanied by significant social

566

Terttu Nevalainen

changes, for example in education and employment opportunities, which intensified the day-to-day contacts among the growing urban population in the second half of the 20th century. The linguistic consequences of these developments were basically of two kinds. A number of distinct local features were levelled in Texas and Oklahoma. In both states, the post-war population has increasingly abandoned features such as r-lessness in post-vocalic contexts in words like forty, and replaced local lexis by standard expressions, such as snap bean by green bean. At the same time, the unrounding of the vowel in words like hawk, lost and walk, and yod-dropping in words like Tuesday are spreading. Bailey et al. (1996: 443) found that recent arrivals in both Texas and Oklahoma were leading these developments towards dialect levelling. The authors suggest that the local variety is losing its distinctiveness and becoming more like General American. This is not the whole story, however. Certain distinctive Southern features such as the plural pronoun yall and the quasi-modal fixin to (marking the immediate future) continue to be used by natives, but they are also adopted by newcomers. In view of the relative ease of acquiring lexical features in a second dialect, this is not surprising. These features are not, however, distinct to Texas and Oklahoma, but are expanding throughout the South. On the other hand, features such as the double modal might could and the monophthongal realization of the vowel in words like night are spreading, especially among the native population in Texas. The latter feature correlates strongly with the positive ratings the respondents gave to the question how they viewed Texas as a place to live. The authors conclude that the primary motive for the expansion of monophthongal /ai/ is identity with Texas as a place to live. These findings suggest that the post-war history of Oklahoma and Texas is largely, but not entirely, one of convergence of the native and newcomer populations in increasingly urban environments. Migrants from Karelia The resettlement of the Finnish Karelian population – 400,000 people – from the area ceded to the Soviet Union after the Second World War is similarly a story of linguistic convergence but one in which the Karelian settlers asymmetrically converged with the Finnish mainstream and local varieties. Leskinen (1980) investigated the linguistic accommodation of Karelian settlers from three regions in three southern Finnish communities. His results reveal the basic patterns of Karelians’ shifting to Finnish: the younger settlers accommodated more than the older, older men more than older women, and those in urban environments more than those in sparsely-populated areas. Those who retained their native variety best were settled in areas where Karelians were in the majority. As is typical of dialect contact situations, the most marked Karelian dialect features disappeared first while those shared with the variety spoken in the area of settlement were retained, although

Sociopragmatics of language change

567

both were nonstandard. The features that were among the first to be lost include the labialized 3rd person present tense suffix: juoksoo (‘she/he/it runs’) as opposed to the general Finnish juoksee, and the syncopated participial form antant (‘given’) as opposed to antanut. By contrast, features that were preserved the longest include, for example, unstressed pronominal forms such as mie (‘I’; for minä) and sie (sg. ‘you’; for sinä); these forms were also found in the local dialect of one of the regions of immigration studied. Migrants to Helsinki The opposite method was used by Nuolijärvi (1994), who investigated the linguistic effects of migration into the Finnish capital, Helsinki, from two distinct rural areas between 1965 and 1974. The subjects were all born between 1943 and 1950, and had moved to Helsinki from southern Ostrobothnia and northern Savo. The post-war decades comprised an intense period of urbanization in Finland, and internal migration to the capital peaked in the decade studied. Nuolijärvi found that all those who moved to the city were affected, and acquired a form of language that differed in various ways from the language of their parents. Analysing the migrants, she detected a strong effect of gender: women accommodated more than men, and young women in particular adapted to the way of speaking of their new environment or used dialect-neutral forms. The prestige of the dialect of origin also mattered. Speakers with a western, Ostrobothnian dialect background generally retained their native dialect to a greater extent than those coming from the more negatively stereotyped Savo dialect region of eastern Finland. Some dialect features were preserved better than others, even among Ostrobothnian migrants; well preserved were, for example, the epenthetic vowel in words like kylymä (‘cold’, for the common colloquial and standard kylmä) and the intervocalic /h/: talohon (‘to the house’) as opposed to taloon. At the other extreme, those migrants from Savo who adapted most used hardly any dialectal forms at all; the few exceptions include the occasional use of the geminate in forms like makkaa (for makaa, ‘she/he/it lies’). Nuolijärvi (1994: 167–169) concludes that typical adapters aimed at a balance of standard and common colloquial forms, a compromise language that could be used in any situation. African American migration to the North The three studies discussed show that it is possible to trace the linguistic effects of migration on both the community and the migrants, and make assumptions about the motive forces behind them. Social dialectologists relate them to social motivations and identity practices, typically, social integration combined with continuity in life (e.g. Andersson and Thelander 1994; see further section 5). Although processes of change can be observed, in most cases the question of their actuation

568

Terttu Nevalainen

and transition from the individual to the community remains unanswered. This issue is addressed by van Herk (2008), who is looking for the linguistic triggers of the Northern Cities Shift (NCS), an ongoing chain shift in vowel realizations in urban areas between Madison, Wisconsin and upstate New York (Labov 1994). The questions that occupy van Herk (2008: 155) include: how might local identity practices relate to similar linguistic processes across such a large area? Why should a vowel subsystem that has remained stable for many centuries suddenly shift? Why should only part of the area with the appropriate prior linguistic system be involved, and why should the shift be absent or delayed among African Americans, rural speakers or Canadians? The answer van Herk (2008) proposes suggests divergence of the native and the migrant populations as the motivating force that triggered the process in these cities. The Great Migration of millions of African Americans from the South to the cities in the North between the end of World War I and 1960 substantially altered the ethnic composition of these cities. Van Herk proposes that the NCS was triggered by white residents wishing to differentiate their speech from that of the migrants. The evidence van Herk draws on in support of his argument comes from U.S. Census data and descriptions of the Great Migration. Comparing the original vowel systems of the relevant areas with the current ones, he reports strong correlations between participation in the NCS and the speed and degree to which African American populations increased in these communities. Van Herk finds that these correlations provide a good match between the chronological, social and geographical boundaries of the NCS, but acknowledges that more work is needed to place this differentiation process in a broader varietal context and to gain a better understanding of the identity issues involved. 4.3.

Social networks as vehicles for language change

It was pointed out above that social networks provide the breeding-ground for long-term shifts in individual and group-level linguistic usage. Milroy and Gordon (2003: 117) define a social network as “the aggregate of relationships contracted with others” and go on to specify it “as a boundless web of ties which reaches out through social and geographical space linking many individuals, sometimes remotely”. Used to capture the dynamics of speakers’ interactional behaviours, social networks are typically characterized in terms of their structure and content. Direct contact between individuals creates first-order ties, whereas in second-order ties the link is indirect. The density of network ties varies from high to low, and their strength from strong (e.g. between friends and kin in the first-order zone) to weak (between acquaintances). The analysis of social networks is contextualized within macro-level social frameworks, but these frameworks are backgrounded in order to focus on less abstract modes of analysis. Although the group distinctions, or speaker variables, which informed the first linguistic applications of the network

Sociopragmatics of language change

569

model were not interpreted in terms of identity (Milroy 1987: 114–116), social networks have been proposed as the ‘meso-level’ context for identity projection (Auer and Hinskens 2005: 336; see section 5, below). Social network differences have been used to explain why some communities maintain their linguistic norms while others more readily accept external influences. Strong network ties serve as a norm-enforcement mechanism, while mobility typically loosens network ties, and loose-knit groups in general are more susceptible to innovations. People with many weak ties do not necessarily change their own idiolects but are ideal carriers of innovations to other groups and communities. According to Milroy (1987: 202), the people who introduce a new form into a network tend to be persons who are weakly linked to the group, while the early adopters of these innovations are central members of the group, and conform closely to the group norms. In a diachronic network study, Sairio (2009) reconstructs the social networks of Elizabeth Montagu, the renowned London Bluestocking hostess, in four time periods between 1738 and 1778. The question she sets out to answer is the extent to which these networks, as evidenced in the personal correspondence of this literary social circle, transmit linguistic influence. The linguistic features considered include the progressive aspect, preposition stranding vs. pied-piping, and spelling variation in verb forms. In-group network practices could be detected in the use of the progressive and, to some extent, in socially controlled epistolary spellings. However, Sairio (2009: 311–313) concludes that when network ties are strong, linguistic variation is more satisfactorily explained by informants’ social background, especially their gender, but also their social class and register variation. These results connect back to issues to do with language users’ multiple social category characteristics and their variable linguistic indexing, to be discussed in section 6.3.

5.

Social evaluation and identity

5.1.

Acts of identity

One of the influential frameworks that have been put forward to account for linguistic processes of identity projection is the acts of identity model (Le Page and Tabouret-Keller 1985). It is informed by the social identity theory, originally developed by Henri Tajfel and his associates, which attributes to individuals multiple personal and social identities, ranging from the unique to the collective. An individual’s group identities typically include gender, occupation, ethnicity, locality, religion, social class, and nationality, whereas personal identities vary contextually depending on the person’s own knowledge of the membership and the value and emotional significance they attach to it (e.g. Joseph 2004: 76–77). The investi-

570

Terttu Nevalainen

gation of social identities and style is often proposed as a remedy for the neglect in language variation studies of social meaning and the ways in which it is created through language (e.g. Auer 2007; Coupland 2007). The acts of identity model was originally developed to account for language variation and choice in creole contact situations, notably those in Belize and St Lucia in the Caribbean, but also elsewhere. Le Page and Tabouret-Keller (1985: 182–184) suggest that the most important constraint governing linguistic behaviour is individual language users’ positive motivation to behave according to the behavioural patterns of groups they wish to identify with, or negative motivation to dissociate themselves from groups with which they do not wish to be associated. Positive acts bring about linguistic convergence, contributing to an in-group identity in discourse. These acts of identity, as it were, hold the language together. However, for this to happen, the speakers must be able to identify other groups, have access to these groups and be capable of analysing their linguistic behaviour; they must also be motivated to ‘join’ the groups, and have the ability to modify their own behaviour (Le Page and Tabouret-Keller 1985: 182). Talking about language evolution, Le Page and Tabouret-Keller (1985: 200–201) contrast diffuseness, typical of pidgins, which arises from language contact, with focusing, typical of the creole end of the continuum, where the language becomes highly regularized. They note that focusing is a ‘normal’ language situation, in which the linguistic creativity of contact is constrained by daily interaction and the norms of the different social groups become marked with the attributes of those groups. A prestige group’s norm may then be institutionalized as the standard against which the usage of other groups is measured. Rise of post-colonial Englishes Social identity has been used to account for processes of language variation and change at both macro- and micro-levels. At the macro-level, the rise of a new collective identity plays a key role in Schneider’s (2007) five-stage dynamic model of the evolution of postcolonial Englishes. At the foundation stage English is brought to a non-English-speaking territory, where it begins to be used on a regular basis. At this stage the settlers and native population perceive themselves as distinct from the ‘other’ and their relations can vary from friendly to varyingly hostile, as in Singapore and North America, respectively. Their contacts may give rise to dialect levelling, incipient pidginization, and toponymic borrowing. In American English, Schneider (2007: 254–264) argues, this stage stretches from the earliest colonization attempts in 1587 to the stabilization of the first colonies by c. 1670. In the next stage, exonormative stabilization, the settler language becomes the language of administration and education in at least some geographical regions or social strata. Locally born generations develop a hybrid identity. Bilingualism spreads among the indigenous population, and intensified contacts between the

Sociopragmatics of language change

571

settlers and the natives result in lexical changes but may also affect other linguistic levels. In American English, this phase lasted about a hundred years, from the last decades of the 17th century to the Boston Tea Party in 1773, which led up to the colonists declaring independence in 1776 (Schneider 2007: 264–274). The nativization stage is marked by the settlers gradually breaking away from the mother country, both politically and linguistically. New identities are being constructed, and despite any social and cultural differences, both immigrant and indigenous population groups consider themselves permanent residents of the territory. Mutual linguistic accommodation results in a communal language which is a second language for some and a first language for others, but which amalgamates features from the prolonged bilingual contact. Mixed codes of varying stability characterize this nativization stage, and often serve as identity carriers. For American English, the period from 1773 to the first decades of the 19th century witnessed the growing self-confidence of the common American (Schneider 2007: 273–283). The fourth stage, endonormative stabilization, requires both political independence and cultural self-reliance, and is characterized by the acceptance of the local linguistic norm also in formal contexts. The language variety begins to be codified, and the first dictionaries are compiled. In American English, this happened in the course of the 19th century, which was also marked by the flowering of American literary creativity. Schneider (2007: 282–291) dates the sociopolitical cut-off point of the phase to 1898, the Spanish-American War, which heralded the beginning of a growing American presence in world politics. The ultimate stage, when the post-colonial variety is taken for granted, is characterized by differentiation. Various group identities and finer social categorizations override the earlier collective identity of the population and subvarieties emerge, regional and social variation normally increasing with time. This has been happening in American English since the turn of the 20th century, as the linguistic landscape of the United States has continued to diversify as a result of urbanization, increased migration and growing ethnic diversity (Schneider 2007: 291–308; for illustrations, see 4.2.). Rise of the Finnish literary language The recorded history of the Finnish language has been greatly affected by the geopolitical position of its speakers between Sweden and Russia. The Finnish-speaking territory was under Swedish rule for five hundred years until 1809, when it became a self-governing Grand Duchy within the Russian Empire. In 1917 Finland gained its current status as an independent republic. The first texts published in Finnish go back to the 16th century, with the translation of the Bible as the literary landmark of the period, and until the early 19th century, Finnish literature largely consisted of religious texts. Swedish was the language of the administration, education and culture of the realm, and also continued as an official language in Fin-

572

Terttu Nevalainen

land after 1809. Finnish, the mother tongue of the great majority of the population, was granted equal status with Swedish in 1863. By the end of the 19th century Finnish had evolved into a fully-fledged literary language and was used in public life, education, culture and science (e.g. Nordlund 2007: 229–231). Historians argue about the territorial identity that Finns had under the Swedish rule, but are in agreement that in the 19th century linguistic focusing took place in the context of nation building and the construction of a national identity (e.g. Klinge 1990; Nordin 2000; Huss and Lindgren 2010). The establishment of the Finnish literary language in the course of the 19th century was not the result of a prestige group’s usage being institutionalized as a standard, but a conscious creation of the country’s academic elite, educated in Swedish. The earlier literary language, Biblical Finnish, was based on south-western Finnish dialects and heavily influenced by Swedish. Hence, the speakers of other dialects found themselves in a situation that Nordlund (2007) refers to as double diglossia, diglossia on both the national (Finnish vs. Swedish) and local level (vernacular vs. Biblical Finnish). One of the issues that gave rise to heated debates was whether the new literary language should be built on western or eastern dialects. The resulting written standard was a compromise, a mixed dialect regulated from above, which was subsequently taught at school, used in print, and codified in grammar books and dictionaries. Today’s usage continues to be monitored by The Finnish Language Office. 5.2.

Speaker dynamics

From the speaker’s viewpoint, the linguistic outcomes of accommodation are connected with varying levels of situational indeterminacy and uncertainty between interlocutors. Meyerhoff (2001) observes that it is not always clear what the salient features are, and relates degrees of certainty between interlocutors to social psychological motivations for language variation. The notion of certainty has to do with the interlocutors’ identities, positions and motivations, and the ways in which these are linguistically indexed. In-group and out-group contrasts continue to account for a large part of this linguistic variation. A high degree of situational certainty is required for speakers to be able to accentuate in-group similarities and out-group distinctions, in Meyerhoff’s terms, maximizing their fit while at the same time maintaining their individual distinctiveness, or as noted above in 4.2, combining social integration with continuity in life. Approaching the issue from the perspective of linguistic variation, Sankoff (2004: 136) points out that phonological and phonetic features can vary in three respects: the extent to which they carry social meaning, the degree to which speakers are conscious of such variation, and the extent to which speakers can control this level of variation in their speech. Adopting a given feature depends on all three. In situations where the social and linguistic indexing of variant forms is mutually recognized, participants can accentuate the positive, and by doing so accrue

Sociopragmatics of language change

573

social capital, or eliminate the negative and avoid or minimize risk. In order for this to happen, however, they must also recognize that the situation is unstable and that the processes by which social and linguistic variables index each other are not fixed (Meyerhoff 2001: 68). Where there is uncertainty among the interlocutors as to salient group identifications, speakers can test their hypotheses. These various degrees of shared understanding are summarized in Table 3. How they may relate to language change will be discussed next. Table 3.

Degrees of situational certainty reflected in motivations for language variation (based on Meyerhoff 2001: 70)

Situational certainty

Situational instability

Epistemological uncertainty

maximize fit be distinct

accrue social capital avoid risk

test your hypothesis

Maximizing fit The studies discussed in 4.2. are mostly marked by situational certainty. Migrants are by definition not local and their speech can index them as such. Moreover, if they constitute a minority group in their new environment, they are apt to adapt to the majority rule. In both studies on Finnish, those who accommodated became more plurilectal, and contributed to the diffusion of the national standard language or the common colloquial norms, while these norms were not affected by the migrant input. Marked with more situational instability, and therefore probably also motivated by risk avoidance, the Texas and Oklahoma findings also showed dialect levelling and convergence to the general, supralocal American norm by the increasing urban population, native and migrant alike. Dialect levelling is a general trend in similar situations, and well documented for many localities in the past, for example, medieval Madrid and Paris, and 16th-century Amsterdam and London (Nevalainen 2011). Being distinct A number of studies report identity markers through which speakers can assert their loyalty to local values. This is typically the case in close-knit rural communities, where linguistic usage is focused, compared to large cities, which are linguistically more diffuse. One of the classic studies demonstrating this is Labov’s 1963 study of Martha’s Vineyard, which showed how the use of a centralized variant of the diphthong (ay) in words like I, night and wife was connected with indigenousness and native status as an islander. It characterized the speech of the local Chilmark fishermen, who were the most close-knit group on the island. They had a positive attitude to Martha’s Vineyard and were opposed to the incursion of the ser-

574

Terttu Nevalainen

vice economy and tourism on the island. Their increasing centralization of (ay) could be seen as a response to an external social force which threatened their social identity. The use of centralized (ay) and the concomitant centralization of (aw) also spread to other ethnic groups on the island. A study of the Northern Finnish variety spoken in the Tornio Valley reveals a similar strong identity projection. Vaattovaara (2009) studied local and linguistic identity in this region, and found that in the Tornio Valley the use of /h/ in non-initial syllables serves as a powerful local identity marker, which is on the increase among the younger speakers of both sexes. Those with a strong local identity use h-ful forms such as sauhnaan (‘to sauna’; standard Finnish saunaan) and kouhluun (‘to school’; standard Finnish: kouluun) much more frequently than those who identify weakly with the local community and prefer h-less variants. This identity marker is a metathetical development of the common northern Finnish use of /h/ in non-initial syllables, saunhaan and koulhuun, which goes back to earlier intervocalic forms, saunahan, kouluhun. The marker distinguishes the young people of the rural community studied, for example, from the usage of the nearby town of Kemi, and is associated with the genuine, friendly and hospitable inhabitants of the Tornio Valley.3 Accruing social capital and avoiding risk Group identities based on gender and social class have routinely been evoked in interpreting the results of variation studies and the instability of ongoing language change. As Meyerhoff (2001: 63) points out, the use of a particular linguistic variant is taken as “an indication of a speaker’s conscious or unconscious desire to access the privileges and social capital associated with the prototypical users of that variant”. While not denying that linguistic variation is rooted in social structure, frameworks focusing on identity construction typically avoid direct correlations between social structure and language use in terms of identity. Coupland (2007: 112–114) emphasizes the complexity of identity contextualization, and, as part of his model, recognizes three levels of discourse framing which are involved in determining how identities are made relevant or salient in discourse. Socio-cultural framing is placed at the macro-level, and includes constructs such as social class, gender, ethnicity and age, and the socio-cultural values they may index in a particular culture. Genre framing comes at the meso-level, and interacts with the wider socio-cultural frame, consolidating or possibly even contradicting identities foregrounded in that frame. Finally, interpersonal framing is located at the microlevel of individuals, who forge their personal and relational identities linguistically in subtle ways in their spoken interaction within a consolidated genre and community of practice. By using a certain feature, Coupland (2007: 114) notes, a speaker may present him- or herself, for example, as more or less powerful in a given relationship. In studies of ongoing language change, these notions of framing may be

Sociopragmatics of language change

575

used to analyse the ways in which social ambition is reflected in individual linguistic behaviour (Meyerhoff 2001). A diachronic case study: social aspirers in Tudor England Nevalainen and Raumolin-Brunberg (2003) studied a number of grammatical changes in Early Modern English using a corpus of personal correspondence as their data. Personal letters represent an interactive written genre with identifiable writers and recipients, which conforms to certain conventions, a number of them reminiscent of spoken interaction (forms of address, use of first- and second-person pronouns, etc.). The period studied, from the 15th to late 17th century, represents the era before the codification of Standard English, making it possible to explore issues like the indexing of place in the letter writers’ communication. The study of linguistic changes in progress established a set of regional patterns of diffusion of various incoming forms at the macro-level in real time. Despite historical sociolinguists’ limited access to the population at large due to the high levels of illiteracy, some patterns of social stratification could also be identified. The social distinctions established were based on the work of social historians and covered the literate section of the population who were able to communicate in writing. The distinctions made were thus analysts’ categories, but often also coincided with the writers’ self-identification, for example, as gentlemen. Social aspirers were men who in the highly stratified society of the time advanced to gentility or the upper clergy from their merchant, professional or other non-gentry backgrounds (Nevalainen and Raumolin-Brunberg 2003: 32–37). The results on Early Modern English social aspirers bore out previous findings indicating that upwardly mobile people tend to be sensitive to the social connotations of linguistic variants, promoting prestigious forms and avoiding those with low-status associations. Upwardly mobile men were found to lead the change away from the use of multiple negative elements in a clause (they never did nothing) in favour of a single negator followed by nonassertive indefinites (they never did anything; see Figure 1). This usage originated from Late Middle English administrative language and hence from educated circles. Throughout the 16th century, the lower-ranking writers in the corpus used the incoming pattern least. These findings on the decline of the use of multiple negation could perhaps be considered a feature of upwardly mobile men’s professional communities of practice rather than a systematic means of accruing social capital in written interaction. However, the origins of this change in the late 15th century were associated with the Royal Court, which may have indexed the incoming pattern socially. More evidence for these professional men’s linguistic sensitivity comes from the generalization of verbal s (she writes) and the decline of the equivalent th form (she writeth). Figure 2 shows that when the incoming suffix -s started to diffuse across the language community, it was promoted by lower-ranking writers and

576

Terttu Nevalainen

Figure 1. Use of single negation (%) vs. multiple negation (based on Nevalainen and Raumolin-Brunberg 2003: 146)

conspicuously avoided by the upwardly mobile coming from these ranks. Twenty years on, social aspirers were using the incoming form more than either upper- or middle-ranking writers – but still not as much as the lower ranking correspondents. Following Meyerhoff’s (2001) argument, the last decades of the 16th century may have presented more situational uncertainty, and hence resulted in avoidance on the part of this group, whereas this uncertainty appears to have been resolved by the early decades of the 17th century in favour of the incoming form.

Figure 2. Third-person singular suffix -s (%) vs. -th (based on Nevalainen and Raumolin-Brunberg 2003: 145)

Sociopragmatics of language change

6.

Competing and complementary models

6.1.

Identity projection or change-by-accommodation?

577

As a prerequisite for language change, accommodation is basically neutral as to the degree to which the process is automatic and mechanical as opposed to being a conscious choice on the part of interlocutors. However, there are various arguments that build up a tension between automaticity and intentionality (MendozaDenton 2002: 492). Auer and Hinskens (2005) find it useful to distinguish between the change-by-accommodation model, based on the frequency of direct interaction, and the identity-projection model. Reviewing a series of interactional studies they come to the conclusion that speakers tend to converge to a stereotype of an ideal speaker persona or a mental representation of a social group they want to be associated with rather than to their actual interlocutors. What they suggest is decisive for convergence to occur is not the frequency of interaction but one’s attitudinal orientation towards the group one wants to be associated with or dissociated from. Hypercorrection, overshooting the target, can then be explained by the lack of sufficient behavioural data. While they find some evidence for speakers accommodating to each other with respect to ongoing language change, Auer and Hinskens (2005: 351) subscribe to the basic tenet of variationist sociolinguistics that language change can only be observed at the level of the community. However, at the community level, studies focusing on migration patterns suggest that the frequency of interaction also matters. Van Herk (2008: 157) discusses what he calls flight models, that is, flight from stigmatized features. The Great Migration, discussed in section 4.2, above, would have been the type of process that could trigger heightened group identification. The increased salience of out-group members, by their sheer numbers, leads in-group members to focus on the linguistic features that they share with their own group, rather than on their internal differentiation. Studying geographical variation in African American English (AAE) in the early 21st century, Thomas and Yaeger-Dror (2010) report both divergence and convergence. The pattern that emerges shows that local African American English communities throughout the country have adopted phonological features of the locally dominant vernaculars while maintaining some distinctly nonlocal features of their own. These processes depend on various factors, such as population size, the degree of the integration/segregation of the communities, and individual speakers’ social networks. As community integration seems to have been reduced since the 1970s, Thomas and Yaeger-Dror (2010: 13) suggest that nonlocal features are probably maintained to index social identity. The division of labour between the change-by-accommodation and identity-projection models also causes controversy among sociolinguists who are interested in new dialect formation. It is widely recognized that “New Englishes”, for example, display considerably less regional and social variation than older var-

578

Terttu Nevalainen

ieties. While Schneider’s model of the rise of postcolonial Englishes (2007; section 5.1) builds on identity projection, Trudgill (2008) argues that many processes of new dialect formation represent dialect mixing and are the inevitable result of dialect contact. He supports his argument with case studies from the colonial expansions of various European languages. Instead of appealing to the role of national identities in this pre-16th-century period, Trudgill maintains that dialect mixtures must have resulted from quasi-automatic accommodation processes below the level of conscious awareness. 6.2.

Contact or imitation?

It has been suggested that perceived certainty of in-group and out-group contrasts, and the salience of the linguistic features involved, are also a precondition for processes like “dialect-hopping”. It is based on the gravity model, which specifies that an innovation may hop from one urban centre to another, bypassing the countryside in between. This model is based on ideas developed in economics, geography and sociology on how innovations spread, taking account of the population sizes of the communities in interaction and the distances between the urban centres (Chambers and Trudgill 1998: 166–186). Dialect hopping is seen as a means of aiding the diffusion of a linguistic innovation, as the features involved may be used to index urbanness or other such attributes related to innovation desirable to certain groups across the population. Trudgill (1986: 54–57) illustrates the complexity of the process by discussing the merger of the fricatives /θ/ and /f/, and of /ð/ and /d/ (word-initially) and /v/ (elsewhere) in East Anglian English. Both are well-known features of London English. According to Trudgill’s two Norwich surveys, speakers born in 1957 do not have these mergers at all, while those born in 1968 use especially the fricative /θ/ extensively in all urban centres in East Anglia. While the findings are well-documented, the problem of adolescents’ adopting this feature from working-class Londoners is not easy to account for. Trudgill does not expect extensive face-to-face accommodation to have taken place as the two groups are unlikely to have had much direct immediate contact with each other. The explanation must be sought elsewhere. The spread of the features could be the result of imitation rather than accommodation, which Trudgill finds to be the case, for example, in British pop singers’ widespread adoption of features of American English pronunciation. One of the possible explanations Trudgill also considers is internal migration of small groups of people or individuals into Norwich from the London area, and the return to Norwich of families temporarily resident in London. Moreover, /θ/ and /ð/ are good candidates for merger since they are phonologically unusual (Trudgill 1986: 57). As shown in Table 1, mergers are easy second-dialect features to acquire throughout one’s lifespan. However, the diffusion of a linguistic change is likely to have multiple mechanisms. As noted above, speakers may converge to a feature corresponding to their mental representation of a particular social

Sociopragmatics of language change

579

group. Here the growing interest in media impact will throw more light on global influences and their local appropriation. 6.3.

Social cognition

Discussing language change, and sound change in particular, sociolinguists often make a distinction between change from below and change from above the level of conscious awareness (Labov 1994: 78, 300–302). Many researchers also connect this distinction with positions in the socio-economic hierarchy. Labov argues that a change from above is introduced by the dominant social class (which need not be the highest) and is often modelled on other, higher-prestige communities. By contrast, change from below may be introduced by any social class, but typically involves an “interior” social group in the hierarchy (lower middle class, upper working class, skilled workers; teachers, salespeople, merchants, clerks, artisans, mechanics, etc.). Changes from below begin as local patterns and, generalizing throughout the group, may for others become associated with the group’s values. The linguistic variable undergoing change may remain below the level of social awareness as an indicator which indexes the group but is not commented on or necessarily even recognized by native speakers.4 As the change progresses, it usually becomes a marker and, varying both socially and stylistically, acquires a degree of social awareness. Ultimately the variable may become a stereotype, which is well-known and subject to comment in the speech community (Labov 2001: 196). This categorization is not restricted to language changes in progress but applies to linguistic features more generally. Labov (2010: 2–4) accounts for such effects in terms of social cognition, to which he attributes the systematic differentiation of change by gender and social class. Linguistic forms are hence apt to become associated with a range of social evaluations such as formal/informal, older/younger, higher/lower, rural/urban, and local/outsider. These forms index such polarities, and can serve as means of identity projection and part of an indexical system that embeds ideology in language. It is worth noting that, at least as far as ongoing sound-changes are concerned, empirical studies indicate that most speakers fall in between such recognized polarities. Moreover, speakers can shift their behaviour as their orientation towards these polar targets changes (Labov 2001: 463). Emphasizing the discoursal nature of the indexing of linguistic forms, Eckert (2008: 454, 464) proposes the concept of indexical field in order to account for the interpretation and potential reinterpretations of the meaning of a linguistic variable. An indexical field associates social meaning with a linguistic feature in a constellation of ideologically linked meanings. A given feature associated with the social evaluation of a group of speakers (marker) may become internalized to index specific elements of character in their own dialectal variability, creating a fluid indexical field. The social evaluation of the velar vs. apical realization of the suffix -ing in American English, for example,

580

Terttu Nevalainen

may fluctuate between polarities such as formal – relaxed; educated – uneducated; effortful – easy-going/lazy; articulate/pretentious – inarticulate/unpretentious. The first evaluation in each pair is associated with the velar variant. Eckert (2008: 466–467) concludes: “[s]ince the same variable will be used to make ideological moves by different people, in different situations, and to different purposes, its meaning in practice will not be uniform across the population”. Hence even stable sociolinguistic variables such as -ing are subject to semantic variation and change.

7.

Methodological envoy

This chapter began by contrasting speaker innovation and language change. Innovation only becomes a language change when it is adopted by other speakers. Analysing language change hence follows what Chambers (2002: 365) calls the aggregate principle in that group results are more revealing than the results for an individual member of the group. This observation has been made repeatedly by researchers who work within different theoretical frameworks. Considering the accommodation theory, Auer and Hinskens (2005) place language change at the macro-level of the speech community. Even social-networks, which they suggest provide the meso-level context for incipient language change, may be approached from the macro-level perspective in research into migration and dialect contact. The macro-level of socio-cultural framing similarly forms part of comprehensive models of identity projection, such as that proposed by Coupland (2007). The study of language change as a community phenomenon has obvious methodological implications. Considering the limits of linguistic ethnography, Sealey (2007: 655) writes: “[…] ethnography, then, can reveal in detail how speakers negotiate the constraints and enablements attendant on their particular situated experience, but it cannot account for the pre-existing structural properties and powers which are experienced as constraints and enablements by these social actors: different kinds of research methods are needed to explore this dimension of social reality.” The studies of language change as a communal process which I have used to illustrate the topics addressed in this chapter have mostly been quantitative, and based on sociolinguistic interviews or electronic corpora. Referring to these quantitative methods, Sealey (2007: 652) makes the point succinctly: “this research reveals how ‘what is frequently said’ can become a constraint on what, ‘although we are in principle free to say whatever we want,’ subsequent speakers can say.”

Sociopragmatics of language change

581

Acknowledgement My work on this chapter was supported by the Academy of Finland Academy Professorship scheme (2010–14), which I gratefully acknowledge.

Notes 1. Actuation in the linguistic sense is, however, only part of the story for those who view language change as the outcome of multiple causes, involving processes in which not only linguistic but also communicative, social and cognitive factors have major roles to play. Over the past decades, various usage-based approaches to language variation and change have gained momentum. These often cross disciplinary boundaries and build on models and methods developed in fields such as evolutionary biology, social psychology and the study of cognition (e.g., Croft 2000). See further the contributions by Harder and Trousdale to this volume. 2. For sociolinguistic studies on language variation and change in children and adolescents, see e.g. Roberts (2002), Foulkes, Docherty, and Watt (2005), and Smith, Durham, and Fortune (2007). 3. Vaattovaara (2009) attributes the overall preservation of h-ful forms in the region to a broader regional dialect identity with Meänkieli, which is spoken on the Swedish side of the River Tornio. It originally formed part of the same dialect as the Tornio Valley variety studied by Vaattovaara. The two were separated as a result of the Hamina peace treaty in 1809, as Sweden ceded to Russia its Finnish territories east of the Tornio and Muonio Rivers, which became the autonomous Grand Duchy of Finland (see 5.1.). Suppressed for a long time, Meänkieli is now a recognized minority language in Sweden. 4. Indexical factors are, of course, all-pervasive, ranging from cross-speaker differences (e.g. in age, gender, ethnicity, socio-economic status, regional background, and individual identity) to intraspeaker effects related to attitude, emotion, and a variety of contextspecific factors subsumed under “style” (Foulkes 2010: 7).

References Andersson, Roger and Mats Thelander 1994 Internal migration, biography formation and linguistic change. In: Bengt Nordberg (ed.), The Sociolinguistics of Urbanization: The Case of the Nordic Countries, 51–86. Berlin: Walter de Gruyter. Archer, Dawn 2005 Questions and Answers in the English Courtroom (1640–1760). Amsterdam/ Philadelphia: John Benjamins. Auer, Peter 2007 Introduction. In: Peter Auer (ed.), Style and Social Identities: Alternative Approaches to Linguistic Heterogeneity, 1–24. Berlin/New York: Mouton de Gruyter.

582

Terttu Nevalainen

Auer, Peter and Frans Hinskens 2005 The role of interpersonal accommodation in a theory of language change. In: Peter Auer, Frans Hinskens and Paul Kerswill (eds.), Dialect Change: Convergence and Divergence in European Languages, 335–357. Cambridge: Cambridge University Press. Bailey, Guy 2002 Real and apparent time. In: J.K. Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.), The Handbook of Language Variation and Change, 312–332. Malden, MA/Oxford, UK: Blackwell. Bailey, Guy, Tom Wikle, Jan Tillery and Lori Sand 1996 The linguistic consequences of catastrophic events: An example from the American Southwest. In: Jennifer Arnold, Renée Blake, Brad Davidson, Scott Schwenter and Julie Solomon (eds.), Sociolinguistic Variation: Data, Theory, and Analysis, 435–451. Stanford, CA: CSLI Publications. Chambers, J.K. 2002 Patterns of variation including change. In: J.K. Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.), The Handbook of Language Variation and Change, 349–372. Malden, MA/Oxford, UK: Blackwell. Chambers, J.K. and Peter Trudgill 1998 [1980] Dialectology, 2nd edition. Cambridge: Cambridge University Press. Coupland, Nikolas 2001 Introduction: Sociolinguistic theory and social theory. In: Nikolas Coupland, Srikant Sarangi and Christopher N. Candlin (eds.), Sociolinguistics and Social Theory, 1–25. Harlow/London: Longman. Coupland, Nikolas 2007 Style: Language Variation and Identity. Cambridge: Cambridge University Press. Croft, William 2000 Explaining Language Change: An Evolutionary Approach. London: Longman. Culpeper, Jonathan 2009 Historical sociopragmatics. Journal of Historical Pragmatics 10(2): 179–186. Eckert, Penelope 2008 Variation and the indexical field. Journal of Sociolinguistics 12(4): 453–476. Foulkes, Paul 2010 Exploring social-indexical knowledge: A long past but a short history. Laboratory Phonology 1: 5–39. Foulkes, Paul, Gerard Docherty and Dominic Watt 2005 Phonological variation in child-directed speech. Language 81(1): 177–206. Giles, Howard, Nikolas Coupland and Justine Coupland 1991 Accommodation theory: Communication, context and consequence. In: Howard Giles, Justine Coupland and Nikolas Coupland (eds.), Contexts of Accommodation: Developments in Applied Sociolinguistics, 1–68. Cambridge: Cambridge University Press. Herold, Ruth 1997 Solving the actuation problem: Merger and immigration in eastern Pennsylvania. Language Variation and Change 9: 165–189.

Sociopragmatics of language change

583

Hudson, Richard A. 1996 Sociolinguistics, 2nd edition. Cambridge: Cambridge University Press. Huss, Leena and Anna-Riitta Lindgren 2010 Scandinavia. In: Joshua A. Fishman and Ofelia García (eds.), Handbook of Language and Ethnic Identity: Disciplinary and Regional Perspectives, Volume 1, 255–268. New York: Oxford University Press. Joseph, John E. 2004 Language and Identity: National, Ethnic, Religious. Houndmills/New York: Palgrave Macmillan. Kerswill, Paul 1996 Children, adolescents, and language change. Language Variation and Change 8: 177–202. Klinge, Matti 1990 Let Us Be Finns: Essays on History. Helsinki: Otava. Labov, William 1963 The social motivation of a sound change. Word 18: 1–42. Labov, William 1994 Principles of Linguistic Change, Volume 1: Internal Factors. Oxford, UK/ Cambridge, MA: Blackwell. Labov, William 2001 Principles of Linguistic Change, Volume 2: Social Factors. Oxford, UK/Cambridge, MA: Blackwell. Labov, William 2010 Principles of Linguistic Change, Volume 3: Cognitive and Cultural Factors. Oxford, UK/Cambridge, MA: Blackwell. Lass, Roger 1997 Historical Linguistics and Language Change. Cambridge: Cambridge University Press. Le Page, Robert B. and Andrée Tabouret-Keller 1985 Acts of Identity: Creole-based Approaches to Language and Ethnicity. Cambridge/New York: Cambridge University Press. Leskinen, Heikki 1980 Wie verschwindet ein Dialekt? Beobachtungen zur heutigen Umgangssprache der karelischen Umsiedler. In: Heikki Leskinen (ed.), Heutige Wege der finnischen Dialektologie, 67–91. (Studia Fennica 24.) Helsinki: Suomalaisen Kirjallisuuden Seura. Mendoza-Denton, Norma 2002 Language and identity. In: J.K. Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.), The Handbook of Language Variation and Change, 475–499. Malden, MA/Oxford, UK: Blackwell. Meyerhoff, Miriam 2001 Dynamics of differentiation: On social psychology and cases of language variation. In: Nikolas Coupland, Srikant Sarangi and Christopher N. Candlin (eds.), Sociolinguistics and Social Theory, 61–87. Harlow/London: Longman. Milroy, James 1992 Linguistic Variation and Change: On the Historical Sociolinguistics of English. Oxford: Blackwell.

584

Terttu Nevalainen

Milroy, James 2003 On the role of the speaker in language change. In: Raymond Hickey (ed.), Motives for Language Change, 143–157. Cambridge: Cambridge University Press. Milroy, James, and Lesley Milroy 1985 Linguistic change, social network and speaker innovation. Journal of Linguistics 21: 339–384. Milroy, Lesley 1987 Language and Social Networks, 2nd edition. Oxford: Blackwell. Milroy, Lesley and Matthew Gordon 2003 Sociolinguistics: Method and Interpretation. Malden, MA: Blackwell. Nevalainen, Terttu 2011 Historical sociolinguistics. In: Ruth Wodak, Barbara Johnstone and Paul Kerswill (eds.), The Sage Handbook of Sociolinguistics, 279–295. London: Sage Publications. Nevalainen, Terttu and Helena Raumolin-Brunberg 2003 Historical Sociolinguistics: Language Change in Tudor and Stuart England. London: Pearson Education. Nordberg, Bengt (ed.) 1994 The Sociolinguistics of Urbanization: The Case of the Nordic Countries. Berlin: Walter de Gruyter. Nordin, Jonas 2000 Ett fattigt men fritt folk: Nationell och politisk självbild i Sverige från sen stormaktstid till slutet av frihetstiden [A People of Poverty and Liberty: National and Political Self-Image in Sweden from the Late Age of Greatness to the End of the Age of Liberty (c. 1660–1772)]. Stockholm: Bokförlag Symposion. Nordlund, Taru 2007 Double diglossia: Lower class writing in 19th-century Finland. Multilingua 26: 229–246. Nuolijärvi, Pirkko 1994 On the interlinkage of sociolinguistic background variables. In: Bengt Nordberg (ed.), The Sociolinguistics of Urbanization: The Case of the Nordic Countries, 149–170. Berlin: Walter de Gruyter. Roberts, Julie 2002 Child language variation. In: J.K. Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.), The Handbook of Language Variation and Change, 333–348. Malden, MA/Oxford, UK: Blackwell. Sairio, Anni 2009 Language and Letters of the Bluestocking Network: Sociolinguistic Issues in Eighteenth-century Epistolary English. (Mémoires de la Société Néophilologique de Helsinki 75.) Helsinki: Société Néophilologique. Sankoff, Gillian 2004 Adolescents, young adults and the critical period: Two case studies from Seven Up. In: Carmen Fought (ed.), Sociolinguistic Variation, 121–139. Oxford: Oxford University Press. Sankoff, Gillian 2005 Cross-sectional and longitudinal studies in sociolinguistics. In: Ulrich Ammon, Norbert Dittmar, Klaus J. Mattheier and Peter Trudgill (eds.), Socio-

Sociopragmatics of language change

585

linguistics. An International Handbook of the Science of Language and Society, Volume 2, 1003–1013. Berlin: Mouton de Gruyter. Schneider, Edgar W. 2007 Postcolonial English: Varieties Around the World. Cambridge/New York: Cambridge University Press. Sealey, Alison 2007 Linguistic ethnography in realist perspective. Journal of Sociolinguistics 11 (5): 641–660. Smith, Jennifer, Mercedes Durham and Liane Fortune 2007 “Mam, my trousers is fa’ in doon!”: Community, caregiver, and child in the acquisition of variation in a Scottish dialect. Language Variation and Change 19 (1): 63–99. Thomas, Erik R. and Malcah Yaeger-Dror 2010 Introduction. In: Malcah Yaeger-Dror and Erik R. Thomas (eds.), African American English Speakers and Their Participation in Local Sound Changes, (Publications of the American Dialect Society 94 [1]) 1–20. Durham, N.C.: Duke University Press. Trudgill, Peter 1986 Dialects in Contact. Oxford: Basil Blackwell. Trudgill, Peter 2008 Colonial dialect contact in the history of European languages: On the irrelevance of identity to new-dialect formation. Language in Society 37 (2): 241–254. Vaattovaara, Johanna 2009 Meän tapa puhua: Tornionlaakso pellolaisnuorten subjektiivisena paikkana ja murrealueena [Our way of talking: The Tornio Valley as a perceptual space and dialect area]. Helsinki: Suomalaisen Kirjallisuuden Seura. Van Herk, Gerard 2008 Fear of black phonology: The Northern Cities Shift as linguistic white flight. University of Pennsylvania Working Papers in Linguistics 14(2): 155–161. Vogelaer, Gunther De 2006 Actuation, diffusion, and universals: Change in the pronominal system in Dutch dialects. Zeitschrift für Dialektologie und Linguistik 73(3): 259–274. Weinreich, Uriel, William Labov and Marvin Herzog 1968 Empirical foundations for a theory of language change. In: Winifred P. Lehmann and Yakov Malkiel (eds.), Directions for Historical Linguistics, 95–188. Austin: University of Texas Press.

21.

The semantics of pragmatic expressions Maj-Britt Mosegaard Hansen

1.

Introduction

The term “pragmatic expression” or “pragmatic marker” (Fraser 1996) is often used to denote linguistic items or constructions which do not contribute to the truth-conditions of the discourse segments hosting them, but whose principal function is to signal aspects of the relationship between that utterance and its co(n)text (e.g. (1) and (2)), between the utterance and its speaker (e.g. (3)), and/or between the speaker and the hearer (e.g. (4)): (1) (2) (3) (4)

Please don’t do that! It wasn’t the best performance I’ve seen. Still, it wasn’t bad. Frankly, you’re a pig! How can I help you, Reverend?

Although the pragmatic expressions exemplified in (1) to (4) all happen to be single words, discourse functions of the types mentioned are, as a number of scholars have pointed out, frequently fulfilled by multi-word expressions, which may or may not be formally variable to a certain extent (Fischer 2006: 4–5), cf. (5) and (6): (5) If you don’t mind my saying so, you really should stop drinking now. (6) I just didn’t wanna spend any more time there, you know (what I mean)? While such multi-word expressions can no doubt be considered to be lexicalized to at least some extent, the potential for formal variability shows that pragmatic expressions do not belong exclusively to the lexicon, but that syntactic templates may be involved. Furthermore, the types of meanings that are conveyed by pragmatic expressions can also be communicated by a range of other expressive means, not all of which belong to language. Thus, for instance, a purely syntactic means, such as the French left-dislocation construction in (7) (cf. Lambrecht 1994), or a phonological feature like the French “pre-pausal” schwa in (8) (cf. Hansen and Hansen 2003), can be used to signal the relationship between an utterance and its context, while gestures, facial expressions and proxemics clearly add important information about speaker/hearer, speaker/utterance, and utterance/co(n)text relations in face-to-face interaction (Kerbrat-Orecchioni 1992: 41–43, 76–80): (7) … c’était surtout le bruit c’était les sirènes qui donnaient l’alarme – et on en – ce bruit dans les oreilles je l’ai entendu au moins pendant dix ans (from Blasco-Dulbecco 1999: 253)

588

Maj-Britt Mosegaard Hansen

‘… it was mainly the noise it was the sirens sounding the alarm – and we hea – this noise in my ears I kept hearing it for at least ten years’ (8) ah non non c’est pas les salaires moi je suis pas du tout au traitement/, moi je suis au service command[´ ]s// (from Hansen and Hansen 2003: 103) ‘oh no no it’s not the payroll I’m not at all in the payroll department, I’m in the orders department’ Different though they are, all these expressive means can therefore arguably be considered types of pragmatic markers (if not of pragmatic expressions). In this chapter, however, the focus will henceforth be on (more or less) lexicalized items.

2.

Drawing the semantics/pragmatics boundary

We observed above that a commonly evoked criterion for classifying a linguistic item as a pragmatic expression is non-truth-conditionality. This being the case, the title of the present chapter raises the question of the borderline between semantics and pragmatics, for given the widespread definition of semantics as the truth-conditional content of sentences and utterances (e.g. Gazdar 1979: 2), how can nontruth-conditional expressions meaningfully be said to have a semantics at all? To some scholars, there is at least a subset of pragmatic expressions that do indeed not have a semantics (Aijmer 2002: 22), but only a pragmatics. That view, however, leads to a (quasi) paradox in so far as pragmatics, besides being understood as non-truth-conditional meaning, is also traditionally understood as context-induced, inferrable, hence changeable and crucially defeasible, meaning (e.g. Lyons 1977: 591), whereas semantics is taken to represent context-independent, coded, i.e. relatively stable, meaning. Thus, I may utter (9) upon entering the office of a colleague, whose window is wide open, intending to convey a request that the window be closed: (9) It’s so cold in here! On the other hand, if I have reluctantly agreed to accompany a friend on a journey across Greenland by dog-sled, and utter (9) as we are about to go to sleep in our tent the first evening, I may intend to communicate that I want to go home. Nothing about (9) itself tells us which of these two interpretations (and there are, of course, many more potential interpretations) is the correct one; that is entirely determined by the context of utterance, and hence a matter of pragmatic meaning. In both cases, moreover, the suggested interpretation can be cancelled, either explicitly or by features of the co(n)text. Now, consider (10) to (13):

The semantics of pragmatic expressions

(10) (11) (12) (13)

589

Emmeline left Greenland because it was too cold there. It must be cold in Greenland, because Emmeline left almost immediately. Let’s get out of here! Because the noise is driving me crazy. Do you want to meet at the Art Gallery? Because they have an exhibition by that painter you say you really like.

According to a by now fairly well established tradition, (10) represents a “semantic”, i.e. truth-conditional, use of because, whereas (11) to (13) instantiate “pragmatic”, i.e. non-truth-conditional, uses of that same connective (cf. van Dijk 1979). That the two types of uses differ is not in doubt and can be confirmed by various tests – principally whether the because-clause can fall under the scope of negation or become the focus of a cleft sentence (see further sect. 3 below). There is, however, no sense in which the interpretation of because is any more context-dependent in the latter three examples than in (10). Thus, on pain of rendering the discourse meaningless (or at least quite bizarre), the “pragmatic” interpretation cannot, in any of these examples, be canceled leaving only a strictly “semantic” interpretation. Moreover, not only is there a close relation between interpretations of because in its so-called “pragmatic” uses (roughly paraphrasable as “what follows is intended as a justification of the preceding discourse segment”) and in the so-called “semantic” (i.e. properly causal) use, but the “pragmatic” interpretation appears quite stable across (11) to (13), despite the dissimilarity of the contexts. In other words, the “pragmatic” meaning of because is not a matter of mere usage, inferred anew on each relevant occasion; rather, it seems to be an established polyfunctionality, which must somehow be coded in the language system (see further sect. 5 below). Clearly, the context in which because is used will determine its specific reading in any given instance, but that is true of all polyfunctional items and concerns the truth-conditional reading and the non-truth-conditional one equally. Of far greater significance is the fact that, in virtue of marking its host clause as a justification, the use of a pragmatic marker like because may in and of itself attribute certain features to the situational context. Thus, an English speaker who happened to casually overhear (14) would infer that the alternative means of transportation envisaged was walking, whereas the minimally different utterance (15) would seem to make sense only if it is inferred that the alternative is for the hearer to make use of motorized transportation: (14) You should take your bike, because it’s pretty far. (15) You should take your bike, because it isn’t very far. In this, pragmatic markers are quite unlike genuinely context-dependent meanings such as conversational implicatures. In fact, pragmatic markers seem to fulfil all the criteria for conventional implicature, as defined by Grice (1975), i.e.:

590

Maj-Britt Mosegaard Hansen

– they add a non-propositional, but also non-defeasible, layer of meaning to their host utterance; – that meaning is detachable, in the sense that if the linguistic item conveying it is replaced by a truth-conditionally equivalent item, the conventional implicature will disappear; – the conventional implicature attached to a given expression must be learned; it cannot be inferred on the basis of the Cooperative Principle and its attendant Conversational Maxims. Although, in keeping with the mainstream Anglo-American tradition, Grice considered his conventional implicatures to belong to pragmatics, I will follow a tradition that is well-established in Continental European linguistics (e.g. Ducrot 1980), and which is also represented by certain Anglo-American approaches to the study of meaning, such as Frame Semantics (Fillmore 1985) and Relevance Theory (e.g. Blakemore 1987), whereby the distinction between semantics and pragmatics is not in the first instance a distinction between truth-conditional and non-truth-conditional elements of meaning, but between elements of meaning that appear to be coded in linguistic expressions and those that arise inferentially as a result of the interaction between coded meanings and the specific context of their occurrence. On this view, so-called “pragmatic expressions” not only have a semantics, they actually belong to the study of semantics, or at least to the study of the semantics/ pragmatics interface, as opposed to the study of pragmatics proper. 3.

Content-level vs context-level expressions

A consequence of drawing the semantics/pragmatics boundary in this way is that the term “pragmatic expression” itself becomes infelicitous. Furthermore, as we have just seen with because, many such expressions are polyfunctional between readings that are propositional and readings that are not. In other words, in many cases it is not a given linguistic item in itself which is either propositional (“semantic”) or non-propositional (“pragmatic”) in nature, but rather different uses of it. Henceforth, I will therefore – as proposed in Hansen (2008) – use the term “content-level use” to refer to any use of a linguistic item in which the meaning of that item bears saliently either on a state-of-affairs in some real or imagined world referred to in its host clause or on the relation between that state-of-affairs and other (real or imagined) states-of-affairs. The contrasting term “context-level use” will refer to any use of a linguistic item in which that item primarily expresses the speaker’s comment on the relation between a described state-of-affairs and the discourse itself (including, but not limited to, the way it is represented linguistically) or on the wider speech situation (including, but not limited to, the subjective attitudes to the state-of-affairs in questions that may be entertained by the speaker, the hearer, or some relevant third party).

The semantics of pragmatic expressions

591

Importantly, some linguistic items may be used simultaneously at the contentlevel and at the context-level. Saliently, this is the case in languages that have a T/V-distinction in their system of pronominal address (cf. Brown and Gilman 1972). Thus, in (16), the underlined subject pronouns clearly have referential, i.e. content-level, meaning, but assuming that they refer in both cases to a single addressee, the choice of tu as opposed to vous (or vice versa) makes a difference only at the context-level, where it provides an indication of the nature of the relationship between the speaker and the hearer: (16) Veux-tu2sg/voulez-vous2pl encore du café? ‘Would you like some more coffee?’ A different type of case is represented, for instance, by approximatives like almost and barely, as exemplified in (17) and (18) (adapted from Anscombre and Ducrot 1983: 80): (17) It’s almost dark – you’d better turn on your headlights! (18) It’s barely dark – you need only turn on your sidelights. From a truth-conditional point of view, almost conveys the meaning ‘not yet quite’. Truth-conditionally, therefore, ‘almost p’ entails ¬p. Barely, on the other hand, conveys the meaning ‘only just’; hence, ‘barely p’ entails p. Yet of the mini-discourses in (19) to (22), only (19) and (20) appear natural, whereas (21) and (22), although truth-conditionally equivalent to (17) and (18), are odd: (19) (20) (21) (22)

It’s dark – you’d better turn on your headlights! It’s not dark – you need only turn on your sidelights. #It’s not dark – you’d better turn on your headlights! #It’s dark – you need only turn on your sidelights.

It would appear that, in addition to their contribution to the content-level of host utterances, the lexemes almost and barely systematically convey context-level rhetorical/argumentational elements of meaning that do not merely conflict with, but effectively override, their content-level meaning. The upshot of the above discussion is that, while (non-)truth-conditionality is undoubtedly an important criterion for determining the level of discourse at which a given linguistic item functions, it is neither a necessary nor a sufficient one. This does not mean that we have to rely purely on intuition. Many, if not most, linguistic items seem to have adverbial status when functioning at the contextlevel, and for adverbials, Dik et al. (1990) propose a set of heuristic tests that can be used to distinguish between content-level and context-level uses (or, as they call them, “representational-level” vs “interpersonal-level” adverbials). The tests are based on the assumption that adverbials form part of the informational content of their host clause when used at the content-level, but not when used at the contextlevel. This assumption gives rise to two predictions:

592 1.

Maj-Britt Mosegaard Hansen

Context-level adverbials will fall outside the focus structure of the host clause. Unlike content-level adverbials, they should therefore not occur as answers to wh-questions in (23), in non-metalinguistic contrastive contexts (24), nor in clefts (25):

(23) A. How did John go home? B. *However/By bus. (24) *John didn’t however, but rather evidently, go home./John didn’t go home by bus, but rather on foot. (25) *It was however that John went home./It was by bus that John went home. 2.

Content-level adverbials will form an information unit with their host clause. Adverbial and host should therefore fall under a single unified intonation contour (26), the adverbial should be within the scope of pro-forms (27), and it should be possible to question the contents of the entire host clause, including the adverbial (28):1

(26) Albert left in a huff./Albert left, thank God. (27) Albert left in a huff, and so did Max. (= Max left in a huff.)/Albert left, thank God, and so did Max. (≠ Max left, thank God.) (28) Did Albert leave in a huff?/*Did Albert leave, thank God? 4.

The nature of context-level meaning

If we accept that linguistic items and constructions functioning at the context-level do have a semantics, and not just a pragmatics, the obvious next question becomes how to account for that semantics, given that context-level meanings are not referential in nature. In the past three or four decades it has become increasingly widely accepted that context-level meanings are essentially procedural or, to use a different term, instructional, in nature (e.g. Ducrot et al. 1980; Blakemore 1987; Hansen 1998a, 2008; Traugott and Dasher 2002). Context-level items are thus seen as providing processing instructions to the hearer, indicating how the contents of the host clause as a whole should be contextualized so as to be integrated into a coherent mental representation of the discourse. Thus, for instance, the approximators almost and barely can each be understood as instructing hearers to reconstruct a particular argumentational scale of the general form . In the case of almost, the scale will take the form , and in the case of barely, it will take the form . In both cases, the further implication for the hearer is that a sentence S containing one of these approximators must be processed as supporting the same type of conclusion as a sentence S’ containing no approximator, but having the opposite truth-value. At the same time, S must be processed as constituting a weaker argument than S’ for the type of conclusion in question.

The semantics of pragmatic expressions

593

In other words, taking “It is dark” as p and “It is not dark” as ¬p, (17) instructs the hearer to reconstruct the scale , S being “It is almost dark” and S’ being “It is dark”. (18), on the other hand, instructs the hearer to reconstruct the scale , S being “It is barely dark” and S’ being “It is not dark”. In neither case is there an entailment relation between S and S’; indeed, the two have opposite truth values. However, in the case of (17), both S and S’ will serve as arguments for the conclusion “The hearer should turn on his/her headlights”, S’ being the stronger argument. Similarly, in the case of (18), both S and S’ will support the conclusion “The hearer need only turn on his/her sidelights”, but S’ will again constitute the stronger argument. In other words, despite being logically incompatible, the two members of each scale are rhetorically co-oriented, and the scales evoked must be seen as part of the linguistically coded meanings of almost and barely, respectively. The semantic instructions conveyed by a given context-level item need not involve the reconstruction of a specific rhetorical scale; some items may require the construction of a purely contextual ad hoc scale (cf. Hirschberg 1991), or – as is frequently the case – they may not involve scales at all, but convey other types of instructions. The focus particle even, for instance, essentially instructs the hearer to process the focused entity as occupying (one of) the very highest rung(s) of a scale whose nature is entirely context-dependent and which must therefore be actively constructed (as opposed to merely reconstructed) by the hearer (cf. Kay 1990): thus, the scales constructed upon hearing (29) will probably be quite different depending on whether that utterance is processed in a context where it is immediately preceded by (30) or by (31): (29) It even made it onto the local news the other night. (30) Max’s business is really successful. (31) You wouldn’t believe the trouble Max is having with his neighbors! An adverbial like please in (1), on the other hand, does not involve scalarity at all, but essentially instructs the hearer to process the host utterance as a polite request, while – as we saw in sect. 1 above – because in (11) to (13) instructs the hearer to understand the contents of its host clause as constituting, in the eyes of the speaker, a justification of the preceding speech act. In all these cases, however, the instructions encoded in the semantics of context-level expressions can be seen as pointing hearers to particular – more or less schematic – frames of interpretation for the utterances hosting such expressions. It is not only context-level items which can be seen as having the property of conveying essentially procedural meaning in this way: function words such as auxiliaries or determiners, inflectional morphemes, and grammatical constructions have been described as fulfilling similar tasks, including signaling how referential expressions contained in the host clause should be combined, how referents may be identified, etc. Thus, for instance, Heim (1982: 276) proposes the following initial

594

Maj-Britt Mosegaard Hansen

formulation of the semantics of definite vs. indefinite noun phrases: “For every indefinite, start a new card [in your mental discourse file]; for every definite, update a suitable old card.” Thus, the instructional view can straightforwardly account for the fact that one and the same linguistic item may in some cases – as we saw above – function simultaneously on the content- and on the context-level of discourse. The concept of meanings as instructional affords them significant synchronic and diachronic flexibility, in as much as even the most precise of instructions will leave some room for personal creativity on the part of the instructee. As Ducrot et al. (1980: 33) put it, “il y a mille façons possibles d’obéir aux instructions véhiculées par [une] phrase” (‘there are a thousand different ways to comply with the instructions conveyed by [a] sentence’). In respect of linguistic meaning, the way a given instruction is carried out will be constrained not only by any other instructions conveyed by the sentence or discourse, but also by the larger context, as well as by the fact that no two people have exactly identical mental grammars, let alone identical encyclopedic knowledge and experience on which to draw in identifying and fleshing out relevant frames of interpretation. The result of carrying out the set of semantic instructions conveyed by a given utterance will be a mental representation of what the hearer takes the speaker’s intended meaning to be. Speaker’s meaning being a pragmatic notion, the form that such a representation takes can only be constrained, and not fully determined, by the instructions provided by the semantics. On the assumption that the meanings conveyed by context-level expressions are purely instructional in nature, the frequently observed polyfunctionality of these items thus becomes entirely predictable.

5.

How to account for the polyfunctionality of context-level expressions

A great many researchers working on context-level expressions have observed that not only do the latter often have more than one possible context-level function, but those very same expressions often also have content-level uses whose meanings bear at least some resemblance to their context-level meanings. Thus, we saw above that the English conjunction because has both content-level and contextlevel uses. Alternatively, the language may include a content-level expression which, although clearly distinct from the context-level expression, is nevertheless reminiscent of it in aspects of its form and meaning, cf. the (somewhat archaic) Danish adjective vist and the modal particle vist in (32) and (33): (32) (Danish) Det er ganske vist. ‘It is quite certain.’

The semantics of pragmatic expressions

595

(33) (Danish) Peter kommer vist i morgen Peter come – PRES .3 SG PARTICLE tomorrow ‘Peter is coming tomorrow, I have reason to believe.’ Being empirically quite extensive, this polyfunctionality clearly calls for a principled explanation, of which there are essentially three types: The simplest – and least popular – solution is to assume that the apparent unicity of form is, in fact, deceptive, and that the different interpretations of, say, English because or Danish vist are actually associated with different lexemes that just happen to be homophonous/homographic. Given the extent of observed contentlevel/context-level polyfunctionality, the explanatory power of chance homonymy fails to impress, however, and it is, indeed, frequently ruled out altogether by the type of discourse context in which polyfunctionality is first observed diachronically (see sect. 6 below). Researchers within the field are currently divided between the two remaining options, namely, to assume either that content-level and context-level uses of formally identical items are monosemous, or that they are polysemous.2 Monosemy preserves the one-form-one-meaning hypothesis, and may – very roughly – be said to be preferred by scholars with structuralist and/or formalist inclinations. The assumption here is that all the different readings of a polyfunctional item or construction can all be subsumed under a single – schematic – core meaning, fleshed out in different ways by features of the contexts in which the monosemous item occurs. Monosemy is descriptively parsimonious, and it therefore represents an attractive option at the strictly synchronic level. The principal problem with this approach is to formulate the presumed unitary core meaning in a way that is precise enough not to generate unacceptable readings (cf. König 1991: 175) and concrete enough to be actually learnable even by immature language users. The assumption of monosemy becomes more difficult to sustain in approaches to contextlevel expressions which seek – alongside synchronic descriptions of their meaning and functions – to trace the diachronic origins of the latter (see sect. 6). Essentially, monosemists must assume that, whenever a new reading of a given expression becomes conventionalized, the unified semantic representation that language users have of that expression must have undergone a qualitative change to allow for that new reading to be generated. Not only is it intuitively unlikely that such conceptual reorganization, which may need to be fairly extensive, should take place on a regular basis, but the monosemy approach completely obscures the step-by-step process by which meaning extensions demonstrably take place in diachrony. The remaining alternative, polysemy, – generally preferred by functionalists and cognitivists – consists in assuming that different discourse functions do, to at least some extent, reflect different coded meanings, and that these meanings are somehow related, such that some meanings can be conceived of as grounded in

596

Maj-Britt Mosegaard Hansen

other, more basic ones. Typically, synchronic polysemy will be the result of diachronic sense extensions, and we may therefore expect the range of possible synchronic relations between senses to be largely identical to the range of mechanisms for sense extension. In other words, I take it that synchronically existing senses of a given item may be related by metaphor, metonymy, generalization, specialization etc. (cf. section 6.6.1). Assuming polysemy allows us to forge a link between diachrony and synchrony and to propose generalizations about possible and likely paths of change. The major problem posed by polysemy has to do with the degree of granularity, and consequently the number, of the different coded senses posited by a particular semantic description. Do we, for instance, wish to say that the two different tokens of the French modal particle bien in (34) and (35) instantiate two different senses because the first token marks an explicit refutation of the previous utterance, whereas the second token does not? Or is it preferable to analyze both tokens simply as marking their host utterance as argumentationally counteroriented with respect to some contextually evoked proposition, actual refutation being merely a possible feature of the contexts of occurrence of this particle (cf. Hansen 1998b)? (34) A. Pourquoi n’as-tu pas posté ma lettre? B. Mais j’ai bien posté ta lettre! ‘A. Why didn’t you mail my letter? B. But I DID mail your letter!’ (35) As-tu bien posté ma lettre? (Approx.:) ‘Did you remember to mail my letter?’ Personally, I prefer to adhere to Foolen’s (1993: 64) principle of “methodological minimalism”, whereby distinct coded senses should not be “multiplied beyond necessity”, such that if a particular reading of a given item is straightforwardly derived from another reading using features of the context and pragmatic principles of general application, then the assumption should be that those two readings instantiate one and the same coded meaning. While it seems clear that no single test is adequate to systematically distinguish ambiguity from vagueness in linguistic items (cf. Geeraerts 1993; Tuggy 1993), a variety of heuristic tools, discussed in Hansen (2008: 37–39), are nevertheless available for this purpose, which together may aid researchers in deciding between actual polysemy and mere contextually determined variation in interpretation, both at the synchronic and at the diachronic level. 6.

The evolution of context-level expressions

In the vast majority of cases where diachronic analyses of expressions that are polyfunctional between content-level and context-level uses have been carried out, the content-level uses of the expressions in question have so far been shown to be

The semantics of pragmatic expressions

597

at least as old as, and typically older than, the corresponding context-level uses. This suggests that there is a strong cross-linguistic tendency for context-level expressions to develop out of existing content-level expressions. In the past couple of decades, the idea, first suggested by Grice ([1975]1989: 39) that “it may not be impossible for what starts life, so to speak, as a conversational implicature to become conventionalized” has provided the impetus for renewed theorizing on the topic of meaning change. Pioneered by the work of Elizabeth Traugott, semantic change is currently seen, to a large extent, as “arising out of the pragmatic uses to which speakers or writers and addressees or readers put language” (Traugott and Dasher 2002: xi). Thus, a large number of observed meaning changes appear to instantiate a set of semantic/pragmatic tendencies, set out in a total of nine diachronic clines, of which the following five appear most significant (Traugott and Dasher 2002: 281):3 1. Meanings tend to become increasingly subjective (i.e. increasingly grounded in the speaker’s subjective perspective), and possibly even intersubjective (i.e. explicitly grounded in the relationship between speaker and hearer). 2. Meanings that were conceptual at the outset tend to become increasingly procedural in nature. 3. Items and constructions that originally had scope within the host proposition tend to progressively enlarge that scope, potentially even up to the level of the discourse. 4. Meanings that were truth-conditional at the outset tend to become non-truthconditional. 5. Meanings that originally made reference to the described event progressively come to refer to the speech event itself. These types of change will clearly correlate with changes from content-level to context-level meaning, as defined in sect. 3 above, and a great deal of recent discussion in the field of diachronic semantics and pragmatics relates directly or indirectly to this approach, focusing on a handful of interrelated key issues (e.g. Eckardt 2003; Dostie 2004; Koch 2004; Visconti 2005; Hansen and Waltereit 2006; Hansen 2008; Hansen and Visconti, eds. 2009; Schwenter and Waltereit 2010). 6.1

Types of inferences, contexts, and mechanisms of change

Researchers have in the first instance been concerned with elucidating the nature of the pragmatic inferences that lead to semantic change, the cognitive mechanisms that underlie those inferences, and the contexts in which they are triggered. It should be noted that the phenomena discussed in this section are not specific to the rise of pragmatic markers, as opposed to other types of meaning change. However, to a significant degree, it is the desire to understand the diachronic develop-

598

Maj-Britt Mosegaard Hansen

ment of coded context-level meanings that has provided the impetus for renewed interest in the study of meaning change, a field which for decades was largely dormant. Traugott and Dasher’s (2002) Invited Inferencing Theory of Semantic Change (IITSC) assumes that conversational implicature4 is central to the rise of contextlevel meanings, and these authors follow Levinson (1995: 95, 2000: 263) in proposing that change proceeds from particularized conversational implicatures (PCI or “invited inferences”, IIN, as they prefer to call them) through generalized implicatures/generalized invited inferences to conventionalized implicatures, along the cline in (36): (36) PCI(IIN) > GCI(GIIN) > coded meaning Hansen and Waltereit (2006) argue that this sequence is actually rare, and that the role of different types of implicature in meaning change is a good deal more nuanced than the IITSC suggests. Based on the assumption that particularized conversational implicatures are “prototypically in the communicative foreground of messages, whereas generalized conversational implicatures are prototypically backgrounded” (Hansen and Waltereit 2006: 235), these authors instead posit the existence of two further clines, most cases of semantic change being either instances of a PCI semanticizing directly, cf. (37), or of a GCI semanticizing, but only after having being foregrounded as a PCI, cf. (38): (37) PCI f coded meaning (38) GCI f PCI f coded meaning In those cases where a PCI does turn into a GCI, Hansen and Waltereit argue that it will most likely not become fully semanticized, such that the third stage of (36) will typically not be instantiated. The issue of the nature of the pragmatic entities involved in change is addressed from a different angle by Schwenter and Waltereit (2010). They outline the evolution of the additive particle too and some of its counterparts in Spanish and German, to show how counter-argumentative uses of the particle in novel discourse contexts can override the additive presupposition that are normally associated with it. Accommodating that presupposition becomes too costly and leads to reanalysis of the meaning contribution of too as primarily adversative. In a similar vein, Eckardt (2009) shows that, antedating the emergence of the contemporary senses of certain words, one finds uses of those words in which the presuppositions of their source meanings are violated, in the sense that they contradict common world knowledge. According to Eckardt, such utterances create a pragmatic overload: under a conservative interpretation, they will trigger presuppositions that the listener cannot accommodate. Thus, the listener can either (a) refuse to interpret the utterance at all, (b) attempt to reconceptualize the world such that the presupposition is consistent, or (c) hypothesize a new meaning for

The semantics of pragmatic expressions

599

parts of the utterance, including the item that gave rise to the problematic presupposition. The observation that the impetus for reinterpretation may be the impossibility of accommodating a presupposition is an important one; it does not, however, entail that presupposition constitutes an alternative to implicature at this level of description, for it seems that, once a given presupposition has been identified as incompatible with the context of utterance, any reinterpretations that this gives rise to cannot themselves be presuppositions, but must rather be conversational implicatures. In other words, the stage of interpretation where hearers attempt – and fail – to accommodate is logically prior to, and does not replace, the stage where reinterpretation – via (particularized) implicature – takes place. With respect to the inferential mechanisms of change, there is general agreement that diachronic sense change is based on patterns of association either between aspects of different linguistic signs or between linguistic signs and aspects of the cognitive or real-world entities that they represent. The two principal patterns of association identified in the literature are similarity and contiguity, both of which may apply on the content side and on the expression side of signs undergoing change. Similarity – particularly on the content side, where it manifests itself as metaphor – has traditionally been regarded as the single most important mechanism, but in the past decade, contiguity – or metonymy, in a broad sense of the word – has increasingly come to be recognized as equally important, indeed perhaps conceptually more basic than metaphor (e.g. Hopper and Traugott 1993; Blank 1997; Barcelona, ed., 2000). The reason is that where metaphor relies on a qualitative leap from one conceptual domain to another, and is based on factual incompatibility between the two domains, metonymy is based on the simultaneous or sequential co-presence of elements within a single conceptual frame, and the elements in question are therefore mutually compatible. In consequence, meanings may not only slide almost imperceptibly from “literal” to metonymic or vice versa within even a short stretch of text (cf. (39)), but in many cases, language users need not even choose between two possible meanings, given that both may be germane to the interpretation of a particular utterance (cf. (40)): (39) I ate the entire box of chocolates and then disposed of it. (40) While John spent the afternoon at the pub, Jane worked hard to finish her report. (f (a) the two states-of-affairs took place simultaneously, and (b) there is a contextually relevant contrast between them) The effect of metonymy thus lies in a more or less obvious figure/ground shift, i.e. a reversal of normally foregrounded and normally backgrounded elements of the interpretation of a given item (Koch 2001; Waltereit 2002; Hansen and Waltereit 2006). The notion of metonymic change employed in much recent work is, as men-

600

Maj-Britt Mosegaard Hansen

tioned above, a very broad one, which includes the conventionalization of all sorts of contextually invited inferences (Traugott and Dasher 2002: 78–81). For instance, when used “literally”, the imperative of the verb look and its equivalents in a number of other languages (e.g. Italian guarda!, French regarde(z)!, Spanish ¡mira!) suggest that the object that the hearer is directed to look at is worthy of immediate attention (Waltereit 2002). Speakers may exploit that background suggestion by using such imperatives when there is, in fact, no particularly noteworthy object in the environment, because they wish instead to draw attention to the contents of their own upcoming discourse. If this happens with sufficient frequency, the erstwhile content-level imperative will be recruited as a discourse marker functioning at the context-level. As implied by the discussion above, it is a prerequisite of change that the meaning contribution made by a given linguistic item or construction be reinterpreted, and, as a number of scholars have observed, the possibility of reinterpretation often seems to rely on a particular type of context, known as “bridging contexts” (cf. Evans and Wilkins 2000; Heine 2002). Bridging contexts are characterized by the fact that they allow for two different interpretations of an item: an established “source” interpretation, and an innovative “target” interpretation. Slightly revising the model of Heine (2002: 86), Hansen (2008: 62) proposes that the following four contextual stages are relevant in accounting for how existing linguistic items acquire new meanings: – Stage 1: The item in question appears only in contexts where it clearly has its source meaning. – Stage 2: The item starts to appear in bridging contexts, where, in addition to its source meaning, it allows hearers to make an inference to an alternative – target – meaning. At this stage, the target meaning may, however, remain backgrounded with respect to the source meaning. – Stage 3: The item appears in “switch” contexts, where the target meaning is clearly intended and conflicts with the source meaning. – Stage 4: The target meaning is fully conventionalized alongside the source meaning; hence, the item has become polysemous. While the bridging contexts of Stage 2 seem to be a prerequisite for change, their existence does not in and of itself constitute evidence that change is, in fact, taking place: while bridging contexts allow for innovative interpretations, such potential interpretations need not actually be brought to bear by hearers. It is thus only at Stage 3 that we can be certain that reinterpretation has taken place. (41) to (43) illustrate Stages 1–3 in the evolution of the French connective puisque ( French témoin (‘witness’) came about as suggested by Koch (2004: 16f), namely via the (necessary) metonymical link between the existence of testimony and that of a witness, in contexts where utterances such as (44) would be produced, then it is unlikely that it would have been speaker-driven: (44) [Judge:] Audiamus testimonium proximum! ‘Let’s hear the next piece of testimony’ Even more obviously, while the change responsible for the two currently competing meanings of the Danish noun bjørnetjeneste (lit.: ‘bear favor’), ‘a favor that has unintended negative results’ > ‘a really big favor’, can be attributed to the nontransparency of that noun, and hence its potential ambiguity in contexts such as (45), it is highly implausible that speakers familiar with the original, negativelyloaded, meaning could be held responsible for the rise of the innovative, positively-loaded, interpretation. In this case, there can be little doubt that the process must have been driven by hearers who were unfamiliar with the intended meaning and who were therefore forced to make conjectures: (45) Peter overtalte chefen til at give mig opgaven. Det var sgu en bjørnetjeneste! Lit.: ‘Peter persuaded the boss to give me the task. That was some bear-favor!’ This does not mean that change is exclusively hearer-driven, but simply that hearers play as important a role as speaker do in language change. For innovations to become part of the shared linguistic system, we need two things: reinterpretation and propagation. Speakers can innovate as much as they please, but if no one notices, their innovations will not take hold. In other words, hearers need to reinterpret what is said to them as having a new meaning. Propagation is, of course, carried out by speakers, but only by speakers who have previously reinterpreted something in their role as hearers. And because propagation is a gradual process, the hearers of the propagating speakers must continue to reinterpret what they hear on a large number of individual occasions before the new meaning can become entrenched. Emphasizing the role of hearers in semantic/pragmatic change may be perceived as incompatible with positing subjectification as an important cross-linguistic tendency, but it need not be. If we see negotiation of meanings within the speaker-hearer dyad as a very important factor in communication; if we admit that when performing speech acts, speakers seek to have some sort of effect upon their hearers; and if we admit further that human beings tend to be more interested in each other than in most other aspects of the world (as evidenced by a number of grammatical hierarchies, where human referents are privileged over non-human

The semantics of pragmatic expressions

603

ones, or where the 1st and the 2nd person are privileged over the 3rd person, e.g. accessibility to subject/object assignment, to relativization, etc.), then it does not appear surprising that hearers should be interested in the point of view of speakers and predisposed to reinterpret linguistic items as expressing precisely that point of view. Moreover, such a predisposition will surely only be reinforced by the gradual realization that most of us come to in the course of our life-times that one and the same state-of-affairs in the world can be perceived very differently by different people, and that any individual speaker can therefore ultimately only express his or her subjective take on any individual state-of-affairs. Nature of the changes Finally, there is some debate about the relationship between the diachronic clines posited by Traugott and Dasher (2002), a debate that ultimately leads to the question whether the development of context-level meanings out of content-level meanings qualifies as an instance of grammaticalization, or whether a different type of process – dubbed “pragmaticalization” by Erman and Kotsinas (1993) – is involved. With respect to the latter question, those who argue that context-level expressions are grammaticalized focus on the fact that their evolution will typically feature the decategorialization of the (content-level) source item, some degree of phonological reduction of that source item, as well as subjectification and increased “procedurality” of its content. Those scholars who prefer to describe context-level expressions as “pragmaticalized”, on the other hand, do so on the basis of Lehmann’s (1985) six classic parameters of grammaticalization, viz. phonological and/or semantic attrition, paradigmatization, obligatorification, scope reduction, syntagmatic coalescence, and syntagmatic fixation. According to Lehmann’s model, grammaticalized items will exhibit a high degree of several of these characteristics. As noted by Waltereit (2002: 1005), Eckardt (2003: 42) and Hansen (2008: 57–58), context-level expressions tend, as a class, not to fulfill Lehmann’s criteria to any great extent. Rather, they typically exhibit scope increase, greater syntactic freedom, optionality, and strengthening of their pragmatic import. The issue of the interrelation of Traugott and Dasher’s clines is part and parcel of this debate, and concerns in particular the interrelation between subjectification and scope, the central question being whether subjectification is necessarily part of, and unique to, the grammaticalization process, which – if true – raises questions about the scope of grammaticalized items. According to Traugott (1995) and Brinton and Traugott (2005), subjectification is at the very least a strong tendency in grammaticalization, particularly in its early stages (Traugott 1995: 47), but it is nevertheless an independent process (Brinton and Traugott 2005: 109; Traugott 2010). Company Company (2006) argues that subjectification is not just a seman-

604

Maj-Britt Mosegaard Hansen

tic-pragmatic phenomenon, but actually constitutes a specific type of syntactic change as well, namely one that is characterized by scope increase and syntactic isolation, and she conceives of subjectification as a subtype of grammaticalization. Interestingly, what this author calls subjectification would seem to be largely equivalent to what others describe as pragmaticalization, were it not for the fact that she deems the syntactic changes mentioned criterial, as opposed to just typical, of subjectification. However, the existence of items like the so-called modal particles that are a salient feature of the Continental Germanic languages (cf. (33) above) appear to provide a strong argument against such a view, for while modal particles are indubitably subjectified as compared to their source items and also exhibit scope increase, they are nevertheless syntactically highly constrained. This debate raises the issue of the exact understanding of what is implied by notions such as grammaticalization and pragmaticalization: are these labels for specific processes of change, in which case they have independent theoretical status, or are they largely just convenient shorthands for the cumulative results of sets of frequently converging, but essentially independent, changes that linguistic items and constructions can undergo? If the former, we would expect there to be a rather strict separation between grammaticalized and pragmaticalized items. If the latter, we should not be surprised to find cases where either or both label(s) might seem appropriate. This would appear to be the case, for instance, with the Germanic modal particles mentioned above, which are characterized by scope increase, optionality, and pragmatic strengthening, but also – unlike discourse markers, for instance – by paradigmaticization and syntagmatic fixation.

7.

Whither the study of the semantics of context-level expressions?

At the synchronic level, more work is needed on the specific types of instructions that context-level items can convey, and on how and to what extent those types of instructions may differ from those conveyed by content-level functional items, not just in terms of their content, but also in terms of how semantic instructions coded by different linguistic items interact in the production of complete utterance interpretations. One avenue of research that has been growing in importance in recent decades is the study of the polyphonic or dialogic properties of context-level items, i.e. the extent to which context-level meanings code the points of view of distinct discourse individuals within one and the same utterance (e.g. Ducrot 1984; Nølke, Fløttum, and Norén 2004). While the polyphonic/dialogic approach to linguistic meaning does not limit itself to the context level, context-level items have thus far proved to be a particularly fertile area of study. Related to the issue of polyphony and dialogicity, at the diachronic level, recent work by Detges and Waltereit (2009) suggests that the different types of meanings conveyed by different classes of context-level items may be attributable to system-

The semantics of pragmatic expressions

605

atic differences in the ways these items arise diachronically. Based on a comparison of the evolution of two cognate adverbs, French bien, which functions like a modal particle, indicating the speaker’s attitude to the speech act expressed by its host utterance, and the Spanish discourse marker bien, whose function is to coordinate the joint construal of discourse, Detges and Waltereit hypothesize that modal particles in general arise from the negotiation of common ground in discourse, whereas discourse markers arise in contexts where the object of negotiation is a subsequent conversational move. As pointed out by an anonymous referee, adjectives and adverbs of the ‘good’/‘well’-type seem to develop into pragmatic markers not just in Romance, but in many other languages, as well. Another example of such polygenetic change would be perception verbs, which as mentioned in sect. 6.1 above are likewise a common source of pragmatic markers. Patterns like these suggest that there are, as suggested by Fleischman (1998: 47), certain pragmatic functions that are (more or less) universal, and which are therefore powerful sources of language change. The number and nature of such (quasi) universal functions deserves closer scrutiny. Current theorizing about the type of pragmatic inferencing that may trigger semantic change (cf. section 6.1) thus raises the further issue of the role and scope of the Uniformitarian Principle – i.e. the assumption that the processes underlying instances of language change in the past were essentially the same as those that can be seen to operate today (e.g. Labov 1994: 21) – in diachronic semantics and pragmatics. While this principle can probably be uncontroversially applied in the study of phonological change, for instance, given that the physiological properties of speech have not changed, it becomes more problematic in the analysis of meaning change; for, as a number of scholars have argued, speech acts and events, norms of politeness, principles of text structure and conversational routines are by no means directly comparable, neither across contemporary cultures nor across different historical stages of a given culture or society (e.g. Arnovick 1999; Jacobs and Jucker 1995; Scollon and Scollon 2001; Wierzbicka 1991, 2006). To the extent that this is true, how can we be justified in assuming that patterns of inference, and hence the fundamental pragmatic entities underlying them, such as presuppositions and implicatures, were similar to those that we take to be operative in contemporary Western discourse? In those cases where the nature of both an older “source” meaning and a newer “target” meaning of a given linguistic item or construction is well-established, it seems legitimate to suppose that the motivations for and mechanisms of extension/change which would most plausibly get us from source to target according to contemporary patterns of inference are likely to have been likewise instrumental in bringing about the change in the past.5 As soon as reconstruction is involved to a more significant extent, we are, however, on shakier ground. For this reason, as current theorizing about diachronic semantics and pragmatics develops and is refined, there can be little doubt that the field would benefit from being informed to a greater extent by insights from “pragmaphilology” (Jac-

606

Maj-Britt Mosegaard Hansen

obs and Jucker 1995) or “historical discourse analysis proper” (Brinton 2001), i.e. the synchronic study of discourse-pragmatic structures and functions and their corresponding means of expression in older texts. At present, however, there appears to be relatively little overlap among practioners of the two disciplines.

Notes 1. Although (23) – (28) make explicit a typical contrast between content- and context-level adverbial with respect to these tests, it is important to note that the predicted test results are not bi-implications. It was observed above that some items can function simultaneously at both levels; hence, it does not follow from prediction no. 1 that any item falling outside the focus structure of its host clause must function exclusively at the contextlevel, nor will an item with context-level meaning necessarily fail the tests in (26) – (28). Thus, the approximator almost, despite having truth-conditional meaning, fails the former set of tests, while passing the second set, in spite of its important rhetorical properties: (a) How did William die of an overdose? B. *Almost. (b) *William didn’t almost die of an overdose. (c) *It was almost that William died of an overdose.

2.

3.

4.

5.

(d) William almost died of an overdose. (e) William almost died of an overdose, and so did Ivan. (f) Did William almost die of an overdose? Thus, Fischer’s (ed.) (2006) state-of-the-art volume on discourse particles is structured in two parts, depending on whether the contributing authors assume a basically “polysemist” or a basically “monosemist” stance. These clines represent a revision of three tendencies posited by Traugott (1989: 34–35), viz.: I. Meanings based in the external described situation > meanings based in the internal (evaluative/perceptual/cognitive) described situation. II. Meanings based in the external or internal described situation > meanings based in the textual or metalinguistic situation. III. Meanings tend to become increasingly based in the speaker’s subjective belief state/attitude toward the proposition. As the name of their model suggests, Traugott and Dasher (2002) speak of “invited inferences” rather than of “conversational implicature”. The essential difference, however, appears to abide in the fact that the IITSC puts more explicit emphasis on speaker/hearer interaction and negotiation than do either Grice’s ([1975]1989) original model of implicature or Levinson’s (1995, 2000) theory of generalized implicatures. Even this cannot be taken for granted, however: when scholars rely on their semanticpragmatic intuitions, competing mechanisms may sometimes be invoked with equal probability to account for one and the same instance of change.

The semantics of pragmatic expressions

607

References Aijmer, Karin 2002 English Discourse Particles. Evidence from a Corpus. Amsterdam: John Benjamins. Anscombre, Jean-Claude and Oswald Ducrot 1983 L’argumentation dans la langue. Brussels: Mardaga. Arnovick, Leslie K. 1999 Diachronic Pragmatics. Amsterdam: John Benjamins. Barcelona, Antonio (ed.) 2000 Metaphor and Metonymy at the Crossroads. A Cognitive Perspective. Berlin: Mouton de Gruyter. Blakemore, Diane 1987 Semantic Constraints on Relevance. Oxford: Blackwell. Blank, Andreas 1997 Prinzipien des lexikalischen Bedeutungswandels am Beispiel romanischen Sprachen. Tübingen: Max Niemeyer. Blasco-Dulbecco, Mylène 1999 Les dislocations en français contemporain. (Les français parlés – Textes et études 1.) Paris: Honoré Champion. Brinton, Laurel J. 2001 Historical discourse analysis. In: Deborah Schiffrin, Deborah Tannen and Heidi E. Hamilton (eds.), Handbook of Discourse Analysis, 138–160. Oxford: Blackwell. Brinton, Laurel J. and Elizabeth Closs Traugott 2005 Lexicalization and Language Change. (Research Surveys in Linguistics.) Cambridge: Cambridge University Press. Brown, R. and A. Gilman 1972 The pronouns of power and solidarity. In: Pier Paolo Giglioli (ed.), Language and Social Context, 252–282. Harmondsworth: Penguin. Company Company, Concepción 2006 Subjectification of verbs into discourse markers: semantic-pragmatic change only? Belgian Journal of Linguistics 20: 97–121. Detges, Ulrich and Richard Waltereit 2009 Diachronic pathways and pragmatic strategies. Different types of pragmatic particles from a diachronic point of view. In: Maj-Britt Mosegaard Hansen and Jacqueline Visconti (eds.), Current Trends in Diachronic Semantics and Pragmatics, 43–61. (Studies in Pragmatics 7.) Bingley: Emerald. Diewald, Gabriele 2002 A model for relevant types of contexts in grammaticalization. In: Ilse Wischer and Gabriele Diewald (eds.), New Reflections on Grammaticalization, 103–120. Amsterdam: John Benjamins. van Dijk, Teun A. 1979 Pragmatic connectives. Journal of Pragmatics 3: 447–456. Dik, Simon C., Kees Hengeveld, Elseline Vester and Co Vet 1990 The hierarchical structure of the clause and the typology of adverbial satellites. In: Jan Nuyts, A. Machtelt Bolkestein and Co Vet (eds.), Layers and Levels of Representation in Language Theory, 25–70. Amsterdam: John Benjamins.

608

Maj-Britt Mosegaard Hansen

Dostie, Gaétane 2004 Pragmaticalisation et marqueurs discursifs. Analyse sémantique et traitement lexicologique. Brussels: De Boeck/Duculot. Ducrot, Oswald 1980 Analyses pragmatiques. Communications 32: 11–60. Ducrot, Oswald 1984 Le dire et le dit. Paris: Editions de Minuit. Ducrot, Oswald et al. 1980 Les mots du discours. Paris: Editions de Minuit. Eckardt, Regine 2003 The Structure of Change. Meaning Change Under Reanalysis. Habilitationthesis. Berlin: Humboldt Universität. (Revised version published in 2006 by Oxford University Press.) Eckardt, Regine 2009 APO: Avoid pragmatic overload. In: Maj-Britt Mosegaard Hansen and Jacqueline Visconti (eds.), Current Trends in Diachronic Semantics and Pragmatics, 21–41. (Studies in Pragmatics 7.) Bingley: Emerald. Erman, Britt and Ulla-Britt Kotsinas 1993 Pragmaticalization: the case of ba’ and you know. Studier i Modern Språkvetenskap 10: 76–93. Evans, Nicholas and David Wilkins 2000 In the mind’s ear: the semantic extensions of perception verbs in Australian languages. Language 76: 546–592. Fillmore, Charles 1985 Frames and the semantics of understanding. Quaderni di semantica VI(2): 222–254. Fischer, Kerstin 2006 Towards an understanding of the spectrum of approaches to discourse particles: introduction to the volume. In: Kerstin Fischer (ed.), Approaches to Discourse Particles, 1–20. (Studies in Pragmatics 1.) Oxford: Elsevier. Fischer, Kerstin (ed.) 2006 Approaches to Discourse Particles. (Studies in Pragmatics 1.) Oxford: Elsevier. Fleischman, Suzanne 1998 Des jumeaux du discours. La linguistique 34(2): 31–47. Foolen, Ad 1993 De betekenis van partikels. Doctoral thesis, University of Nijmegen. Fraser, Bruce 1996 Pragmatic markers. Pragmatics 6(2): 167–190. Gazdar, Gerald 1979 Pragmatics: Implicature, Presupposition and Logical Form. New York: Academic Press. Geeraerts, Dirk 1993 Vagueness’ puzzles, polysemy’s vagaries. Cognitive Linguistics 4(3): 223– 272. Grice, Paul [1975]1989 Logic and conversation. In: Paul Grice, Studies in the Way of Words, 22–40. Cambridge, MA: Harvard University Press.

The semantics of pragmatic expressions

609

Hansen, Anita Berit and Maj-Britt Mosegaard Hansen 2003 Le /´/ prépausal et l’interaction. In: Anita Berit Hansen and Maj-Britt Mosegaard Hansen (eds.), Structures linguistiques et interactionnelles dans le français parlé, 89–109. (Etudes Romanes 54.) Copenhagen: Museum Tusculanum. Hansen, Maj-Britt Mosegaard 1998a The Function of Discourse Particles. A Study with Special Reference to Spoken Standard French. (Pragmatics and Beyond New Series 53.) Amsterdam: John Benjamins. Hansen, Maj-Britt Mosegaard 1998b La grammaticalisation de l’interaction, ou, Pour une approche polysémique de l’adverbe bien. Revue de Sémantique et Pragmatique 4: 111–138. Hansen, Maj-Britt Mosegaard 2008 Particles at the Semantics/Pragmatics Interface: Synchronic and Diachronic Issues. A Study with Special Reference to the French Phasal Adverbs. (Current Research in the Semantics-Pragmatics Interface 19.) Oxford/Bingley: Elsevier/Emerald. Hansen, Maj-Britt Mosegaard and Jacqueline Visconti (eds.) 2009 Current Trends in Diachronic Semantics and Pragmatics. (Studies in Pragmatics 7.) Bingley: Emerald. Hansen, Maj-Britt Mosegaard and Richard Waltereit 2006 GCI theory and language change. Acta Lingvistica Hafniensia 38: 235–268. Heim, Irene 1982 The semantics of definite and indefinite noun phrases. Ph.D. thesis. University of Massachusetts, Amherst, MA. Heine, Bernd 2002 On the role of context in grammaticalization. In: Ilse Wischer and Gabriele Diewald (eds.), New Reflections on Grammaticalization, 103–120. Amsterdam: John Benjamins. Hirschberg, Julia 1991 A Theory of Scalar Implicature. New York: Garland. Hopper, Paul J. and Elizabeth Closs Traugott 1993 Grammaticalization. Cambridge: Cambridge University Press. Jacobs, Andreas and Andreas H. Jucker 1995 The historical perspective in pragmatics. In: Andreas H. Jucker (ed.), Historical Pragmatics, 3–33. Amsterdam: John Benjamins. Kay, Paul 1990 Even. Linguistics and Philosophy 13: 59–111. Kerbrat-Orecchioni, Catherine 1992 Les interactions verbales, Volume II. Paris: Armand Colin. Koch, Peter 2001 Metonymy: unity in diversity. Journal of Historical Pragmatics 2: 201–244. Koch, Peter 2004 Metonymy between pragmatics, reference, and diachrony. Metaphorik.de 7: 6–54. www.metaphorik.de König, Ekkehard 1991 The Meaning of Focus Particles. London: Routledge.

610

Maj-Britt Mosegaard Hansen

Labov, William 1994 Principles of Linguistic Change. Internal Factors. Oxford: Blackwell. Lambrecht, Knud 1994 Information Structure and Sentence Form. Cambridge: Cambridge University Press. Lehmann, Christian 1985 Grammaticalization: synchronic variation and diachronic change. Lingua e Stile XX(3): 303–318. Levinson, Stephen C. 1995 Three levels of meaning. In: Frank R. Palmer (ed.), Grammar and Meaning, 90–115. Cambridge: Cambridge University Press. Levinson, Stephen C. 2000 Presumptive Meanings. The Theory of Generalized Conversational Implicature. Cambridge, MA: MIT Press. Lyons, John 1977 Semantics, Volume II. Cambridge: Cambridge University Press. Nølke, Henning, Kjersti Fløttum and Coco Norén 2004 ScaPoLine. La théorie scandinave de la polyphonie linguistique. Paris: Kimé. Pons Bordería, Salvador 2006 From pragmatics to semantics: esto es in formulaic expressions. In: Maj-Britt Mosegaard Hansen and Ken Turner (eds.), Explorations in the Semantics/ Pragmatics Interface. Special issue of Acta Lingvistica Hafniensia 38: 180–206. Schwenter, Scott A. and Richard Waltereit 2010 Presupposition accommodation and language change. In: Kristin Davidse, Lieven Vandelanotte and Hubert Cuyckens (eds.), Subjectification, Intersubjectification and Grammaticalization, 75–102. (Topics in English Linguistics.) Berlin: De Gruyter Mouton Scollon, Ron and Suzanne Wong Scollon 2001 Intercultural Communication, 2nd edition. Oxford: Blackwell. Torres Cacoullos, Rena and Scott A. Schwenter 2007 Towards an operational notion of subjectification. In: Rebecca T. Cover and Yuni Kim (eds.), 31st Annual Meeting of the Berkeley Linguistics Society, 347–358. Berkeley, CA: Berkeley Linguistics Society. Traugott, Elizabeth Closs 1989 On the rise of epistemic meanings in English: an example of subjectification in semantic change. Language 57: 33–65. Traugott, Elizabeth Closs 1995 Subjectification in grammaticalization. In: Dieter Stein (ed.), Subjectivity and Subjectivisation, 31–54. Cambridge: Cambridge University Press. Traugott, Elizabeth Closs 2010 Revisiting subjectification and intersubjectification. In: Kristin Davidse, Lieven Vandelanotte and Hubert Cuyckens (eds.), Subjectification, Intersubjectification and Grammaticalization, 29–71. (Topics in English Linguistics.) Berlin: De Gruyter Mouton. Traugott, Elizabeth Closs and Richard B. Dasher 2002 Regularity in Semantic Change. (Cambridge Studies in Linguistics 97.) Cambridge: Cambridge University Press.

The semantics of pragmatic expressions

611

Tuggy, David 1993 Ambiguity, polysemy, and vagueness. Cognitive Linguistics 4(3): 273–290. Visconti, Jacqueline 2005 On the origins of scalar particles in Italian. Journal of Historical Pragmatics 6(2): 237–261. Waltereit, Richard 2002 Imperatives, interruption in conversation, and the rise of discourse markers: a study of Italian guarda. Linguistics 40(5): 987–1010. Wierzbicka, Anna 1991 Cross-Cultural Pragmatics. Berlin: Mouton de Gruyter. Wierzbicka, Anna 2006 English. Meaning and Culture. Oxford: Oxford University Press.

About the authors

Suzanne Beeke is Senior Lecturer in acquired language and communication disorders and Head of the Language and Communication Research Department at University College London, UK. She is a speech pathologist with over 18 years’ experience of carrying out research into aphasia, a language disorder most commonly caused by stroke. She has significant expertise in the use of Conversation Analysis to investigate aphasia, having received her doctorate in this area from the University of London. Her research seeks theoretical understandings of aphasia in the lexical, grammatical, prosodic and non-verbal features of everyday conversations in the home. In addition, she is evaluating a rehabilitation programme that directly targets the conversation skills of people with aphasia by working with them to increase their understanding of aphasia and the mechanics of conversation, and to develop strategies to aid taking a turn-at-talk in an interaction. She is currently leading a project funded by the UK Economic and Social Research Council that will create a website to support speech pathologists who wish to carry out conversation-based rehabilitation programmes with people with aphasia and their families. This work is being carried out in partnership with Connect, the communication disability network, a UK charity for people living with aphasia. Geert Brône is Assistant Professor in cognitive discourse analysis and German linguistics at Lessius University College Antwerp and an affiliate researcher at the University of Leuven, Belgium. He received his PhD from the University of Leuven with a dissertation that developed a cognitive-linguistic and discoursesemantic approach to verbal humour. He has been a visiting scholar at the University of California, San Diego. His research interests include the interface between cognitive grammar and (interactional) discourse, linguistic creativity, cognitive poetics, multimodal aspects of interaction (including gesture and gaze analysis) and empirical methods for interaction/discourse analysis. He has published a monograph (Peter Lang, 2010), three edited volumes (John Benjamins, Mouton de Gruyter), a special issue of the journal Humor, as well as a series of paper in journals and edited volumes (including Journal of Pragmatics, Cognitive Linguistics, Humor, Discourse Processes and Style). Anneloes Canestrelli is a PhD candidate at the Utrecht Institute of Linguistics OTS. She received her Research MA in Linguistics in Utrecht in 2008 and is currently working on her dissertation on the processing of causal relations and con-

614

About the authors

nectives. Her research interests include discourse processing, psycholinguistics and eye-tracking. Robyn Carston is Professor of Linguistics and Graduate Tutor at University College London and Research Coordinator at the Centre for the Study of Mind in Nature, University of Oslo. Her main research interests are in pragmatics, semantics, relevance theory, non-literal meaning, and metarepresentation. Her approach is interdisciplinary, drawing on cognitive scientific approaches to language and communication and on the philosophy of language. She has published a monograph Thoughts and Utterances: the Pragmatics of Explicit Communication (Blackwell, 2002) and is preparing a collection of her more philosophically-oriented papers to be published under the title Pragmatics and Semantic Content, by Oxford University Press. She is currently working on two main projects: ‘Understanding metaphor: ad hoc concepts and imagined worlds’ (supported by the Leverhulme Trust) and ‘Word Meaning – What it is and What it isn’t’ (supported by the Arts and Humanities Research Council, UK). Louise Cummings is Professor of Linguistics at Nottingham Trent University in the UK. She conducts research in pragmatics and clinical linguistics, as well as a number of philosophical disciplines. She is the author of Pragmatics: A Multidisciplinary Perspective (Edinburgh University Press, 2005), Clinical Linguistics (Edinburgh University Press, 2008), Clinical Pragmatics (Cambridge University Press, 2009), Rethinking the BSE Crisis (Springer, 2010) and Communication Disorders (Palgrave Macmillan, 2012). She recently edited The Routledge Pragmatics Encyclopedia (Routledge, 2010) and is currently editor of the Handbook of Communication Disorders (Cambridge University Press). She has held Visiting Fellowships in the Department of Philosophy at Harvard University and in the Centre for Research in the Arts, Social Sciences and Humanities (CRASSH) at Cambridge University. Louise Cummings is the special issues editor of the Journal of Pragmatics and is on the editorial board of Pragmatics and Pragmatics and Society. Alice Deignan is a Reader in Applied Linguistics in the School of Education, University of Leeds, UK. She received her doctorate from the University of Birmingham, UK. Her research interests concern aspects of non-literal meaning in naturally-occurring texts, especially metaphor and metonymy, but also pragmatic meaning. She uses corpus linguistic techniques and text analysis in her work. She is the author of Metaphor and Corpus Linguistics (John Benjamins, 2005) and coedited Researching and Applying Metaphor in the Real World (with Graham Low, Zazie Todd and Lynne Cameron, John Benjamins, 2010). She is co-author, with Jeannette Littlemore and Elena Semino, of Figurative Language, Genre and Register (Cambridge University Press, forthcoming), and is currently working on a monograph on figurative language and collocation.

About the authors

615

Małgorzata Fabiszak is a University Professor at Adam Mickiewicz University (AMU) in Poznan, Poland. She received her doctorate and her Habilitation from AMU. She has been invited to give lectures and workshops at the University of Wolverhampton, UK, University Las Palmas de Gran Canaria, Spain, University Complutense, Madrid, Spain, and at Lancaster University, UK. Her research interests concern three major areas: corpus-based cognitive linguistics, metaphor theory and critical historical discourse analysis, in particular in connection with with memory and identity studies. Her monographs include The concept of joy in Old and Middle English. A semantic analysis (Piła: WSB, 2001), Language and Meaning. Cognitive and functional perspectives (ed. Peter Lang, 2007), A conceptual metaphor approach to war discourse and its implications (AMU Press, 2007). She has also published articles in Poetica, Language Sciences and Poznan Studies in Contemporary Linguistics. Nadège Foudon is currently working at a school for autistic children and teaching at EESP in Lausanne and at the University of Fribourg (Switzerland). She received her PhD (on language acquisition in autistic children) from the University of Lyon2 (France). She also holds a postdoctoral position at Octogone-CERPP (University of Toulouse, France) on the early diagnosis of autism. Her research interests concern different aspects of autism, joint attention, language play and emotions. Rachel Giora is Professor of Linguistics at Tel Aviv University. Her research areas include discourse coherence, informativeness, discourse relevance, cognitive pragmatics, language and ideology, women and language, and the language of literary texts. Her recent work focuses on the psycholinguistics and neurolinguistics of figurative language (irony, sarcasm, jokes, and metaphor), context effects, optimal innovation and aesthetic pleasure, discourse negation, default non-literal interpretation, and the notion of salience, relevant also to autism, context, and metaphor. She has published about 100 articles, and her book On Our Mind: Salience, Context, and Figurative Language was published by Oxford University Press in 2003. Together with Patrick Hanks she edited 6 volumes for Routledge’s Critical Concepts series entitled Metaphor and Figurative Language. Alison Hall is a British Academy Postdoctoral Research Fellow in Linguistics at University College London. She received her doctorate from UCL in 2009. She has also been a postdoctoral researcher in the Context-Dependence, Perspective, and Relativity group at Institut Nicod in Paris. Her research interests are in theoretical pragmatics, philosophy of language, and semantics, with a focus on the semanticspragmatics distinction, pragmatic enrichment, unarticulated constituents, subsentential utterances and ellipsis, lexical pragmatics and the nature of ad hoc concepts, and relevance theory.

616

About the authors

Maj-Britt Mosegaard Hansen has been Professor of French Language and Linguistics at the University of Manchester, United Kingdom, since 2007. She was previously Assistant /Associate Professor at the University of Copenhagen, Denmark, where she obtained both her PhD (1996) and her Higher Doctorate (2008). Her research interests include synchronic and diachronic semantics and pragmatics, grammaticalization, functional and cognitive linguistics, interactional linguistics, discourse analysis, and Peircean semiotics. She is the author of two monographs, The Function of Discourse Particles. A Study with Special Reference to Spoken Standard French (Pragmatics & Beyond New Series 53, John Benjamins, 1998) and Particles at the Semantics/Pragmatics Interface: Synchronic and Diachronic Issues. A Study with Special Reference to the French Phasal Adverbs (Current Research in the Semantics-Pragmatics Interface 19, Elsevier/Emerald, 2008), and has published numerous journal articles and book chapters on topics such as aspectual, discourse, and modal particles, negation, interrogative structures, reported speech, and schwa-tagging. Peter Harder is Full Professor of English Linguistics at the University of Copenhagen, Denmark. He received his Dr phil (Habilitation) in 1996 and became a member of the Royal Danish Academy of Sciences and Letters in 1999. He was Visiting Scholar at the University of California, San Diego in 1994, and Visiting Professor at Berkeley in 2003. His research interests include the semantics of grammar, language politics and normativity, grammaticalization and linguistic theory. His main works are Functional Semantics (1996) and Meaning in Mind and Society (2010), both at Mouton de Gruyter. William Horton is an Associate Professor in the Department of Psychology at Northwestern University. He received his PhD in Cognitive Psychology from the University of Chicago. His research interests lie primarily within the domain of experimental pragmatics, with a particular interest in issues related to narrative comprehension, figurative language, conversational perspective taking, common ground, reference, and age-related changes in discourse processing. He is also a member of the Governing Board of the Society for Text and Discourse. Yan Huang (PhD Cambridge, DPhil Oxford) is Professor of Linguistics and Head of Department of the Department of Applied Language Studies and Linguistics at the University of Auckland. He has previously taught linguistics at the universities of Cambridge, Oxford and Reading, where he was Professor of Theoretical Linguistics. His main research interests are in pragmatics, especially the pragmaticssemantics interface and pragmatics-syntax interface including anaphora. His books include The Syntax and Pragmatics of Anaphora (Cambridge University Press, 1994, re-issued in 2007), Anaphora: A Cross-Linguistic Study (Oxford University Press, 2000), Pragmatics (Oxford University Press, 2007) and The Oxford

About the authors

617

Dictionary of Pragmatics (Oxford University Press, 2012). His Pragmatics has been, and is being, translated into a number of languages. He has also published numerous articles and reviews in leading international journals of linguistics. He is on the editorial board of a number of international journals and research monograph series. He has been invited to lecture at around 100 universities and research institutes in many countries in Europe, North America, Asia, Australasia, and North Africa. Istvan Kecskes is Full Professor of Linguistics and Communication at the State University of New York, Albany, USA. He holds a PhD in Linguistics from Kossuth University, Debrecen, an academic degree from the Hungarian Academy of Sciences and an honorary degree and professorship from Zhejiang University, China. His latest publications include five books (Foreign language and mother tongue, Erlbaum, 2000, with T. Papp; Situation-bound utterances in L1 and L2, Mouton, 2003; Cognitive aspects of bilingualism, Springer, 2007, with L. Albertazzi; Explorations in pragmatics, Mouton, 2007, with L. Horn; Intention, common ground and the egocentric speaker-hearer, Mouton, 2008, co-edited with J. Mey) and numerous articles published in prestigious journals such as Journal of Pragmatics, Pragmatics & Cognition, International Journal of Bilingualism, Second Language Research, International Journal of Multilingualism, Intercultural Pragmatics, WORD and others. Professor Kecskes is the founding editor of the linguistics journal Intercultural Pragmatics and the Mouton Series in Pragmatics published by Mouton de Gruyter: Berlin/New York, as well as the new bilingual (Chinese-English) journal CASLAR (Chinese as a Second Language Research). He sits on the editorial board of the Journal of Pragmatics (Elsevier) Pragmatics & Society (Benjamins), International Journal of Multilingualism (Taylor & Francis), Intercultural English (Singapore) and the Journal of Foreign Languages (Waiguoyu) published in China. Dr. Kecskes is also on the editorial board of four book series: Pragmatics, Philosophy and Psychology (Springer), Series in Pragmatics (Cambridge Scholarly Publishing), Pragmatic Interfaces (Equinox), and Általános Nyelvészeti Tanulmányok (Hungarian Academy of Sciences). Brooke Lea, PhD, is Professor and Chair of the Psychology Department at Macalester College in St Paul, MN. He received a BA in English from Haverford College before earning his PhD in cognitive psychology at New York University where he studied psychological models of deduction with Prof Martin Braine. After NYU, Brooke spent two years at the University of Massachusetts-Amherst as a postdoctoral fellow. At UMass, Brooke developed his interests in reading and discourse processing. He has been a Visiting Professor at the University of Minnesota’s Center for Cognitive Science, where he continues to be a member. His publications concern both logical and pragmatic comprehension processes that occur during the reading of narrative texts and poetry; he has also co-authored two sta-

618

About the authors

tistics textbooks. A recent NSF grant supports his interest in using eye-movement methodologies to address numerous psycholinguistic questions. Sabine Manificat is a child psychiatrist at Centre Hospitalier Saint Jean de Dieu, Lyon, France. Her clinical domain of interest is autism spectrum disorders. She was involved for 15 years in Quality of Life Research (in SCRIPT INSERM team), and developed QoL questionnaires for children and adolescents. She is currently in charge of various outpatient services offering support for ASD children and adolescents and their families. Jacques Moeschler is Full Professor of French Linguistics at the University of Geneva. He received his doctorate from the University of Neuchâtel, Switzerland. As a post-doctoral student he was a Visiting Researcher at the Ecole des Hautes Etudes en Sciences Sociales, Paris, the University of California, Los Angeles, Georgetown University, Washington DC and the University of California, Berkeley. He has been Visiting Professor at the Universitad Autonoma de Madrid, the Université Paris-Diderot, the Ecole Polytechnique Fédérale de Lausanne and the Università degli Studi, Verona. His research interests mainly concern semantics and pragmatics, and more precisely inferential and cognitive pragmatics. His publications are about discourse and argumentation, tenses, discourse connectives, causality, negation and logical words. His monographs include Dire et contredire (Peter Lang, 1982), Argumentation et conversation (Hatier, 1985), Modélisation du dialogue (Hermès, 1989) and Théorie pragmatique et pragmatique conversationnelle (Armand, Colin, 1996); with Anne Reboul, the Dictionnnaire encyclopédique de pragmatique (Seuil, 1994), La pragmatique aujourd’hui (Seuil, 1998), Pragmatique du discours (Armand Colin, 1998); with Antoine Auchlin Introduction à la linguistique contemporaine (Armand Colin, 1997, 2003, 2009); with Sandrine Zufferey, Initiation à la linguistique française (Armand Colin, 2010). He is the editor of one issue of Intercultural Pragmatics on The French Tradition in Pragmatics (2006, vol. 3, n° 4). Terttu Nevalainen is Professor and Chair of English Philology at the University of Helsinki and Director of the Research Unit for Variation, Contacts and Change in English, a National Centre of Excellence funded by the Academy of Finland (2000–12). She has studied at University College London and been a Visiting Scholar at the Universities of Cambridge and Sheffield. Her research interests include historical sociolinguistics, corpus linguistics, and language change. She is the author of over 100 journal articles and book chapters, of An Introduction to Early Modern English (Edinburgh University Press, 2006) and Historical Sociolinguistics; Language Change in Tudor and Stuart England (Pearson Longman, 2003, with H. Raumolin-Brunberg), co-editor of several thematic volumes on language variation and change, the English editor of Neuphilologische Mitteilungen,

About the authors

619

and the editor-in-chief of the Oxford Studies in the History of English. Dr Nevalainen is one of the original compilers of the Helsinki Corpus of English Texts and the director of the research project “Sociolinguistics and Language History”. The Corpus of Early English Correspondence, compiled by this project, provides the data for a large number of publications on English historical sociolinguistics. Daniela O’Neill is Full Professor of Psychology at the University of Waterloo, Canada. She received her doctorate from Stanford University. Her research concerns the interplay between children’s developing understanding of mind and their communicative development, with a focus on early pragmatic language development. Her research has included studies of toddlers’ adaptation of their requests to a parent’s knowledge state; children’s gesture use in word learning and storytelling; topic initiations in peer-to-peer conversation; the ability of children to “step into the shoes” of story characters; parent-child talk during picture-book sharing; and relations between children’s early narrative ability and mathematical and abstract thinking. She is the developer of the Language Use Inventory (LUI), a parent-questionnaire and assessment tool designed to help identify delays or impairment in children’s early language use. Anne Reboul is Senior Researcher at the National Center for Scientific Research (CNRS, France) and co-director of her present laboratory. She has a PhD in Language Sciences (E.H.E.S.S., Paris, 1984), a PhD in Philosophy (University of Geneva, Switzerland, 1990) and a Habilitation thesis (University of Strasbourg, France, 1992). She has been Visiting Scholar at UC Berkeley (1985), Honorary Research Fellow at the Department of Linguistics and Phonetics, University College London (1984–1986), Visiting Professor at Tokyo University of Foreign Studies (2010–2011), and has held various research positions, first at FNSRS (Switzerland, 1986–1992), before joining CNRS in 1993. She was awarded the CNRS Bronze Medal for Linguistics in 1997. Her research interests centre on pragmatics, mostly in the theoretical framework of Relevance Theory. She has worked on many topics and her recent interests have been the acquisition of language in autistic children, as well as the empirical investigation of the processes implicated in the recovery of implicitly communicated information in both implicatures and presuppositions. She has written several books, including Le Dictionnaire encyclopédique de pragmatique (with J. Moeschler, Le Seuil, 1992), Pragmatique du discours (with J. Moeschler, Armand Colin, 1998), La pragmatique aujourd’hui (with J. Moeschler, Le Seuil, 1998) and Langage et cognition humaine (Presses Universitaires de Grenoble, 2007). Ted Sanders is Full Professor of Discourse Studies and Dutch Language Use at Utrecht University. His research concentrates on discourse structure and coherence in discourse representation and language acquisition, as well as the cognitive pro-

620

About the authors

cesses involved in discourse production and understanding. He headed a substantial so-called VICI research project on Causality and Subjectivity in Discourse and Cognition, funded by the Dutch organization for scientific research (NWO, 2006–2011). Together with co-authors he published in edited books and in journals, such as Cognitive Linguistics, Discourse Processes, Journal of Child Language, Journal of Pragmatics, Reading and Writing, Text, and Written Communication. He co-edited special issues of Cognitive Linguistics (2001), Discourse Processes (2004), Information Design Journal (2007) and Journal of Pragmatics (2012), as well as the book volumes Text representation: Linguistic and psycholinguistic approaches (Benjamins, 2001) and Causal categories in discourse and cognition (Mouton de Gruyter, 2009). He is currently Head of the Department of Dutch in Utrecht. During the spring semester of 2007 Sanders was a Visiting Scholar at the Department of Linguistics and the Institute for Cognitive Studies, University of California, Berkeley. Hans-Jörg Schmid is Full Professor of English Linguistics at Munich University, Germany, where he also obtained his PhD and his Habilitation. His main research interests are in cognitive linguistics, lexical semantics, word-formation and pragmatics as well as corpus linguistics and quantitative methods. His book publications include: English abstract nouns as conceptual shells. From corpus to cognition (Mouton, 2000); An introduction to cognitive linguistics (Pearson Longman, 2nd edition 2006, co-authored with Friedrich Ungerer; translated into Korean, Japanese and Chinese); Windows to the mind: Metaphor, metonymy, and conceptual blending (Walter de Gryuter, 2010; co-edited with Sandra Handl); Cognitive foundations of linguistic usage patterns. Empirical approaches (Mouton de Gruyter, 2011; co-edited with Susanne Handl); English morphology and word formation. An introduction (Erich Schmidt, 2011). His recent work focuses on the interplay of cognitive and sociopragmatic factors in the emergence and change of linguistic systems. Murray Singer is Professor of Psychology at the University of Manitoba in Winnipeg, Canada. He received his doctorate from Carnegie Mellon University in Pittsburgh. He has also served as Visiting Scholar or Research Associate at the University of California, San Diego, the University of Colorado, and Stanford University. His specialty field is adult language comprehension, with emphasis on inference processing, validation of discourse congruence, text retrieval, question answering, and individual differences in comprehension. He has also scrutinized the cognition and metacognition of discourse retrieval and item recognition. He has served as editor of the Canadian Journal of Experimental Psychology and as President of the Canadian Society for Brain, Behaviour, and Cognitive Science. He is Fellow of the latter society, the American Psychological Association, Division 3, and the Society for Text and Discourse. He is author of Psychology of language:

About the authors

621

An introduction to sentence and discourse processes (Erlbaum, 1990) and co-editor of Reading and language processing (with J. Henderson and F. Ferreira; Erlbaum, 1995). John Taylor was Senior Lecturer in Linguistics at the University of Otago, New Zealand until his retirement in 2010. He previously taught in South Africa (University of the Witwatersrand) and in Germany. He graduated in Romance Languages and Linguistics at the University of Cambridge and received his doctorate with a dissertation in acoustic phonetics at the University of Trier. His research interests are in Cognitive Grammar, especially in relation to lexical semantics, syntactic analysis, and phonology. He is co-editor of Language and the cognitive construal of the world (1995) and Cognitive approaches to lexical semantics (2003), published by Mouton de Gruyter. His monographs include Linguistic categorization (3rd edition 2003), Possessives in English (1996), Cognitive Grammar (2002), and The mental corpus (2012), all published by Oxford University Press. Graeme Trousdale is Senior Lecturer at the University of Edinburgh. He has published a number of articles on variation and change in English in journals such as English Language and Linguistics and English World Wide. He is the author of An introduction to English sociolinguistics (2010), and the co-editor of Constructional approaches to English grammar (ed. with Nikolas Gisborne, de Gruyter, 2008), Gradience, gradualness and grammaticalization (ed. with Elizabeth Closs Traugott, Benjamins, 2010), and the forthcoming Oxford handbook of construction grammar (ed. with Thomas Hoffmann, Oxford University Press, 2012).

622

Subject index

623

Subject index

A abductive change 514 accessibility 9, 14, 73, 202–203, 206–207, 241, 248, 391, 395, 422, 494, 603 accommodation 18, 384, 559–566, 571–572, 577–578, 580 – long-term accommodation 564–565 – pragmatic accommodation 398 – presupposition accommodation 384 – short-term accommodation 565 acquisition – lexical acquisition 318, 320, 322–326, 329 – of pragmatic competence 11, 14–15, 261, 264–270, 272, 299 see also language acquisition activation 13, 16, 86, 95–96, 101, 158–161, 175, 182, 202–206, 218, 220–221, 234–235, 239, 241, 466, 472, 474, 477–478, 484, 487–488, 490 active zone 13, 137, 143, 155–157, 159, 161 act see acts of identity, communicative act, indirect speech act, speech act acts of identity 559, 569–570 actuation 560–561, 567, 581 – actuation problem 561 agrammatism 16, 346–347, 353–359, 362–364 ambiguity 14, 26, 50, 77, 154, 237, 304, 361, 411–413, 420, 458, 478, 482, 496, 602 – ambiguity resolution 14, 38, 142, 235–236, 303, 309, 377 – lexical ambiguity 137, 268, 303, 309, 466 anaphora 138, 143, 273, 511 – anaphora resolution 9–10, 12, 88 and-conjunction 58, 62, 67, 78 – causal interpretation of 51–52, 57–58 animacy 233

aphasia 11, 16, 292–293, 308, 317, 345–359, 362–363 – Broca’s aphasia 345, 353, 355 – fluent aphasia 294, 308, 353, 356–357 – Wernicke’s aphasia 345, 357 appropriateness 346, 486–487 approximator 544, 546, 592, 606 ASD see autistic spectrum disorders Asperger syndrome 15, 292, 295, 297, 301, 303–304, 308, 317–320, 323, 326–329, 331–332, 335–339 audience design 9, 387, 396 autism 11, 15–16, 266, 268, 271, 292, 294–295, 297–301, 303–309, 317–339 autistic spectrum disorders 11, 15–16, 266, 292, 295, 297, 304–309, 318–320, 322, 330–335, 337–339 availability principle 38, 64, 66–67 – domain availability principle 139 B banter 9 bilingualism 570–571 blending see conceptual blending BNC see British National Corpus bridging inferences see inferences British National Corpus 169, 444, 543–544 Broca see aphasia broadening 61, 70 C CA see conversation analysis calculability 414–415 cancellability 52, 414–415 cataphor 205 categorization 123, 126, 128–129, 162, 472 causal – causal bridge 90 – causal connective see connective – causal inference see inference – causal proposition 87 – causal relation 201, 210, 212–220

624

Subject index

centering theory 203–205 change-by-accommodation 577 clinical pragmatics 5, 291, 298, 307 COCA see Corpus of Contemporary American English cognitive effort 381, 388, 390, 392, 422 cognitive environment 380–381 cognitive linguistics 4–5, 13–14, 124, 142, 144, 157, 175, 179–181, 220, 222–223, 464, 470, 474, 478, 484, 490, 492, 494, 510–511, 515, 519, 525, 534, 553 cognitive model 125–126, 141, 143, 471, 475–476, 481 cognitive pragmatic 3, 10–11, 13, 15, 17–18, 56–57, 221, 263, 266, 270–272, 274, 276–277, 463–464, 484, 507, 533–534, 548–550, 553 cognitive principle 8, 10–13, 15, 23, 37, 278, 477 – of relevance 4, 27–28, 428 coherence 9, 14, 86–90, 96, 107, 201–203, 210–222, 449, 469 – coherence relations 14, 202, 210–214, 217, 220, 222 – referential coherence 14, 202–210, 220, 222 – relational coherence 14, 202, 210–213, 219 cohesion 9, 14, 201–202, 210, 273, 306, 336 cohesive tie 88, 202 collaboration 5, 102, 219, 510, 522 collocation 167–168, 234–236, 455–457, 510, 514, 551 common ground 5, 9, 16, 59, 177, 186, 190–193, 303, 325, 375, 382–398, 487, 493, 605 common knowledge 191, 329, 378, 382 communication accommodation see accommodation communication accommodation theory 563 communicative act 13, 48, 76, 123, 186, 295, 378, 490 communicative intention 7, 55, 71, 294–295, 297, 301, 306, 336, 375–376, 384–385, 387, 408, 426 communicative principle of relevance 27–30, 41, 428

community membership 382–383 compositionality 62, 77, 155 comprehension 15, 56, 61, 76, 127, 153, 235, 237, 299, 304, 308–309, 387, 390, 422, 475, 488, 507 – and context 384 – and encyclopaedic knowledge 176–178, 192–193 – and production 42, 176, 269, 272, 390, 507 – and salience 235 – auditory comprehension 345, 364 – discourse comprehension 12–13, 85–108 – idiom comprehension 301–302 – narrative comprehension 206–207 – of implicatures 17, 297, 406, 409, 425 – of non-literal language 17, 70, 87, 255, 268–269, 272, 293, 488 – online comprehension 66, 69–70, 216–219 – reading comprehension 345 – speech act comprehension 295 – text comprehension 217–218 – utterance comprehension 42 comprehension heuristic (relevance-theoretic) 29–31, 422 comprehension time 88, 90 conceptual blending 6, 12–13, 135–136, 138, 143, 464, 475–476, 480–482, 491 conceptual domain see domain conceptual integration 135, 470, 481, 491, 493 see also conceptual blending conceptual metaphor see metaphor conceptual network see network conceptual pact 385–386, 394 conceptualization 126, 129, 158, 222, 397, 478, 482, 526, 533 conformity 559 connective 18, 51–52, 101, 202, 210–220, 222–223, 427, 538, 589, 600 – additive connective 213–214, 219 – adversative connective 220 – causal connective 213–214, 217, 219–220, 222 – contrastive connective 222 – logical connective 51, 101

Subject index construal 3–19, 131, 222, 373, 423, 435, 466, 468, 470, 476, 481–483, 493, 533–536, 542–544, 547–554, 605 construct 548, 550 construction grammar 180, 478, 482–483, 514–516, 524, 554 see also macro-construction, meso-construction, micro-construction constructional pattern 483, 515 constructionalization 18, 533–534, 542–543, 546, 548–553 construction-integration model 13, 86, 96, 106 content-level 590–592, 594–597, 600–601, 603–604 context 27–29, 39–40, 48, 51, 58–61, 91–92, 153, 155, 158, 160, 168, 179–180, 233, 235–242, 293–294, 298, 413, 421–424, 438, 440, 489, 507, 542, 547, 553, 592, 594 – and pragmatic disorders 302–305, 332–334 – communicative context 261, 296, 381–385, 390 – conversational context 267, 375 – critical context 547 – discourse context 94–95, 205, 447, 454, 475–476 – implicit context 38 – interactive/interactional context 142, 354 – isolating context 547 – linguistic context 8–9, 153, 159, 169, 188, 376, 412 see also cotext – physical context 376 – private context 188–189, 191 – sequential context 349–350, 354 – situational context 8–9, 142, 177–178, 188, 191–192, 510, 589 – social context 262, 294, 439, 509, 561, 565 context of culture 9 context-dependent 4, 6, 12, 15, 50, 53, 124, 139, 425, 589, 593 context-free meaning see meaning context-level expression 18–19, 590–598, 600–601, 603–604, 606 context-level meaning see meaning

625

context-sensitivity 47, 56, 62, 79, 451, 457 contextual effect 27–28, 380–381 contextual expressions 397–398 contextual implication 12, 27, 29–30, 71, 78, 380, 421–422, 426 contextualism 36 convention 268, 270, 363, 377–378, 421, 428, 474, 514 conventional implicature see implicatures conventionality 18, 40, 139, 143, 156, 191, 233–234, 271, 415–416, 449, 475–476, 488, 494, 514, 517, 533, 543, 548, 595, 597–598, 600 conversation analysis 11, 16, 176, 345–351, 353–355, 357, 359, 362–363, 511 – methodology of 346, 349–350, 357 conversational coaching 347 conversational implicature see implicature conversational maxim see maxim of conversation cooperative principle 9, 26, 32, 37, 129, 411, 485–496, 590 coordination problem 377–378 copresence heuristics 381, 383, 387 Corpus of Contemporary American English 142, 169, 540 cotext 8–10, 452 see also linguistic context critical discourse analysis 17, 134, 205, 439–440 critical metaphor analysis 447 critical period hypothesis 328–329 cue phrases 202, 210–211, 215–218 cultural knowledge 7, 9, 269, 273 cultural model 9–10, 14, 134, 175, 177–193, 479 D Danish 594, 595, 602 deduction 42, 97, 100, 101, 422 deductive reasoning 13, 85, 97–103, 422, 514 deductive bridges 98 default inference see inference definite description 59, 66 definite reference see reference deictic expression 15, 124–126, 135, 293, 299

626

Subject index

deixis 9, 38, 126, 135, 138, 143–144, 272, 298–299, 405 dementia 296, 303, 306, 309, 347, 363 detachability 415, 590 determinacy 415, 423 direct access view 15, 233, 238–240, 244, 246–249, 251, 254 discourse analysis 133, 307, 438, 508–510, 606 discourse comprehension 12–13, 93, 97 discourse framing 131, 574 discourse markers 18, 354, 538, 604–605 discourse processing 5, 100, 203, 207, 210, 212–213, 216, 218, 220–222 discourse topic 89, 202, 205, 207 disorder, pragmatic 5, 11, 14–15, 266, 291–309, 317–318, 336–337, 345–348, 353, 359, 362–363 domain – cognitive/conceptual domain 10, 12, 128, 132, 137–138, 140, 153, 180 – domain activation 158–162 – domain availability principle 139 – domain-general processing 393–396 – domain highlighting 137 – domain restriction 58–59, 70 – source domain 132, 444, 478 – target domain 132, 135, 478 Dutch 212, 214–215, 218, 525–526 E Early Modern English 575 egocentrism 177–178, 190, 193 elaborative inference see inference embedding test 78 embodiment 13, 25–26, 92, 104–105, 123, 126–127, 135, 141, 164, 168, 475–476 emergence – of common ground 190, 193, 398 – of meaning from discourse 457 – of new constructions 533 – of structure/grammar from usage 6, 10–11, 509–518, 521, 524, 553 emergent grammar 18, 511–512, 516, 519 emergentist 11, 262, 508, 512, 518, 521 encoding 61–62, 73, 77–78, 88–89, 91–92, 94–96, 107, 204, 300, 333, 375, 395, 516, 534–535

encyclopaedic knowledge 10, 12, 14, 16, 73, 130, 137, 141, 158–159, 175–176, 178–181, 192, 482 English 7, 35, 91, 127, 129, 160, 185, 213–214, 235, 267, 269, 351–357, 376, 409, 418–419, 424, 426, 444, 448, 452, 455, 513, 521–522, 524–525, 534, 537, 540–541, 543–544, 546, 549, 552, 554, 562, 570–571, 575, 577–579, 589, 594–595 énoncé 427 enrichment 9–10, 12, 30, 47, 57–60, 68, 70, 76–77, 334, 425, 523, 535 entailment 34, 72–74, 78, 87, 151, 155, 417, 425, 444–445, 447, 593 entrenchment 10, 17, 143, 159, 169, 474, 507, 514, 523, 549, 551, 602 ERP 94, 104, 220, 240, 477 ethnomethodology 348, 511 evaluative framing 131, 133 evaluative function see functions (of language) evolutionary perspective 6, 55, 509–510, 517, 520–521, 581 executive function theory 309 experiential grounding 126–127, 181, 233, 264, 511 see also gestalt experimental evidence 39, 167, 211, 390–391, 526 experimental pragmatics 5, 38–39, 406 explicature 9–10, 12, 29, 31, 38, 47–51, 53, 55–78, 425, 427 explicit-implicit distinction 55, 63, 68 expression minimization 33 eyetracking 203, 205–206, 212, 389, 477 F face 270, 324, 443–444, 448, 456, 560–561 face-work 561 figurative language 9–11, 17, 437–455, 457–458, 470 – density of figurative language 438, 443, 449 Finnish 354, 357, 562, 566–567, 571–574, 581 fixed expression 185–186, 235, 440, 443–444, 448, 473–474 flouting 26, 296, 411, 426, 485

Subject index fMRI 90, 127, 220–221 folk model 175 formulaic language 185–186, 355 frame 9, 13–14, 123, 131, 135, 140–141, 143, 180–182, 450, 465, 468, 475–477, 480, 494, 515, 561, 593–594 frame semantics 129–130, 483 frame-shifting 131, 466, 476–477, 481, 492 frequency 139, 166, 233–234, 382, 449, 473, 488, 507, 513–516, 519, 522, 525, 577, 600 full constraint model (of common ground) 387, 389 functions (of language) – evaluative function 17, 437, 443–445, 448, 450, 457, 486 – ideational function 440–443, 449–453, 457 – ideological function 439, 443, 448, 468, 580 – interpersonal function 17, 274, 347, 440–445, 447–448, 450–451, 457, 511, 563, 574, 591 – textual function 17, 165, 210–211, 217, 440, 442, 448–451, 457, 466, 468–469, 471, 511–512, 547, 606 G garden-path sentence 104 general theory of verbal humour 17, 464, 467–470, 474, 476, 478, 483, 492–494 generalized implicature see implicature generalized invited inference see inference generative grammar 130, 339, 464, 508, 511, 515 genre 91, 135, 216, 218, 438–439, 441, 446, 448, 451, 454–455, 467, 469, 561, 574–575, 601 gestalt 142, 477, 494 – experiential gestalt 142, 494 given information 88–89, 106, 299, 309 graded informativeness 472 graded salience 15, 139, 473, 487–488 – graded salience hypothesis 233, 235, 238–240, 244, 247–248, 251, 254 grammatical form 456 grammaticalization 18, 516–517, 533–542, 547–549, 552–553, 603–604

627

grounding 105, 126, 142, 383–385, 398, 482, 492 grounding criterion 385 GTVH see general theory of verbal humour H Hebrew 247 hidden indexical see indexical Horn-scale 32, 34 host class expansion 544, 548, 551, 553 humour 9–11, 17, 267–268, 304, 306, 334, 463–471, 474–484, 488, 492–494 Hungarian 537, 540 I ICM see idealized cognitive model ideal language philosophy 55 idealized cognitive model 13–14, 123–126, 129, 131, 134–135, 141–143, 175 ideology 445, 579 idiom 301–302 see also fixed expression idiom principle 185, 510 illocutionary act 129, 140–142, 306, 405, 425 image schema 13, 125–129, 132, 143 impairment 14–16, 261–262, 266, 268, 277, 291–295, 297, 300–301, 304–308, 318, 321, 328, 332, 336–338, 345–347, 353–354, 356–357, 359, 362–363 implicated conclusion 29–30, 48, 73–74, 422, 425–426 implicated premise 29–30, 73, 78, 425 implicature 49, 74, 405 – conventional implicature 11, 16, 78, 405–410, 413–414, 416, 425, 589, 590 – conversational implicature 9, 11, 25–26, 30, 32, 35, 38–41, 47, 49–50, 53, 63, 67, 139, 296, 375, 376, 388, 405, 410, 412–414, 425, 479, 589, 597–599, 606 – generalized implicature 30, 40, 64, 71, 413, 423, 425–427, 598, 606 – particularized implicature 30, 39–40, 63–64, 71, 296, 413, 423, 427, 598 – potential implicature 421 – Q-implicature 32–36, 41 – R-implicature 33

628

Subject index

– scalar implicature 32, 34, 38–39, 41, 277, 297, 335, 406, 416–420, 424, 427 – strong implicature 423 – weak implicature 421, 423–424 implicatureF 26, 426 implicatureO 26, 426 implicit meaning see meaning inappropriateness 237, 486–487, 492 indefinite reference see reference indeterminacy 572 indexical 50, 54, 57–58, 63, 70, 76–77, 302, 539, 541–542, 579 indexicals 47, 51, 53, 56–59, 62, 74–75, 77 – covert indexicals 57–58, 77 – hidden indexicals 59, 77 indirect 4, 16, 47, 64, 73, 140, 142, 268–270, 275, 294–296, 329, 332, 385, 420, 464, 479, 485, 488–489, 494, 568 indirect speech act 4, 16, 142, 269, 294, 296, 329, 420 inference 13, 18, 29, 37, 39, 40–41, 53, 64, 66, 69–71, 74–75, 85–90, 93–108, 134, 139, 141, 211–212, 215, 219, 268, 334, 376–377, 380–381, 388, 406, 408, 421–422, 479, 483, 522, 547, 548, 600, 605 – bridging inference 13, 88–94, 96, 99, 108 – causal inference 90–93, 102, 421 – deductive inference 97–103 – default inference 39, 426 – elaborative inference 13, 94–96, 99–100, 104–105 – figurative inference 87 – generalized invited inference 543, 548, 551–552, 598 – invited inference 18, 533, 543, 546–552, 598, 600, 606 see also invited inferencing (theory of semantic change) – pragmatic inference 29, 40–41, 62, 65–66, 68, 71, 97, 99–100, 102, 266, 299, 425, 484, 597 – predictive inference 95 – scalar inference 78 inferencing 5, 9–10, 12, 16, 107, 139–140, 143, 294, 543, 605

informative intention 407–408, 425–426 innovation 18, 177, 473–475, 492, 494, 560–562, 569, 578, 601–602 instantiation 13, 33, 141, 153, 155, 160, 163, 186–188, 192–193, 483, 598 intention-reading 9, 15, 266 interpersonal function see functions (of language) interpretive framework 351–352 invited inference see inference invited inferencing (theory of semantic change) 598 I-principle 34–35, 420–421, 428 irony 9–11, 17, 241, 244, 246–248, 254–255, 266–268, 293, 301, 329, 332–333, 336, 376, 405, 437, 443, 445, 458, 463–464, 475, 481, 484–494 Italian 168, 540, 600 J joint action 384 joint attention 9, 266, 324–325, 328 joke 17, 131, 267–268, 275, 356, 466–469, 471–472, 474–477, 480 K knowledge representation 17 see also common knowledge, cultural knowledge, encyclopaedic knowledge, mutual knowledge, shared knowledge, world knowledge L language acquisition 15, 213–214, 261–272, 317–329, 337–338, 508, 513 see also acquisition of pragmatic competence language change 11, 18, 510, 533–534, 536, 539, 541, 543, 550, 553, 559–565, 568, 573–574, 577, 579–581, 602, 605 language comprehension see comprehension language development 11, 15, 261, 276, 278, 317 language production see production Late Middle English 575 Latin 538, 540–541, 600, 602

Subject index left-dislocation 587 left-hemisphere damage 293, 297 lexicalization 18, 300, 517, 533–535, 537–542, 553 lexicon 162, 175, 337–338, 417, 425, 454, 482, 511, 516, 518–519, 540, 587 see also mental lexicon LHD see left-hemisphere damage linguistic copresence 382–383, 385 linguistic cotext 8–9 logic 41, 49, 52, 97–98, 102–103, 124, 335, 389, 405–406, 408, 410–411, 414, 425, 468 logical form 29–30, 51, 55, 58, 62–64, 69, 71, 73, 75, 77, 407, 409, 422 logical mechanism 468, 470 logical square 418–419 long-term accommodation see accommodation loosening 538 M macro-construction 543, 551 Manner (maxim of conversation) 26, 31–32, 411, 420 marked informativeness 471–474, 494 maxim of conversation 12, 25–26, 31–37, 50, 53, 67, 139, 207, 269, 278, 296–298, 306, 335, 346, 376, 381, 386, 411–415, 418, 420–422, 426–428, 463, 465, 485–487, 590 see also Manner, Quality, Quantity, Relation Middle English 358–359, 540–541, 544, 546–547, 551, 575 meaning – context-free meaning 47, 53, 55 – context-level meaning 592, 594, 597–598, 601, 603–604, 606 – implicit meaning 8–9, 29, 38, 47, 87, 106, 133, 155–156, 181, 211–213, 217–220, 234, 330, 394, 396, 406, 425, 444, 446, 489, 550, 562 – non-literal meaning 8–9, 11, 16–17, 19, 139, 240, 373, 435 – non-natural meaning 405–407, 409, 425–426 – speaker meaning 36–37, 42, 53–54, 63, 69

629

– truth-conditional meaning 410, 417, 588, 606 – utterance meaning 49, 64, 141 – utterance-type meaning 40 – word meaning 13, 56, 60–62, 74, 77, 130, 152, 165, 167, 180, 272, 305, 447, 477 meaning construction 178, 188 meaning negotiation 383 meaning-in-context 3–8, 10–11, 13–14, 16–17, 373, 505, 533–534, 547–548, 550 memory 13, 78, 86–87, 98, 100–101, 106–107, 123, 135, 161, 166–167, 184–185, 190, 210, 213, 317, 382, 387, 394–397, 469, 475–477, 480 memory-based (text processing) 86, 101, 106–107, 394–397 mental lexicon 33, 152–157, 159, 165–168, 179, 233–234, 397, 471, 473–474 mental model 85, 175–177, 181–182, 188 mental space 14, 126, 135–136, 215, 222–223, 464, 468, 475, 478, 480, 482, 490–492 mental state 3, 8–9, 265, 267, 270, 273, 302, 329–330, 336, 394 mentalese 103 meso-construction 543, 548 metaphor 5, 9, 12–13, 17, 71, 87, 123, 127, 131–135, 137, 141, 143, 157, 236–237, 239–241, 248–251, 254, 293, 300–301, 329, 332–333, 336, 437–451, 453–454, 456–458, 464, 470, 472, 478, 481–483, 493, 554, 596, 599 – conceptual metaphor 17, 127, 131–135, 143, 440, 447, 451, 454, 478, 493 – generative metaphor 133 – primary metaphor 132 – systematic metaphor 449, 451 metonymy 9, 12–13, 17, 59, 70–71, 123, 125, 137–141, 143, 154–159, 437, 440–441, 445, 449, 458, 464, 470, 478–482, 554, 596, 599 micro-construction 548, 551–552 mindful pragmatics 15, 263, 265, 271–272, 276–277 minimal proposition 74–75 minimalism 36, 55, 79, 596

630

Subject index

misinterpretation 480 modality 129, 136, 142–143 Modern English 536, 538, 540 Modified Occam’s Razor 417 modularity 398 modus ponens 99–100 monosemy 595 motivation hypothesis 325–326 multi-word expressions 18, 587 see also fixed expressions mutual knowledge 16, 102, 293, 375, 377–383, 386–390, 393, 395, 397 mutual knowledge paradox 378–379, 381 mutual manifestness 380, 383 mutual understanding 11, 16, 345, 348, 351, 375, 384–385, 396 N narrative strategy 468–469 narrowing 61–62, 69 negation 34–36, 67–68, 76, 98, 100–101, 130, 236, 252, 254, 335, 406, 410, 417, 419, 427, 488, 575–576, 589 negative particular 418–419, 427 neo-Gricean 12, 25, 30–31, 34, 36, 38–42, 72, 74, 420, 422–423, 428, 486 network – conceptual network 10, 12–13, 123, 158–159, 167, 482 – constructional network 548, 551–552 – knowledge network 175, 180 – metaphorical network 447, 451 – propositional network 85–86, 105 – social network 18, 539, 559–565, 568–569, 577, 580 new information 27, 88–89, 267, 299–300, 383, 422 non-calculable 414 non-detachable 414–415 non-literal language 8, 15, 17, 71, 233, 268–269, 293, 300–301, 306, 464, 485 non-literal meaning see meaning non-natural meaning see meaning non-verbal communication 15, 55, 233, 262, 265–266, 276, 297, 305–307, 318, 334, 345, 352, 359, 480 Norwegian 363

O Old English 129, 535–538, 540–541, 546–547, 549 optimal audience design 16, 387 optimal innovation 236, 471, 473–477, 480, 483, 492, 494 optimal relevance 27, 29, 41, 57, 68–69, 76, 330, 380, 422, 428 ordinary language 51–52, 89, 97, 105 P parsing 55, 86, 104, 203, 206 partial constraint model 391 particle 97, 593–594, 596, 598, 604–606 particularized implicature see implicature perspective adjustment model 394 perspective taking 390, 392 philosophy of language 36, 49–50, 56, 405, 408–409, 425 phonology 124, 135, 164–167, 169, 235, 268, 339, 405, 541, 548, 563, 572, 577, 587, 603, 605 physical copresence 382–383 pidgin 570 politeness 142, 255, 270, 294, 297, 329, 560–561, 605 polyfunctionality 19, 589, 594–595 polysemy 9, 13–14, 19, 129, 142, 152–155, 159, 164, 166, 168, 235–237, 239, 255, 451, 466, 551, 595–596, 600 – polysemy resolution 14, 236 potential implicature see implicature pragmatic accommodation see accommodation pragmatic competence 5, 7, 9–12, 14–15, 23, 138, 263–265, 273–274, 298, 353–355, 357, 363 pragmatic expressions 11, 18, 587–588, 590 pragmatic inference see inference pragmatic marker 218, 587–589, 597, 605 pragmatic processes 4–5, 11, 16, 18, 47, 55–56, 62–63, 69, 78, 326, 330, 337, 398 pragmatically enriched 29, 31, 34, 36, 38, 64, 75 predictive inference see inference presupposition 9, 130, 144, 299–300,

Subject index 307, 309, 384, 405, 484, 492, 598–599, 601 presupposition accommodation see accommodation pretence (theory of irony) 484, 489–491 primary metaphor see metaphor priming 124, 206, 249, 393–394, 396, 467 principle of relevance 4, 12, 27–28, 37, 41, 380–381, 384, 422, 428 privileged information 388 procedural processing 19, 78, 176, 470, 539, 541–542, 592–593, 597 procedurality 603 production 3, 34, 42, 100, 136, 158, 176, 185, 193, 247–248, 268–269, 272, 296, 299–300, 320, 326, 330–331, 345, 348, 387, 390–391, 397, 455, 492, 507, 520, 550, 604 profiling 137, 478, 533, 546 pronominal/pronoun resolution 206 propagation 560–561, 602 proposition 12, 47–48, 50–61, 63–64, 66–71, 73–79, 86–87, 89, 99–100, 104, 108, 130, 144, 383, 408–410, 412, 421, 423–426, 596–597, 606 propositional logic 13 prototype 32, 123–125, 143, 162, 167, 472, 494 prototypical structure 13, 123–124 punch line 465–469, 471–472, 474–477, 493 Q Q-implicature see implicature Q-principle 31–35, 420, 422, 427–428 qualia 160 Quality (maxim of conversation) 26, 31, 34, 411 Quantity (maxim of conversation) 26, 31–32, 386, 411, 420, 427–428 R radial structure 124–125 radical construction grammar 516 reading time 94, 104, 203, 205, 209, 211–213, 215–219, 240–241, 244–245, 250, 477, 482

631

reanalysis 62, 477, 537, 541–543, 548–550, 552, 598 reasoning 9–10, 12–13, 16, 42, 48, 55, 75, 97–99, 101–102, 175, 181, 214–215, 268, 376–377, 388, 412, 425, 427, 454, 470, 481, 548 recipient design 191, 387 reference 14, 38, 105, 137, 156, 160, 202–206, 299, 376–377, 382, 386–387, 391, 396, 405–406–409, 412, 421, 550–551 – definite reference 376, 382 – indefinite reference 89, 205, 537, 575 – reference assignment 59, 69, 334 – reference chain 139 – reference resolution 30 – reference tracking 5, 9 reference point 141, 156, 181, 272, 479, 536 referring expression 203, 376, 382, 385–386, 388–389, 391, 395–396 register 32, 448, 451, 453–455, 457–458, 493, 569 Relation (maxim of conversation) 25–26, 32, 36, 411, 420 relevance 4–5, 12, 25–26, 37, 68, 207, 233, 244, 248–249, 251, 254, 330, 333, 381, 405, 421–423, 425–428, 472, 489, 590 see also cognitive principle of relevance, communicative principle of relevance, principle of relevance relevance theory 5, 12, 15, 25–26, 28–31, 37–40, 42, 47–49, 51, 55–56, 63, 68–69, 71–72, 74–75, 78, 233, 244, 248–249, 251, 254, 264, 292, 307, 330, 333, 421–423, 425–426, 489, 590 repair 267, 307, 349–351, 356–359, 361–363, 450 resonance 90, 101, 106–107, 241–242, 247, 395–396, 481 riddle 267–268, 468 right-hemisphere damage 15, 292, 296, 299, 301, 304–305 R-implicature see implicature routinization 13, 182, 514 R-principle 31–33, 420, 428 Russian 127, 144

632

Subject index

S salience 10–14, 155, 161–162, 178, 193, 204, 233, 235, 237–242, 244–252, 254, 390, 392, 471, 473–474, 488, 535, 577–578 sarcasm 17, 267–269, 304, 306, 464, 475, 481, 490, 494 satisficing 390 scalar implicature see implicature scalar inference see inference scanning 533 scenario 65–66, 98, 134, 136, 140–142, 163, 175, 185–186, 188, 190, 192, 208–210, 293, 301, 444–445, 448, 491, 517–518 schema 100, 135, 141, 181, 309, 382 see also image-schema schizophrenia 297–298, 301, 303–306, 308 script 9, 14, 192, 382, 464–468, 470, 476, 478, 494 script opposition 467–468, 478, 492 self-monitoring 357, 559 semantic content 49–50, 52–54, 67, 74–76, 137, 158–159, 356, 441, 457 semantic minimization 33 semantic prosody 154 semantic script theory 463–465 semanticization 543, 548, 551 semantic-pragmatic interface 18, 152, 588, 590 sense disambiguation 9–10, 12 shared intentionality 324–325, 338 shared knowledge 5, 9, 11, 13, 16–17, 123, 177, 190, 193, 268, 294, 375, 378–379, 381–383, 392, 396–397, 601 short-term accommodation see accommodation Sign Language 135 situational context see context SLI see specific langaue impairment social cognition 188, 265, 274, 337, 579 social competence 7, 9, 265, 273 social identity 18, 560–561, 569, 574, 577 social network see network social norm 7, 9, 490 socialization 13, 123, 183

social pragmatics 15, 263, 269, 273–277 socio-cognitive approach 175–178, 180, 182, 186, 189–192 sociopragmatic approach 11, 18, 324, 560–563 source domain see domain space structuring model 475–477, 480 Spanish 77, 571, 598, 600, 605 speaker innovation 560–562, 580 speaker meaning see meaning speaker presupposition 383 specific language impairment 15, 294–295, 297, 328, 332, 335 specificity 48, 186–187, 223, 326, 525, 534, 536, 543, 549 speech act 15, 51, 75, 77–78, 92, 128–129, 131, 140–143, 187, 214, 269, 291, 293–296, 306, 308, 330, 346, 405, 414, 485, 593, 602, 605 spoken discourse 223, 437, 439, 450 standard pragmatic model 238–240, 244, 247–249, 251, 254, 485 stimulus onset asynchrony 235 strengthening 27, 35, 246, 421–422, 549–550, 603–604 strong implicature see implicature structure building framework 206 subjectification 534–536, 542–543, 550, 552–554, 602–604 subjective culture 183–184 subjectivity 214, 223, 534–536, 543, 554 subjectivization 18 suppression hypothesis 472 surface form 34, 77, 85, 106, 469 syllogism 13, 98, 100, 102 synonymy 165 systematic metaphor see metaphor T target domain see domain textual see functions (of language) 17, 165, 210–211, 217, 440, 442, 448–451, 457, 466, 468–469, 471, 511–512, 547, 606 thematic link 14 thematic development 205 thematic discontinuity 207 thematic role 483

Subject index theory of mind 16, 262, 265–266, 294–295, 297, 306, 309, 329–331, 333, 335–337, 388 ToM see theory of mind traumatic brain injury 293, 305–306, 308 truth-condition 38–39, 47, 50–56, 59, 64–67, 71, 76–79, 409–410, 413, 415, 417, 424–425, 428, 587–591, 597, 606 truth-conditional content 39, 47, 50, 55–56, 59, 64–67, 71, 588 truth-conditional meaning see meaning truth-value 409, 592 U unarticulated constituents 56–57, 60, 62, 70, 77 underdeterminacy 63 usage-based approach 11, 13, 18, 477, 507–511, 514–519, 521–526, 533, 543, 560, 581 utterance content 47, 56, 63 utterance meaning see meaning utterance-type meaning see meaning

633

V validation 86, 93, 102, 165, 492 viewpoint 18, 480–481, 533–536, 543, 550, 572 visual negation 252–254 W weak central coherence 309 weak implicature see implicature Wernicke see aphasia what is implicated 39, 41, 76, 408–409, 413 what is meant 3, 12, 17, 19, 26, 248, 268, 408, 491 what is said 3, 12, 17, 19, 26, 29, 31, 35, 38–39, 41, 50, 52–55, 62, 64–69, 73–76, 248, 268, 375, 408–410, 413–415, 423 word meaning see meaning working memory 86, 91, 101, 103, 106, 209, 396 world knowledge 7–10, 55, 85–86, 88, 93–94, 96, 101–103, 152, 180, 208, 522, 598

634

Name index

Name index

A Abbeduto, Leonard 265, 267, 295 Abelson, Robert 90, 141, 175, 382, 465, 494 Abrahamsen, Eileen P. 301 Adams, Catherine 262 Aijmer, Karin vi, 588 Albrecht, Jason E. 86, 92–93, 96, 106–107 Alexander, Richard J. 478 Alkabets-Zlozover, Inbar 236, 249 Allan, Keith 270 Allen, Doris A. 292, 308 Allwood, Jens 41 Almor, Amit 220 Aloni, Shulamit 250 Altmann, Gerry 55 Alvesson, Mats 176 Andersen, Elaine S. 275, 514 Andersen, Gisle vi Andersen, Henning 514 Anderson, Alan Ross 41 Anderson, Anne 208, 394 Anderson, Anthony 85, 91 Anderson, John R. 105 Anderson, Richard C. 94, 161, 163 Andersson, Roger 567 Anscombre, Jean-Claude 591 Antilla, Raimo 537, 540 Antonopoulou, Eleni 471, 482–483, 493 Apperly, Ian. A. 388, 392 Archakis, Argiris 469 Archer, Dawn 561 Ariel, Mira 5, 14, 202, 206, 220, 396 Arnold, Jennifer E. 203–204, 206, 391 Arnovick, Leslie K. 605 Arnseth, Hans Christian 190 Arppe, Antti 526 Aschkenazi, Keren 236, 249, 252 Aslin, Richard N. 323 Asperger, Hans 15, 292, 295, 297, 301, 303–304, 308, 317–320, 323, 326–329, 331–332, 335–339 Astington, Janet W. 265–266, 269

Atkins, B. T. Sue 153, 180 Atkinson, John M. 349 Atlas, Jay D. 32, 41–42, 420–421, 428 Attardo, Salvatore 17, 463–470, 474, 478, 484–487, 489–490, 492–494 Atwood, Tony 332, 339 Aubert, Sandrine 305 Auer, Peter 18, 565, 569–570, 577, 580 Austin, John v, 51, 262, 330, 346, 405, 408–409, 487 B Bach, Kent 28–29, 38, 41–42, 54, 76, 78, 376, 379, 409, 424, 428 Bache, Carl 511, 522 Baddeley, Alan D. 94 Baicchi, Annalisa 141–142 Bailey, Guy 565–566 Bailey, Karl G. D. 106 Baillet, Susan D. 90, 212–213 Baird, Jodie A. 267 Balaban, Noga 251–252 Baldwin, Dare A. 266 Baltaxe, Christine A. 300 Bandura, Albert 190 Bara, Bruno 3, 6, 268 Barcelona, Antonio 156, 479, 599 Barclay, J. Richard 92, 105 Bargh, John A. 393–394 Barghouti, Omar 251 Bar-Hillel, Yehoshua 507 Barker, Gregory 223 Barlow, Jack M. 443 Barlow, Michael 507 Baron, Anne vi Baron-Cohen, Simon 303, 324, 331, 388 Barr, Dale 178, 190, 388–390 Barresi, Barbara 345, 353 Barsalou, Laurence 104, 162–163, 494 Bartak, Lawrence 309 Bartlett, Brendan J. 218 Barton, Michelle E. 266

Name index Bartsch, Andreas J. 221 Bartsch, Karen 266–267 Bastiaanse, Roelien 353, 364 Bates, Elisabeth 262, 265–266, 269, 325, 356, 516 Bazin, Nadine 303–304 Bearman, Peter 338 Becker, Guido 86, 91, 93 Beeke, Suzanne 5, 16, 293, 345, 347–348, 353–357, 359, 363–364 Beeman, William 270 Bell, Alan 165 Belnap, Noel 41 Benczes, Réka 161 Benveniste, Emile 427 Bergen, Benjamin 91, 475, 482–484, 493 Bergen, Doris 267 Berger, Peter. L. 183 Bergs, Alexander 542, 553 Berkum, Jos J. A. van 104 Bern, Hyman 212 Berndt, Rita S. 353 Bernicot, Josie 268–269 Bernstein, Richard. J. 176 Bertus, Eugenie L. 91 Bestgen, Yves 207, 223 Bezuidenhout, Anne 61 Biber, Douglas 523 Bierwisch, Manfred 160, 168 Binsted, Kim 483–484, 493 Bishop, Dorothy V.M. 5, 292, 305, 308 Black, John B. 212 Blakemore, Diane 27, 78, 590, 592 Blank, Andreas 599 Blasco-Dulbecco, Mylène 587 Blijzer, Femke De 215 Bloch, Steven 348, 353 Bloom, Lois 213, 262, 275 Bloom, Paul 319, 323, 325 Blum-Kulka, Shoshana 276 Bluth, George J. 217–218 Bock, J. Kathryn 309 Bodenhausen, Galen 189 Body, Richard 347, 474 Bollinger, Susan ix Borg, Emma 31, 55, 74, 76, 79, 426 Bosch, Peter 61 Bouillon 160

635

Bower, Gordon H. 92 Boyd, Richard 441 Boye, Kasper 517, 519 Braine, Martin D. S. 97, 100–101, 103, 517 Brandt, David M. 217–218 Bransford, John D. 89, 92, 105, 221 Breheny, Richard 5 Brennan, Susan E. 204, 384–386, 391, 394 Brenneise-Sarshad, Robin 303 Bresnan-Enders, Maura ix Brewer, William F. 383 Brinton, Bonnie 262, 275 Brinton, Laurel J. 537–539, 541–542, 549, 603, 606 Brisard, Frank, Steven Frisson 239–240, 250 Britton, Bruce K. 211 Brock, Alexander 468, 474 Broek, Paul van den 212 Brône, Geert 17, 463, 467, 470–471, 474–475, 478–482, 493 Brooke-Rose, Cristina 456 Brown, Colin M. 86, 104, 107 Brown, Gillian 438 Brown, John S. 176 Brown, Mary 490 Brown, Penelope 270 Brown, Polly 90, 212–213 Brown, R. 591 Brown-Schmidt, Sarah 16, 390, 393 Brugman, Claudia 157, 169 Brüne, Martin 305 Bruner, Jerome 262, 271, 273 Bryson, Susan E. 324 Bublitz, Wolfram v, vi, ix, 3–5, 41 Bucciarelli, Monica 268–269 Bunge, Mario 177, 190 Burton-Roberts, Noel 29 Butler, Christopher S. 511 Buttrick, Samuel 387, 390 Bybee, Joan 18, 166, 214–215, 513–515, 517, 521, 533, 553 Byrne, Ruth M. 101 C Caballero, Rosario 440–441, 451, 454 Calvo, Manuel G. 95 Camaioni, Luigia 266, 325

636

Name index

Cameron, Catherine A. 267 Cameron, E. Leslie 267 Cameron, Lynne 5, 17, 134, 440–444, 449–451, 453–454, 456–458 Campion, Nicolas 100 Canestrelli, Anneloes R. 12, 14, 201, 215 Capelli, Carol. A. 269 Cappelen, Herman 55, 58, 74–75, 79 Capps, Lisa 271, 309 Carlomagno, Sergio 303 Carlson, Greg N. 95 Carlson, Thomas B. 381, 386–387, 390 Carpendale, Jeremy 266 Carpenter, Patricia A. 103–104, 106, 203 Carreiras, Manuel 91 Carston, Robyn 12–13, 48, 55, 58, 61–62, 64, 68, 71, 75–79, 87, 175, 406, 421, 426, 428 Castillo, M. Dolores 95 Cattaneo, Luigi 337 Ceponiene, Rita 323, 337 Chafe, Wallace L. 208, 267 Chambers, J.K. 578, 580 Chaminaud, Stephanie 268–269 Chandler, Michael J. 266, 483 Charniak, Eugene 94 Charteris-Black, Jonathan 134, 446–448 Chartrand, Tanya L. 393–394 Cheng, Patricia W. 90, 97 Chevallier, Coralie 335 Chiarello, Christine 249–250 Chierchia, Gennaro 39–40 Chilton, Paul 134, 439 Chomsky, Noam 55, 169, 407, 510, 524 Chong, Jennifer L. 93 Choueka, Yaakov 153 Chunyo, Gitana 327 Cienki, Alan 493 Clancy, Patricia M. 270 Clark, Eve V. 213, 264, 266, 269, 271–272, 397 Clark, Herbert 5, 16, 88, 106, 191, 206, 213, 377, 379, 381–387, 390, 393–394, 397, 420, 477, 481, 487, 489–491 Clark, Tedra F. 305 Clausner, Timothy 533 Cleland, Alexandra A. 393–394 Clifford, James 512

Code, Chris 355 Coelho, Carl A. 293, 308 Coffin, Caroline 446 Cohen, L. Jonathan 67, 424 Cole, Peter 405 Colle, Livia 268, 309, 336 Colston, Herbert L. 127, 178, 190, 237, 239, 463, 484 Coltheart, Max 345 Company Company, Concepción 603 Cook, Anne E. 96 Cook, Guy 445 Cooper, Judith 328 Copeland, David E. 86, 92 Corbett, Albert 89 Corcoran, Rhiannon 297 Corts, Daniel P. 449–450 Cosmides, Leda 55, 97 Costerman, Jean 207 Coulson, Seana 5, 17, 131, 136, 144, 255, 463–468, 471–472, 475–478, 480–482, 490–492, 494 Coulthard, Malcolm 439, 510 Couper-Kuhlen, Elizabeth 223 Coupland, Justine 18, 387, 563–564 Coupland, Nikolas 18, 387, 559, 561, 563–564, 570, 574, 580 Cowan, William B. 393 Cox, James R. 97 Cozijn, Reinier 212 Crain, Stephen 103, 106 Croft, William 123, 137–138, 158, 180, 510, 515–516, 518, 520, 533–534, 542, 554, 581 Cross, David 270, 331 Cruse, Alan D. 123, 137, 180, 515, 534 Culler, Jonathan 512 Culpeper, Jonathan 561 Curcó, Carmen 474, 494 Curtis, Mary E. 88 Cutica, Ilaria 293 Cutting, Cooper 237 D D’Agostino, Paul R. 96 D’Andrade, Roy 175, 179, 181 Da˛browska, Ewa 521 Dahl, Östen 517

Name index Dahlgren, Svenolof 309 Dale, Philip 266 Damico, Jack 348, 351, 353, 355 Dancygier, Barbara 223 Daneman, Meredyth 105 Dasher, Richard B. 18, 543, 546, 548–549, 551–552, 592, 597–598, 600, 603, 606 Davies, Mark 25, 169, 540, 543 Davis, G. Albyn 346–347, 362 De Smet, Hendrik 550, 552, 554 Deacon, Terrence 520 Dean, Elizabeth 346 Deane, Paul 157 Decarie, Therese Gouin 272 Degand, Liesbeth 214, 216–217, 223 Deignan, Alice 5, 17, 87, 134, 437, 440–441, 444–445, 450–452, 455–457 Dell, Gary S. 391 Demorest, Amy 269 Denis, Michel 94 Dennett, Daniel 331 Detges, Ulrich 604–605 Dewell, Robert 157 Dias, Maria da 98 Dickerson, Paul 348 Diewald, Gabriele 542, 547, 553, 601 Dijk, Teun A. van 86, 105, 134, 163, 589 Dik, Simon C. 591 DiMaggio, Paul 181 Dipper, Lucy T. 292 Dirven, René 6 Docherty, Gerard 581 Dolitsky, Marlene 467 Donald, Merlin 271, 355–356 Donnellan, Keith 66, 376 Dostie, Gaétane 597 Drager, Katie 167 Drew, Paul 347 Drummond, Sakina S. 295 Du Bois, John 241, 251, 512 Dubois, Daniele 94 Duchan, Judy 207, 348, 351 Ducrot, Oswald 410, 590–592, 594, 604 Duffy, Susan A. 90, 203, 206, 213, 234 Duguid, Paul 176 Dunham, Philip J. 324 Dunn, Judy 266–267, 274–275 Dunn, Michelle 326

637

Durham, Mercedes 581 Durkheim, Emile 183, 521 Durleman, Stephanie 328 E Eales, Martin 336 Eckardt, Regine 597–598, 601, 603 Eckert, Penelope 579–580 Edwards, Susan 356, 364 Ehrlich, Kate 88 Eilan, Naomi 324 Eisele, Julie A. 300 Elbourne, Paul 77 Ellis, Andrew W. 355 Emerich, David 334 Erbaugh, Mary S. 270 Erickson, Thomas D. 105 Ericsson, K. Anders 469 Erman, Britt 603 Ervin-Tripp, Susan 264, 269, 275 Escandell-Vidal, Victoria 276 Estevez, Adelina 95 Evans, Nicholas 600 Evans, Vyvyan 13, 99, 123, 152, 157, 159, 175, 179–180 Evers-Vermeul, Jaqueline 213–214 Eviatar, Zohar 235, 239 F Fabiszak, Małgorzata 13–14, 123, 134, 470 Faerch, Claus 265 Fairclough, Norman 134, 439 Fauconnier, Gilles 6, 135–136, 138, 144, 175, 222, 468, 470, 475–476, 480–482, 492 Faust, Mark E. 106 Fecica, Agnieszka M. 272–273 Fein, Ofer 234–236, 239–241, 244–246, 248–252, 254–255, 485, 488 Feldman, Carol 271 Feldman, Michael. S. 188 Ferraro, Vittoria 106 Ferreira, Fernanda 91, 94, 106 Ferreira, Victor S. 391–392 Ferretti, Todd 236, 240–241 Feyaerts, Kurt 467, 470–471, 474, 478–482, 493

638

Name index

Feyereisen, Pierre 303 Filippova, Eva 269 Fillmore, Charles 89, 130–131, 141, 153, 175, 180, 465, 482, 515, 523, 526, 590 Fincher-Kiefer, Rebecca 96 Fischer, Kerstin 606 Fischer, Olga 542, 587 Fitch, W. Tecumseh 407 Fleischman, Suzanne 512, 605 Fletcher, Charles R. 86 Flewellyn, Laura 293 Fløttum, Kjersti 604 Fodor, Jerry 55, 74, 103, 105, 233–234, 237, 330, 398, 464 Foolen, Ad 596 Forceville, Charles 482 Ford, Celia E. 223, 514, 523 Fortune, Liane 581 Foster, Susan H. 264, 269 Foudon, Nadège 15, 292, 294, 317, 320 Foulkes, Paul 581 Fox, Danny 427–428 France, Lisa 70, 320, 338, 411 Frank, Stefan L. 97 Franks, Bridget A. 97 Franks, Jeffery J. 92, 105 Fraser, Bruce 587 Frattali, Carol M. 346 Frazier, Lyn 87, 104, 235 Frege, Gottlob 50, 53, 55, 405 Fried, Mirjam 543 Friedman, Marilyn W. 204 Frith, Christopher D. 297, 317, 329, 331–332, 388 Fritz, Anna S. 266 Fromm, Davida 346 Fujiki, Martin 262, 275 ´ Fyrberg, Åsa 305 G Gabora, Liane 478, 482 Gal, Meir 253 Gallagher, Tanya M. 261–262, 273 Gallerani, Catherine 267 Ganzi, Jonathan 255 Garman, Michael 356 Garnham, Alan 380 Garnica, Olga K. 272

Garrett, Merrill 39 Garrod, Simon 5, 89, 91–92, 94, 203, 208, 393–394, 396 Garst, Jennifer 189 Garvey, Catherine 267 Gastbeg, Holly Zajac 323 Gazdar, Gerald 42, 398, 405, 410, 416–417, 421, 426, 428, 588 Gee, James. P. 182 Gee, John Paul 272 Geeraerts, Dirk ix, 6, 20, 123–124, 181, 223, 521, 525, 596 Geis, Michael 420, 546 Gentner, Dedre 441 Gentner, Donald R. 441 Gernsbacher, Morton A. 5, 106, 203, 205–207, 237 Gerrig, Richard 377, 387, 394–397, 448, 489 Geurts, Bart 42 Gherardi, Silvia 176 Giacalone Ramat, Anna 540 Gibbs, Raymond 5, 15, 17, 127, 139, 153, 168, 233, 237–239, 244, 246, 251, 254, 268, 380–381, 384–386, 437, 439–440, 448, 463, 479, 484–485, 488–489 Gibson, James J. 520 Gigerenzer, Gerd 390 Giles, Howard 18, 387, 563–564 Gilliom, Laura 204–205, 220 Gillund, Gary 395 Gilman, A. 591 Giora, Rachel 5, 13–15, 17, 87, 139, 178, 233–236, 239–241, 244–252, 254–255, 463–464, 466–467, 471–478, 480, 483–485, 487–490, 492, 494 Girouard, Pascale, C. 272 Givón, Talmy 14, 202, 204, 517 Glasersfeld, Ernst von 177 Glenberg, Arthur M. 92, 104–105, 494 Glenn, Phillip J. 463 Glucksberg, Sam 5, 236, 252, 490 Goatly, Andrew 451 Goffman, Erving 136, 185, 273 Gohl, Christine 223 Goldberg, Adele 482–483, 515–516, 554 Golding, Jonathan M. 91, 94, 223, 545 Goldinger 167

Name index Goldstein 390, 463 Golinkoff, Roberta Michnick 322–323 Goodglass, Harold 345, 353, 355, 357 Goodwin, Charles 349, 351–352, 363 Goodwin, Marjorie H. 330 Gopnik, Alison 89, 204–205, 220, 561, 568 Grady, Joseph 132, 144, 478 Graesser, Arthur C. 5, 90–91, 93, 105–107, 201, 224, 463, 474 Grandin, Temple 331, 333 Grant, Julia 108 Green, David 450 Green, Georgia M. 265 Green, Ginnie 347 Green, Melanie 123 Greene, Steven B. 86 Greenfield, Patricia M. 300 Greenspan, Steven L. 92, 161 Grice, H. P. 4–5, 12, 25–26, 31–33, 36–39, 41–42, 47, 49–50, 52–56, 62–64, 67–68, 71–72, 74, 76, 78, 87, 92–93, 129, 139, 191, 207, 238, 244, 246, 251, 269, 276, 296–297, 330, 346, 375–377, 381, 388, 396, 405–418, 422, 425–426, 428, 463, 465, 485, 589–590, 597, 606 Gries, Stefan Th. 153, 524 Griggs, Richard A. 97 Grodner, Daniel 39, 390 Grondelaers, Stefan 525 Grossman, James B. 266 Grosz, Barbara 204–205, 220 Gülich, Elisabeth 441–442, 453–454 Gumperz, John J. 265, 275 Gundel, Jeannette 202, 376–377, 396 Gunlogson, Christine 16, 390 Gunter, Thomas C. 255 Gur, Inbal 248 Guzman, Alexandria E. 95 H Haberlandt, Karl 211, 213, 222, 224 Hagoort, Peter 86, 104, 107 Hahnel, Andrea 86, 91, 93 Haigh, Matthew 104 Haiman, John 494 Hala, Suzanne 266 Hale, Courtney M. 309, 336

639

Hall, Alison 12–13, 58, 60, 71, 76–77, 87, 421 Halldorson, Michael 86, 89, 93, 102, 107 Halle, Morris 166, 169, 601 Haller, Sven 221 Halliday, Michael A. K. 88–89, 201, 440, 450–452, 457, 510–512, 522 Hamblin, Jennifer L. 237 Hamby, S. 356 Hamilton, Craig A. 134 Hamrick, Phillip 471 Hanks, William F. 144 Hanna, Joy E. 65, 390–391 Hannon, Brenda 105 Hansen, Anita Berit 587–588, 599 Hansen, Maj-Britt Mosegaard 18, 587–588, 590, 592, 596–601, 603 Happé, Francesca 271, 333, 336 Harder, Peter 6, 13, 18, 507, 511, 517, 519–520, 581 Hare, Mary L. 514, 523 Harkness, Dana 93 Harnish, Robert 39, 379, 421 Harris, Paul L. 265–267, 272 Harris, Zellig S. 508 Hasan, Ruqaiya 88–89, 201, 451–452, 511 Haspelmath, Martin 537 Hasson, Uri 236, 252 Hauser, Marc D. 407 Haviland, Susan E. 88, 420 Hawthorne, John 58 Hay, Jennifer 167 Heath, Shirley Brice 272 Hedberg, Nancy 202, 376–377, 396 Heeschen, Claus 353–354 Heim, Irene 593 Heine, Bernd 517, 521, 547, 600 Helasvuo, Marja-Liisa 351 Heller, Daphna 390 Hempelmann, Christian F. 468, 470, 478 Heritage, John 347, 349 Herman, Vimala 135 Hernández, Lorena 138–139 Herold, Ruth 561 Herring, Susan vii Herzog, Marvin 561 Hiebert, Elfrieda H. 97, 103

640

Name index

Hilpert, Martin 553 Himmelmann, Nikolaus P. 537–539, 541–542, 546, 548–553 Hinskens, Frans 18, 565, 569, 577, 580 Hintzman, Douglas L. 167, 395 Hirschberg, Julia 42, 593 Hirsh-Pasek, Kathryn 322–323 Hobbs 210 Hoey, Michael 153 Hoffman, Dustin 6 Hoffman, James E. 94, 221 Hofstadter, Douglas 463, 478, 482 Höger, Claudia ix Holland, Audrey 345–347 Holland, Dorothy 175 Hollos, Marida 270 Holmes, Amanda 272–273 Holyoak, Keith J. 90, 97 Hopper, Paul 18, 511–517, 519, 521–522, 537, 549, 599 Horn, Laurence R. 12, 17, 25, 31–36, 39, 41–42, 405–406, 417–421, 427–428 Horton, William S. 16, 190, 375, 377, 387–388, 392, 394–397, 487 House, Juliane 276 Howlin, Patricia 319, 327, 338 Huang, Yan 4, 12, 25–26, 29–30, 32–36, 38–41, 49, 152, 375, 421, 426, 486 Huang, Yi Ting 277 Huber-Okrainec, Joelene 301 Huddleston, Rodney 546 Hudson, Richard 162, 204, 559, 561 Hudson-D’Zmura, Susan 204 Hünig, Wolfgang K. 482 Hünnemeier, Friederike 521 Huss, Leena 572 Hustinx, M. 89, 205 Hutchby, Ian 348–350 Hutchins, Edwin 482 Hutchinson, Thomas A. 266 Hymes, Dell 265 Hyönä, Jukka 216–217 I Ifantidou, Elly 51, 77 Ingram, David 275 Ireson, Judy 348 Isard, Stephen 106

J Jackendoff, Ray 161 Jacobs, Andreas 605 Jakobson, Roman 166 Jarvella, Robert J. 86 Jaswal, Vikram J. 267 Jefferson, Gail 348, 350, 387, 511 Jescheniak, Jörg D. 206 Johnson, Fiona 355–356 Johnson, Keith 165 Johnson, Marcia K. 89, 221 Johnson, Mark 17, 126–127, 129, 131–133, 137, 141, 175, 437, 440–441, 465, 478 Johnson-Laird, Philip N. 85, 97, 101, 175 Jolliffe, Therese 303 Jordan, Rita R. 298–299 Joseph, John E. 569 Joshi, Aravind 14, 203 Jucker, Andreas H. v, vii, 605–606 Jurafsky, Dan 515 Just, Marcel A. 90, 100, 103–104, 106, 203, 212, 217, 252, 261, 521 K Kacinik, Natalie A. 249–250 Kagan, Aura 347–348, 363 Kahneman, Daniel 98 Kamalski, Judith 215, 218 Kameyama, Megumi 204 Kanner, Leo 15, 317–318, 326 Kaplan, David 56, 76 Kaplan, Edith 345, 353, 355, 357 Karlson, Barbara ix Karmiloff-Smith, Annette 327 Kärreman, Dan 176 Karttunen, Lauri 409–410 Kasari, Connie 325 Kaschak, Michael P. 105 Kasher, Asa 3, 297 Kashua, Sayed 247 Kasper, Gabriele 265, 276 Katz, Albert 178, 190, 236, 240–241 Katz, Jerrold J. 464 Kaufer, David S. 485 Kaufman, Ronie 245 Kaup, Barbara 252, 254 Kay, Janice 345 Kay, Paul 482, 515, 593

Name index Kecskes, Istvan 14, 175–178, 182, 185, 188–191, 465 Keefe, Dennis E. 87, 90, 95–96 Keenan, Janice M. 90–91, 94, 212–213 Kehler, Andrew 210 Keller, Andrée 569–570 Kemmer, Suzanne 507 Kempff, Henk J. 93, 218 Kempson, Ruth 405 Kendrick, Carol 267 Kennedy, Katherine M. 267 Kennison, Sheila.M. 205 Kerbrat-Orecchioni, Catherine 587 Kernan, Claudia 264, 269 Kerswill, Paul 562–563 Keysar, Boaz 16, 178, 190, 388–392, 394–395 Khlentzos, Drew 103 Kiang, Michael 301 Kidd, Evan 104 Kihara, Yoshihiko 490–491 Kilgarriff, Adam 153 King, Jeffrey C. 58, 77–78 King, Marissa 338 Kintsch, Walter 5, 13, 86, 89, 92, 95–96, 105–106, 108, 163, 217–218, 469 Kirchner, D. 346 Kishner, Jeffrey 153 Kjelgaard, Margaret M. 321, 328 Klin, Ami 323 Klin, Celia M. 93, 95, 99 Klinge, Matti 572 Klippi, Anu 347, 351, 354, 361 Koch, Peter 597, 599, 602 Koenig, Melissa A. 267 Koestler, Arthur 463, 478, 481 Kogan, Michael D. 318 Koller, Veronika 446 Kolodner, Janet L. 93 König, Ekkehard 595 Konwitschny, Wolfgang ix Korta, Kepa 39, 42 Kotsinas, Ulla-Britt 603 Kotthoff, Helga 248, 468, 474 Kövecses, Zoltán 123, 138, 437, 440, 478 Krauss, Robert M. 385 Kreuz, Roger J. 490–491 Kripke, Saul 376

641

Kristiansen, Gitte 6 Kronenfeld, David B. 179, 183, 185, 187, 190 Kronmüller, Edmundo 389 Kruger, Ann C. 266 Kruley, Peter 494 Kruschke, John K. 167 Krzeszowski, Tomasz Paweł 127–129 Kuiken, Don 474 Kusbit, Gail W. 106 Kutas, Marta 91, 472, 475–477 Kuteva, Tania 517 Kyratzis, Amy 275 L Laadan, Daphna 239, 241, 245–246, 254 Laakso, Minna 347, 350–351, 357, 361 Labov, William 508–509, 523, 561–562, 568, 573, 579, 605 Lachs, Lorin 166 Lahey, Margaret 262 Lakin, Jessica L. 394 Lakoff, George 13, 17, 124–126, 131–133, 137, 141, 152–153, 157, 168–169, 175, 181, 437, 440–441, 444–445, 465, 468, 478 Lamb, Michael 265 Lambrecht, Knud 587 Landau, Barbara 327 Langacker, Ronald 13, 18, 137, 141, 144, 155–156, 158–159, 161–162, 164, 169, 180, 214, 465, 475, 479, 482, 507, 514–515, 533–537, 542, 546, 549, 554 Langlotz, Andreas 474 Langston, William E. 494 Lass, Roger 560 Laval, Virginia 268–269 Le Page, Robert B. 569–570 Lea, R. Brooke 12, 93, 98–102 Leacock, Claudia 153 Lee, Anthony 309 Lee, Benny P. H. 379 Leech, Geoffrey 123, 141 Legrenzi, Maria S. 97 Legrenzi, Paolo 537–538, 540, 603 Leinonen, Eva 265, 292 Lentz, Leo 218 Leonard, Laurence B. 320, 338

642

Name index

Lepore, Ernie 55, 74–75, 79 Lesgold, Alan M. 88 Leskinen, Heikki 566 Leslie, Alan M. 331, 388 Lesser, Ruth 345–348, 350 Levine, William H. 95 Levinson, Stephen 12, 17, 25, 28, 32–36, 38–42, 72, 78, 270, 276, 299, 405, 417, 420–421, 426–428, 598, 606 Levinthal, Daniel 177 Levy, Elena T. 273 Levy, Gideon 242, 247, 250–251 Lewis, Charlie 266 Lewis, David K. 377–378, 381, 384, 397 Liddell, Scott K. 135 Limber, John E. 96 Lin, Shuhong 388 Lind, Marianne 351, 353, 363 Lindgren, Anna-Riitta 572 Lindsay, Jane 350 Littlemore, Jeannette 17, 440–441, 451 Livnat, Elad 255 Lloyd, Barbara B. 123 Locher, Miriam A. vii Loftus, Elizabeth F. 105 Logan, Gordon D. 393 Lonergan, Julia 440 Long, Debra 93, 106, 463, 474 Lopez-Cóuso, María José 554 Lorch, Elizabeth P. 216–218 Lorch, Robert F. 86, 107, 216–218 Loth, Eva 309 Loukusa, Soile 292, 304, 334 Louw, Bill 168 Louwerse, Max M. 201 Loveland, Katherine A. 308 Low, Graham 132, 450 Lowerre, George F. 97 Lubinski, R. 351 Lucariello, Joan 269 Lucas, Margery M. 95 Luckmann, Thomas 183 Lüdtke, Jana 252 Lum, Carmel C. 355 Lundmark, Carita 481–482 Lusignan, Serge 153 Lutz, Mark F. 107 Luzzi, Simona 306

Lyon, Jon G. 320, 348 Lyons, John 91, 169, 510, 536, 588 M MacDonald, C. Maryellen 100, 252 Mackie, Nina 347 Macrae, C. Neil 189 MacWhinney, Brian 206, 214, 262, 509, 513, 520 Madden, Carol J. 91, 269, 353 Magliano, Joseph P. 90–91, 93 Mahon, Merle 348 Main, Rebecca M. 267, 275 Maio, Sara Di 468, 470, 478 Mak, Willem M. 215 Makin, Valarie S. 167 Malinowski, Bronislaw 510 Malone, Lauren S. 267 Mann, William C. 210 Mannes, Suzanne 94, 221 Manochiopinig, S. 346 Marcos, Haydee 269 Marín-Arrese, Juana 482 Markman, Ellen M. 123, 267, 272 Marmaridou, Sophia S.A. 126, 129, 144 Marshall, Catherine R. 377, 379, 381–383 Marshall, Jane 353 Marslen-Wilson, William D. 523 Martí, Luisa 58, 62, 77 Martin, Ingerith 301 Martin, Janice 537 Martin, Rod A. 467 Martín-Loeches 221 Marx, Traci 275 Masci, Michela 448 Mason, Robert A. 90 Mate, David 42 Matthiessen, Christian M.I.M. 522 Mattson, Mark E. 105 Maury, Pascale 219 Maxim, Jane 347–348, 353, 356–357 Mazali, Rela 250–251 McCaleb, Peggy 300 McClelland, James L. 108, 524 McDaniel, Mark 87, 90, 95–96 McDonald, Janet L. 206 McDonald, Skye 296, 301 McGhee, Paul E. 267

Name index McGlone, Chadd. 216 McKoon, Gail 86, 91, 94–95, 97, 106, 395 McMichael, Kipp 166 McNamara, Danielle S. 106, 201, 218 McSpadden, Merrill 106 McTear, Michael F. 264, 346 McVicker, Sally 347, 363 Medin, Douglas L. 167 Meillet, Antoine 516 Meltzoff, Andrew N. 330 Menn, Lise 346 Merkin, Susan 275 Mervis, Caroline 123, 162, 272 Metzing, Charles 386 Mey, Jacob 187–188 Meyer, Bonnie J. F. 217–218 Meyerhoff, Miriam 572–576 Meyers, Kristina 449–450 Meytes, Dalia 240 Miall, David. S. 474 Michaelis, Laura A. 483 Miller, George 106, 153 Miller, James R. 95 Miller, Jim 185 Millis, Keith K. 91, 201, 212, 217, 223–224 Milroy, James 560–561 Milroy, Lesley 346–347, 350, 560–561, 568–569 Mindolovich, Catherine 269 Minsky, Marvin 175, 465 Mio, Jeffrey S. 463, 474 Moeschler, Jacques 12, 16–17, 49, 87, 139, 296, 405, 426–428 Moon, Rosamund E. 440, 443–444, 448–450 Moore, Chris 324 Morgan, Jerry 405 Morris, Charles W. v, 405 Morris, Robin K. 234 Morrow, Daniel G. 91–92 Mross, Ernest F. 95 Mulder, Gerben 216 Mulkay, Michael 467 Mulligan, Elizabeth J. 93, 101 Münte, Thomas F. 91 Murphy, Gregory 169, 387 Murray, John D. 95, 99, 216 Mürter, Johannes ix

643

Musolff, Andreas 134, 444–445 Musolino, Julien 327 Myers, Jerome L. 86, 90, 93, 95–97, 99, 106, 213, 395 N Nakagawa, Noreen 269 Nakamura, Keiko 270, 275, 490 Nassar, Natasha 338 Nayak, Nandini. P. 237 Neale, Stephen 53 Nelson, Katherine 262, 264, 275 Nerlich, Brigitte 134 Nevalainen, Terttu 18, 559, 573, 575–576 Newell, Allen 105 Newport, Elissa L. 323 Nikiforidou, Kiki 471, 482–483, 493 Ninio, Anat 262, 265, 269, 273–274, 278 Nippold, Marilyn A. 268 Nølke, Henning 604 Noordman, Leo 93, 201, 210–213, 215, 217–218 Norbury, Courtenay 301, 303, 309, 332 Nordin, Jonas 572 Nordlund, Taru 572 Norén, Coco 604 Norrick, Neal v, 3–5 Noveck, Ira 5, 39, 65–66, 78, 103, 406 Nunan, David 439 Nunberg, Geoffrey 70, 169 Nuolijärvi, Pirkko 567 O O’Brien, David P. 86, 92–93, 96–97, 101, 103, 106–107 O’Connor, Mary Catherine 482, 515 O’Halloran, Kieran 446 O’Neill, Daniela K. 5, 15, 261, 265, 267, 270–273, 275, 619 Oakley, Todd 136, 144, 475, 481 Ochs, Elinor 207, 275 Oelschlaeger, Mary 348, 351, 353 Olsher, David 351 Olson, David R. 265 Oostendorp, Herre van 92 Oppy, Brian J. 106 Ortony, Andrew 161, 163, 250 Otten, Marte 95

644

Name index

P Pacht, Jeremy M. 234 Pagin, Peter 77 Pagliuca, William 18, 517, 521, 533, 553 Palma, Paul De 494 Pander Maat, Henk 201, 204–205, 214 Panizza, D. 39 Panther, Klaus-Uwe 139–141, 479 Papafragou, Anna 129, 277 Papagno, Costanza 301 Paris, Scott G. 97, 573 Parker, Mark 347 Parr, Susie 348 Partington, Alan 248 Pascual, Esther 136 Patterson, Courtney 236, 240 Pawley, Andrew K. 514 Pearce, Michelle J. 272 Pearce, Samantha 296 Peirce, C.S. v Peleg, Orna 234–236, 239 Penn, Claire 346 Pentland, Brian T. 188 Perfetti, Charles A. 94, 107 Perkins, Lisa 346–348, 350, 359 Perkins, Michael 262, 363 Perkins, Revere 18, 517, 521, 533, 553 Perner, Josef 266, 331, 380, 388 Perrett, Daniel 92 Perry, John 39, 42, 65 Peters, Stanley 409–410 Pexman, Penny 241, 333–334 Pick, Jennifer L. 272, 309 Pickar, Joanne 265 Pickering, Martin J. 214–215, 393–394, 396 Pierrehumbert, Janet 167 Pietroski, Paul 61, 77 Pijnacker, Judith 297, 335 Pinker, Steven 103 Pisoni, David 166 Pollard, Charles J. 204 Pollio, Howard R. 443, 449, 478 Pons Bordería, Salvador 601 Pons, Francisco 266 Port, Robert 166 Potts, Christopher 425 Potts, George R. 86, 91, 94

Pound, Carole 348 Povinelli, Daniel 337 Premack, David 16, 331, 388 Prince, Ellen 14, 203 Prins, Ronald 353 Prizant, Barry M. 266, 300, 326 Prutting, Carol A. 262, 265, 273, 346 Pullum, Geoffrey K. 546 Pustejovsky, James 160 Putnam, Hilary 42 Q Qualls, Constance Dean 301 Quinn, Naomi 175 Quirk, Randolph 544 R Radden, Günter 138, 479 Rader, Aaron W. 100 Radford, Julie 348 Radvansky, Gabriel 86, 90, 92, 107, 163 Rall, Jaime 272 Ramsberger, Gail 346 Ramsden, Gina 264, 346 Rapin, Isabelle 292, 308, 326 Raskin, Victor 17, 463–469, 476, 479, 494 Ratcliff, Roger 86, 91, 94–95, 97, 106, 395 Raumolin-Brunberg, Helena 575–576 Rayner, Keith 87, 104, 203, 206, 234–235 Reboul, Anne 5, 15, 39, 264–266, 292, 294, 317, 427 Recanati, François 29–30, 38, 53, 55–56, 58, 61, 64–66, 70, 77–78, 412, 424, 428 Reddy, Michael 180 Reddy, Vasudevi 267 Reder, Lynne M. 106 Reed, V. 346 Rees, Norma S. 262 Regel, Stefanie 255 Regier, Terry 127 Reihlen, Markus 177, 188–189 Reiser, Brian J. 100 Rerup, Claus 177 Revlin, Russell 86, 93, 102 Reznik, J. Steven 324 Ricard, Marcelle 272 Richardson, Daniel C. 127 Rinaldi, Maria Christina 301

Name index Rinaldi, Wendy 304 Rinck, Mike 86, 91–93 Ringberg, Torsten 177, 188–189 Ripich, Danielle N. 295 Rips, Lance J. 97, 101, 103 Rispens, Judith 364 Ritchie, Graeme 470–471, 480 Rizzella, Michelle L. 96, 106 Rizzolatti, Giacomo 330 Roazzi, Antonio 98 Roberts, Celia 439, 581 Rodrigues, Amariles A. 98 Roenfeldt, Barbara 357 Rogers, Erin S. 391 Rohrer, Tim 127 Rollins, Pamela R. 295, 308 Rolnik, Guy 248 Rorty, Richard 507 Rosch, Eleanor 123, 162, 175 Rosenbloom, Lewis 292, 308 Ross, Brian. H. 167 Roßnagel, Christian S. 392 Roth, Steven F. 88, 455 Rowan, Lynne E. 300 Rowe, Victoria 134 Ruch, Willibald 469 Rudzka-Ostyn, Brygida 478 Ruiz de Mendoza-Ibáñez, Francisco-José 138–139, 141–142 Rumain, Barbara 97, 100, 103 Rumelhart, David E. 108, 524 Ruppenhofer, Josef 515 Russell, Bertrand 50, 53, 55, 405 Ryckebusch, Celine 269 Ryder, Nuala 297 S Sacks, Harvey 348, 350, 387, 511 Sadock, Jerrold M. 414–416 Saeed, John 124 Saffran, Jenny R. 323 Sag, Ivan 70 Sage, L. Graham vii Sainson, Claire 305 Sairio, Anni 569 Salmon, Nathan 76 Sanchez, Rebecca P. 216, 218 Sandberg, Annika Dahlgren 309

645

Sanders, José 215, 223 Sanders, Ted 5, 12, 14, 201, 203–205, 210–214, 216–218, 222 Sandra, Dominiek 239–240, 250 Sanford, Anthony 5, 89, 91–92, 105–106, 203, 208 Sankoff, Gillian 562, 572 Santa Ana, Otto 446 Sarangi, Srikant 439 Sarno, Martha Taylor 346 Sauerland, Uli 42 Saul, Jennifer 41–42, 474 Saussure, Ferdinand de 164, 183, 407, 521 Saylor, Megan M. 267 Scandura, Joseph M. 97 Scearce, Kimberly A. 89, 204 Schaefer, Edward F. 384–385 Schaeken, Walter 101 Schank, Roger 85, 90, 141, 175, 382, 465, 494 Schegloff, Emanuel 348–350, 353–354, 361, 363, 387, 511 Schelling, Thomas 377 Schiefelbusch, Richard L. 265 Schiffer, Stephen R. 377, 379, 382 Schiltz, Kolja 91 Schleich, Michelle C. 91 Schmalhofer, Franz 87, 90, 94, 96 Schmid, Hans-Jörg 3, 13, 41, 143, 554 Schmidt, Sarah 393 Schmitzer, A 304 Schneider, Edgar W. 570–571, 578 Schneider, Klaus P. v, vi Schober, Michael F. 390–391 Schoen, Lawrence M. 161 Schön, Donald 133, 446 Schooler, Jonathan W. 106 Schreuder, Robert 168, 387, 390 Schuler, Adriana L. 326 Schutz, Alfred 41 Schwartz, Tamir 244, 485, 488 Schwenter, Scott A. 597–598, 601 Schwint, Christopher A. 236 Scollon, Ron 605 Scollon, Suzanne Wong 605 Scott, Fiona J. 318 Scott, Paula J. 207

646

Name index

Sealey, Alison 580 Searle, John 4–5, 56, 61, 92, 139–141, 168, 238, 262, 330, 346, 376, 405, 409 Seely, Mark R. 106 Segal, Erwin M. 207 Semino, Elena 17, 439–441, 444, 446, 448, 451 Sengul, C. J. 206, 272 Shatz, Marilyn 266–267, 275 Sheard, C. 346 Sherrod, Lonnie 265 Shiffrin, Deborah 508 Shiffrin, Richard M. 395 Shinjo, Makiko 90, 213 Shore, Bradd 181 Short-Meyerson, Katherine J. 265, 267 Shroyer, Suzanne 205 Shwe, Helen 267 Sigman, Marian 271, 325 Silverman, David 349 Simmel, Georg 183–184 Simmons, Tamra P. 295, 347–348, 352, 355 Simmons-Mackie, Nina 348, 352, 355 Simon, Herbert A. 5, 105, 359, 361–362, 364, 390 Simons, Wim H. G. 89, 205 Sinclair, John 185, 439, 456, 510, 514–515 Singer, Murray 12, 86, 89–91, 93–94, 96, 102, 106–108 Sinha, Chris 136 Sinigaglia, Corrado 330 Sitnikova, Tatiana 303–304 Skarakis, Elizabeth 300 Skinner, Christine 346 Slobin, Dan 275 Sloutsky, Vladimir 100 Smallwood, Jonathan 106 Smith, Benita R. 265 Smith, Edward E. 167 Smith, Jennifer 581 Smith, K. Aaron. 513 Smith, Neil U. 379 Snedeker, Jesse 277 Snow, Catherine E. 262, 264–265, 269, 273–274, 278 Soames, Scott 76 Soar, Naomi R. 355 Solheim, Ivar 190

Solomon, Susan K. 89 Sorjonen, Marja-Liisa 351 Soulières, Isabelle 323 Speelman, Dirk 223, 525 Sperber, Dan 4–6, 12, 15, 17, 25–28, 30, 37–42, 48, 55, 58, 64–65, 69, 71–72, 74, 76–78, 102, 142, 189, 233, 238, 251, 254, 264, 276, 292, 330, 379–381, 383–384, 405–406, 421–426, 428, 485–487, 489, 491 Spooren, Wilbert 201, 210–212, 222–223 Spyridakis, Jan H. 217 St. George, Marie 94 St. James, Paula 268 Stalnaker, Robert. 191, 383–384, 398, 405, 412 Standal, Timothy C. 217 Stanfield, Robert A. 92, 105 Stanley, Jason 58–59, 62, 77–78 Steedman, Mark 55, 106 Steen, Gerard 456–457 Stefanowitsch, Anatol 524 Stein, Dieter vii Stelma, Juurd 442–444, 449 Stewart, Andrew J 93, 104 Strawson, Peter F. 51, 76, 409 Stroop, John R. 393 Stubbs, Michael 510 Stukker, Ninke, Ted Sanders 214 Suh, S. Y. 90 Sullivan, Kate 338 Surian, Luca 297, 335 Sweetser, Eve 129, 202, 215, 223, 535, 549 Swinney, David A. 55, 86, 244 Syder, Frances H. 514 Szabó, Zoltan Gendler 77 Szwedek, Aleksander 144 T Taavitsainen, Irma vii Taboada, Maite 210 Tager-Flusberg, Helen 268, 309, 320–321, 328, 336, 338 Talmy, Leonard 129, 144 Tamir, Ariella 240 Tanenhaus, Michael 16, 95, 204, 389–391 Tannen, Deborah 186, 463

Name index Tantalou, Niki 277 Taylor, John R. 6, 12–15, 124, 137, 151–152, 155, 157, 162, 164, 168, 470 Teisserenc, Amélie 219 Tényi, Tamás 297 Terkourafi, Marina 539 Thelander, Mats 567 Thomas, Erik R. 214, 545, 577 Thompson, Emma 6 Thompson, Linda 265 Thompson, Sandra 210, 512, 519, 521–522 Thornburg, Linda 139, 141, 479 Thornbury, Scott 438 Till, Robert E. 95 Tirassa, Maurizio 42 Tognini-Bonelli, Elena 524 Tomasello, Michael 18, 262, 265–266, 323–325, 513 Tomlin, Russell S. 204 Tooby, John 55 Torres Cacoullos, Rena 601 Trabasso, Thomas 90, 212 Traugott, Elizabeth Closs 18, 535–539, 541–543, 545–554, 592, 597–601, 603, 606 Travis, Charles 52, 56 Traxler, Matthew J. 93, 214–215 Treyens, James C. 383 Triandis, Harry C. 183 Trosborg, Anna vii Trousdale, Graeme 18, 517, 533, 539, 542, 549, 581 Trudgill, Peter 564, 578 Trueswell, John 389–391 Tsakona, Villy 469 Tuggy, David 596 Tulving, Endel 93, 382 Turner, Ken v Turner, Mark 6, 135–136, 138, 144, 468, 475–476, 481–482 Tversky, Amos 98 Tyack, Dorothy 275 Tyler, Andrea 152, 157 U Ungerer, Friedrich 13 Urbach, Thomas P. 475–476

647

V Vaattovaara, Johanna 574, 581 Vaid, Jyotsna 467, 472 van Berkum 86, 95, 104, 107 van den Broek, Paul 90, 107 Van Herk, Gerard 568, 577 Van Lancker Sidtis, Diana 363 Van Teeffellen, Toine 446 Vandaele, Jeroen 467 Vanhalle, C. 296 Varley, Rosemary 299 Varner, Kathleen R. 106 Veale, Tony 467, 470–471, 478–479, 482 Vega-Moreno 268 Verhagen, Arie 214, 223, 513, 519 Verschueren, Jeff 265 Verstraete, Jean-Christophe 550, 552, 554 Virtanen, Tuija vii, 207 Visconti, Jacqueline 597, 601 Vlassova, Tatiana 128 Vogelaer, Gunther De 561 Volterra, Virginia 266, 325 von Fintel, Kai 77 Vonk, Jennifer 337 Vonk, Wietske 89, 93, 205, 207, 211, 218 Vygotsky, Lev. S. 176 W Wade, Evelyn 275 Walker, E. C. 14, 203 Waltereit, Richard 597–601, 603–605 Walton, Jennifer L. 93 Wang, Jinrui 6 Wardlow Lane, Liane 392 Wason, Peter C. 97 Watson, Julanne 331 Watt, Dominic 581 Wedgwood, Daniel 75 Weiner, Bernard 90 Weiner, E. Judith 494 Weinert, Regina 185 Weingartner, Kristin M. 95 Weinheimer, Sidney 385 Weinreich, Uriel 561 Weitzner-Lin, B. 351 Wekker, Herman C. 524 Wellman, Henry M. 266–267

648

Name index

Wetherby, Amy M. 262, 266 Wheeler, Kathryn 91 White, Michael 451 Whitney, Paul 94 Whitten, Shannon N. 91 Whitworth, Anne 347–348 Wieman, Leslie A. 267 Wierzbicka, Anna 159, 605 Wilcox, M. Jeanne 346–347, 362 Wilkins, David 600 Wilkinson, Ray 347–350, 353, 356–357 Williams, Donna 331, 333, 338 Williams, John. N. 236, 239 Wilson, Deirdre 4, 6, 12, 15, 17, 25–30, 37–38, 40–42, 48, 55, 58, 61, 63–64, 69, 71–72, 74, 76–78, 92, 102, 142, 189, 233, 238, 251, 254, 264, 276, 292, 330, 377, 379–381, 383–384, 405, 421–426, 428, 485–487, 489, 491, 545 Wimmer, Heinz 266, 331 Winkler, Isvan 323 Winner, Ellen 269 Wirz, Sheila L. 346 Wischer, Ilse 541 Wittgenstein, Ludwig v, 51, 176, 507 Wodak, Ruth 134

Wode, Henning 275 Woodruff, Guy 16, 331, 388 Wooffitt, Robin 348–350 Woolf, Philip 441 Wray, Alison 185, 363, 456, 510 Y Yaeger-Dror, Malcah 577 Yamaguchi, Haruhiko 474 Yang, Chin Lung 94, 104 Yates, Wendy 347, 357 Yaxley, Richard H. 92, 105 Young, Carole J. 218 Yule, George 438 Z Zacharski, Ron 202, 376–377, 396 Zhang, Fenghui 175–176, 178, 190–191 Ziatas, Kathryn 295 Ziemke, Tom 267, 275 Zinken, Jörg 134, 144 Zufferey, Sandrine 328 Zurif, Edgar 356 Zwaan, Rolf 5, 90–93, 105, 163, 201, 209–210, 252, 254 Zwicky, Arnold 420, 546